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 Subclasses may also override suitable() if necessary, but ensure the function
410 signature is preserved and that this function imports everything it needs
411 (except other extractors), so that lazy_extractors works correctly
413 _GEO_BYPASS attribute may be set to False in order to disable
414 geo restriction bypass mechanisms for a particular extractor.
415 Though it won't disable explicit geo restriction bypass based on
416 country code provided with geo_bypass_country.
418 _GEO_COUNTRIES attribute may contain a list of presumably geo unrestricted
419 countries for this extractor. One of these countries will be used by
420 geo restriction bypass mechanism right away in order to bypass
421 geo restriction, of course, if the mechanism is not disabled.
423 _GEO_IP_BLOCKS attribute may contain a list of presumably geo unrestricted
424 IP blocks in CIDR notation for this extractor. One of these IP blocks
425 will be used by geo restriction bypass mechanism similarly
428 The _WORKING attribute should be set to False for broken IEs
429 in order to warn the users and skip the tests.
434 _x_forwarded_for_ip
= None
436 _GEO_COUNTRIES
= None
437 _GEO_IP_BLOCKS
= None
441 'any': 'Use --cookies, --username and --password or --netrc to provide account credentials',
443 'Use --cookies for the authentication. '
444 'See https://github.com/ytdl-org/youtube-dl#how-do-i-pass-cookies-to-youtube-dl for how to pass cookies'),
445 'password': 'Use --username and --password or --netrc to provide account credentials',
448 def __init__(self
, downloader
=None):
449 """Constructor. Receives an optional downloader."""
451 self
._x
_forwarded
_for
_ip
= None
452 self
._printed
_messages
= set()
453 self
.set_downloader(downloader
)
456 def _match_valid_url(cls
, url
):
457 # This does not use has/getattr intentionally - we want to know whether
458 # we have cached the regexp for *this* class, whereas getattr would also
459 # match the superclass
460 if '_VALID_URL_RE' not in cls
.__dict
__:
461 cls
._VALID
_URL
_RE
= re
.compile(cls
._VALID
_URL
)
462 return cls
._VALID
_URL
_RE
.match(url
)
465 def suitable(cls
, url
):
466 """Receives a URL and returns True if suitable for this IE."""
467 # This function must import everything it needs (except other extractors),
468 # so that lazy_extractors works correctly
469 return cls
._match
_valid
_url
(url
) is not None
472 def _match_id(cls
, url
):
473 return cls
._match
_valid
_url
(url
).group('id')
476 def get_temp_id(cls
, url
):
478 return cls
._match
_id
(url
)
479 except (IndexError, AttributeError):
484 """Getter method for _WORKING."""
487 def initialize(self
):
488 """Initializes an instance (authentication, etc)."""
489 self
._printed
_messages
= set()
490 self
._initialize
_geo
_bypass
({
491 'countries': self
._GEO
_COUNTRIES
,
492 'ip_blocks': self
._GEO
_IP
_BLOCKS
,
495 self
._real
_initialize
()
498 def _initialize_geo_bypass(self
, geo_bypass_context
):
500 Initialize geo restriction bypass mechanism.
502 This method is used to initialize geo bypass mechanism based on faking
503 X-Forwarded-For HTTP header. A random country from provided country list
504 is selected and a random IP belonging to this country is generated. This
505 IP will be passed as X-Forwarded-For HTTP header in all subsequent
508 This method will be used for initial geo bypass mechanism initialization
509 during the instance initialization with _GEO_COUNTRIES and
512 You may also manually call it from extractor's code if geo bypass
513 information is not available beforehand (e.g. obtained during
514 extraction) or due to some other reason. In this case you should pass
515 this information in geo bypass context passed as first argument. It may
516 contain following fields:
518 countries: List of geo unrestricted countries (similar
520 ip_blocks: List of geo unrestricted IP blocks in CIDR notation
521 (similar to _GEO_IP_BLOCKS)
524 if not self
._x
_forwarded
_for
_ip
:
526 # Geo bypass mechanism is explicitly disabled by user
527 if not self
.get_param('geo_bypass', True):
530 if not geo_bypass_context
:
531 geo_bypass_context
= {}
533 # Backward compatibility: previously _initialize_geo_bypass
534 # expected a list of countries, some 3rd party code may still use
536 if isinstance(geo_bypass_context
, (list, tuple)):
537 geo_bypass_context
= {
538 'countries': geo_bypass_context
,
541 # The whole point of geo bypass mechanism is to fake IP
542 # as X-Forwarded-For HTTP header based on some IP block or
545 # Path 1: bypassing based on IP block in CIDR notation
547 # Explicit IP block specified by user, use it right away
548 # regardless of whether extractor is geo bypassable or not
549 ip_block
= self
.get_param('geo_bypass_ip_block', None)
551 # Otherwise use random IP block from geo bypass context but only
552 # if extractor is known as geo bypassable
554 ip_blocks
= geo_bypass_context
.get('ip_blocks')
555 if self
._GEO
_BYPASS
and ip_blocks
:
556 ip_block
= random
.choice(ip_blocks
)
559 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(ip_block
)
560 self
._downloader
.write_debug(
561 '[debug] Using fake IP %s as X-Forwarded-For' % self
._x
_forwarded
_for
_ip
)
564 # Path 2: bypassing based on country code
566 # Explicit country code specified by user, use it right away
567 # regardless of whether extractor is geo bypassable or not
568 country
= self
.get_param('geo_bypass_country', None)
570 # Otherwise use random country code from geo bypass context but
571 # only if extractor is known as geo bypassable
573 countries
= geo_bypass_context
.get('countries')
574 if self
._GEO
_BYPASS
and countries
:
575 country
= random
.choice(countries
)
578 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(country
)
579 self
._downloader
.write_debug(
580 'Using fake IP %s (%s) as X-Forwarded-For' % (self
._x
_forwarded
_for
_ip
, country
.upper()))
582 def extract(self
, url
):
583 """Extracts URL information and returns it in list of dicts."""
588 self
.write_debug('Extracting URL: %s' % url
)
589 ie_result
= self
._real
_extract
(url
)
590 if ie_result
is None:
592 if self
._x
_forwarded
_for
_ip
:
593 ie_result
['__x_forwarded_for_ip'] = self
._x
_forwarded
_for
_ip
594 subtitles
= ie_result
.get('subtitles')
595 if (subtitles
and 'live_chat' in subtitles
596 and 'no-live-chat' in self
.get_param('compat_opts', [])):
597 del subtitles
['live_chat']
599 except GeoRestrictedError
as e
:
600 if self
.__maybe
_fake
_ip
_and
_retry
(e
.countries
):
603 except ExtractorError
as e
:
604 video_id
= e
.video_id
or self
.get_temp_id(url
)
605 raise ExtractorError(
606 e
.msg
, video_id
=video_id
, ie
=self
.IE_NAME
, tb
=e
.traceback
, expected
=e
.expected
, cause
=e
.cause
)
607 except compat_http_client
.IncompleteRead
as e
:
608 raise ExtractorError('A network error has occurred.', cause
=e
, expected
=True, video_id
=self
.get_temp_id(url
))
609 except (KeyError, StopIteration) as e
:
610 raise ExtractorError('An extractor error has occurred.', cause
=e
, video_id
=self
.get_temp_id(url
))
612 def __maybe_fake_ip_and_retry(self
, countries
):
613 if (not self
.get_param('geo_bypass_country', None)
615 and self
.get_param('geo_bypass', True)
616 and not self
._x
_forwarded
_for
_ip
618 country_code
= random
.choice(countries
)
619 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(country_code
)
620 if self
._x
_forwarded
_for
_ip
:
622 'Video is geo restricted. Retrying extraction with fake IP %s (%s) as X-Forwarded-For.'
623 % (self
._x
_forwarded
_for
_ip
, country_code
.upper()))
627 def set_downloader(self
, downloader
):
628 """Sets the downloader for this IE."""
629 self
._downloader
= downloader
631 def _real_initialize(self
):
632 """Real initialization process. Redefine in subclasses."""
635 def _real_extract(self
, url
):
636 """Real extraction process. Redefine in subclasses."""
641 """A string for getting the InfoExtractor with get_info_extractor"""
642 return cls
.__name
__[:-2]
646 return compat_str(type(self
).__name
__[:-2])
649 def __can_accept_status_code(err
, expected_status
):
650 assert isinstance(err
, compat_urllib_error
.HTTPError
)
651 if expected_status
is None:
653 elif callable(expected_status
):
654 return expected_status(err
.code
) is True
656 return err
.code
in variadic(expected_status
)
658 def _request_webpage(self
, url_or_request
, video_id
, note
=None, errnote
=None, fatal
=True, data
=None, headers
={}, query={}
, expected_status
=None):
660 Return the response handle.
662 See _download_webpage docstring for arguments specification.
664 if not self
._downloader
._first
_webpage
_request
:
665 sleep_interval
= float_or_none(self
.get_param('sleep_interval_requests')) or 0
666 if sleep_interval
> 0:
667 self
.to_screen('Sleeping %s seconds ...' % sleep_interval
)
668 time
.sleep(sleep_interval
)
670 self
._downloader
._first
_webpage
_request
= False
673 self
.report_download_webpage(video_id
)
674 elif note
is not False:
676 self
.to_screen('%s' % (note
,))
678 self
.to_screen('%s: %s' % (video_id
, note
))
680 # Some sites check X-Forwarded-For HTTP header in order to figure out
681 # the origin of the client behind proxy. This allows bypassing geo
682 # restriction by faking this header's value to IP that belongs to some
683 # geo unrestricted country. We will do so once we encounter any
684 # geo restriction error.
685 if self
._x
_forwarded
_for
_ip
:
686 if 'X-Forwarded-For' not in headers
:
687 headers
['X-Forwarded-For'] = self
._x
_forwarded
_for
_ip
689 if isinstance(url_or_request
, compat_urllib_request
.Request
):
690 url_or_request
= update_Request(
691 url_or_request
, data
=data
, headers
=headers
, query
=query
)
694 url_or_request
= update_url_query(url_or_request
, query
)
695 if data
is not None or headers
:
696 url_or_request
= sanitized_Request(url_or_request
, data
, headers
)
698 return self
._downloader
.urlopen(url_or_request
)
699 except network_exceptions
as err
:
700 if isinstance(err
, compat_urllib_error
.HTTPError
):
701 if self
.__can
_accept
_status
_code
(err
, expected_status
):
702 # Retain reference to error to prevent file object from
703 # being closed before it can be read. Works around the
704 # effects of <https://bugs.python.org/issue15002>
705 # introduced in Python 3.4.1.
712 errnote
= 'Unable to download webpage'
714 errmsg
= '%s: %s' % (errnote
, error_to_compat_str(err
))
716 raise ExtractorError(errmsg
, sys
.exc_info()[2], cause
=err
)
718 self
.report_warning(errmsg
)
721 def _download_webpage_handle(self
, url_or_request
, video_id
, note
=None, errnote
=None, fatal
=True, encoding
=None, data
=None, headers
={}, query={}
, expected_status
=None):
723 Return a tuple (page content as string, URL handle).
725 See _download_webpage docstring for arguments specification.
727 # Strip hashes from the URL (#1038)
728 if isinstance(url_or_request
, (compat_str
, str)):
729 url_or_request
= url_or_request
.partition('#')[0]
731 urlh
= self
._request
_webpage
(url_or_request
, video_id
, note
, errnote
, fatal
, data
=data
, headers
=headers
, query
=query
, expected_status
=expected_status
)
735 content
= self
._webpage
_read
_content
(urlh
, url_or_request
, video_id
, note
, errnote
, fatal
, encoding
=encoding
)
736 return (content
, urlh
)
739 def _guess_encoding_from_content(content_type
, webpage_bytes
):
740 m
= re
.match(r
'[a-zA-Z0-9_.-]+/[a-zA-Z0-9_.-]+\s*;\s*charset=(.+)', content_type
)
742 encoding
= m
.group(1)
744 m
= re
.search(br
'<meta[^>]+charset=[\'"]?([^\'")]+)[ /\'">]',
745 webpage_bytes[:1024])
747 encoding = m.group(1).decode('ascii')
748 elif webpage_bytes.startswith(b'\xff\xfe'):
755 def __check_blocked(self, content):
756 first_block = content[:512]
757 if ('<title>Access to this site is blocked</title>' in content
758 and 'Websense' in first_block):
759 msg = 'Access to this webpage has been blocked by Websense filtering software in your network.'
760 blocked_iframe = self._html_search_regex(
761 r'<iframe src="([^
"]+)"', content,
762 'Websense information URL
', default=None)
764 msg += ' Visit
%s for more details
' % blocked_iframe
765 raise ExtractorError(msg, expected=True)
766 if '<title
>The URL you requested has been blocked
</title
>' in first_block:
768 'Access to this webpage has been blocked by Indian censorship
. '
769 'Use a VPN
or proxy
server (with --proxy
) to route around it
.')
770 block_msg = self._html_search_regex(
771 r'</h1
><p
>(.*?
)</p
>',
772 content, 'block message
', default=None)
774 msg += ' (Message
: "%s")' % block_msg.replace('\n', ' ')
775 raise ExtractorError(msg, expected=True)
776 if ('<title
>TTK
:: Доступ к ресурсу ограничен
</title
>' in content
777 and 'blocklist
.rkn
.gov
.ru
' in content):
778 raise ExtractorError(
779 'Access to this webpage has been blocked by decision of the Russian government
. '
780 'Visit http
://blocklist
.rkn
.gov
.ru
/ for a block reason
.',
783 def _webpage_read_content(self, urlh, url_or_request, video_id, note=None, errnote=None, fatal=True, prefix=None, encoding=None):
784 content_type = urlh.headers.get('Content
-Type
', '')
785 webpage_bytes = urlh.read()
786 if prefix is not None:
787 webpage_bytes = prefix + webpage_bytes
789 encoding = self._guess_encoding_from_content(content_type, webpage_bytes)
790 if self.get_param('dump_intermediate_pages
', False):
791 self.to_screen('Dumping request to
' + urlh.geturl())
792 dump = base64.b64encode(webpage_bytes).decode('ascii
')
793 self._downloader.to_screen(dump)
794 if self.get_param('write_pages
', False):
795 basen = '%s_%s' % (video_id, urlh.geturl())
796 trim_length = self.get_param('trim_file_name
') or 240
797 if len(basen) > trim_length:
798 h = '___
' + hashlib.md5(basen.encode('utf
-8')).hexdigest()
799 basen = basen[:trim_length - len(h)] + h
800 raw_filename = basen + '.dump
'
801 filename = sanitize_filename(raw_filename, restricted=True)
802 self.to_screen('Saving request to
' + filename)
803 # Working around MAX_PATH limitation on Windows (see
804 # http://msdn.microsoft.com/en-us/library/windows/desktop/aa365247(v=vs.85).aspx)
805 if compat_os_name == 'nt
':
806 absfilepath = os.path.abspath(filename)
807 if len(absfilepath) > 259:
808 filename = '\\\\?
\\' + absfilepath
809 with open(filename, 'wb
') as outf:
810 outf.write(webpage_bytes)
813 content = webpage_bytes.decode(encoding, 'replace
')
815 content = webpage_bytes.decode('utf
-8', 'replace
')
817 self.__check_blocked(content)
821 def _download_webpage(
822 self, url_or_request, video_id, note=None, errnote=None,
823 fatal=True, tries=1, timeout=5, encoding=None, data=None,
824 headers={}, query={}, expected_status=None):
826 Return the data of the page as a string.
829 url_or_request -- plain text URL as a string or
830 a compat_urllib_request.Requestobject
831 video_id -- Video/playlist/item identifier (string)
834 note -- note printed before downloading (string)
835 errnote -- note printed in case of an error (string)
836 fatal -- flag denoting whether error should be considered fatal,
837 i.e. whether it should cause ExtractionError to be raised,
838 otherwise a warning will be reported and extraction continued
839 tries -- number of tries
840 timeout -- sleep interval between tries
841 encoding -- encoding for a page content decoding, guessed automatically
842 when not explicitly specified
843 data -- POST data (bytes)
844 headers -- HTTP headers (dict)
845 query -- URL query (dict)
846 expected_status -- allows to accept failed HTTP requests (non 2xx
847 status code) by explicitly specifying a set of accepted status
848 codes. Can be any of the following entities:
849 - an integer type specifying an exact failed status code to
851 - a list or a tuple of integer types specifying a list of
852 failed status codes to accept
853 - a callable accepting an actual failed status code and
854 returning True if it should be accepted
855 Note that this argument does not affect success status codes (2xx)
856 which are always accepted.
861 while success is False:
863 res = self._download_webpage_handle(
864 url_or_request, video_id, note, errnote, fatal,
865 encoding=encoding, data=data, headers=headers, query=query,
866 expected_status=expected_status)
868 except compat_http_client.IncompleteRead as e:
870 if try_count >= tries:
872 self._sleep(timeout, video_id)
879 def _download_xml_handle(
880 self, url_or_request, video_id, note='Downloading XML
',
881 errnote='Unable to download XML
', transform_source=None,
882 fatal=True, encoding=None, data=None, headers={}, query={},
883 expected_status=None):
885 Return a tuple (xml as an compat_etree_Element, URL handle).
887 See _download_webpage docstring for arguments specification.
889 res = self._download_webpage_handle(
890 url_or_request, video_id, note, errnote, fatal=fatal,
891 encoding=encoding, data=data, headers=headers, query=query,
892 expected_status=expected_status)
895 xml_string, urlh = res
896 return self._parse_xml(
897 xml_string, video_id, transform_source=transform_source,
901 self, url_or_request, video_id,
902 note='Downloading XML
', errnote='Unable to download XML
',
903 transform_source=None, fatal=True, encoding=None,
904 data=None, headers={}, query={}, expected_status=None):
906 Return the xml as an compat_etree_Element.
908 See _download_webpage docstring for arguments specification.
910 res = self._download_xml_handle(
911 url_or_request, video_id, note=note, errnote=errnote,
912 transform_source=transform_source, fatal=fatal, encoding=encoding,
913 data=data, headers=headers, query=query,
914 expected_status=expected_status)
915 return res if res is False else res[0]
917 def _parse_xml(self, xml_string, video_id, transform_source=None, fatal=True):
919 xml_string = transform_source(xml_string)
921 return compat_etree_fromstring(xml_string.encode('utf
-8'))
922 except compat_xml_parse_error as ve:
923 errmsg = '%s: Failed to parse XML
' % video_id
925 raise ExtractorError(errmsg, cause=ve)
927 self.report_warning(errmsg + str(ve))
929 def _download_json_handle(
930 self, url_or_request, video_id, note='Downloading JSON metadata
',
931 errnote='Unable to download JSON metadata
', transform_source=None,
932 fatal=True, encoding=None, data=None, headers={}, query={},
933 expected_status=None):
935 Return a tuple (JSON object, URL handle).
937 See _download_webpage docstring for arguments specification.
939 res = self._download_webpage_handle(
940 url_or_request, video_id, note, errnote, fatal=fatal,
941 encoding=encoding, data=data, headers=headers, query=query,
942 expected_status=expected_status)
945 json_string, urlh = res
946 return self._parse_json(
947 json_string, video_id, transform_source=transform_source,
951 self, url_or_request, video_id, note='Downloading JSON metadata
',
952 errnote='Unable to download JSON metadata
', transform_source=None,
953 fatal=True, encoding=None, data=None, headers={}, query={},
954 expected_status=None):
956 Return the JSON object as a dict.
958 See _download_webpage docstring for arguments specification.
960 res = self._download_json_handle(
961 url_or_request, video_id, note=note, errnote=errnote,
962 transform_source=transform_source, fatal=fatal, encoding=encoding,
963 data=data, headers=headers, query=query,
964 expected_status=expected_status)
965 return res if res is False else res[0]
967 def _parse_json(self, json_string, video_id, transform_source=None, fatal=True):
969 json_string = transform_source(json_string)
971 return json.loads(json_string)
972 except ValueError as ve:
973 errmsg = '%s: Failed to parse JSON
' % video_id
975 raise ExtractorError(errmsg, cause=ve)
977 self.report_warning(errmsg + str(ve))
979 def _parse_socket_response_as_json(self, data, video_id, transform_source=None, fatal=True):
980 return self._parse_json(
981 data[data.find('{'):data.rfind('}
') + 1],
982 video_id, transform_source, fatal)
984 def _download_socket_json_handle(
985 self, url_or_request, video_id, note='Polling socket
',
986 errnote='Unable to poll socket
', transform_source=None,
987 fatal=True, encoding=None, data=None, headers={}, query={},
988 expected_status=None):
990 Return a tuple (JSON object, URL handle).
992 See _download_webpage docstring for arguments specification.
994 res = self._download_webpage_handle(
995 url_or_request, video_id, note, errnote, fatal=fatal,
996 encoding=encoding, data=data, headers=headers, query=query,
997 expected_status=expected_status)
1001 return self._parse_socket_response_as_json(
1002 webpage, video_id, transform_source=transform_source,
1005 def _download_socket_json(
1006 self, url_or_request, video_id, note='Polling socket
',
1007 errnote='Unable to poll socket
', transform_source=None,
1008 fatal=True, encoding=None, data=None, headers={}, query={},
1009 expected_status=None):
1011 Return the JSON object as a dict.
1013 See _download_webpage docstring for arguments specification.
1015 res = self._download_socket_json_handle(
1016 url_or_request, video_id, note=note, errnote=errnote,
1017 transform_source=transform_source, fatal=fatal, encoding=encoding,
1018 data=data, headers=headers, query=query,
1019 expected_status=expected_status)
1020 return res if res is False else res[0]
1022 def report_warning(self, msg, video_id=None, *args, only_once=False, **kwargs):
1023 idstr = format_field(video_id, template='%s: ')
1024 msg = f'[{self.IE_NAME}
] {idstr}{msg}
'
1026 if f'WARNING
: {msg}
' in self._printed_messages:
1028 self._printed_messages.add(f'WARNING
: {msg}
')
1029 self._downloader.report_warning(msg, *args, **kwargs)
1031 def to_screen(self, msg, *args, **kwargs):
1032 """Print msg to screen, prefixing it with '[ie_name
]'"""
1033 self._downloader.to_screen('[%s] %s' % (self.IE_NAME, msg), *args, **kwargs)
1035 def write_debug(self, msg, *args, **kwargs):
1036 self._downloader.write_debug('[%s] %s' % (self.IE_NAME, msg), *args, **kwargs)
1038 def get_param(self, name, default=None, *args, **kwargs):
1039 if self._downloader:
1040 return self._downloader.params.get(name, default, *args, **kwargs)
1043 def report_drm(self, video_id, partial=False):
1044 self.raise_no_formats('This video
is DRM protected
', expected=True, video_id=video_id)
1046 def report_extraction(self, id_or_name):
1047 """Report information extraction."""
1048 self.to_screen('%s: Extracting information
' % id_or_name)
1050 def report_download_webpage(self, video_id):
1051 """Report webpage download."""
1052 self.to_screen('%s: Downloading webpage
' % video_id)
1054 def report_age_confirmation(self):
1055 """Report attempt to confirm age."""
1056 self.to_screen('Confirming age
')
1058 def report_login(self):
1059 """Report attempt to log in."""
1060 self.to_screen('Logging
in')
1062 def raise_login_required(
1063 self, msg='This video
is only available
for registered users
',
1064 metadata_available=False, method='any
'):
1065 if metadata_available and self.get_param('ignore_no_formats_error
'):
1066 self.report_warning(msg)
1067 if method is not None:
1068 msg = '%s. %s' % (msg, self._LOGIN_HINTS[method])
1069 raise ExtractorError(msg, expected=True)
1071 def raise_geo_restricted(
1072 self, msg='This video
is not available
from your location due to geo restriction
',
1073 countries=None, metadata_available=False):
1074 if metadata_available and self.get_param('ignore_no_formats_error
'):
1075 self.report_warning(msg)
1077 raise GeoRestrictedError(msg, countries=countries)
1079 def raise_no_formats(self, msg, expected=False, video_id=None):
1080 if expected and self.get_param('ignore_no_formats_error
'):
1081 self.report_warning(msg, video_id)
1082 elif isinstance(msg, ExtractorError):
1085 raise ExtractorError(msg, expected=expected, video_id=video_id)
1087 # Methods for following #608
1089 def url_result(url, ie=None, video_id=None, video_title=None):
1090 """Returns a URL that points to a page that should be processed"""
1091 # TODO: ie should be the class used for getting the info
1092 video_info = {'_type
': 'url
',
1095 if video_id is not None:
1096 video_info['id'] = video_id
1097 if video_title is not None:
1098 video_info['title
'] = video_title
1101 def playlist_from_matches(self, matches, playlist_id=None, playlist_title=None, getter=None, ie=None):
1103 self.url_result(self._proto_relative_url(getter(m) if getter else m), ie)
1105 return self.playlist_result(
1106 urls, playlist_id=playlist_id, playlist_title=playlist_title)
1109 def playlist_result(entries, playlist_id=None, playlist_title=None, playlist_description=None, **kwargs):
1110 """Returns a playlist"""
1111 video_info = {'_type
': 'playlist
',
1113 video_info.update(kwargs)
1115 video_info['id'] = playlist_id
1117 video_info['title
'] = playlist_title
1118 if playlist_description is not None:
1119 video_info['description
'] = playlist_description
1122 def _search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
1124 Perform a regex search on the given string, using a single or a list of
1125 patterns returning the first matching group.
1126 In case of failure return a default value or raise a WARNING or a
1127 RegexNotFoundError, depending on fatal, specifying the field name.
1129 if isinstance(pattern, (str, compat_str, compiled_regex_type)):
1130 mobj = re.search(pattern, string, flags)
1133 mobj = re.search(p, string, flags)
1137 _name = self._downloader._color_text(name, 'blue
')
1141 # return the first matching group
1142 return next(g for g in mobj.groups() if g is not None)
1143 elif isinstance(group, (list, tuple)):
1144 return tuple(mobj.group(g) for g in group)
1146 return mobj.group(group)
1147 elif default is not NO_DEFAULT:
1150 raise RegexNotFoundError('Unable to extract
%s' % _name)
1152 self.report_warning('unable to extract
%s' % _name + bug_reports_message())
1155 def _html_search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
1157 Like _search_regex, but strips HTML tags and unescapes entities.
1159 res = self._search_regex(pattern, string, name, default, fatal, flags, group)
1161 return clean_html(res).strip()
1165 def _get_netrc_login_info(self, netrc_machine=None):
1168 netrc_machine = netrc_machine or self._NETRC_MACHINE
1170 if self.get_param('usenetrc
', False):
1172 netrc_file = compat_expanduser(self.get_param('netrc_location
') or '~
')
1173 if os.path.isdir(netrc_file):
1174 netrc_file = os.path.join(netrc_file, '.netrc
')
1175 info = netrc.netrc(file=netrc_file).authenticators(netrc_machine)
1176 if info is not None:
1180 raise netrc.NetrcParseError(
1181 'No authenticators
for %s' % netrc_machine)
1182 except (IOError, netrc.NetrcParseError) as err:
1183 self.report_warning(
1184 'parsing
.netrc
: %s' % error_to_compat_str(err))
1186 return username, password
1188 def _get_login_info(self, username_option='username
', password_option='password
', netrc_machine=None):
1190 Get the login info as (username, password)
1191 First look for the manually specified credentials using username_option
1192 and password_option as keys in params dictionary. If no such credentials
1193 available look in the netrc file using the netrc_machine or _NETRC_MACHINE
1195 If there's no info available
, return (None, None)
1198 # Attempt to use provided username and password or .netrc data
1199 username = self.get_param(username_option)
1200 if username is not None:
1201 password = self.get_param(password_option)
1203 username, password = self._get_netrc_login_info(netrc_machine)
1205 return username, password
1207 def _get_tfa_info(self, note='two-factor verification code'):
1209 Get the two
-factor authentication info
1210 TODO
- asking the user will be required
for sms
/phone verify
1211 currently just uses the command line option
1212 If there
's no info available, return None
1215 tfa = self.get_param('twofactor
')
1219 return compat_getpass('Type
%s and press
[Return
]: ' % note)
1221 # Helper functions for extracting OpenGraph info
1223 def _og_regexes(prop):
1224 content_re = r'content
=(?
:"([^"]+?
)"|\'([^\']+?)\'|\s*([^\s"\'=<>`
]+?
))'
1225 property_re = (r'(?
:name|
property)=(?
:\'og
[:-]%(prop)s\'|
"og[:-]%(prop)s"|\s
*og
[:-]%(prop)s\b)'
1226 % {'prop': re.escape(prop)})
1227 template = r'<meta
[^
>]+?
%s[^
>]+?
%s'
1229 template % (property_re, content_re),
1230 template % (content_re, property_re),
1234 def _meta_regex(prop):
1235 return r'''(?isx)<meta
1236 (?=[^>]+(?:itemprop|name|property|id|http-equiv)=(["\']?)%s\1)
1237 [^>]+?content=(["\'])(?P<content>.*?)\2''' % re.escape(prop)
1239 def _og_search_property(self, prop, html, name=None, **kargs):
1240 prop = variadic(prop)
1242 name = 'OpenGraph
%s' % prop[0]
1245 og_regexes.extend(self._og_regexes(p))
1246 escaped = self._search_regex(og_regexes, html, name, flags=re.DOTALL, **kargs)
1249 return unescapeHTML(escaped)
1251 def _og_search_thumbnail(self, html, **kargs):
1252 return self._og_search_property('image
', html, 'thumbnail URL
', fatal=False, **kargs)
1254 def _og_search_description(self, html, **kargs):
1255 return self._og_search_property('description
', html, fatal=False, **kargs)
1257 def _og_search_title(self, html, **kargs):
1258 return self._og_search_property('title
', html, **kargs)
1260 def _og_search_video_url(self, html, name='video url
', secure=True, **kargs):
1261 regexes = self._og_regexes('video
') + self._og_regexes('video
:url
')
1263 regexes = self._og_regexes('video
:secure_url
') + regexes
1264 return self._html_search_regex(regexes, html, name, **kargs)
1266 def _og_search_url(self, html, **kargs):
1267 return self._og_search_property('url
', html, **kargs)
1269 def _html_search_meta(self, name, html, display_name=None, fatal=False, **kwargs):
1270 name = variadic(name)
1271 if display_name is None:
1272 display_name = name[0]
1273 return self._html_search_regex(
1274 [self._meta_regex(n) for n in name],
1275 html, display_name, fatal=fatal, group='content
', **kwargs)
1277 def _dc_search_uploader(self, html):
1278 return self._html_search_meta('dc
.creator
', html, 'uploader
')
1280 def _rta_search(self, html):
1281 # See http://www.rtalabel.org/index.php?content=howtofaq#single
1282 if re.search(r'(?ix
)<meta\s
+name
="rating"\s
+'
1283 r' content
="RTA-5042-1996-1400-1577-RTA"',
1288 def _media_rating_search(self, html):
1289 # See http://www.tjg-designs.com/WP/metadata-code-examples-adding-metadata-to-your-web-pages/
1290 rating = self._html_search_meta('rating
', html)
1302 return RATING_TABLE.get(rating.lower())
1304 def _family_friendly_search(self, html):
1305 # See http://schema.org/VideoObject
1306 family_friendly = self._html_search_meta(
1307 'isFamilyFriendly
', html, default=None)
1309 if not family_friendly:
1318 return RATING_TABLE.get(family_friendly.lower())
1320 def _twitter_search_player(self, html):
1321 return self._html_search_meta('twitter
:player
', html,
1322 'twitter card player
')
1324 def _search_json_ld(self, html, video_id, expected_type=None, **kwargs):
1325 json_ld_list = list(re.finditer(JSON_LD_RE, html))
1326 default = kwargs.get('default
', NO_DEFAULT)
1327 # JSON-LD may be malformed and thus `fatal` should be respected.
1328 # At the same time `default` may be passed that assumes `fatal=False`
1329 # for _search_regex. Let's simulate the same behavior here
as well
.
1330 fatal
= kwargs
.get('fatal', True) if default
is NO_DEFAULT
else False
1332 for mobj
in json_ld_list
:
1333 json_ld_item
= self
._parse
_json
(
1334 mobj
.group('json_ld'), video_id
, fatal
=fatal
)
1335 if not json_ld_item
:
1337 if isinstance(json_ld_item
, dict):
1338 json_ld
.append(json_ld_item
)
1339 elif isinstance(json_ld_item
, (list, tuple)):
1340 json_ld
.extend(json_ld_item
)
1342 json_ld
= self
._json
_ld
(json_ld
, video_id
, fatal
=fatal
, expected_type
=expected_type
)
1345 if default
is not NO_DEFAULT
:
1348 raise RegexNotFoundError('Unable to extract JSON-LD')
1350 self
.report_warning('unable to extract JSON-LD %s' % bug_reports_message())
1353 def _json_ld(self
, json_ld
, video_id
, fatal
=True, expected_type
=None):
1354 if isinstance(json_ld
, compat_str
):
1355 json_ld
= self
._parse
_json
(json_ld
, video_id
, fatal
=fatal
)
1359 if not isinstance(json_ld
, (list, tuple, dict)):
1361 if isinstance(json_ld
, dict):
1364 INTERACTION_TYPE_MAP
= {
1365 'CommentAction': 'comment',
1366 'AgreeAction': 'like',
1367 'DisagreeAction': 'dislike',
1368 'LikeAction': 'like',
1369 'DislikeAction': 'dislike',
1370 'ListenAction': 'view',
1371 'WatchAction': 'view',
1372 'ViewAction': 'view',
1375 def extract_interaction_type(e
):
1376 interaction_type
= e
.get('interactionType')
1377 if isinstance(interaction_type
, dict):
1378 interaction_type
= interaction_type
.get('@type')
1379 return str_or_none(interaction_type
)
1381 def extract_interaction_statistic(e
):
1382 interaction_statistic
= e
.get('interactionStatistic')
1383 if isinstance(interaction_statistic
, dict):
1384 interaction_statistic
= [interaction_statistic
]
1385 if not isinstance(interaction_statistic
, list):
1387 for is_e
in interaction_statistic
:
1388 if not isinstance(is_e
, dict):
1390 if is_e
.get('@type') != 'InteractionCounter':
1392 interaction_type
= extract_interaction_type(is_e
)
1393 if not interaction_type
:
1395 # For interaction count some sites provide string instead of
1396 # an integer (as per spec) with non digit characters (e.g. ",")
1397 # so extracting count with more relaxed str_to_int
1398 interaction_count
= str_to_int(is_e
.get('userInteractionCount'))
1399 if interaction_count
is None:
1401 count_kind
= INTERACTION_TYPE_MAP
.get(interaction_type
.split('/')[-1])
1404 count_key
= '%s_count' % count_kind
1405 if info
.get(count_key
) is not None:
1407 info
[count_key
] = interaction_count
1409 def extract_video_object(e
):
1410 assert e
['@type'] == 'VideoObject'
1411 author
= e
.get('author')
1413 'url': url_or_none(e
.get('contentUrl')),
1414 'title': unescapeHTML(e
.get('name')),
1415 'description': unescapeHTML(e
.get('description')),
1416 'thumbnail': url_or_none(e
.get('thumbnailUrl') or e
.get('thumbnailURL')),
1417 'duration': parse_duration(e
.get('duration')),
1418 'timestamp': unified_timestamp(e
.get('uploadDate')),
1419 # author can be an instance of 'Organization' or 'Person' types.
1420 # both types can have 'name' property(inherited from 'Thing' type). [1]
1421 # however some websites are using 'Text' type instead.
1422 # 1. https://schema.org/VideoObject
1423 'uploader': author
.get('name') if isinstance(author
, dict) else author
if isinstance(author
, compat_str
) else None,
1424 'filesize': float_or_none(e
.get('contentSize')),
1425 'tbr': int_or_none(e
.get('bitrate')),
1426 'width': int_or_none(e
.get('width')),
1427 'height': int_or_none(e
.get('height')),
1428 'view_count': int_or_none(e
.get('interactionCount')),
1430 extract_interaction_statistic(e
)
1434 item_type
= e
.get('@type')
1435 if expected_type
is not None and expected_type
!= item_type
:
1437 if item_type
in ('TVEpisode', 'Episode'):
1438 episode_name
= unescapeHTML(e
.get('name'))
1440 'episode': episode_name
,
1441 'episode_number': int_or_none(e
.get('episodeNumber')),
1442 'description': unescapeHTML(e
.get('description')),
1444 if not info
.get('title') and episode_name
:
1445 info
['title'] = episode_name
1446 part_of_season
= e
.get('partOfSeason')
1447 if isinstance(part_of_season
, dict) and part_of_season
.get('@type') in ('TVSeason', 'Season', 'CreativeWorkSeason'):
1449 'season': unescapeHTML(part_of_season
.get('name')),
1450 'season_number': int_or_none(part_of_season
.get('seasonNumber')),
1452 part_of_series
= e
.get('partOfSeries') or e
.get('partOfTVSeries')
1453 if isinstance(part_of_series
, dict) and part_of_series
.get('@type') in ('TVSeries', 'Series', 'CreativeWorkSeries'):
1454 info
['series'] = unescapeHTML(part_of_series
.get('name'))
1455 elif item_type
== 'Movie':
1457 'title': unescapeHTML(e
.get('name')),
1458 'description': unescapeHTML(e
.get('description')),
1459 'duration': parse_duration(e
.get('duration')),
1460 'timestamp': unified_timestamp(e
.get('dateCreated')),
1462 elif item_type
in ('Article', 'NewsArticle'):
1464 'timestamp': parse_iso8601(e
.get('datePublished')),
1465 'title': unescapeHTML(e
.get('headline')),
1466 'description': unescapeHTML(e
.get('articleBody')),
1468 elif item_type
== 'VideoObject':
1469 extract_video_object(e
)
1470 if expected_type
is None:
1474 video
= e
.get('video')
1475 if isinstance(video
, dict) and video
.get('@type') == 'VideoObject':
1476 extract_video_object(video
)
1477 if expected_type
is None:
1481 return dict((k
, v
) for k
, v
in info
.items() if v
is not None)
1484 def _hidden_inputs(html
):
1485 html
= re
.sub(r
'<!--(?:(?!<!--).)*-->', '', html
)
1487 for input in re
.findall(r
'(?i)(<input[^>]+>)', html
):
1488 attrs
= extract_attributes(input)
1491 if attrs
.get('type') not in ('hidden', 'submit'):
1493 name
= attrs
.get('name') or attrs
.get('id')
1494 value
= attrs
.get('value')
1495 if name
and value
is not None:
1496 hidden_inputs
[name
] = value
1497 return hidden_inputs
1499 def _form_hidden_inputs(self
, form_id
, html
):
1500 form
= self
._search
_regex
(
1501 r
'(?is)<form[^>]+?id=(["\'])%s\
1[^
>]*>(?P
<form
>.+?
)</form
>' % form_id,
1502 html, '%s form
' % form_id, group='form
')
1503 return self._hidden_inputs(form)
1506 regex = r' *((?P
<reverse
>\
+)?
(?P
<field
>[a
-zA
-Z0
-9_]+)((?P
<separator
>[~
:])(?P
<limit
>.*?
))?
)?
*$
'
1508 default = ('hidden
', 'aud_or_vid
', 'hasvid
', 'ie_pref
', 'lang
', 'quality
',
1509 'res
', 'fps
', 'codec
:vp9
.2
', 'size
', 'br
', 'asr
',
1510 'proto
', 'ext
', 'hasaud
', 'source
', 'format_id
') # These must not be aliases
1511 ytdl_default = ('hasaud
', 'lang
', 'quality
', 'tbr
', 'filesize
', 'vbr
',
1512 'height
', 'width
', 'proto
', 'vext
', 'abr
', 'aext
',
1513 'fps
', 'fs_approx
', 'source
', 'format_id
')
1516 'vcodec
': {'type': 'ordered
', 'regex
': True,
1517 'order
': ['av0?
1', 'vp0?
9.2', 'vp0?
9', '[hx
]265|he?vc?
', '[hx
]264|avc
', 'vp0?
8', 'mp4v|h263
', 'theora
', '', None, 'none
']},
1518 'acodec
': {'type': 'ordered
', 'regex
': True,
1519 'order
': ['opus
', 'vorbis
', 'aac
', 'mp?
4a?
', 'mp3
', 'e?a?c
-?
3', 'dts
', '', None, 'none
']},
1520 'proto
': {'type': 'ordered
', 'regex
': True, 'field
': 'protocol
',
1521 'order
': ['(ht|f
)tps
', '(ht|f
)tp$
', 'm3u8
.+', '.*dash
', 'ws|websocket
', '', 'mms|rtsp
', 'none
', 'f4
']},
1522 'vext
': {'type': 'ordered
', 'field
': 'video_ext
',
1523 'order
': ('mp4
', 'webm
', 'flv
', '', 'none
'),
1524 'order_free
': ('webm
', 'mp4
', 'flv
', '', 'none
')},
1525 'aext
': {'type': 'ordered
', 'field
': 'audio_ext
',
1526 'order
': ('m4a
', 'aac
', 'mp3
', 'ogg
', 'opus
', 'webm
', '', 'none
'),
1527 'order_free
': ('opus
', 'ogg
', 'webm
', 'm4a
', 'mp3
', 'aac
', '', 'none
')},
1528 'hidden
': {'visible': False, 'forced': True, 'type': 'extractor', 'max': -1000},
1529 'aud_or_vid
': {'visible
': False, 'forced
': True, 'type': 'multiple
',
1530 'field
': ('vcodec
', 'acodec
'),
1531 'function
': lambda it: int(any(v != 'none
' for v in it))},
1532 'ie_pref
': {'priority': True, 'type': 'extractor'},
1533 'hasvid
': {'priority': True, 'field': 'vcodec', 'type': 'boolean', 'not_in_list': ('none',)},
1534 'hasaud
': {'field': 'acodec', 'type': 'boolean', 'not_in_list': ('none',)},
1535 'lang
': {'convert': 'ignore', 'field': 'language_preference'},
1536 'quality
': {'convert': 'float_none', 'default': -1},
1537 'filesize
': {'convert': 'bytes'},
1538 'fs_approx
': {'convert': 'bytes', 'field': 'filesize_approx'},
1539 'id': {'convert': 'string', 'field': 'format_id'},
1540 'height
': {'convert': 'float_none'},
1541 'width
': {'convert': 'float_none'},
1542 'fps
': {'convert': 'float_none'},
1543 'tbr
': {'convert': 'float_none'},
1544 'vbr
': {'convert': 'float_none'},
1545 'abr
': {'convert': 'float_none'},
1546 'asr
': {'convert': 'float_none'},
1547 'source
': {'convert': 'ignore', 'field': 'source_preference'},
1549 'codec
': {'type': 'combined', 'field': ('vcodec', 'acodec')},
1550 'br
': {'type': 'combined', 'field': ('tbr', 'vbr', 'abr'), 'same_limit': True},
1551 'size
': {'type': 'combined', 'same_limit': True, 'field': ('filesize', 'fs_approx')},
1552 'ext
': {'type': 'combined', 'field': ('vext', 'aext')},
1553 'res
': {'type': 'multiple
', 'field
': ('height
', 'width
'),
1554 'function
': lambda it: (lambda l: min(l) if l else 0)(tuple(filter(None, it)))},
1556 # Most of these exist only for compatibility reasons
1557 'dimension
': {'type': 'alias', 'field': 'res'},
1558 'resolution
': {'type': 'alias', 'field': 'res'},
1559 'extension
': {'type': 'alias', 'field': 'ext'},
1560 'bitrate
': {'type': 'alias', 'field': 'br'},
1561 'total_bitrate
': {'type': 'alias', 'field': 'tbr'},
1562 'video_bitrate
': {'type': 'alias', 'field': 'vbr'},
1563 'audio_bitrate
': {'type': 'alias', 'field': 'abr'},
1564 'framerate
': {'type': 'alias', 'field': 'fps'},
1565 'language_preference
': {'type': 'alias', 'field': 'lang'}, # not named as 'language
' because such a field exists
1566 'protocol
': {'type': 'alias', 'field': 'proto'},
1567 'source_preference
': {'type': 'alias', 'field': 'source'},
1568 'filesize_approx
': {'type': 'alias', 'field': 'fs_approx'},
1569 'filesize_estimate
': {'type': 'alias', 'field': 'size'},
1570 'samplerate
': {'type': 'alias', 'field': 'asr'},
1571 'video_ext
': {'type': 'alias', 'field': 'vext'},
1572 'audio_ext
': {'type': 'alias', 'field': 'aext'},
1573 'video_codec
': {'type': 'alias', 'field': 'vcodec'},
1574 'audio_codec
': {'type': 'alias', 'field': 'acodec'},
1575 'video
': {'type': 'alias', 'field': 'hasvid'},
1576 'has_video
': {'type': 'alias', 'field': 'hasvid'},
1577 'audio
': {'type': 'alias', 'field': 'hasaud'},
1578 'has_audio
': {'type': 'alias', 'field': 'hasaud'},
1579 'extractor
': {'type': 'alias', 'field': 'ie_pref'},
1580 'preference
': {'type': 'alias', 'field': 'ie_pref'},
1581 'extractor_preference
': {'type': 'alias', 'field': 'ie_pref'},
1582 'format_id
': {'type': 'alias', 'field': 'id'},
1587 def _get_field_setting(self, field, key):
1588 if field not in self.settings:
1589 self.settings[field] = {}
1590 propObj = self.settings[field]
1591 if key not in propObj:
1592 type = propObj.get('type')
1594 default = 'preference
' if type == 'extractor
' else (field,) if type in ('combined
', 'multiple
') else field
1595 elif key == 'convert
':
1596 default = 'order
' if type == 'ordered
' else 'float_string
' if field else 'ignore
'
1598 default = {'type': 'field', 'visible': True, 'order': [], 'not_in_list': (None,)}.get(key, None)
1599 propObj[key] = default
1602 def _resolve_field_value(self, field, value, convertNone=False):
1607 value = value.lower()
1608 conversion = self._get_field_setting(field, 'convert
')
1609 if conversion == 'ignore
':
1611 if conversion == 'string
':
1613 elif conversion == 'float_none
':
1614 return float_or_none(value)
1615 elif conversion == 'bytes':
1616 return FileDownloader.parse_bytes(value)
1617 elif conversion == 'order
':
1618 order_list = (self._use_free_order and self._get_field_setting(field, 'order_free
')) or self._get_field_setting(field, 'order
')
1619 use_regex = self._get_field_setting(field, 'regex
')
1620 list_length = len(order_list)
1621 empty_pos = order_list.index('') if '' in order_list else list_length + 1
1622 if use_regex and value is not None:
1623 for i, regex in enumerate(order_list):
1624 if regex and re.match(regex, value):
1625 return list_length - i
1626 return list_length - empty_pos # not in list
1627 else: # not regex or value = None
1628 return list_length - (order_list.index(value) if value in order_list else empty_pos)
1630 if value.isnumeric():
1633 self.settings[field]['convert
'] = 'string
'
1636 def evaluate_params(self, params, sort_extractor):
1637 self._use_free_order = params.get('prefer_free_formats
', False)
1638 self._sort_user = params.get('format_sort
', [])
1639 self._sort_extractor = sort_extractor
1641 def add_item(field, reverse, closest, limit_text):
1642 field = field.lower()
1643 if field in self._order:
1645 self._order.append(field)
1646 limit = self._resolve_field_value(field, limit_text)
1649 'closest
': False if limit is None else closest,
1650 'limit_text
': limit_text,
1652 if field in self.settings:
1653 self.settings[field].update(data)
1655 self.settings[field] = data
1658 tuple(field for field in self.default if self._get_field_setting(field, 'forced
'))
1659 + (tuple() if params.get('format_sort_force
', False)
1660 else tuple(field for field in self.default if self._get_field_setting(field, 'priority
')))
1661 + tuple(self._sort_user) + tuple(sort_extractor) + self.default)
1663 for item in sort_list:
1664 match = re.match(self.regex, item)
1666 raise ExtractorError('Invalid format sort string
"%s" given by extractor
' % item)
1667 field = match.group('field
')
1670 if self._get_field_setting(field, 'type') == 'alias
':
1671 field = self._get_field_setting(field, 'field
')
1672 reverse = match.group('reverse
') is not None
1673 closest = match.group('separator
') == '~
'
1674 limit_text = match.group('limit
')
1676 has_limit = limit_text is not None
1677 has_multiple_fields = self._get_field_setting(field, 'type') == 'combined
'
1678 has_multiple_limits = has_limit and has_multiple_fields and not self._get_field_setting(field, 'same_limit
')
1680 fields = self._get_field_setting(field, 'field
') if has_multiple_fields else (field,)
1681 limits = limit_text.split(":") if has_multiple_limits else (limit_text,) if has_limit else tuple()
1682 limit_count = len(limits)
1683 for (i, f) in enumerate(fields):
1684 add_item(f, reverse, closest,
1685 limits[i] if i < limit_count
1686 else limits[0] if has_limit and not has_multiple_limits
1689 def print_verbose_info(self, write_debug):
1691 write_debug('Sort order given by user
: %s' % ', '.join(self._sort_user))
1692 if self._sort_extractor:
1693 write_debug('Sort order given by extractor
: %s' % ', '.join(self._sort_extractor))
1694 write_debug('Formats
sorted by
: %s' % ', '.join(['%s%s%s' % (
1695 '+' if self._get_field_setting(field, 'reverse
') else '', field,
1696 '%s%s(%s)' % ('~
' if self._get_field_setting(field, 'closest
') else ':',
1697 self._get_field_setting(field, 'limit_text
'),
1698 self._get_field_setting(field, 'limit
'))
1699 if self._get_field_setting(field, 'limit_text
') is not None else '')
1700 for field in self._order if self._get_field_setting(field, 'visible
')]))
1702 def _calculate_field_preference_from_value(self, format, field, type, value):
1703 reverse = self._get_field_setting(field, 'reverse
')
1704 closest = self._get_field_setting(field, 'closest
')
1705 limit = self._get_field_setting(field, 'limit
')
1707 if type == 'extractor
':
1708 maximum = self._get_field_setting(field, 'max')
1709 if value is None or (maximum is not None and value >= maximum):
1711 elif type == 'boolean
':
1712 in_list = self._get_field_setting(field, 'in_list
')
1713 not_in_list = self._get_field_setting(field, 'not_in_list
')
1714 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
1715 elif type == 'ordered
':
1716 value = self._resolve_field_value(field, value, True)
1718 # try to convert to number
1719 val_num = float_or_none(value, default=self._get_field_setting(field, 'default
'))
1720 is_num = self._get_field_setting(field, 'convert
') != 'string
' and val_num is not None
1724 return ((-10, 0) if value is None
1725 else (1, value, 0) if not is_num # if a field has mixed strings and numbers, strings are sorted higher
1726 else (0, -abs(value - limit), value - limit if reverse else limit - value) if closest
1727 else (0, value, 0) if not reverse and (limit is None or value <= limit)
1728 else (0, -value, 0) if limit is None or (reverse and value == limit) or value > limit
1729 else (-1, value, 0))
1731 def _calculate_field_preference(self, format, field):
1732 type = self._get_field_setting(field, 'type') # extractor, boolean, ordered, field, multiple
1733 get_value = lambda f: format.get(self._get_field_setting(f, 'field
'))
1734 if type == 'multiple
':
1735 type = 'field
' # Only 'field
' is allowed in multiple for now
1736 actual_fields = self._get_field_setting(field, 'field
')
1738 value = self._get_field_setting(field, 'function
')(get_value(f) for f in actual_fields)
1740 value = get_value(field)
1741 return self._calculate_field_preference_from_value(format, field, type, value)
1743 def calculate_preference(self, format):
1744 # Determine missing protocol
1745 if not format.get('protocol
'):
1746 format['protocol
'] = determine_protocol(format)
1748 # Determine missing ext
1749 if not format.get('ext
') and 'url
' in format:
1750 format['ext
'] = determine_ext(format['url
'])
1751 if format.get('vcodec
') == 'none
':
1752 format['audio_ext
'] = format['ext
'] if format.get('acodec
') != 'none
' else 'none
'
1753 format['video_ext
'] = 'none
'
1755 format['video_ext
'] = format['ext
']
1756 format['audio_ext
'] = 'none
'
1757 # if format.get('preference
') is None and format.get('ext
') in ('f4f
', 'f4m
'): # Not supported?
1758 # format['preference
'] = -1000
1760 # Determine missing bitrates
1761 if format.get('tbr
') is None:
1762 if format.get('vbr
') is not None and format.get('abr
') is not None:
1763 format['tbr
'] = format.get('vbr
', 0) + format.get('abr
', 0)
1765 if format.get('vcodec
') != "none" and format.get('vbr
') is None:
1766 format['vbr
'] = format.get('tbr
') - format.get('abr
', 0)
1767 if format.get('acodec
') != "none" and format.get('abr
') is None:
1768 format['abr
'] = format.get('tbr
') - format.get('vbr
', 0)
1770 return tuple(self._calculate_field_preference(format, field) for field in self._order)
1772 def _sort_formats(self, formats, field_preference=[]):
1775 format_sort = self.FormatSort() # params and to_screen are taken from the downloader
1776 format_sort.evaluate_params(self._downloader.params, field_preference)
1777 if self.get_param('verbose
', False):
1778 format_sort.print_verbose_info(self._downloader.write_debug)
1779 formats.sort(key=lambda f: format_sort.calculate_preference(f))
1781 def _check_formats(self, formats, video_id):
1783 formats[:] = filter(
1784 lambda f: self._is_valid_url(
1786 item='%s video format
' % f.get('format_id
') if f.get('format_id
') else 'video
'),
1790 def _remove_duplicate_formats(formats):
1794 if f['url
'] not in format_urls:
1795 format_urls.add(f['url
'])
1796 unique_formats.append(f)
1797 formats[:] = unique_formats
1799 def _is_valid_url(self, url, video_id, item='video
', headers={}):
1800 url = self._proto_relative_url(url, scheme='http
:')
1801 # For now assume non HTTP(S) URLs always valid
1802 if not (url.startswith('http
://') or url.startswith('https
://')):
1805 self._request_webpage(url, video_id, 'Checking
%s URL
' % item, headers=headers)
1807 except ExtractorError as e:
1809 '%s: %s URL
is invalid
, skipping
: %s'
1810 % (video_id, item, error_to_compat_str(e.cause)))
1813 def http_scheme(self):
1814 """ Either "http:" or "https:", depending on the user's preferences
"""
1817 if self.get_param('prefer_insecure', False)
1820 def _proto_relative_url(self, url, scheme=None):
1823 if url.startswith('//'):
1825 scheme = self.http_scheme()
1830 def _sleep(self, timeout, video_id, msg_template=None):
1831 if msg_template is None:
1832 msg_template = '%(video_id)s: Waiting for %(timeout)s seconds'
1833 msg = msg_template % {'video_id': video_id, 'timeout': timeout}
1837 def _extract_f4m_formats(self, manifest_url, video_id, preference=None, quality=None, f4m_id=None,
1838 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1839 fatal=True, m3u8_id=None, data=None, headers={}, query={}):
1840 manifest = self._download_xml(
1841 manifest_url, video_id, 'Downloading f4m manifest',
1842 'Unable to download f4m manifest',
1843 # Some manifests may be malformed, e.g. prosiebensat1 generated manifests
1844 # (see https://github.com/ytdl-org/youtube-dl/issues/6215#issuecomment-121704244)
1845 transform_source=transform_source,
1846 fatal=fatal, data=data, headers=headers, query=query)
1848 if manifest is False:
1851 return self._parse_f4m_formats(
1852 manifest, manifest_url, video_id, preference=preference, quality=quality, f4m_id=f4m_id,
1853 transform_source=transform_source, fatal=fatal, m3u8_id=m3u8_id)
1855 def _parse_f4m_formats(self, manifest, manifest_url, video_id, preference=None, quality=None, f4m_id=None,
1856 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1857 fatal=True, m3u8_id=None):
1858 if not isinstance(manifest, compat_etree_Element) and not fatal:
1861 # currently yt-dlp cannot decode the playerVerificationChallenge as Akamai uses Adobe Alchemy
1862 akamai_pv = manifest.find('{http://ns.adobe.com/f4m/1.0}pv-2.0')
1863 if akamai_pv is not None and ';' in akamai_pv.text:
1864 playerVerificationChallenge = akamai_pv.text.split(';')[0]
1865 if playerVerificationChallenge.strip() != '':
1869 manifest_version = '1.0'
1870 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/1.0}media')
1872 manifest_version = '2.0'
1873 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/2.0}media')
1874 # Remove unsupported DRM protected media from final formats
1875 # rendition (see https://github.com/ytdl-org/youtube-dl/issues/8573).
1876 media_nodes = remove_encrypted_media(media_nodes)
1880 manifest_base_url = get_base_url(manifest)
1882 bootstrap_info = xpath_element(
1883 manifest, ['{http://ns.adobe.com/f4m/1.0}bootstrapInfo', '{http://ns.adobe.com/f4m/2.0}bootstrapInfo'],
1884 'bootstrap info', default=None)
1887 mime_type = xpath_text(
1888 manifest, ['{http://ns.adobe.com/f4m/1.0}mimeType', '{http://ns.adobe.com/f4m/2.0}mimeType'],
1889 'base URL', default=None)
1890 if mime_type and mime_type.startswith('audio/'):
1893 for i, media_el in enumerate(media_nodes):
1894 tbr = int_or_none(media_el.attrib.get('bitrate'))
1895 width = int_or_none(media_el.attrib.get('width'))
1896 height = int_or_none(media_el.attrib.get('height'))
1897 format_id = '-'.join(filter(None, [f4m_id, compat_str(i if tbr is None else tbr)]))
1898 # If <bootstrapInfo> is present, the specified f4m is a
1899 # stream-level manifest, and only set-level manifests may refer to
1900 # external resources. See section 11.4 and section 4 of F4M spec
1901 if bootstrap_info is None:
1903 # @href is introduced in 2.0, see section 11.6 of F4M spec
1904 if manifest_version == '2.0':
1905 media_url = media_el.attrib.get('href')
1906 if media_url is None:
1907 media_url = media_el.attrib.get('url')
1911 media_url if media_url.startswith('http://') or media_url.startswith('https://')
1912 else ((manifest_base_url or '/'.join(manifest_url.split('/')[:-1])) + '/' + media_url))
1913 # If media_url is itself a f4m manifest do the recursive extraction
1914 # since bitrates in parent manifest (this one) and media_url manifest
1915 # may differ leading to inability to resolve the format by requested
1916 # bitrate in f4m downloader
1917 ext = determine_ext(manifest_url)
1919 f4m_formats = self._extract_f4m_formats(
1920 manifest_url, video_id, preference=preference, quality=quality, f4m_id=f4m_id,
1921 transform_source=transform_source, fatal=fatal)
1922 # Sometimes stream-level manifest contains single media entry that
1923 # does not contain any quality metadata (e.g. http://matchtv.ru/#live-player).
1924 # At the same time parent's media entry in set-level manifest may
1925 # contain it. We will copy it from parent in such cases.
1926 if len(f4m_formats) == 1:
1929 'tbr': f.get('tbr') or tbr,
1930 'width': f.get('width') or width,
1931 'height': f.get('height') or height,
1932 'format_id': f.get('format_id') if not tbr else format_id,
1935 formats.extend(f4m_formats)
1938 formats.extend(self._extract_m3u8_formats(
1939 manifest_url, video_id, 'mp4', preference=preference,
1940 quality=quality, m3u8_id=m3u8_id, fatal=fatal))
1943 'format_id': format_id,
1944 'url': manifest_url,
1945 'manifest_url': manifest_url,
1946 'ext': 'flv' if bootstrap_info is not None else None,
1952 'preference': preference,
1957 def _m3u8_meta_format(self, m3u8_url, ext=None, preference=None, quality=None, m3u8_id=None):
1959 'format_id': '-'.join(filter(None, [m3u8_id, 'meta'])),
1963 'preference': preference - 100 if preference else -100,
1965 'resolution': 'multiple',
1966 'format_note': 'Quality selection URL',
1969 def _extract_m3u8_formats(self, *args, **kwargs):
1970 fmts, subs = self._extract_m3u8_formats_and_subtitles(*args, **kwargs)
1972 self.report_warning(bug_reports_message(
1973 "Ignoring subtitle tracks found in the HLS manifest; "
1974 "if any subtitle tracks are missing,"
1978 def _extract_m3u8_formats_and_subtitles(
1979 self, m3u8_url, video_id, ext=None, entry_protocol='m3u8_native',
1980 preference=None, quality=None, m3u8_id=None, note=None,
1981 errnote=None, fatal=True, live=False, data=None, headers={},
1984 res = self._download_webpage_handle(
1986 note='Downloading m3u8 information' if note is None else note,
1987 errnote='Failed to download m3u8 information' if errnote is None else errnote,
1988 fatal=fatal, data=data, headers=headers, query=query)
1993 m3u8_doc, urlh = res
1994 m3u8_url = urlh.geturl()
1996 return self._parse_m3u8_formats_and_subtitles(
1997 m3u8_doc, m3u8_url, ext=ext, entry_protocol=entry_protocol,
1998 preference=preference, quality=quality, m3u8_id=m3u8_id,
1999 note=note, errnote=errnote, fatal=fatal, live=live, data=data,
2000 headers=headers, query=query, video_id=video_id)
2002 def _parse_m3u8_formats_and_subtitles(
2003 self, m3u8_doc, m3u8_url, ext=None, entry_protocol='m3u8_native',
2004 preference=None, quality=None, m3u8_id=None, live=False, note=None,
2005 errnote=None, fatal=True, data=None, headers={}, query={},
2007 formats, subtitles = [], {}
2009 if '#EXT-X-FAXS-CM:' in m3u8_doc: # Adobe Flash Access
2010 return formats, subtitles
2012 has_drm = re.search(r'#EXT-X-SESSION-KEY:.*?URI="skd://', m3u8_doc)
2014 def format_url(url):
2015 return url if re.match(r'^https?://', url) else compat_urlparse.urljoin(m3u8_url, url)
2017 if self.get_param('hls_split_discontinuity', False):
2018 def _extract_m3u8_playlist_indices(manifest_url=None, m3u8_doc=None):
2020 if not manifest_url:
2022 m3u8_doc = self._download_webpage(
2023 manifest_url, video_id, fatal=fatal, data=data, headers=headers,
2024 note=False, errnote='Failed to download m3u8 playlist information')
2025 if m3u8_doc is False:
2027 return range(1 + sum(line.startswith('#EXT-X-DISCONTINUITY') for line in m3u8_doc.splitlines()))
2030 def _extract_m3u8_playlist_indices(*args, **kwargs):
2034 # 1. https://tools.ietf.org/html/draft-pantos-http-live-streaming-21
2035 # 2. https://github.com/ytdl-org/youtube-dl/issues/12211
2036 # 3. https://github.com/ytdl-org/youtube-dl/issues/18923
2038 # We should try extracting formats only from master playlists [1, 4.3.4],
2039 # i.e. playlists that describe available qualities. On the other hand
2040 # media playlists [1, 4.3.3] should be returned as is since they contain
2041 # just the media without qualities renditions.
2042 # Fortunately, master playlist can be easily distinguished from media
2043 # playlist based on particular tags availability. As of [1, 4.3.3, 4.3.4]
2044 # master playlist tags MUST NOT appear in a media playlist and vice versa.
2045 # As of [1, 4.3.3.1] #EXT-X-TARGETDURATION tag is REQUIRED for every
2046 # media playlist and MUST NOT appear in master playlist thus we can
2047 # clearly detect media playlist with this criterion.
2049 if '#EXT-X-TARGETDURATION' in m3u8_doc: # media playlist, return as is
2051 'format_id': '-'.join(map(str, filter(None, [m3u8_id, idx]))),
2052 'format_index': idx,
2055 'protocol': entry_protocol,
2056 'preference': preference,
2059 } for idx in _extract_m3u8_playlist_indices(m3u8_doc=m3u8_doc)]
2061 return formats, subtitles
2064 last_stream_inf = {}
2066 def extract_media(x_media_line):
2067 media = parse_m3u8_attributes(x_media_line)
2068 # As per [1, 4.3.4.1] TYPE, GROUP-ID and NAME are REQUIRED
2069 media_type, group_id, name = media.get('TYPE'), media.get('GROUP-ID'), media.get('NAME')
2070 if not (media_type and group_id and name):
2072 groups.setdefault(group_id, []).append(media)
2073 # <https://tools.ietf.org/html/rfc8216#section-4.3.4.1>
2074 if media_type == 'SUBTITLES':
2075 # According to RFC 8216 §4.3.4.2.1, URI is REQUIRED in the
2076 # EXT-X-MEDIA tag if the media type is SUBTITLES.
2077 # However, lack of URI has been spotted in the wild.
2078 # e.g. NebulaIE; see https://github.com/yt-dlp/yt-dlp/issues/339
2079 if not media.get('URI'):
2081 url = format_url(media['URI'])
2084 'ext': determine_ext(url),
2086 if sub_info['ext'] == 'm3u8':
2087 # Per RFC 8216 §3.1, the only possible subtitle format m3u8
2088 # files may contain is WebVTT:
2089 # <https://tools.ietf.org/html/rfc8216#section-3.1>
2090 sub_info['ext'] = 'vtt'
2091 sub_info['protocol'] = 'm3u8_native'
2092 lang = media.get('LANGUAGE') or 'und'
2093 subtitles.setdefault(lang, []).append(sub_info)
2094 if media_type not in ('VIDEO', 'AUDIO'):
2096 media_url = media.get('URI')
2098 manifest_url = format_url(media_url)
2100 'format_id': '-'.join(map(str, filter(None, (m3u8_id, group_id, name, idx)))),
2101 'format_note': name,
2102 'format_index': idx,
2103 'url': manifest_url,
2104 'manifest_url': m3u8_url,
2105 'language': media.get('LANGUAGE'),
2107 'protocol': entry_protocol,
2108 'preference': preference,
2110 'vcodec': 'none' if media_type == 'AUDIO' else None,
2111 } for idx in _extract_m3u8_playlist_indices(manifest_url))
2113 def build_stream_name():
2114 # Despite specification does not mention NAME attribute for
2115 # EXT-X-STREAM-INF tag it still sometimes may be present (see [1]
2116 # or vidio test in TestInfoExtractor.test_parse_m3u8_formats)
2117 # 1. http://www.vidio.com/watch/165683-dj_ambred-booyah-live-2015
2118 stream_name = last_stream_inf.get('NAME')
2121 # If there is no NAME in EXT-X-STREAM-INF it will be obtained
2122 # from corresponding rendition group
2123 stream_group_id = last_stream_inf.get('VIDEO')
2124 if not stream_group_id:
2126 stream_group = groups.get(stream_group_id)
2127 if not stream_group:
2128 return stream_group_id
2129 rendition = stream_group[0]
2130 return rendition.get('NAME') or stream_group_id
2132 # parse EXT-X-MEDIA tags before EXT-X-STREAM-INF in order to have the
2133 # chance to detect video only formats when EXT-X-STREAM-INF tags
2134 # precede EXT-X-MEDIA tags in HLS manifest such as [3].
2135 for line in m3u8_doc.splitlines():
2136 if line.startswith('#EXT-X-MEDIA:'):
2139 for line in m3u8_doc.splitlines():
2140 if line.startswith('#EXT-X-STREAM-INF:'):
2141 last_stream_inf = parse_m3u8_attributes(line)
2142 elif line.startswith('#') or not line.strip():
2145 tbr = float_or_none(
2146 last_stream_inf.get('AVERAGE-BANDWIDTH')
2147 or last_stream_inf.get('BANDWIDTH'), scale=1000)
2148 manifest_url = format_url(line.strip())
2150 for idx in _extract_m3u8_playlist_indices(manifest_url):
2151 format_id = [m3u8_id, None, idx]
2152 # Bandwidth of live streams may differ over time thus making
2153 # format_id unpredictable. So it's better to keep provided
2156 stream_name = build_stream_name()
2157 format_id[1] = stream_name if stream_name else '%d' % (tbr if tbr else len(formats))
2159 'format_id': '-'.join(map(str, filter(None, format_id))),
2160 'format_index': idx,
2161 'url': manifest_url,
2162 'manifest_url': m3u8_url,
2165 'fps': float_or_none(last_stream_inf.get('FRAME-RATE')),
2166 'protocol': entry_protocol,
2167 'preference': preference,
2170 resolution = last_stream_inf.get('RESOLUTION')
2172 mobj = re.search(r'(?P<width>\d+)[xX](?P<height>\d+)', resolution)
2174 f['width'] = int(mobj.group('width'))
2175 f['height'] = int(mobj.group('height'))
2176 # Unified Streaming Platform
2178 r'audio.*?(?:%3D|=)(\d+)(?:-video.*?(?:%3D|=)(\d+))?', f['url'])
2180 abr, vbr = mobj.groups()
2181 abr, vbr = float_or_none(abr, 1000), float_or_none(vbr, 1000)
2186 codecs = parse_codecs(last_stream_inf.get('CODECS'))
2188 audio_group_id = last_stream_inf.get('AUDIO')
2189 # As per [1, 4.3.4.1.1] any EXT-X-STREAM-INF tag which
2190 # references a rendition group MUST have a CODECS attribute.
2191 # However, this is not always respected, for example, [2]
2192 # contains EXT-X-STREAM-INF tag which references AUDIO
2193 # rendition group but does not have CODECS and despite
2194 # referencing an audio group it represents a complete
2195 # (with audio and video) format. So, for such cases we will
2196 # ignore references to rendition groups and treat them
2197 # as complete formats.
2198 if audio_group_id and codecs and f.get('vcodec') != 'none':
2199 audio_group = groups.get(audio_group_id)
2200 if audio_group and audio_group[0].get('URI'):
2201 # TODO: update acodec for audio only formats with
2203 f['acodec'] = 'none'
2204 if not f.get('ext'):
2205 f['ext'] = 'm4a' if f.get('vcodec') == 'none' else 'mp4'
2209 progressive_uri = last_stream_inf.get('PROGRESSIVE-URI')
2212 del http_f['manifest_url']
2214 'format_id': f['format_id'].replace('hls-', 'http-'),
2216 'url': progressive_uri,
2218 formats.append(http_f)
2220 last_stream_inf = {}
2221 return formats, subtitles
2223 def _extract_m3u8_vod_duration(
2224 self, m3u8_vod_url, video_id, note=None, errnote=None, data=None, headers={}, query={}):
2226 m3u8_vod = self._download_webpage(
2227 m3u8_vod_url, video_id,
2228 note='Downloading m3u8 VOD manifest' if note is None else note,
2229 errnote='Failed to download VOD manifest' if errnote is None else errnote,
2230 fatal=False, data=data, headers=headers, query=query)
2232 return self._parse_m3u8_vod_duration(m3u8_vod or '', video_id)
2234 def _parse_m3u8_vod_duration(self, m3u8_vod, video_id):
2235 if '#EXT-X-PLAYLIST-TYPE:VOD' not in m3u8_vod:
2239 float(line[len('#EXTINF:'):].split(',')[0])
2240 for line in m3u8_vod.splitlines() if line.startswith('#EXTINF:'))) or None
2243 def _xpath_ns(path, namespace=None):
2247 for c in path.split('/'):
2248 if not c or c == '.':
2251 out.append('{%s}%s' % (namespace, c))
2252 return '/'.join(out)
2254 def _extract_smil_formats_and_subtitles(self, smil_url, video_id, fatal=True, f4m_params=None, transform_source=None):
2255 smil = self._download_smil(smil_url, video_id, fatal=fatal, transform_source=transform_source)
2261 namespace = self._parse_smil_namespace(smil)
2263 fmts = self._parse_smil_formats(
2264 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
2265 subs = self._parse_smil_subtitles(
2266 smil, namespace=namespace)
2270 def _extract_smil_formats(self, *args, **kwargs):
2271 fmts, subs = self._extract_smil_formats_and_subtitles(*args, **kwargs)
2273 self.report_warning(bug_reports_message(
2274 "Ignoring subtitle tracks found in the SMIL manifest; "
2275 "if any subtitle tracks are missing,"
2279 def _extract_smil_info(self, smil_url, video_id, fatal=True, f4m_params=None):
2280 smil = self._download_smil(smil_url, video_id, fatal=fatal)
2283 return self._parse_smil(smil, smil_url, video_id, f4m_params=f4m_params)
2285 def _download_smil(self, smil_url, video_id, fatal=True, transform_source=None):
2286 return self._download_xml(
2287 smil_url, video_id, 'Downloading SMIL file',
2288 'Unable to download SMIL file', fatal=fatal, transform_source=transform_source)
2290 def _parse_smil(self, smil, smil_url, video_id, f4m_params=None):
2291 namespace = self._parse_smil_namespace(smil)
2293 formats = self._parse_smil_formats(
2294 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
2295 subtitles = self._parse_smil_subtitles(smil, namespace=namespace)
2297 video_id = os.path.splitext(url_basename(smil_url))[0]
2301 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
2302 name = meta.attrib.get('name')
2303 content = meta.attrib.get('content')
2304 if not name or not content:
2306 if not title and name == 'title':
2308 elif not description and name in ('description', 'abstract'):
2309 description = content
2310 elif not upload_date and name == 'date':
2311 upload_date = unified_strdate(content)
2314 'id': image.get('type'),
2315 'url': image.get('src'),
2316 'width': int_or_none(image.get('width')),
2317 'height': int_or_none(image.get('height')),
2318 } for image in smil.findall(self._xpath_ns('.//image', namespace)) if image.get('src')]
2322 'title': title or video_id,
2323 'description': description,
2324 'upload_date': upload_date,
2325 'thumbnails': thumbnails,
2327 'subtitles': subtitles,
2330 def _parse_smil_namespace(self, smil):
2331 return self._search_regex(
2332 r'(?i)^{([^}]+)?}smil$', smil.tag, 'namespace', default=None)
2334 def _parse_smil_formats(self, smil, smil_url, video_id, namespace=None, f4m_params=None, transform_rtmp_url=None):
2336 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
2337 b = meta.get('base') or meta.get('httpBase')
2349 media = smil.findall(self._xpath_ns('.//video', namespace)) + smil.findall(self._xpath_ns('.//audio', namespace))
2350 for medium in media:
2351 src = medium.get('src')
2352 if not src or src in srcs:
2356 bitrate = float_or_none(medium.get('system-bitrate') or medium.get('systemBitrate'), 1000)
2357 filesize = int_or_none(medium.get('size') or medium.get('fileSize'))
2358 width = int_or_none(medium.get('width'))
2359 height = int_or_none(medium.get('height'))
2360 proto = medium.get('proto')
2361 ext = medium.get('ext')
2362 src_ext = determine_ext(src)
2363 streamer = medium.get('streamer') or base
2365 if proto == 'rtmp' or streamer.startswith('rtmp'):
2371 'format_id': 'rtmp-%d' % (rtmp_count if bitrate is None else bitrate),
2373 'filesize': filesize,
2377 if transform_rtmp_url:
2378 streamer, src = transform_rtmp_url(streamer, src)
2379 formats[-1].update({
2385 src_url = src if src.startswith('http') else compat_urlparse.urljoin(base, src)
2386 src_url = src_url.strip()
2388 if proto == 'm3u8' or src_ext == 'm3u8':
2389 m3u8_formats = self._extract_m3u8_formats(
2390 src_url, video_id, ext or 'mp4', m3u8_id='hls', fatal=False)
2391 if len(m3u8_formats) == 1:
2393 m3u8_formats[0].update({
2394 'format_id': 'hls-%d' % (m3u8_count if bitrate is None else bitrate),
2399 formats.extend(m3u8_formats)
2400 elif src_ext == 'f4m':
2405 'plugin': 'flowplayer-3.2.0.1',
2407 f4m_url += '&' if '?' in f4m_url else '?'
2408 f4m_url += compat_urllib_parse_urlencode(f4m_params)
2409 formats.extend(self._extract_f4m_formats(f4m_url, video_id, f4m_id='hds', fatal=False))
2410 elif src_ext == 'mpd':
2411 formats.extend(self._extract_mpd_formats(
2412 src_url, video_id, mpd_id='dash', fatal=False))
2413 elif re.search(r'\.ism/[Mm]anifest', src_url):
2414 formats.extend(self._extract_ism_formats(
2415 src_url, video_id, ism_id='mss', fatal=False))
2416 elif src_url.startswith('http') and self._is_valid_url(src, video_id):
2420 'ext': ext or src_ext or 'flv',
2421 'format_id': 'http-%d' % (bitrate or http_count),
2423 'filesize': filesize,
2428 for medium in smil.findall(self._xpath_ns('.//imagestream', namespace)):
2429 src = medium.get('src')
2430 if not src or src in srcs:
2436 'format_id': 'imagestream-%d' % (imgs_count),
2438 'ext': mimetype2ext(medium.get('type')),
2441 'width': int_or_none(medium.get('width')),
2442 'height': int_or_none(medium.get('height')),
2443 'format_note': 'SMIL storyboards',
2448 def _parse_smil_subtitles(self, smil, namespace=None, subtitles_lang='en'):
2451 for num, textstream in enumerate(smil.findall(self._xpath_ns('.//textstream', namespace))):
2452 src = textstream.get('src')
2453 if not src or src in urls:
2456 ext = textstream.get('ext') or mimetype2ext(textstream.get('type')) or determine_ext(src)
2457 lang = textstream.get('systemLanguage') or textstream.get('systemLanguageName') or textstream.get('lang') or subtitles_lang
2458 subtitles.setdefault(lang, []).append({
2464 def _extract_xspf_playlist(self, xspf_url, playlist_id, fatal=True):
2465 xspf = self._download_xml(
2466 xspf_url, playlist_id, 'Downloading xpsf playlist',
2467 'Unable to download xspf manifest', fatal=fatal)
2470 return self._parse_xspf(
2471 xspf, playlist_id, xspf_url=xspf_url,
2472 xspf_base_url=base_url(xspf_url))
2474 def _parse_xspf(self, xspf_doc, playlist_id, xspf_url=None, xspf_base_url=None):
2476 'xspf': 'http://xspf.org/ns/0/',
2477 's1': 'http://static.streamone.nl/player/ns/0',
2481 for track in xspf_doc.findall(xpath_with_ns('./xspf:trackList/xspf:track', NS_MAP)):
2483 track, xpath_with_ns('./xspf:title', NS_MAP), 'title', default=playlist_id)
2484 description = xpath_text(
2485 track, xpath_with_ns('./xspf:annotation', NS_MAP), 'description')
2486 thumbnail = xpath_text(
2487 track, xpath_with_ns('./xspf:image', NS_MAP), 'thumbnail')
2488 duration = float_or_none(
2489 xpath_text(track, xpath_with_ns('./xspf:duration', NS_MAP), 'duration'), 1000)
2492 for location in track.findall(xpath_with_ns('./xspf:location', NS_MAP)):
2493 format_url = urljoin(xspf_base_url, location.text)
2498 'manifest_url': xspf_url,
2499 'format_id': location.get(xpath_with_ns('s1:label', NS_MAP)),
2500 'width': int_or_none(location.get(xpath_with_ns('s1:width', NS_MAP))),
2501 'height': int_or_none(location.get(xpath_with_ns('s1:height', NS_MAP))),
2503 self._sort_formats(formats)
2508 'description': description,
2509 'thumbnail': thumbnail,
2510 'duration': duration,
2515 def _extract_mpd_formats(self, *args, **kwargs):
2516 fmts, subs = self._extract_mpd_formats_and_subtitles(*args, **kwargs)
2518 self.report_warning(bug_reports_message(
2519 "Ignoring subtitle tracks found in the DASH manifest; "
2520 "if any subtitle tracks are missing,"
2524 def _extract_mpd_formats_and_subtitles(
2525 self, mpd_url, video_id, mpd_id=None, note=None, errnote=None,
2526 fatal=True, data=None, headers={}, query={}):
2527 res = self._download_xml_handle(
2529 note='Downloading MPD manifest' if note is None else note,
2530 errnote='Failed to download MPD manifest' if errnote is None else errnote,
2531 fatal=fatal, data=data, headers=headers, query=query)
2537 mpd_base_url = base_url(urlh.geturl())
2539 return self._parse_mpd_formats_and_subtitles(
2540 mpd_doc, mpd_id, mpd_base_url, mpd_url)
2542 def _parse_mpd_formats(self, *args, **kwargs):
2543 fmts, subs = self._parse_mpd_formats_and_subtitles(*args, **kwargs)
2545 self.report_warning(bug_reports_message(
2546 "Ignoring subtitle tracks found in the DASH manifest; "
2547 "if any subtitle tracks are missing,"
2551 def _parse_mpd_formats_and_subtitles(
2552 self, mpd_doc, mpd_id=None, mpd_base_url='', mpd_url=None):
2554 Parse formats
from MPD manifest
.
2556 1. MPEG
-DASH Standard
, ISO
/IEC
23009-1:2014(E
),
2557 http
://standards
.iso
.org
/ittf
/PubliclyAvailableStandards
/c065274_ISO_IEC_23009
-1_2014.zip
2558 2. https
://en
.wikipedia
.org
/wiki
/Dynamic_Adaptive_Streaming_over_HTTP
2560 if not self.get_param('dynamic_mpd', True):
2561 if mpd_doc.get('type') == 'dynamic':
2564 namespace = self._search_regex(r'(?i)^{([^}]+)?}MPD$', mpd_doc.tag, 'namespace', default=None)
2567 return self._xpath_ns(path, namespace)
2569 def is_drm_protected(element):
2570 return element.find(_add_ns('ContentProtection')) is not None
2572 def extract_multisegment_info(element, ms_parent_info):
2573 ms_info = ms_parent_info.copy()
2575 # As per [1, 5.3.9.2.2] SegmentList and SegmentTemplate share some
2576 # common attributes and elements. We will only extract relevant
2578 def extract_common(source):
2579 segment_timeline = source.find(_add_ns('SegmentTimeline'))
2580 if segment_timeline is not None:
2581 s_e = segment_timeline.findall(_add_ns('S'))
2583 ms_info['total_number'] = 0
2586 r = int(s.get('r', 0))
2587 ms_info['total_number'] += 1 + r
2588 ms_info['s'].append({
2589 't': int(s.get('t', 0)),
2590 # @d is mandatory (see [1, 5.3.9.6.2, Table 17, page 60])
2591 'd': int(s.attrib['d']),
2594 start_number = source.get('startNumber')
2596 ms_info['start_number'] = int(start_number)
2597 timescale = source.get('timescale')
2599 ms_info['timescale'] = int(timescale)
2600 segment_duration = source.get('duration')
2601 if segment_duration:
2602 ms_info['segment_duration'] = float(segment_duration)
2604 def extract_Initialization(source):
2605 initialization = source.find(_add_ns('Initialization'))
2606 if initialization is not None:
2607 ms_info['initialization_url'] = initialization.attrib['sourceURL']
2609 segment_list = element.find(_add_ns('SegmentList'))
2610 if segment_list is not None:
2611 extract_common(segment_list)
2612 extract_Initialization(segment_list)
2613 segment_urls_e = segment_list.findall(_add_ns('SegmentURL'))
2615 ms_info['segment_urls'] = [segment.attrib['media'] for segment in segment_urls_e]
2617 segment_template = element.find(_add_ns('SegmentTemplate'))
2618 if segment_template is not None:
2619 extract_common(segment_template)
2620 media = segment_template.get('media')
2622 ms_info['media'] = media
2623 initialization = segment_template.get('initialization')
2625 ms_info['initialization'] = initialization
2627 extract_Initialization(segment_template)
2630 mpd_duration = parse_duration(mpd_doc.get('mediaPresentationDuration'))
2631 formats, subtitles = [], {}
2632 stream_numbers = {'audio': 0, 'video': 0}
2633 for period in mpd_doc.findall(_add_ns('Period')):
2634 period_duration = parse_duration(period.get('duration')) or mpd_duration
2635 period_ms_info = extract_multisegment_info(period, {
2639 for adaptation_set in period.findall(_add_ns('AdaptationSet')):
2640 adaption_set_ms_info = extract_multisegment_info(adaptation_set, period_ms_info)
2641 for representation in adaptation_set.findall(_add_ns('Representation')):
2642 representation_attrib = adaptation_set.attrib.copy()
2643 representation_attrib.update(representation.attrib)
2644 # According to [1, 5.3.7.2, Table 9, page 41], @mimeType is mandatory
2645 mime_type = representation_attrib['mimeType']
2646 content_type = representation_attrib.get('contentType', mime_type.split('/')[0])
2648 codecs = representation_attrib.get('codecs', '')
2649 if content_type not in ('video', 'audio', 'text'):
2650 if mime_type == 'image/jpeg':
2651 content_type = mime_type
2652 elif codecs.split('.')[0] == 'stpp':
2653 content_type = 'text'
2655 self.report_warning('Unknown MIME type %s in DASH manifest' % mime_type)
2659 for element in (representation, adaptation_set, period, mpd_doc):
2660 base_url_e = element.find(_add_ns('BaseURL'))
2661 if base_url_e is not None:
2662 base_url = base_url_e.text + base_url
2663 if re.match(r'^https?://', base_url):
2665 if mpd_base_url and base_url.startswith('/'):
2666 base_url = compat_urlparse.urljoin(mpd_base_url, base_url)
2667 elif mpd_base_url and not re.match(r'^https?://', base_url):
2668 if not mpd_base_url.endswith('/'):
2670 base_url = mpd_base_url + base_url
2671 representation_id = representation_attrib.get('id')
2672 lang = representation_attrib.get('lang')
2673 url_el = representation.find(_add_ns('BaseURL'))
2674 filesize = int_or_none(url_el.attrib.get('{http://youtube.com/yt/2012/10/10}contentLength') if url_el is not None else None)
2675 bandwidth = int_or_none(representation_attrib.get('bandwidth'))
2676 if representation_id is not None:
2677 format_id = representation_id
2679 format_id = content_type
2681 format_id = mpd_id + '-' + format_id
2682 if content_type in ('video', 'audio'):
2684 'format_id': format_id,
2685 'manifest_url': mpd_url,
2686 'ext': mimetype2ext(mime_type),
2687 'width': int_or_none(representation_attrib.get('width')),
2688 'height': int_or_none(representation_attrib.get('height')),
2689 'tbr': float_or_none(bandwidth, 1000),
2690 'asr': int_or_none(representation_attrib.get('audioSamplingRate')),
2691 'fps': int_or_none(representation_attrib.get('frameRate')),
2692 'language': lang if lang not in ('mul', 'und', 'zxx', 'mis') else None,
2693 'format_note': 'DASH %s' % content_type,
2694 'filesize': filesize,
2695 'container': mimetype2ext(mime_type) + '_dash',
2696 'manifest_stream_number': stream_numbers[content_type]
2698 f.update(parse_codecs(codecs))
2699 stream_numbers[content_type] += 1
2700 elif content_type == 'text':
2702 'ext': mimetype2ext(mime_type),
2703 'manifest_url': mpd_url,
2704 'filesize': filesize,
2706 elif content_type == 'image/jpeg':
2707 # See test case in VikiIE
2708 # https://www.viki.com/videos/1175236v-choosing-spouse-by-lottery-episode-1
2710 'format_id': format_id,
2712 'manifest_url': mpd_url,
2713 'format_note': 'DASH storyboards (jpeg)',
2717 if is_drm_protected(adaptation_set) or is_drm_protected(representation):
2719 representation_ms_info = extract_multisegment_info(representation, adaption_set_ms_info)
2721 def prepare_template(template_name, identifiers):
2722 tmpl = representation_ms_info[template_name]
2723 # First of, % characters outside $...$ templates
2724 # must be escaped by doubling for proper processing
2725 # by % operator string formatting used further (see
2726 # https://github.com/ytdl-org/youtube-dl/issues/16867).
2732 in_template = not in_template
2733 elif c == '%' and not in_template:
2735 # Next, $...$ templates are translated to their
2736 # %(...) counterparts to be used with % operator
2737 if representation_id is not None:
2738 t = t.replace('$RepresentationID$', representation_id)
2739 t = re.sub(r'\$(%s)\$' % '|'.join(identifiers), r'%(\1)d', t)
2740 t = re.sub(r'\$(%s)%%([^$]+)\$' % '|'.join(identifiers), r'%(\1)\2', t)
2741 t.replace('$$', '$')
2744 # @initialization is a regular template like @media one
2745 # so it should be handled just the same way (see
2746 # https://github.com/ytdl-org/youtube-dl/issues/11605)
2747 if 'initialization' in representation_ms_info:
2748 initialization_template = prepare_template(
2750 # As per [1, 5.3.9.4.2, Table 15, page 54] $Number$ and
2751 # $Time$ shall not be included for @initialization thus
2752 # only $Bandwidth$ remains
2754 representation_ms_info['initialization_url'] = initialization_template % {
2755 'Bandwidth': bandwidth,
2758 def location_key(location):
2759 return 'url' if re.match(r'^https?://', location) else 'path'
2761 if 'segment_urls' not in representation_ms_info and 'media' in representation_ms_info:
2763 media_template = prepare_template('media', ('Number', 'Bandwidth', 'Time'))
2764 media_location_key = location_key(media_template)
2766 # As per [1, 5.3.9.4.4, Table 16, page 55] $Number$ and $Time$
2767 # can't be used at the same time
2768 if '%(Number' in media_template and 's' not in representation_ms_info:
2769 segment_duration = None
2770 if 'total_number' not in representation_ms_info and 'segment_duration' in representation_ms_info:
2771 segment_duration = float_or_none(representation_ms_info['segment_duration'], representation_ms_info['timescale'])
2772 representation_ms_info['total_number'] = int(math.ceil(float(period_duration) / segment_duration))
2773 representation_ms_info['fragments'] = [{
2774 media_location_key: media_template % {
2775 'Number': segment_number,
2776 'Bandwidth': bandwidth,
2778 'duration': segment_duration,
2779 } for segment_number in range(
2780 representation_ms_info['start_number'],
2781 representation_ms_info['total_number'] + representation_ms_info['start_number'])]
2783 # $Number*$ or $Time$ in media template with S list available
2784 # Example $Number*$: http://www.svtplay.se/klipp/9023742/stopptid-om-bjorn-borg
2785 # Example $Time$: https://play.arkena.com/embed/avp/v2/player/media/b41dda37-d8e7-4d3f-b1b5-9a9db578bdfe/1/129411
2786 representation_ms_info['fragments'] = []
2789 segment_number = representation_ms_info['start_number']
2791 def add_segment_url():
2792 segment_url = media_template % {
2793 'Time': segment_time,
2794 'Bandwidth': bandwidth,
2795 'Number': segment_number,
2797 representation_ms_info['fragments'].append({
2798 media_location_key: segment_url,
2799 'duration': float_or_none(segment_d, representation_ms_info['timescale']),
2802 for num, s in enumerate(representation_ms_info['s']):
2803 segment_time = s.get('t') or segment_time
2807 for r in range(s.get('r', 0)):
2808 segment_time += segment_d
2811 segment_time += segment_d
2812 elif 'segment_urls' in representation_ms_info and 's' in representation_ms_info:
2814 # Example: https://www.youtube.com/watch?v=iXZV5uAYMJI
2815 # or any YouTube dashsegments video
2818 timescale = representation_ms_info['timescale']
2819 for s in representation_ms_info['s']:
2820 duration = float_or_none(s['d'], timescale)
2821 for r in range(s.get('r', 0) + 1):
2822 segment_uri = representation_ms_info['segment_urls'][segment_index]
2824 location_key(segment_uri): segment_uri,
2825 'duration': duration,
2828 representation_ms_info['fragments'] = fragments
2829 elif 'segment_urls' in representation_ms_info:
2830 # Segment URLs with no SegmentTimeline
2831 # Example: https://www.seznam.cz/zpravy/clanek/cesko-zasahne-vitr-o-sile-vichrice-muze-byt-i-zivotu-nebezpecny-39091
2832 # https://github.com/ytdl-org/youtube-dl/pull/14844
2834 segment_duration = float_or_none(
2835 representation_ms_info['segment_duration'],
2836 representation_ms_info['timescale']) if 'segment_duration' in representation_ms_info else None
2837 for segment_url in representation_ms_info['segment_urls']:
2839 location_key(segment_url): segment_url,
2841 if segment_duration:
2842 fragment['duration'] = segment_duration
2843 fragments.append(fragment)
2844 representation_ms_info['fragments'] = fragments
2845 # If there is a fragments key available then we correctly recognized fragmented media.
2846 # Otherwise we will assume unfragmented media with direct access. Technically, such
2847 # assumption is not necessarily correct since we may simply have no support for
2848 # some forms of fragmented media renditions yet, but for now we'll use this fallback.
2849 if 'fragments' in representation_ms_info:
2851 # NB: mpd_url may be empty when MPD manifest is parsed from a string
2852 'url': mpd_url or base_url,
2853 'fragment_base_url': base_url,
2855 'protocol': 'http_dash_segments' if mime_type != 'image/jpeg' else 'mhtml',
2857 if 'initialization_url' in representation_ms_info:
2858 initialization_url = representation_ms_info['initialization_url']
2859 if not f.get('url'):
2860 f['url'] = initialization_url
2861 f['fragments'].append({location_key(initialization_url): initialization_url})
2862 f['fragments'].extend(representation_ms_info['fragments'])
2864 # Assuming direct URL to unfragmented media.
2866 if content_type in ('video', 'audio') or mime_type == 'image/jpeg':
2868 elif content_type == 'text':
2869 subtitles.setdefault(lang or 'und', []).append(f)
2871 return formats, subtitles
2873 def _extract_ism_formats(self, *args, **kwargs):
2874 fmts, subs = self._extract_ism_formats_and_subtitles(*args, **kwargs)
2876 self.report_warning(bug_reports_message(
2877 "Ignoring subtitle tracks found in the ISM manifest; "
2878 "if any subtitle tracks are missing,"
2882 def _extract_ism_formats_and_subtitles(self, ism_url, video_id, ism_id=None, note=None, errnote=None, fatal=True, data=None, headers={}, query={}):
2883 res = self._download_xml_handle(
2885 note='Downloading ISM manifest' if note is None else note,
2886 errnote='Failed to download ISM manifest' if errnote is None else errnote,
2887 fatal=fatal, data=data, headers=headers, query=query)
2894 return self._parse_ism_formats_and_subtitles(ism_doc, urlh.geturl(), ism_id)
2896 def _parse_ism_formats_and_subtitles(self, ism_doc, ism_url, ism_id=None):
2898 Parse formats
from ISM manifest
.
2900 1. [MS
-SSTR
]: Smooth Streaming Protocol
,
2901 https
://msdn
.microsoft
.com
/en
-us
/library
/ff469518
.aspx
2903 if ism_doc.get('IsLive') == 'TRUE':
2906 duration = int(ism_doc.attrib['Duration'])
2907 timescale = int_or_none(ism_doc.get('TimeScale')) or 10000000
2911 for stream in ism_doc.findall('StreamIndex'):
2912 stream_type = stream.get('Type')
2913 if stream_type not in ('video', 'audio', 'text'):
2915 url_pattern = stream.attrib['Url']
2916 stream_timescale = int_or_none(stream.get('TimeScale')) or timescale
2917 stream_name = stream.get('Name')
2918 stream_language = stream.get('Language', 'und')
2919 for track in stream.findall('QualityLevel'):
2920 fourcc = track.get('FourCC') or ('AACL' if track.get('AudioTag') == '255' else None)
2921 # TODO: add support for WVC1 and WMAP
2922 if fourcc not in ('H264', 'AVC1', 'AACL', 'TTML'):
2923 self.report_warning('%s is not a supported codec' % fourcc)
2925 tbr = int(track.attrib['Bitrate']) // 1000
2926 # [1] does not mention Width and Height attributes. However,
2927 # they're often present while MaxWidth and MaxHeight are
2928 # missing, so should be used as fallbacks
2929 width = int_or_none(track.get('MaxWidth') or track.get('Width'))
2930 height = int_or_none(track.get('MaxHeight') or track.get('Height'))
2931 sampling_rate = int_or_none(track.get('SamplingRate'))
2933 track_url_pattern = re.sub(r'{[Bb]itrate}', track.attrib['Bitrate'], url_pattern)
2934 track_url_pattern = compat_urlparse.urljoin(ism_url, track_url_pattern)
2940 stream_fragments = stream.findall('c')
2941 for stream_fragment_index, stream_fragment in enumerate(stream_fragments):
2942 fragment_ctx['time'] = int_or_none(stream_fragment.get('t')) or fragment_ctx['time']
2943 fragment_repeat = int_or_none(stream_fragment.get('r')) or 1
2944 fragment_ctx['duration'] = int_or_none(stream_fragment.get('d'))
2945 if not fragment_ctx['duration']:
2947 next_fragment_time = int(stream_fragment[stream_fragment_index + 1].attrib['t'])
2949 next_fragment_time = duration
2950 fragment_ctx['duration'] = (next_fragment_time - fragment_ctx['time']) / fragment_repeat
2951 for _ in range(fragment_repeat):
2953 'url': re.sub(r'{start[ _]time}', compat_str(fragment_ctx['time']), track_url_pattern),
2954 'duration': fragment_ctx['duration'] / stream_timescale,
2956 fragment_ctx['time'] += fragment_ctx['duration']
2960 format_id.append(ism_id)
2962 format_id.append(stream_name)
2963 format_id.append(compat_str(tbr))
2965 if stream_type == 'text':
2966 subtitles.setdefault(stream_language, []).append({
2970 'manifest_url': ism_url,
2971 'fragments': fragments,
2972 '_download_params': {
2973 'stream_type': stream_type,
2974 'duration': duration,
2975 'timescale': stream_timescale,
2977 'language': stream_language,
2978 'codec_private_data': track.get('CodecPrivateData'),
2981 elif stream_type in ('video', 'audio'):
2983 'format_id': '-'.join(format_id),
2985 'manifest_url': ism_url,
2986 'ext': 'ismv' if stream_type == 'video' else 'isma',
2990 'asr': sampling_rate,
2991 'vcodec': 'none' if stream_type == 'audio' else fourcc,
2992 'acodec': 'none' if stream_type == 'video' else fourcc,
2994 'fragments': fragments,
2995 'has_drm': ism_doc.find('Protection') is not None,
2996 '_download_params': {
2997 'stream_type': stream_type,
2998 'duration': duration,
2999 'timescale': stream_timescale,
3000 'width': width or 0,
3001 'height': height or 0,
3003 'language': stream_language,
3004 'codec_private_data': track.get('CodecPrivateData'),
3005 'sampling_rate': sampling_rate,
3006 'channels': int_or_none(track.get('Channels', 2)),
3007 'bits_per_sample': int_or_none(track.get('BitsPerSample', 16)),
3008 'nal_unit_length_field': int_or_none(track.get('NALUnitLengthField', 4)),
3011 return formats, subtitles
3013 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):
3014 def absolute_url(item_url):
3015 return urljoin(base_url, item_url)
3017 def parse_content_type(content_type):
3018 if not content_type:
3020 ctr = re.search(r'(?P<mimetype>[^/]+/[^;]+)(?:;\s*codecs="?(?P<codecs>[^"]+))?', content_type)
3022 mimetype, codecs = ctr.groups()
3023 f = parse_codecs(codecs)
3024 f['ext'] = mimetype2ext(mimetype)
3028 def _media_formats(src, cur_media_type, type_info={}):
3029 full_url = absolute_url(src)
3030 ext = type_info.get('ext') or determine_ext(full_url)
3032 is_plain_url = False
3033 formats = self._extract_m3u8_formats(
3034 full_url, video_id, ext='mp4',
3035 entry_protocol=m3u8_entry_protocol, m3u8_id=m3u8_id,
3036 preference=preference, quality=quality, fatal=False)
3038 is_plain_url = False
3039 formats = self._extract_mpd_formats(
3040 full_url, video_id, mpd_id=mpd_id, fatal=False)
3045 'vcodec': 'none' if cur_media_type == 'audio' else None,
3047 return is_plain_url, formats
3050 # amp-video and amp-audio are very similar to their HTML5 counterparts
3051 # so we wll include them right here (see
3052 # https://www.ampproject.org/docs/reference/components/amp-video)
3053 # For dl8-* tags see https://delight-vr.com/documentation/dl8-video/
3054 _MEDIA_TAG_NAME_RE = r'(?:(?:amp|dl8(?:-live)?)-)?(video|audio)'
3055 media_tags = [(media_tag, media_tag_name, media_type, '')
3056 for media_tag, media_tag_name, media_type
3057 in re.findall(r'(?s)(<(%s)[^>]*/>)' % _MEDIA_TAG_NAME_RE, webpage)]
3058 media_tags.extend(re.findall(
3059 # We only allow video|audio followed by a whitespace or '>'.
3060 # Allowing more characters may end up in significant slow down (see
3061 # https://github.com/ytdl-org/youtube-dl/issues/11979, example URL:
3062 # http://www.porntrex.com/maps/videositemap.xml).
3063 r'(?s)(<(?P<tag>%s)(?:\s+[^>]*)?>)(.*?)</(?P=tag)>' % _MEDIA_TAG_NAME_RE, webpage))
3064 for media_tag, _, media_type, media_content in media_tags:
3069 media_attributes = extract_attributes(media_tag)
3070 src = strip_or_none(media_attributes.get('src'))
3072 _, formats = _media_formats(src, media_type)
3073 media_info['formats'].extend(formats)
3074 media_info['thumbnail'] = absolute_url(media_attributes.get('poster'))
3076 for source_tag in re.findall(r'<source[^>]+>', media_content):
3077 s_attr = extract_attributes(source_tag)
3078 # data-video-src and data-src are non standard but seen
3079 # several times in the wild
3080 src = strip_or_none(dict_get(s_attr, ('src', 'data-video-src', 'data-src')))
3083 f = parse_content_type(s_attr.get('type'))
3084 is_plain_url, formats = _media_formats(src, media_type, f)
3086 # width, height, res, label and title attributes are
3087 # all not standard but seen several times in the wild
3090 for lbl in ('label', 'title')
3091 if str_or_none(s_attr.get(lbl))
3093 width = int_or_none(s_attr.get('width'))
3094 height = (int_or_none(s_attr.get('height'))
3095 or int_or_none(s_attr.get('res')))
3096 if not width or not height:
3098 resolution = parse_resolution(lbl)
3101 width = width or resolution.get('width')
3102 height = height or resolution.get('height')
3104 tbr = parse_bitrate(lbl)
3113 'format_id': s_attr.get('label') or s_attr.get('title'),
3115 f.update(formats[0])
3116 media_info['formats'].append(f)
3118 media_info['formats'].extend(formats)
3119 for track_tag in re.findall(r'<track[^>]+>', media_content):
3120 track_attributes = extract_attributes(track_tag)
3121 kind = track_attributes.get('kind')
3122 if not kind or kind in ('subtitles', 'captions'):
3123 src = strip_or_none(track_attributes.get('src'))
3126 lang = track_attributes.get('srclang') or track_attributes.get('lang') or track_attributes.get('label')
3127 media_info['subtitles'].setdefault(lang, []).append({
3128 'url': absolute_url(src),
3130 for f in media_info['formats']:
3131 f.setdefault('http_headers', {})['Referer'] = base_url
3132 if media_info['formats'] or media_info['subtitles']:
3133 entries.append(media_info)
3136 def _extract_akamai_formats(self, *args, **kwargs):
3137 fmts, subs = self._extract_akamai_formats_and_subtitles(*args, **kwargs)
3139 self.report_warning(bug_reports_message(
3140 "Ignoring subtitle tracks found in the manifests; "
3141 "if any subtitle tracks are missing,"
3145 def _extract_akamai_formats_and_subtitles(self, manifest_url, video_id, hosts={}):
3146 signed = 'hdnea=' in manifest_url
3148 # https://learn.akamai.com/en-us/webhelp/media-services-on-demand/stream-packaging-user-guide/GUID-BE6C0F73-1E06-483B-B0EA-57984B91B7F9.html
3149 manifest_url = re.sub(
3150 r'(?:b=[\d,-]+|(?:__a__|attributes)=off|__b__=\d+)&?',
3151 '', manifest_url).strip('?')
3156 hdcore_sign = 'hdcore=3.7.0'
3157 f4m_url = re.sub(r'(https?://[^/]+)/i/', r'\1/z/', manifest_url).replace('/master.m3u8', '/manifest.f4m')
3158 hds_host = hosts.get('hds')
3160 f4m_url = re.sub(r'(https?://)[^/]+', r'\1' + hds_host, f4m_url)
3161 if 'hdcore=' not in f4m_url:
3162 f4m_url += ('&' if '?' in f4m_url else '?') + hdcore_sign
3163 f4m_formats = self._extract_f4m_formats(
3164 f4m_url, video_id, f4m_id='hds', fatal=False)
3165 for entry in f4m_formats:
3166 entry.update({'extra_param_to_segment_url': hdcore_sign})
3167 formats.extend(f4m_formats)
3169 m3u8_url = re.sub(r'(https?://[^/]+)/z/', r'\1/i/', manifest_url).replace('/manifest.f4m', '/master.m3u8')
3170 hls_host = hosts.get('hls')
3172 m3u8_url = re.sub(r'(https?://)[^/]+', r'\1' + hls_host, m3u8_url)
3173 m3u8_formats, m3u8_subtitles = self._extract_m3u8_formats_and_subtitles(
3174 m3u8_url, video_id, 'mp4', 'm3u8_native',
3175 m3u8_id='hls', fatal=False)
3176 formats.extend(m3u8_formats)
3177 subtitles = self._merge_subtitles(subtitles, m3u8_subtitles)
3179 http_host = hosts.get('http')
3180 if http_host and m3u8_formats and not signed:
3181 REPL_REGEX = r'https?://[^/]+/i/([^,]+),([^/]+),([^/]+)\.csmil/.+'
3182 qualities = re.match(REPL_REGEX, m3u8_url).group(2).split(',')
3183 qualities_length = len(qualities)
3184 if len(m3u8_formats) in (qualities_length, qualities_length + 1):
3186 for f in m3u8_formats:
3187 if f['vcodec'] != 'none':
3188 for protocol in ('http', 'https'):
3190 del http_f['manifest_url']
3192 REPL_REGEX, protocol + r'://%s/\g<1>%s\3' % (http_host, qualities[i]), f['url'])
3194 'format_id': http_f['format_id'].replace('hls-', protocol + '-'),
3196 'protocol': protocol,
3198 formats.append(http_f)
3201 return formats, subtitles
3203 def _extract_wowza_formats(self, url, video_id, m3u8_entry_protocol='m3u8_native', skip_protocols=[]):
3204 query = compat_urlparse.urlparse(url).query
3205 url = re.sub(r'/(?:manifest|playlist|jwplayer)\.(?:m3u8|f4m|mpd|smil)', '', url)
3207 r'(?:(?:http|rtmp|rtsp)(?P<s>s)?:)?(?P<url>//[^?]+)', url)
3208 url_base = mobj.group('url')
3209 http_base_url = '%s%s:%s' % ('http', mobj.group('s') or '', url_base)
3212 def manifest_url(manifest):
3213 m_url = '%s/%s' % (http_base_url, manifest)
3215 m_url += '?%s' % query
3218 if 'm3u8' not in skip_protocols:
3219 formats.extend(self._extract_m3u8_formats(
3220 manifest_url('playlist.m3u8'), video_id, 'mp4',
3221 m3u8_entry_protocol, m3u8_id='hls', fatal=False))
3222 if 'f4m' not in skip_protocols:
3223 formats.extend(self._extract_f4m_formats(
3224 manifest_url('manifest.f4m'),
3225 video_id, f4m_id='hds', fatal=False))
3226 if 'dash' not in skip_protocols:
3227 formats.extend(self._extract_mpd_formats(
3228 manifest_url('manifest.mpd'),
3229 video_id, mpd_id='dash', fatal=False))
3230 if re.search(r'(?:/smil:|\.smil)', url_base):
3231 if 'smil' not in skip_protocols:
3232 rtmp_formats = self._extract_smil_formats(
3233 manifest_url('jwplayer.smil'),
3234 video_id, fatal=False)
3235 for rtmp_format in rtmp_formats:
3236 rtsp_format = rtmp_format.copy()
3237 rtsp_format['url'] = '%s/%s' % (rtmp_format['url'], rtmp_format['play_path'])
3238 del rtsp_format['play_path']
3239 del rtsp_format['ext']
3240 rtsp_format.update({
3241 'url': rtsp_format['url'].replace('rtmp://', 'rtsp://'),
3242 'format_id': rtmp_format['format_id'].replace('rtmp', 'rtsp'),
3245 formats.extend([rtmp_format, rtsp_format])
3247 for protocol in ('rtmp', 'rtsp'):
3248 if protocol not in skip_protocols:
3250 'url': '%s:%s' % (protocol, url_base),
3251 'format_id': protocol,
3252 'protocol': protocol,
3256 def _find_jwplayer_data(self, webpage, video_id=None, transform_source=js_to_json):
3258 r'(?s)jwplayer\((?P<quote>[\'"])[^\'" ]+(?P=quote)\)(?!</script>).*?\.setup\s*\((?P<options>[^)]+)\)',
3262 jwplayer_data = self._parse_json(mobj.group('options'),
3264 transform_source=transform_source)
3265 except ExtractorError:
3268 if isinstance(jwplayer_data, dict):
3269 return jwplayer_data
3271 def _extract_jwplayer_data(self, webpage, video_id, *args, **kwargs):
3272 jwplayer_data = self._find_jwplayer_data(
3273 webpage, video_id, transform_source=js_to_json)
3274 return self._parse_jwplayer_data(
3275 jwplayer_data, video_id, *args, **kwargs)
3277 def _parse_jwplayer_data(self, jwplayer_data, video_id=None, require_title=True,
3278 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
3279 # JWPlayer backward compatibility: flattened playlists
3280 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/api/config.js#L81-L96
3281 if 'playlist' not in jwplayer_data:
3282 jwplayer_data = {'playlist': [jwplayer_data]}
3286 # JWPlayer backward compatibility: single playlist item
3287 # https://github.com/jwplayer/jwplayer/blob/v7.7.0/src/js/playlist/playlist.js#L10
3288 if not isinstance(jwplayer_data['playlist'], list):
3289 jwplayer_data['playlist'] = [jwplayer_data['playlist']]
3291 for video_data in jwplayer_data['playlist']:
3292 # JWPlayer backward compatibility: flattened sources
3293 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/playlist/item.js#L29-L35
3294 if 'sources' not in video_data:
3295 video_data['sources'] = [video_data]
3297 this_video_id = video_id or video_data['mediaid']
3299 formats = self._parse_jwplayer_formats(
3300 video_data['sources'], video_id=this_video_id, m3u8_id=m3u8_id,
3301 mpd_id=mpd_id, rtmp_params=rtmp_params, base_url=base_url)
3304 tracks = video_data.get('tracks')
3305 if tracks and isinstance(tracks, list):
3306 for track in tracks:
3307 if not isinstance(track, dict):
3309 track_kind = track.get('kind')
3310 if not track_kind or not isinstance(track_kind, compat_str):
3312 if track_kind.lower() not in ('captions', 'subtitles'):
3314 track_url = urljoin(base_url, track.get('file'))
3317 subtitles.setdefault(track.get('label') or 'en', []).append({
3318 'url': self._proto_relative_url(track_url)
3322 'id': this_video_id,
3323 'title': unescapeHTML(video_data['title'] if require_title else video_data.get('title')),
3324 'description': clean_html(video_data.get('description')),
3325 'thumbnail': urljoin(base_url, self._proto_relative_url(video_data.get('image'))),
3326 'timestamp': int_or_none(video_data.get('pubdate')),
3327 'duration': float_or_none(jwplayer_data.get('duration') or video_data.get('duration')),
3328 'subtitles': subtitles,
3330 # https://github.com/jwplayer/jwplayer/blob/master/src/js/utils/validator.js#L32
3331 if len(formats) == 1 and re.search(r'^(?:http|//).*(?:youtube\.com|youtu\.be)/.+', formats[0]['url']):
3333 '_type': 'url_transparent',
3334 'url': formats[0]['url'],
3337 self._sort_formats(formats)
3338 entry['formats'] = formats
3339 entries.append(entry)
3340 if len(entries) == 1:
3343 return self.playlist_result(entries)
3345 def _parse_jwplayer_formats(self, jwplayer_sources_data, video_id=None,
3346 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
3349 for source in jwplayer_sources_data:
3350 if not isinstance(source, dict):
3352 source_url = urljoin(
3353 base_url, self._proto_relative_url(source.get('file')))
3354 if not source_url or source_url in urls:
3356 urls.append(source_url)
3357 source_type = source.get('type') or ''
3358 ext = mimetype2ext(source_type) or determine_ext(source_url)
3359 if source_type == 'hls' or ext == 'm3u8':
3360 formats.extend(self._extract_m3u8_formats(
3361 source_url, video_id, 'mp4', entry_protocol='m3u8_native',
3362 m3u8_id=m3u8_id, fatal=False))
3363 elif source_type == 'dash' or ext == 'mpd':
3364 formats.extend(self._extract_mpd_formats(
3365 source_url, video_id, mpd_id=mpd_id, fatal=False))
3367 formats.extend(self._extract_smil_formats(
3368 source_url, video_id, fatal=False))
3369 # https://github.com/jwplayer/jwplayer/blob/master/src/js/providers/default.js#L67
3370 elif source_type.startswith('audio') or ext in (
3371 'oga', 'aac', 'mp3', 'mpeg', 'vorbis'):
3378 height = int_or_none(source.get('height'))
3380 # Often no height is provided but there is a label in
3381 # format like "1080p", "720p SD", or 1080.
3382 height = int_or_none(self._search_regex(
3383 r'^(\d{3,4})[pP]?(?:\b|$)', compat_str(source.get('label') or ''),
3384 'height', default=None))
3387 'width': int_or_none(source.get('width')),
3389 'tbr': int_or_none(source.get('bitrate')),
3392 if source_url.startswith('rtmp'):
3393 a_format['ext'] = 'flv'
3394 # See com/longtailvideo/jwplayer/media/RTMPMediaProvider.as
3395 # of jwplayer.flash.swf
3396 rtmp_url_parts = re.split(
3397 r'((?:mp4|mp3|flv):)', source_url, 1)
3398 if len(rtmp_url_parts) == 3:
3399 rtmp_url, prefix, play_path = rtmp_url_parts
3402 'play_path': prefix + play_path,
3405 a_format.update(rtmp_params)
3406 formats.append(a_format)
3409 def _live_title(self, name):
3410 """ Generate the title
for a live video
"""
3411 now = datetime.datetime.now()
3412 now_str = now.strftime('%Y-%m-%d %H:%M')
3413 return name + ' ' + now_str
3415 def _int(self, v, name, fatal=False, **kwargs):
3416 res = int_or_none(v, **kwargs)
3417 if 'get_attr' in kwargs:
3418 print(getattr(v, kwargs['get_attr']))
3420 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
3422 raise ExtractorError(msg)
3424 self.report_warning(msg)
3427 def _float(self, v, name, fatal=False, **kwargs):
3428 res = float_or_none(v, **kwargs)
3430 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
3432 raise ExtractorError(msg)
3434 self.report_warning(msg)
3437 def _set_cookie(self, domain, name, value, expire_time=None, port=None,
3438 path='/', secure=False, discard=False, rest={}, **kwargs):
3439 cookie = compat_cookiejar_Cookie(
3440 0, name, value, port, port is not None, domain, True,
3441 domain.startswith('.'), path, True, secure, expire_time,
3442 discard, None, None, rest)
3443 self._downloader.cookiejar.set_cookie(cookie)
3445 def _get_cookies(self, url):
3446 """ Return a compat_cookies_SimpleCookie
with the cookies
for the url
"""
3447 req = sanitized_Request(url)
3448 self._downloader.cookiejar.add_cookie_header(req)
3449 return compat_cookies_SimpleCookie(req.get_header('Cookie'))
3451 def _apply_first_set_cookie_header(self, url_handle, cookie):
3453 Apply first Set
-Cookie header instead of the last
. Experimental
.
3455 Some
sites (e
.g
. [1-3]) may serve two cookies under the same name
3456 in Set
-Cookie header
and expect the
first (old
) one to be
set rather
3457 than
second (new
). However
, as of RFC6265 the newer one cookie
3458 should be
set into cookie store what actually happens
.
3459 We will workaround this issue by resetting the cookie to
3460 the first one manually
.
3461 1. https
://new
.vk
.com
/
3462 2. https
://github
.com
/ytdl
-org
/youtube
-dl
/issues
/9841#issuecomment-227871201
3463 3. https
://learning
.oreilly
.com
/
3465 for header, cookies in url_handle.headers.items():
3466 if header.lower() != 'set-cookie':
3468 if sys.version_info[0] >= 3:
3469 cookies = cookies.encode('iso-8859-1')
3470 cookies = cookies.decode('utf-8')
3471 cookie_value = re.search(
3472 r'%s=(.+?);.*?\b[Dd]omain=(.+?)(?:[,;]|$)' % cookie, cookies)
3474 value, domain = cookie_value.groups()
3475 self._set_cookie(domain, cookie, value)
3478 def get_testcases(self, include_onlymatching=False):
3479 t = getattr(self, '_TEST', None)
3481 assert not hasattr(self, '_TESTS'), \
3482 '%s has _TEST and _TESTS' % type(self).__name__
3485 tests = getattr(self, '_TESTS', [])
3487 if not include_onlymatching and t.get('only_matching', False):
3489 t['name'] = type(self).__name__[:-len('IE')]
3492 def is_suitable(self, age_limit):
3493 """ Test whether the extractor
is generally suitable
for the given
3494 age
limit (i
.e
. pornographic sites are
not, all others usually are
) """
3496 any_restricted = False
3497 for tc in self.get_testcases(include_onlymatching=False):
3498 if tc.get('playlist', []):
3499 tc = tc['playlist'][0]
3500 is_restricted = age_restricted(
3501 tc.get('info_dict', {}).get('age_limit'), age_limit)
3502 if not is_restricted:
3504 any_restricted = any_restricted or is_restricted
3505 return not any_restricted
3507 def extract_subtitles(self, *args, **kwargs):
3508 if (self.get_param('writesubtitles', False)
3509 or self.get_param('listsubtitles')):
3510 return self._get_subtitles(*args, **kwargs)
3513 def _get_subtitles(self, *args, **kwargs):
3514 raise NotImplementedError('This method must be implemented by subclasses')
3517 def _merge_subtitle_items(subtitle_list1, subtitle_list2):
3518 """ Merge subtitle items
for one language
. Items
with duplicated URLs
3519 will be dropped
. """
3520 list1_urls = set([item['url'] for item in subtitle_list1])
3521 ret = list(subtitle_list1)
3522 ret.extend([item for item in subtitle_list2 if item['url'] not in list1_urls])
3526 def _merge_subtitles(cls, *dicts, target=None):
3527 """ Merge subtitle dictionaries
, language by language
. """
3531 for lang, subs in d.items():
3532 target[lang] = cls._merge_subtitle_items(target.get(lang, []), subs)
3535 def extract_automatic_captions(self, *args, **kwargs):
3536 if (self.get_param('writeautomaticsub', False)
3537 or self.get_param('listsubtitles')):
3538 return self._get_automatic_captions(*args, **kwargs)
3541 def _get_automatic_captions(self, *args, **kwargs):
3542 raise NotImplementedError('This method must be implemented by subclasses')
3544 def mark_watched(self, *args, **kwargs):
3545 if not self.get_param('mark_watched', False):
3547 if (self._get_login_info()[0] is not None
3548 or self.get_param('cookiefile')
3549 or self.get_param('cookiesfrombrowser')):
3550 self._mark_watched(*args, **kwargs)
3552 def _mark_watched(self, *args, **kwargs):
3553 raise NotImplementedError('This method must be implemented by subclasses')
3555 def geo_verification_headers(self):
3557 geo_verification_proxy = self.get_param('geo_verification_proxy')
3558 if geo_verification_proxy:
3559 headers['Ytdl-request-proxy'] = geo_verification_proxy
3562 def _generic_id(self, url):
3563 return compat_urllib_parse_unquote(os.path.splitext(url.rstrip('/').split('/')[-1])[0])
3565 def _generic_title(self, url):
3566 return compat_urllib_parse_unquote(os.path.splitext(url_basename(url))[0])
3569 def _availability(is_private=None, needs_premium=None, needs_subscription=None, needs_auth=None, is_unlisted=None):
3570 all_known = all(map(
3571 lambda x: x is not None,
3572 (is_private, needs_premium, needs_subscription, needs_auth, is_unlisted)))
3574 'private' if is_private
3575 else 'premium_only' if needs_premium
3576 else 'subscriber_only' if needs_subscription
3577 else 'needs_auth' if needs_auth
3578 else 'unlisted' if is_unlisted
3579 else 'public' if all_known
3582 def _configuration_arg(self, key, default=NO_DEFAULT, casesense=False):
3584 @returns A list of values for the extractor argument given by "key"
3585 or "default" if no such key is present
3586 @param default The default value to return when the key is not present (default: [])
3587 @param casesense When false, the values are converted to lower case
3590 self._downloader.params, ('extractor_args', self.ie_key().lower(), key))
3592 return [] if default is NO_DEFAULT else default
3593 return list(val) if casesense else [x.lower() for x in val]
3596 class SearchInfoExtractor(InfoExtractor):
3598 Base
class for paged search queries extractors
.
3599 They accept URLs
in the format
_SEARCH_KEY(|all|
[0-9]):{query}
3600 Instances should define _SEARCH_KEY
and _MAX_RESULTS
.
3604 def _make_valid_url(cls):
3605 return r'%s(?P<prefix>|[1-9][0-9]*|all):(?P<query>[\s\S]+)' % cls._SEARCH_KEY
3608 def suitable(cls, url):
3609 return re.match(cls._make_valid_url(), url) is not None
3611 def _real_extract(self, query):
3612 mobj = re.match(self._make_valid_url(), query)
3614 raise ExtractorError('Invalid search query "%s"' % query)
3616 prefix = mobj.group('prefix')
3617 query = mobj.group('query')
3619 return self._get_n_results(query, 1)
3620 elif prefix == 'all':
3621 return self._get_n_results(query, self._MAX_RESULTS)
3625 raise ExtractorError('invalid download number %s for query "%s"' % (n, query))
3626 elif n > self._MAX_RESULTS:
3627 self.report_warning('%s returns max %i results (you requested %i)' % (self._SEARCH_KEY, self._MAX_RESULTS, n))
3628 n = self._MAX_RESULTS
3629 return self._get_n_results(query, n)
3631 def _get_n_results(self, query, n):
3632 """Get a specified number of results
for a query
"""
3633 raise NotImplementedError('This method must be implemented by subclasses')
3636 def SEARCH_KEY(self):
3637 return self._SEARCH_KEY