2 from __future__
import unicode_literals
17 from ..compat
import (
18 compat_cookiejar_Cookie
,
19 compat_cookies_SimpleCookie
,
21 compat_etree_fromstring
,
28 compat_urllib_parse_unquote
,
29 compat_urllib_parse_urlencode
,
30 compat_urllib_request
,
32 compat_xml_parse_error
,
34 from ..downloader
import FileDownloader
35 from ..downloader
.f4m
import (
37 remove_encrypted_media
,
68 parse_m3u8_attributes
,
93 class InfoExtractor(object):
94 """Information Extractor class.
96 Information extractors are the classes that, given a URL, extract
97 information about the video (or videos) the URL refers to. This
98 information includes the real video URL, the video title, author and
99 others. The information is stored in a dictionary which is then
100 passed to the YoutubeDL. The YoutubeDL processes this
101 information possibly downloading the video to the file system, among
102 other possible outcomes.
104 The type field determines the type of the result.
105 By far the most common value (and the default if _type is missing) is
106 "video", which indicates a single video.
108 For a video, the dictionaries must include the following fields:
110 id: Video identifier.
111 title: Video title, unescaped.
113 Additionally, it must contain either a formats entry or a url one:
115 formats: A list of dictionaries for each format available, ordered
116 from worst to best quality.
119 * url The mandatory URL representing the media:
120 for plain file media - HTTP URL of this file,
122 for HLS - URL of the M3U8 media playlist,
123 for HDS - URL of the F4M manifest,
125 - HTTP URL to plain file media (in case of
127 - URL of the MPD manifest or base URL
128 representing the media if MPD manifest
129 is parsed from a string (in case of
131 for MSS - URL of the ISM manifest.
133 The URL of the manifest file in case of
135 for HLS - URL of the M3U8 master playlist,
136 for HDS - URL of the F4M manifest,
137 for DASH - URL of the MPD manifest,
138 for MSS - URL of the ISM manifest.
139 * ext Will be calculated from URL if missing
140 * format A human-readable description of the format
141 ("mp4 container with h264/opus").
142 Calculated from the format_id, width, height.
143 and format_note fields if missing.
144 * format_id A short description of the format
145 ("mp4_h264_opus" or "19").
146 Technically optional, but strongly recommended.
147 * format_note Additional info about the format
148 ("3D" or "DASH video")
149 * width Width of the video, if known
150 * height Height of the video, if known
151 * resolution Textual description of width and height
152 * dynamic_range The dynamic range of the video. One of:
153 "SDR" (None), "HDR10", "HDR10+, "HDR12", "HLG, "DV"
154 * tbr Average bitrate of audio and video in KBit/s
155 * abr Average audio bitrate in KBit/s
156 * acodec Name of the audio codec in use
157 * asr Audio sampling rate in Hertz
158 * vbr Average video bitrate in KBit/s
160 * vcodec Name of the video codec in use
161 * container Name of the container format
162 * filesize The number of bytes, if known in advance
163 * filesize_approx An estimate for the number of bytes
164 * player_url SWF Player URL (used for rtmpdump).
165 * protocol The protocol that will be used for the actual
166 download, lower-case.
167 "http", "https", "rtsp", "rtmp", "rtmp_ffmpeg", "rtmpe",
168 "m3u8", "m3u8_native" or "http_dash_segments".
170 Base URL for fragments. Each fragment's path
171 value (if present) will be relative to
173 * fragments A list of fragments of a fragmented media.
174 Each fragment entry must contain either an url
175 or a path. If an url is present it should be
176 considered by a client. Otherwise both path and
177 fragment_base_url must be present. Here is
178 the list of all potential fields:
179 * "url" - fragment's URL
180 * "path" - fragment's path relative to
182 * "duration" (optional, int or float)
183 * "filesize" (optional, int)
184 * preference Order number of this format. If this field is
185 present and not None, the formats get sorted
186 by this field, regardless of all other values.
187 -1 for default (order by other properties),
188 -2 or smaller for less than default.
189 < -1000 to hide the format (if there is
190 another one which is strictly better)
191 * language Language code, e.g. "de" or "en-US".
192 * language_preference Is this in the language mentioned in
194 10 if it's what the URL is about,
195 -1 for default (don't know),
196 -10 otherwise, other values reserved for now.
197 * quality Order number of the video quality of this
198 format, irrespective of the file format.
199 -1 for default (order by other properties),
200 -2 or smaller for less than default.
201 * source_preference Order number for this video source
202 (quality takes higher priority)
203 -1 for default (order by other properties),
204 -2 or smaller for less than default.
205 * http_headers A dictionary of additional HTTP headers
206 to add to the request.
207 * stretched_ratio If given and not 1, indicates that the
208 video's pixels are not square.
209 width : height ratio as float.
210 * no_resume The server does not support resuming the
211 (HTTP or RTMP) download. Boolean.
212 * has_drm The format has DRM and cannot be downloaded. Boolean
213 * downloader_options A dictionary of downloader options as
214 described in FileDownloader
215 RTMP formats can also have the additional fields: page_url,
216 app, play_path, tc_url, flash_version, rtmp_live, rtmp_conn,
217 rtmp_protocol, rtmp_real_time
219 url: Final video URL.
220 ext: Video filename extension.
221 format: The video format, defaults to ext (used for --get-format)
222 player_url: SWF Player URL (used for rtmpdump).
224 The following fields are optional:
226 alt_title: A secondary title of the video.
227 display_id An alternative identifier for the video, not necessarily
228 unique, but available before title. Typically, id is
229 something like "4234987", title "Dancing naked mole rats",
230 and display_id "dancing-naked-mole-rats"
231 thumbnails: A list of dictionaries, with the following entries:
232 * "id" (optional, string) - Thumbnail format ID
234 * "preference" (optional, int) - quality of the image
235 * "width" (optional, int)
236 * "height" (optional, int)
237 * "resolution" (optional, string "{width}x{height}",
239 * "filesize" (optional, int)
240 thumbnail: Full URL to a video thumbnail image.
241 description: Full video description.
242 uploader: Full name of the video uploader.
243 license: License name the video is licensed under.
244 creator: The creator of the video.
245 release_timestamp: UNIX timestamp of the moment the video was released.
246 release_date: The date (YYYYMMDD) when the video was released.
247 timestamp: UNIX timestamp of the moment the video was uploaded
248 upload_date: Video upload date (YYYYMMDD).
249 If not explicitly set, calculated from timestamp.
250 uploader_id: Nickname or id of the video uploader.
251 uploader_url: Full URL to a personal webpage of the video uploader.
252 channel: Full name of the channel the video is uploaded on.
253 Note that channel fields may or may not repeat uploader
254 fields. This depends on a particular extractor.
255 channel_id: Id of the channel.
256 channel_url: Full URL to a channel webpage.
257 location: Physical location where the video was filmed.
258 subtitles: The available subtitles as a dictionary in the format
259 {tag: subformats}. "tag" is usually a language code, and
260 "subformats" is a list sorted from lower to higher
261 preference, each element is a dictionary with the "ext"
263 * "data": The subtitles file contents
264 * "url": A URL pointing to the subtitles file
265 It can optionally also have:
266 * "name": Name or description of the subtitles
267 "ext" will be calculated from URL if missing
268 automatic_captions: Like 'subtitles'; contains automatically generated
269 captions instead of normal subtitles
270 duration: Length of the video in seconds, as an integer or float.
271 view_count: How many users have watched the video on the platform.
272 like_count: Number of positive ratings of the video
273 dislike_count: Number of negative ratings of the video
274 repost_count: Number of reposts of the video
275 average_rating: Average rating give by users, the scale used depends on the webpage
276 comment_count: Number of comments on the video
277 comments: A list of comments, each with one or more of the following
278 properties (all but one of text or html optional):
279 * "author" - human-readable name of the comment author
280 * "author_id" - user ID of the comment author
281 * "author_thumbnail" - The thumbnail of the comment author
283 * "html" - Comment as HTML
284 * "text" - Plain text of the comment
285 * "timestamp" - UNIX timestamp of comment
286 * "parent" - ID of the comment this one is replying to.
287 Set to "root" to indicate that this is a
288 comment to the original video.
289 * "like_count" - Number of positive ratings of the comment
290 * "dislike_count" - Number of negative ratings of the comment
291 * "is_favorited" - Whether the comment is marked as
292 favorite by the video uploader
293 * "author_is_uploader" - Whether the comment is made by
295 age_limit: Age restriction for the video, as an integer (years)
296 webpage_url: The URL to the video webpage, if given to yt-dlp it
297 should allow to get the same result again. (It will be set
298 by YoutubeDL if it's missing)
299 categories: A list of categories that the video falls in, for example
301 tags: A list of tags assigned to the video, e.g. ["sweden", "pop music"]
302 cast: A list of the video cast
303 is_live: True, False, or None (=unknown). Whether this video is a
304 live stream that goes on instead of a fixed-length video.
305 was_live: True, False, or None (=unknown). Whether this video was
306 originally a live stream.
307 live_status: 'is_live', 'is_upcoming', 'was_live', 'not_live' or None (=unknown)
308 If absent, automatically set from is_live, was_live
309 start_time: Time in seconds where the reproduction should start, as
310 specified in the URL.
311 end_time: Time in seconds where the reproduction should end, as
312 specified in the URL.
313 chapters: A list of dictionaries, with the following entries:
314 * "start_time" - The start time of the chapter in seconds
315 * "end_time" - The end time of the chapter in seconds
316 * "title" (optional, string)
317 playable_in_embed: Whether this video is allowed to play in embedded
318 players on other sites. Can be True (=always allowed),
319 False (=never allowed), None (=unknown), or a string
320 specifying the criteria for embedability (Eg: 'whitelist')
321 availability: Under what condition the video is available. One of
322 'private', 'premium_only', 'subscriber_only', 'needs_auth',
323 'unlisted' or 'public'. Use 'InfoExtractor._availability'
325 __post_extractor: A function to be called just before the metadata is
326 written to either disk, logger or console. The function
327 must return a dict which will be added to the info_dict.
328 This is usefull for additional information that is
329 time-consuming to extract. Note that the fields thus
330 extracted will not be available to output template and
331 match_filter. So, only "comments" and "comment_count" are
332 currently allowed to be extracted via this method.
334 The following fields should only be used when the video belongs to some logical
337 chapter: Name or title of the chapter the video belongs to.
338 chapter_number: Number of the chapter the video belongs to, as an integer.
339 chapter_id: Id of the chapter the video belongs to, as a unicode string.
341 The following fields should only be used when the video is an episode of some
342 series, programme or podcast:
344 series: Title of the series or programme the video episode belongs to.
345 series_id: Id of the series or programme the video episode belongs to, as a unicode string.
346 season: Title of the season the video episode belongs to.
347 season_number: Number of the season the video episode belongs to, as an integer.
348 season_id: Id of the season the video episode belongs to, as a unicode string.
349 episode: Title of the video episode. Unlike mandatory video title field,
350 this field should denote the exact title of the video episode
351 without any kind of decoration.
352 episode_number: Number of the video episode within a season, as an integer.
353 episode_id: Id of the video episode, as a unicode string.
355 The following fields should only be used when the media is a track or a part of
358 track: Title of the track.
359 track_number: Number of the track within an album or a disc, as an integer.
360 track_id: Id of the track (useful in case of custom indexing, e.g. 6.iii),
362 artist: Artist(s) of the track.
363 genre: Genre(s) of the track.
364 album: Title of the album the track belongs to.
365 album_type: Type of the album (e.g. "Demo", "Full-length", "Split", "Compilation", etc).
366 album_artist: List of all artists appeared on the album (e.g.
367 "Ash Borer / Fell Voices" or "Various Artists", useful for splits
369 disc_number: Number of the disc or other physical medium the track belongs to,
371 release_year: Year (YYYY) when the album was released.
373 Unless mentioned otherwise, the fields should be Unicode strings.
375 Unless mentioned otherwise, None is equivalent to absence of information.
378 _type "playlist" indicates multiple videos.
379 There must be a key "entries", which is a list, an iterable, or a PagedList
380 object, each element of which is a valid dictionary by this specification.
382 Additionally, playlists can have "id", "title", and any other relevent
383 attributes with the same semantics as videos (see above).
386 _type "multi_video" indicates that there are multiple videos that
387 form a single show, for examples multiple acts of an opera or TV episode.
388 It must have an entries key like a playlist and contain all the keys
389 required for a video at the same time.
392 _type "url" indicates that the video must be extracted from another
393 location, possibly by a different extractor. Its only required key is:
394 "url" - the next URL to extract.
395 The key "ie_key" can be set to the class name (minus the trailing "IE",
396 e.g. "Youtube") if the extractor class is known in advance.
397 Additionally, the dictionary may have any properties of the resolved entity
398 known in advance, for example "title" if the title of the referred video is
402 _type "url_transparent" entities have the same specification as "url", but
403 indicate that the given additional information is more precise than the one
404 associated with the resolved URL.
405 This is useful when a site employs a video service that hosts the video and
406 its technical metadata, but that video service does not embed a useful
407 title, description etc.
410 Subclasses of this one should re-define the _real_initialize() and
411 _real_extract() methods and define a _VALID_URL regexp.
412 Probably, they should also be added to the list of extractors.
414 Subclasses may also override suitable() if necessary, but ensure the function
415 signature is preserved and that this function imports everything it needs
416 (except other extractors), so that lazy_extractors works correctly
418 _GEO_BYPASS attribute may be set to False in order to disable
419 geo restriction bypass mechanisms for a particular extractor.
420 Though it won't disable explicit geo restriction bypass based on
421 country code provided with geo_bypass_country.
423 _GEO_COUNTRIES attribute may contain a list of presumably geo unrestricted
424 countries for this extractor. One of these countries will be used by
425 geo restriction bypass mechanism right away in order to bypass
426 geo restriction, of course, if the mechanism is not disabled.
428 _GEO_IP_BLOCKS attribute may contain a list of presumably geo unrestricted
429 IP blocks in CIDR notation for this extractor. One of these IP blocks
430 will be used by geo restriction bypass mechanism similarly
433 The _WORKING attribute should be set to False for broken IEs
434 in order to warn the users and skip the tests.
439 _x_forwarded_for_ip
= None
441 _GEO_COUNTRIES
= None
442 _GEO_IP_BLOCKS
= None
446 'any': 'Use --cookies, --username and --password, or --netrc to provide account credentials',
448 'Use --cookies-from-browser or --cookies for the authentication. '
449 'See https://github.com/ytdl-org/youtube-dl#how-do-i-pass-cookies-to-youtube-dl for how to manually pass cookies'),
450 'password': 'Use --username and --password, or --netrc to provide account credentials',
453 def __init__(self
, downloader
=None):
454 """Constructor. Receives an optional downloader (a YoutubeDL instance).
455 If a downloader is not passed during initialization,
456 it must be set using "set_downloader()" before "extract()" is called"""
458 self
._x
_forwarded
_for
_ip
= None
459 self
._printed
_messages
= set()
460 self
.set_downloader(downloader
)
463 def _match_valid_url(cls
, url
):
464 # This does not use has/getattr intentionally - we want to know whether
465 # we have cached the regexp for *this* class, whereas getattr would also
466 # match the superclass
467 if '_VALID_URL_RE' not in cls
.__dict
__:
468 cls
._VALID
_URL
_RE
= re
.compile(cls
._VALID
_URL
)
469 return cls
._VALID
_URL
_RE
.match(url
)
472 def suitable(cls
, url
):
473 """Receives a URL and returns True if suitable for this IE."""
474 # This function must import everything it needs (except other extractors),
475 # so that lazy_extractors works correctly
476 return cls
._match
_valid
_url
(url
) is not None
479 def _match_id(cls
, url
):
480 return cls
._match
_valid
_url
(url
).group('id')
483 def get_temp_id(cls
, url
):
485 return cls
._match
_id
(url
)
486 except (IndexError, AttributeError):
491 """Getter method for _WORKING."""
494 def initialize(self
):
495 """Initializes an instance (authentication, etc)."""
496 self
._printed
_messages
= set()
497 self
._initialize
_geo
_bypass
({
498 'countries': self
._GEO
_COUNTRIES
,
499 'ip_blocks': self
._GEO
_IP
_BLOCKS
,
502 self
._real
_initialize
()
505 def _initialize_geo_bypass(self
, geo_bypass_context
):
507 Initialize geo restriction bypass mechanism.
509 This method is used to initialize geo bypass mechanism based on faking
510 X-Forwarded-For HTTP header. A random country from provided country list
511 is selected and a random IP belonging to this country is generated. This
512 IP will be passed as X-Forwarded-For HTTP header in all subsequent
515 This method will be used for initial geo bypass mechanism initialization
516 during the instance initialization with _GEO_COUNTRIES and
519 You may also manually call it from extractor's code if geo bypass
520 information is not available beforehand (e.g. obtained during
521 extraction) or due to some other reason. In this case you should pass
522 this information in geo bypass context passed as first argument. It may
523 contain following fields:
525 countries: List of geo unrestricted countries (similar
527 ip_blocks: List of geo unrestricted IP blocks in CIDR notation
528 (similar to _GEO_IP_BLOCKS)
531 if not self
._x
_forwarded
_for
_ip
:
533 # Geo bypass mechanism is explicitly disabled by user
534 if not self
.get_param('geo_bypass', True):
537 if not geo_bypass_context
:
538 geo_bypass_context
= {}
540 # Backward compatibility: previously _initialize_geo_bypass
541 # expected a list of countries, some 3rd party code may still use
543 if isinstance(geo_bypass_context
, (list, tuple)):
544 geo_bypass_context
= {
545 'countries': geo_bypass_context
,
548 # The whole point of geo bypass mechanism is to fake IP
549 # as X-Forwarded-For HTTP header based on some IP block or
552 # Path 1: bypassing based on IP block in CIDR notation
554 # Explicit IP block specified by user, use it right away
555 # regardless of whether extractor is geo bypassable or not
556 ip_block
= self
.get_param('geo_bypass_ip_block', None)
558 # Otherwise use random IP block from geo bypass context but only
559 # if extractor is known as geo bypassable
561 ip_blocks
= geo_bypass_context
.get('ip_blocks')
562 if self
._GEO
_BYPASS
and ip_blocks
:
563 ip_block
= random
.choice(ip_blocks
)
566 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(ip_block
)
567 self
._downloader
.write_debug(
568 '[debug] Using fake IP %s as X-Forwarded-For' % self
._x
_forwarded
_for
_ip
)
571 # Path 2: bypassing based on country code
573 # Explicit country code specified by user, use it right away
574 # regardless of whether extractor is geo bypassable or not
575 country
= self
.get_param('geo_bypass_country', None)
577 # Otherwise use random country code from geo bypass context but
578 # only if extractor is known as geo bypassable
580 countries
= geo_bypass_context
.get('countries')
581 if self
._GEO
_BYPASS
and countries
:
582 country
= random
.choice(countries
)
585 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(country
)
586 self
._downloader
.write_debug(
587 'Using fake IP %s (%s) as X-Forwarded-For' % (self
._x
_forwarded
_for
_ip
, country
.upper()))
589 def extract(self
, url
):
590 """Extracts URL information and returns it in list of dicts."""
595 self
.write_debug('Extracting URL: %s' % url
)
596 ie_result
= self
._real
_extract
(url
)
597 if ie_result
is None:
599 if self
._x
_forwarded
_for
_ip
:
600 ie_result
['__x_forwarded_for_ip'] = self
._x
_forwarded
_for
_ip
601 subtitles
= ie_result
.get('subtitles')
602 if (subtitles
and 'live_chat' in subtitles
603 and 'no-live-chat' in self
.get_param('compat_opts', [])):
604 del subtitles
['live_chat']
606 except GeoRestrictedError
as e
:
607 if self
.__maybe
_fake
_ip
_and
_retry
(e
.countries
):
610 except UnsupportedError
:
612 except ExtractorError
as e
:
614 'video_id': e
.video_id
or self
.get_temp_id(url
),
617 'expected': e
.expected
,
620 if hasattr(e
, 'countries'):
621 kwargs
['countries'] = e
.countries
622 raise type(e
)(e
.msg
, **kwargs
)
623 except compat_http_client
.IncompleteRead
as e
:
624 raise ExtractorError('A network error has occurred.', cause
=e
, expected
=True, video_id
=self
.get_temp_id(url
))
625 except (KeyError, StopIteration) as e
:
626 raise ExtractorError('An extractor error has occurred.', cause
=e
, video_id
=self
.get_temp_id(url
))
628 def __maybe_fake_ip_and_retry(self
, countries
):
629 if (not self
.get_param('geo_bypass_country', None)
631 and self
.get_param('geo_bypass', True)
632 and not self
._x
_forwarded
_for
_ip
634 country_code
= random
.choice(countries
)
635 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(country_code
)
636 if self
._x
_forwarded
_for
_ip
:
638 'Video is geo restricted. Retrying extraction with fake IP %s (%s) as X-Forwarded-For.'
639 % (self
._x
_forwarded
_for
_ip
, country_code
.upper()))
643 def set_downloader(self
, downloader
):
644 """Sets the downloader for this IE."""
645 self
._downloader
= downloader
647 def _real_initialize(self
):
648 """Real initialization process. Redefine in subclasses."""
651 def _real_extract(self
, url
):
652 """Real extraction process. Redefine in subclasses."""
657 """A string for getting the InfoExtractor with get_info_extractor"""
658 return cls
.__name
__[:-2]
662 return compat_str(type(self
).__name
__[:-2])
665 def __can_accept_status_code(err
, expected_status
):
666 assert isinstance(err
, compat_urllib_error
.HTTPError
)
667 if expected_status
is None:
669 elif callable(expected_status
):
670 return expected_status(err
.code
) is True
672 return err
.code
in variadic(expected_status
)
674 def _request_webpage(self
, url_or_request
, video_id
, note
=None, errnote
=None, fatal
=True, data
=None, headers
={}, query={}
, expected_status
=None):
676 Return the response handle.
678 See _download_webpage docstring for arguments specification.
680 if not self
._downloader
._first
_webpage
_request
:
681 sleep_interval
= self
.get_param('sleep_interval_requests') or 0
682 if sleep_interval
> 0:
683 self
.to_screen('Sleeping %s seconds ...' % sleep_interval
)
684 time
.sleep(sleep_interval
)
686 self
._downloader
._first
_webpage
_request
= False
689 self
.report_download_webpage(video_id
)
690 elif note
is not False:
692 self
.to_screen('%s' % (note
,))
694 self
.to_screen('%s: %s' % (video_id
, note
))
696 # Some sites check X-Forwarded-For HTTP header in order to figure out
697 # the origin of the client behind proxy. This allows bypassing geo
698 # restriction by faking this header's value to IP that belongs to some
699 # geo unrestricted country. We will do so once we encounter any
700 # geo restriction error.
701 if self
._x
_forwarded
_for
_ip
:
702 if 'X-Forwarded-For' not in headers
:
703 headers
['X-Forwarded-For'] = self
._x
_forwarded
_for
_ip
705 if isinstance(url_or_request
, compat_urllib_request
.Request
):
706 url_or_request
= update_Request(
707 url_or_request
, data
=data
, headers
=headers
, query
=query
)
710 url_or_request
= update_url_query(url_or_request
, query
)
711 if data
is not None or headers
:
712 url_or_request
= sanitized_Request(url_or_request
, data
, headers
)
714 return self
._downloader
.urlopen(url_or_request
)
715 except network_exceptions
as err
:
716 if isinstance(err
, compat_urllib_error
.HTTPError
):
717 if self
.__can
_accept
_status
_code
(err
, expected_status
):
718 # Retain reference to error to prevent file object from
719 # being closed before it can be read. Works around the
720 # effects of <https://bugs.python.org/issue15002>
721 # introduced in Python 3.4.1.
728 errnote
= 'Unable to download webpage'
730 errmsg
= '%s: %s' % (errnote
, error_to_compat_str(err
))
732 raise ExtractorError(errmsg
, sys
.exc_info()[2], cause
=err
)
734 self
.report_warning(errmsg
)
737 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):
739 Return a tuple (page content as string, URL handle).
741 See _download_webpage docstring for arguments specification.
743 # Strip hashes from the URL (#1038)
744 if isinstance(url_or_request
, (compat_str
, str)):
745 url_or_request
= url_or_request
.partition('#')[0]
747 urlh
= self
._request
_webpage
(url_or_request
, video_id
, note
, errnote
, fatal
, data
=data
, headers
=headers
, query
=query
, expected_status
=expected_status
)
751 content
= self
._webpage
_read
_content
(urlh
, url_or_request
, video_id
, note
, errnote
, fatal
, encoding
=encoding
)
752 return (content
, urlh
)
755 def _guess_encoding_from_content(content_type
, webpage_bytes
):
756 m
= re
.match(r
'[a-zA-Z0-9_.-]+/[a-zA-Z0-9_.-]+\s*;\s*charset=(.+)', content_type
)
758 encoding
= m
.group(1)
760 m
= re
.search(br
'<meta[^>]+charset=[\'"]?([^\'")]+)[ /\'">]',
761 webpage_bytes[:1024])
763 encoding = m.group(1).decode('ascii')
764 elif webpage_bytes.startswith(b'\xff\xfe'):
771 def __check_blocked(self, content):
772 first_block = content[:512]
773 if ('<title>Access to this site is blocked</title>' in content
774 and 'Websense' in first_block):
775 msg = 'Access to this webpage has been blocked by Websense filtering software in your network.'
776 blocked_iframe = self._html_search_regex(
777 r'<iframe src="([^
"]+)"', content,
778 'Websense information URL
', default=None)
780 msg += ' Visit
%s for more details
' % blocked_iframe
781 raise ExtractorError(msg, expected=True)
782 if '<title
>The URL you requested has been blocked
</title
>' in first_block:
784 'Access to this webpage has been blocked by Indian censorship
. '
785 'Use a VPN
or proxy
server (with --proxy
) to route around it
.')
786 block_msg = self._html_search_regex(
787 r'</h1
><p
>(.*?
)</p
>',
788 content, 'block message
', default=None)
790 msg += ' (Message
: "%s")' % block_msg.replace('\n', ' ')
791 raise ExtractorError(msg, expected=True)
792 if ('<title
>TTK
:: Доступ к ресурсу ограничен
</title
>' in content
793 and 'blocklist
.rkn
.gov
.ru
' in content):
794 raise ExtractorError(
795 'Access to this webpage has been blocked by decision of the Russian government
. '
796 'Visit http
://blocklist
.rkn
.gov
.ru
/ for a block reason
.',
799 def _webpage_read_content(self, urlh, url_or_request, video_id, note=None, errnote=None, fatal=True, prefix=None, encoding=None):
800 content_type = urlh.headers.get('Content
-Type
', '')
801 webpage_bytes = urlh.read()
802 if prefix is not None:
803 webpage_bytes = prefix + webpage_bytes
805 encoding = self._guess_encoding_from_content(content_type, webpage_bytes)
806 if self.get_param('dump_intermediate_pages
', False):
807 self.to_screen('Dumping request to
' + urlh.geturl())
808 dump = base64.b64encode(webpage_bytes).decode('ascii
')
809 self._downloader.to_screen(dump)
810 if self.get_param('write_pages
', False):
811 basen = '%s_%s' % (video_id, urlh.geturl())
812 trim_length = self.get_param('trim_file_name
') or 240
813 if len(basen) > trim_length:
814 h = '___
' + hashlib.md5(basen.encode('utf
-8')).hexdigest()
815 basen = basen[:trim_length - len(h)] + h
816 raw_filename = basen + '.dump
'
817 filename = sanitize_filename(raw_filename, restricted=True)
818 self.to_screen('Saving request to
' + filename)
819 # Working around MAX_PATH limitation on Windows (see
820 # http://msdn.microsoft.com/en-us/library/windows/desktop/aa365247(v=vs.85).aspx)
821 if compat_os_name == 'nt
':
822 absfilepath = os.path.abspath(filename)
823 if len(absfilepath) > 259:
824 filename = '\\\\?
\\' + absfilepath
825 with open(filename, 'wb
') as outf:
826 outf.write(webpage_bytes)
829 content = webpage_bytes.decode(encoding, 'replace
')
831 content = webpage_bytes.decode('utf
-8', 'replace
')
833 self.__check_blocked(content)
837 def _download_webpage(
838 self, url_or_request, video_id, note=None, errnote=None,
839 fatal=True, tries=1, timeout=5, encoding=None, data=None,
840 headers={}, query={}, expected_status=None):
842 Return the data of the page as a string.
845 url_or_request -- plain text URL as a string or
846 a compat_urllib_request.Requestobject
847 video_id -- Video/playlist/item identifier (string)
850 note -- note printed before downloading (string)
851 errnote -- note printed in case of an error (string)
852 fatal -- flag denoting whether error should be considered fatal,
853 i.e. whether it should cause ExtractionError to be raised,
854 otherwise a warning will be reported and extraction continued
855 tries -- number of tries
856 timeout -- sleep interval between tries
857 encoding -- encoding for a page content decoding, guessed automatically
858 when not explicitly specified
859 data -- POST data (bytes)
860 headers -- HTTP headers (dict)
861 query -- URL query (dict)
862 expected_status -- allows to accept failed HTTP requests (non 2xx
863 status code) by explicitly specifying a set of accepted status
864 codes. Can be any of the following entities:
865 - an integer type specifying an exact failed status code to
867 - a list or a tuple of integer types specifying a list of
868 failed status codes to accept
869 - a callable accepting an actual failed status code and
870 returning True if it should be accepted
871 Note that this argument does not affect success status codes (2xx)
872 which are always accepted.
877 while success is False:
879 res = self._download_webpage_handle(
880 url_or_request, video_id, note, errnote, fatal,
881 encoding=encoding, data=data, headers=headers, query=query,
882 expected_status=expected_status)
884 except compat_http_client.IncompleteRead as e:
886 if try_count >= tries:
888 self._sleep(timeout, video_id)
895 def _download_xml_handle(
896 self, url_or_request, video_id, note='Downloading XML
',
897 errnote='Unable to download XML
', transform_source=None,
898 fatal=True, encoding=None, data=None, headers={}, query={},
899 expected_status=None):
901 Return a tuple (xml as an compat_etree_Element, URL handle).
903 See _download_webpage docstring for arguments specification.
905 res = self._download_webpage_handle(
906 url_or_request, video_id, note, errnote, fatal=fatal,
907 encoding=encoding, data=data, headers=headers, query=query,
908 expected_status=expected_status)
911 xml_string, urlh = res
912 return self._parse_xml(
913 xml_string, video_id, transform_source=transform_source,
917 self, url_or_request, video_id,
918 note='Downloading XML
', errnote='Unable to download XML
',
919 transform_source=None, fatal=True, encoding=None,
920 data=None, headers={}, query={}, expected_status=None):
922 Return the xml as an compat_etree_Element.
924 See _download_webpage docstring for arguments specification.
926 res = self._download_xml_handle(
927 url_or_request, video_id, note=note, errnote=errnote,
928 transform_source=transform_source, fatal=fatal, encoding=encoding,
929 data=data, headers=headers, query=query,
930 expected_status=expected_status)
931 return res if res is False else res[0]
933 def _parse_xml(self, xml_string, video_id, transform_source=None, fatal=True):
935 xml_string = transform_source(xml_string)
937 return compat_etree_fromstring(xml_string.encode('utf
-8'))
938 except compat_xml_parse_error as ve:
939 errmsg = '%s: Failed to parse XML
' % video_id
941 raise ExtractorError(errmsg, cause=ve)
943 self.report_warning(errmsg + str(ve))
945 def _download_json_handle(
946 self, url_or_request, video_id, note='Downloading JSON metadata
',
947 errnote='Unable to download JSON metadata
', transform_source=None,
948 fatal=True, encoding=None, data=None, headers={}, query={},
949 expected_status=None):
951 Return a tuple (JSON object, URL handle).
953 See _download_webpage docstring for arguments specification.
955 res = self._download_webpage_handle(
956 url_or_request, video_id, note, errnote, fatal=fatal,
957 encoding=encoding, data=data, headers=headers, query=query,
958 expected_status=expected_status)
961 json_string, urlh = res
962 return self._parse_json(
963 json_string, video_id, transform_source=transform_source,
967 self, url_or_request, video_id, note='Downloading JSON metadata
',
968 errnote='Unable to download JSON metadata
', transform_source=None,
969 fatal=True, encoding=None, data=None, headers={}, query={},
970 expected_status=None):
972 Return the JSON object as a dict.
974 See _download_webpage docstring for arguments specification.
976 res = self._download_json_handle(
977 url_or_request, video_id, note=note, errnote=errnote,
978 transform_source=transform_source, fatal=fatal, encoding=encoding,
979 data=data, headers=headers, query=query,
980 expected_status=expected_status)
981 return res if res is False else res[0]
983 def _parse_json(self, json_string, video_id, transform_source=None, fatal=True):
985 json_string = transform_source(json_string)
987 return json.loads(json_string)
988 except ValueError as ve:
989 errmsg = '%s: Failed to parse JSON
' % video_id
991 raise ExtractorError(errmsg, cause=ve)
993 self.report_warning(errmsg + str(ve))
995 def _parse_socket_response_as_json(self, data, video_id, transform_source=None, fatal=True):
996 return self._parse_json(
997 data[data.find('{'):data.rfind('}
') + 1],
998 video_id, transform_source, fatal)
1000 def _download_socket_json_handle(
1001 self, url_or_request, video_id, note='Polling socket
',
1002 errnote='Unable to poll socket
', transform_source=None,
1003 fatal=True, encoding=None, data=None, headers={}, query={},
1004 expected_status=None):
1006 Return a tuple (JSON object, URL handle).
1008 See _download_webpage docstring for arguments specification.
1010 res = self._download_webpage_handle(
1011 url_or_request, video_id, note, errnote, fatal=fatal,
1012 encoding=encoding, data=data, headers=headers, query=query,
1013 expected_status=expected_status)
1017 return self._parse_socket_response_as_json(
1018 webpage, video_id, transform_source=transform_source,
1021 def _download_socket_json(
1022 self, url_or_request, video_id, note='Polling socket
',
1023 errnote='Unable to poll socket
', transform_source=None,
1024 fatal=True, encoding=None, data=None, headers={}, query={},
1025 expected_status=None):
1027 Return the JSON object as a dict.
1029 See _download_webpage docstring for arguments specification.
1031 res = self._download_socket_json_handle(
1032 url_or_request, video_id, note=note, errnote=errnote,
1033 transform_source=transform_source, fatal=fatal, encoding=encoding,
1034 data=data, headers=headers, query=query,
1035 expected_status=expected_status)
1036 return res if res is False else res[0]
1038 def report_warning(self, msg, video_id=None, *args, only_once=False, **kwargs):
1039 idstr = format_field(video_id, template='%s: ')
1040 msg = f'[{self.IE_NAME}
] {idstr}{msg}
'
1042 if f'WARNING
: {msg}
' in self._printed_messages:
1044 self._printed_messages.add(f'WARNING
: {msg}
')
1045 self._downloader.report_warning(msg, *args, **kwargs)
1047 def to_screen(self, msg, *args, **kwargs):
1048 """Print msg to screen, prefixing it with '[ie_name
]'"""
1049 self._downloader.to_screen('[%s] %s' % (self.IE_NAME, msg), *args, **kwargs)
1051 def write_debug(self, msg, *args, **kwargs):
1052 self._downloader.write_debug('[%s] %s' % (self.IE_NAME, msg), *args, **kwargs)
1054 def get_param(self, name, default=None, *args, **kwargs):
1055 if self._downloader:
1056 return self._downloader.params.get(name, default, *args, **kwargs)
1059 def report_drm(self, video_id, partial=False):
1060 self.raise_no_formats('This video
is DRM protected
', expected=True, video_id=video_id)
1062 def report_extraction(self, id_or_name):
1063 """Report information extraction."""
1064 self.to_screen('%s: Extracting information
' % id_or_name)
1066 def report_download_webpage(self, video_id):
1067 """Report webpage download."""
1068 self.to_screen('%s: Downloading webpage
' % video_id)
1070 def report_age_confirmation(self):
1071 """Report attempt to confirm age."""
1072 self.to_screen('Confirming age
')
1074 def report_login(self):
1075 """Report attempt to log in."""
1076 self.to_screen('Logging
in')
1078 def raise_login_required(
1079 self, msg='This video
is only available
for registered users
',
1080 metadata_available=False, method='any
'):
1081 if metadata_available and self.get_param('ignore_no_formats_error
'):
1082 self.report_warning(msg)
1083 if method is not None:
1084 msg = '%s. %s' % (msg, self._LOGIN_HINTS[method])
1085 raise ExtractorError(msg, expected=True)
1087 def raise_geo_restricted(
1088 self, msg='This video
is not available
from your location due to geo restriction
',
1089 countries=None, metadata_available=False):
1090 if metadata_available and self.get_param('ignore_no_formats_error
'):
1091 self.report_warning(msg)
1093 raise GeoRestrictedError(msg, countries=countries)
1095 def raise_no_formats(self, msg, expected=False, video_id=None):
1096 if expected and self.get_param('ignore_no_formats_error
'):
1097 self.report_warning(msg, video_id)
1098 elif isinstance(msg, ExtractorError):
1101 raise ExtractorError(msg, expected=expected, video_id=video_id)
1103 # Methods for following #608
1105 def url_result(url, ie=None, video_id=None, video_title=None, **kwargs):
1106 """Returns a URL that points to a page that should be processed"""
1107 # TODO: ie should be the class used for getting the info
1108 video_info = {'_type
': 'url
',
1111 video_info.update(kwargs)
1112 if video_id is not None:
1113 video_info['id'] = video_id
1114 if video_title is not None:
1115 video_info['title
'] = video_title
1118 def playlist_from_matches(self, matches, playlist_id=None, playlist_title=None, getter=None, ie=None):
1120 self.url_result(self._proto_relative_url(getter(m) if getter else m), ie)
1122 return self.playlist_result(
1123 urls, playlist_id=playlist_id, playlist_title=playlist_title)
1126 def playlist_result(entries, playlist_id=None, playlist_title=None, playlist_description=None, **kwargs):
1127 """Returns a playlist"""
1128 video_info = {'_type
': 'playlist
',
1130 video_info.update(kwargs)
1132 video_info['id'] = playlist_id
1134 video_info['title
'] = playlist_title
1135 if playlist_description is not None:
1136 video_info['description
'] = playlist_description
1139 def _search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
1141 Perform a regex search on the given string, using a single or a list of
1142 patterns returning the first matching group.
1143 In case of failure return a default value or raise a WARNING or a
1144 RegexNotFoundError, depending on fatal, specifying the field name.
1146 if isinstance(pattern, (str, compat_str, compiled_regex_type)):
1147 mobj = re.search(pattern, string, flags)
1150 mobj = re.search(p, string, flags)
1154 _name = self._downloader._format_err(name, self._downloader.Styles.EMPHASIS)
1158 # return the first matching group
1159 return next(g for g in mobj.groups() if g is not None)
1160 elif isinstance(group, (list, tuple)):
1161 return tuple(mobj.group(g) for g in group)
1163 return mobj.group(group)
1164 elif default is not NO_DEFAULT:
1167 raise RegexNotFoundError('Unable to extract
%s' % _name)
1169 self.report_warning('unable to extract
%s' % _name + bug_reports_message())
1172 def _html_search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
1174 Like _search_regex, but strips HTML tags and unescapes entities.
1176 res = self._search_regex(pattern, string, name, default, fatal, flags, group)
1178 return clean_html(res).strip()
1182 def _get_netrc_login_info(self, netrc_machine=None):
1185 netrc_machine = netrc_machine or self._NETRC_MACHINE
1187 if self.get_param('usenetrc
', False):
1189 netrc_file = compat_expanduser(self.get_param('netrc_location
') or '~
')
1190 if os.path.isdir(netrc_file):
1191 netrc_file = os.path.join(netrc_file, '.netrc
')
1192 info = netrc.netrc(file=netrc_file).authenticators(netrc_machine)
1193 if info is not None:
1197 raise netrc.NetrcParseError(
1198 'No authenticators
for %s' % netrc_machine)
1199 except (IOError, netrc.NetrcParseError) as err:
1200 self.report_warning(
1201 'parsing
.netrc
: %s' % error_to_compat_str(err))
1203 return username, password
1205 def _get_login_info(self, username_option='username
', password_option='password
', netrc_machine=None):
1207 Get the login info as (username, password)
1208 First look for the manually specified credentials using username_option
1209 and password_option as keys in params dictionary. If no such credentials
1210 available look in the netrc file using the netrc_machine or _NETRC_MACHINE
1212 If there's no info available
, return (None, None)
1215 # Attempt to use provided username and password or .netrc data
1216 username = self.get_param(username_option)
1217 if username is not None:
1218 password = self.get_param(password_option)
1220 username, password = self._get_netrc_login_info(netrc_machine)
1222 return username, password
1224 def _get_tfa_info(self, note='two-factor verification code'):
1226 Get the two
-factor authentication info
1227 TODO
- asking the user will be required
for sms
/phone verify
1228 currently just uses the command line option
1229 If there
's no info available, return None
1232 tfa = self.get_param('twofactor
')
1236 return compat_getpass('Type
%s and press
[Return
]: ' % note)
1238 # Helper functions for extracting OpenGraph info
1240 def _og_regexes(prop):
1241 content_re = r'content
=(?
:"([^"]+?
)"|\'([^\']+?)\'|\s*([^\s"\'=<>`
]+?
))'
1242 property_re = (r'(?
:name|
property)=(?
:\'og
[:-]%(prop)s\'|
"og[:-]%(prop)s"|\s
*og
[:-]%(prop)s\b)'
1243 % {'prop': re.escape(prop)})
1244 template = r'<meta
[^
>]+?
%s[^
>]+?
%s'
1246 template % (property_re, content_re),
1247 template % (content_re, property_re),
1251 def _meta_regex(prop):
1252 return r'''(?isx)<meta
1253 (?=[^>]+(?:itemprop|name|property|id|http-equiv)=(["\']?)%s\1)
1254 [^>]+?content=(["\'])(?P<content>.*?)\2''' % re.escape(prop)
1256 def _og_search_property(self, prop, html, name=None, **kargs):
1257 prop = variadic(prop)
1259 name = 'OpenGraph
%s' % prop[0]
1262 og_regexes.extend(self._og_regexes(p))
1263 escaped = self._search_regex(og_regexes, html, name, flags=re.DOTALL, **kargs)
1266 return unescapeHTML(escaped)
1268 def _og_search_thumbnail(self, html, **kargs):
1269 return self._og_search_property('image
', html, 'thumbnail URL
', fatal=False, **kargs)
1271 def _og_search_description(self, html, **kargs):
1272 return self._og_search_property('description
', html, fatal=False, **kargs)
1274 def _og_search_title(self, html, **kargs):
1275 return self._og_search_property('title
', html, **kargs)
1277 def _og_search_video_url(self, html, name='video url
', secure=True, **kargs):
1278 regexes = self._og_regexes('video
') + self._og_regexes('video
:url
')
1280 regexes = self._og_regexes('video
:secure_url
') + regexes
1281 return self._html_search_regex(regexes, html, name, **kargs)
1283 def _og_search_url(self, html, **kargs):
1284 return self._og_search_property('url
', html, **kargs)
1286 def _html_search_meta(self, name, html, display_name=None, fatal=False, **kwargs):
1287 name = variadic(name)
1288 if display_name is None:
1289 display_name = name[0]
1290 return self._html_search_regex(
1291 [self._meta_regex(n) for n in name],
1292 html, display_name, fatal=fatal, group='content
', **kwargs)
1294 def _dc_search_uploader(self, html):
1295 return self._html_search_meta('dc
.creator
', html, 'uploader
')
1297 def _rta_search(self, html):
1298 # See http://www.rtalabel.org/index.php?content=howtofaq#single
1299 if re.search(r'(?ix
)<meta\s
+name
="rating"\s
+'
1300 r' content
="RTA-5042-1996-1400-1577-RTA"',
1305 def _media_rating_search(self, html):
1306 # See http://www.tjg-designs.com/WP/metadata-code-examples-adding-metadata-to-your-web-pages/
1307 rating = self._html_search_meta('rating
', html)
1319 return RATING_TABLE.get(rating.lower())
1321 def _family_friendly_search(self, html):
1322 # See http://schema.org/VideoObject
1323 family_friendly = self._html_search_meta(
1324 'isFamilyFriendly
', html, default=None)
1326 if not family_friendly:
1335 return RATING_TABLE.get(family_friendly.lower())
1337 def _twitter_search_player(self, html):
1338 return self._html_search_meta('twitter
:player
', html,
1339 'twitter card player
')
1341 def _search_json_ld(self, html, video_id, expected_type=None, **kwargs):
1342 json_ld_list = list(re.finditer(JSON_LD_RE, html))
1343 default = kwargs.get('default
', NO_DEFAULT)
1344 # JSON-LD may be malformed and thus `fatal` should be respected.
1345 # At the same time `default` may be passed that assumes `fatal=False`
1346 # for _search_regex. Let's simulate the same behavior here
as well
.
1347 fatal
= kwargs
.get('fatal', True) if default
is NO_DEFAULT
else False
1349 for mobj
in json_ld_list
:
1350 json_ld_item
= self
._parse
_json
(
1351 mobj
.group('json_ld'), video_id
, fatal
=fatal
)
1352 if not json_ld_item
:
1354 if isinstance(json_ld_item
, dict):
1355 json_ld
.append(json_ld_item
)
1356 elif isinstance(json_ld_item
, (list, tuple)):
1357 json_ld
.extend(json_ld_item
)
1359 json_ld
= self
._json
_ld
(json_ld
, video_id
, fatal
=fatal
, expected_type
=expected_type
)
1362 if default
is not NO_DEFAULT
:
1365 raise RegexNotFoundError('Unable to extract JSON-LD')
1367 self
.report_warning('unable to extract JSON-LD %s' % bug_reports_message())
1370 def _json_ld(self
, json_ld
, video_id
, fatal
=True, expected_type
=None):
1371 if isinstance(json_ld
, compat_str
):
1372 json_ld
= self
._parse
_json
(json_ld
, video_id
, fatal
=fatal
)
1376 if not isinstance(json_ld
, (list, tuple, dict)):
1378 if isinstance(json_ld
, dict):
1381 INTERACTION_TYPE_MAP
= {
1382 'CommentAction': 'comment',
1383 'AgreeAction': 'like',
1384 'DisagreeAction': 'dislike',
1385 'LikeAction': 'like',
1386 'DislikeAction': 'dislike',
1387 'ListenAction': 'view',
1388 'WatchAction': 'view',
1389 'ViewAction': 'view',
1392 def extract_interaction_type(e
):
1393 interaction_type
= e
.get('interactionType')
1394 if isinstance(interaction_type
, dict):
1395 interaction_type
= interaction_type
.get('@type')
1396 return str_or_none(interaction_type
)
1398 def extract_interaction_statistic(e
):
1399 interaction_statistic
= e
.get('interactionStatistic')
1400 if isinstance(interaction_statistic
, dict):
1401 interaction_statistic
= [interaction_statistic
]
1402 if not isinstance(interaction_statistic
, list):
1404 for is_e
in interaction_statistic
:
1405 if not isinstance(is_e
, dict):
1407 if is_e
.get('@type') != 'InteractionCounter':
1409 interaction_type
= extract_interaction_type(is_e
)
1410 if not interaction_type
:
1412 # For interaction count some sites provide string instead of
1413 # an integer (as per spec) with non digit characters (e.g. ",")
1414 # so extracting count with more relaxed str_to_int
1415 interaction_count
= str_to_int(is_e
.get('userInteractionCount'))
1416 if interaction_count
is None:
1418 count_kind
= INTERACTION_TYPE_MAP
.get(interaction_type
.split('/')[-1])
1421 count_key
= '%s_count' % count_kind
1422 if info
.get(count_key
) is not None:
1424 info
[count_key
] = interaction_count
1426 def extract_video_object(e
):
1427 assert e
['@type'] == 'VideoObject'
1428 author
= e
.get('author')
1430 'url': url_or_none(e
.get('contentUrl')),
1431 'title': unescapeHTML(e
.get('name')),
1432 'description': unescapeHTML(e
.get('description')),
1433 'thumbnail': url_or_none(e
.get('thumbnailUrl') or e
.get('thumbnailURL')),
1434 'duration': parse_duration(e
.get('duration')),
1435 'timestamp': unified_timestamp(e
.get('uploadDate')),
1436 # author can be an instance of 'Organization' or 'Person' types.
1437 # both types can have 'name' property(inherited from 'Thing' type). [1]
1438 # however some websites are using 'Text' type instead.
1439 # 1. https://schema.org/VideoObject
1440 'uploader': author
.get('name') if isinstance(author
, dict) else author
if isinstance(author
, compat_str
) else None,
1441 'filesize': float_or_none(e
.get('contentSize')),
1442 'tbr': int_or_none(e
.get('bitrate')),
1443 'width': int_or_none(e
.get('width')),
1444 'height': int_or_none(e
.get('height')),
1445 'view_count': int_or_none(e
.get('interactionCount')),
1447 extract_interaction_statistic(e
)
1451 item_type
= e
.get('@type')
1452 if expected_type
is not None and expected_type
!= item_type
:
1454 if item_type
in ('TVEpisode', 'Episode'):
1455 episode_name
= unescapeHTML(e
.get('name'))
1457 'episode': episode_name
,
1458 'episode_number': int_or_none(e
.get('episodeNumber')),
1459 'description': unescapeHTML(e
.get('description')),
1461 if not info
.get('title') and episode_name
:
1462 info
['title'] = episode_name
1463 part_of_season
= e
.get('partOfSeason')
1464 if isinstance(part_of_season
, dict) and part_of_season
.get('@type') in ('TVSeason', 'Season', 'CreativeWorkSeason'):
1466 'season': unescapeHTML(part_of_season
.get('name')),
1467 'season_number': int_or_none(part_of_season
.get('seasonNumber')),
1469 part_of_series
= e
.get('partOfSeries') or e
.get('partOfTVSeries')
1470 if isinstance(part_of_series
, dict) and part_of_series
.get('@type') in ('TVSeries', 'Series', 'CreativeWorkSeries'):
1471 info
['series'] = unescapeHTML(part_of_series
.get('name'))
1472 elif item_type
== 'Movie':
1474 'title': unescapeHTML(e
.get('name')),
1475 'description': unescapeHTML(e
.get('description')),
1476 'duration': parse_duration(e
.get('duration')),
1477 'timestamp': unified_timestamp(e
.get('dateCreated')),
1479 elif item_type
in ('Article', 'NewsArticle'):
1481 'timestamp': parse_iso8601(e
.get('datePublished')),
1482 'title': unescapeHTML(e
.get('headline')),
1483 'description': unescapeHTML(e
.get('articleBody')),
1485 elif item_type
== 'VideoObject':
1486 extract_video_object(e
)
1487 if expected_type
is None:
1491 video
= e
.get('video')
1492 if isinstance(video
, dict) and video
.get('@type') == 'VideoObject':
1493 extract_video_object(video
)
1494 if expected_type
is None:
1498 return dict((k
, v
) for k
, v
in info
.items() if v
is not None)
1500 def _search_nextjs_data(self
, webpage
, video_id
, **kw
):
1501 return self
._parse
_json
(
1503 r
'(?s)<script[^>]+id=[\'"]__NEXT_DATA__[\'"][^
>]*>([^
<]+)</script
>',
1504 webpage, 'next
.js data
', **kw),
1508 def _hidden_inputs(html):
1509 html = re.sub(r'<!--(?
:(?
!<!--).)*-->', '', html)
1511 for input in re.findall(r'(?i
)(<input[^
>]+>)', html):
1512 attrs = extract_attributes(input)
1515 if attrs.get('type') not in ('hidden
', 'submit
'):
1517 name = attrs.get('name
') or attrs.get('id')
1518 value = attrs.get('value
')
1519 if name and value is not None:
1520 hidden_inputs[name] = value
1521 return hidden_inputs
1523 def _form_hidden_inputs(self, form_id, html):
1524 form = self._search_regex(
1525 r'(?
is)<form
[^
>]+?
id=(["\'])%s\1[^>]*>(?P<form>.+?)</form>' % form_id,
1526 html, '%s form' % form_id, group='form')
1527 return self._hidden_inputs(form)
1530 regex = r' *((?P<reverse>\+)?(?P<field>[a-zA-Z0-9_]+)((?P<separator>[~:])(?P<limit>.*?))?)? *$'
1532 default = ('hidden', 'aud_or_vid', 'hasvid', 'ie_pref', 'lang', 'quality',
1533 'res', 'fps', 'hdr:12', 'codec:vp9.2', 'size', 'br', 'asr',
1534 'proto', 'ext', 'hasaud', 'source', 'format_id') # These must not be aliases
1535 ytdl_default = ('hasaud', 'lang', 'quality', 'tbr', 'filesize', 'vbr',
1536 'height', 'width', 'proto', 'vext', 'abr', 'aext',
1537 'fps', 'fs_approx', 'source', 'format_id')
1540 'vcodec': {'type': 'ordered', 'regex': True,
1541 'order': ['av0?1', 'vp0?9.2', 'vp0?9', '[hx]265|he?vc?', '[hx]264|avc', 'vp0?8', 'mp4v|h263', 'theora', '', None, 'none']},
1542 'acodec': {'type': 'ordered', 'regex': True,
1543 'order': ['opus', 'vorbis', 'aac', 'mp?4a?', 'mp3', 'e-?a?c-?3', 'ac-?3', 'dts', '', None, 'none']},
1544 'hdr': {'type': 'ordered', 'regex': True, 'field': 'dynamic_range',
1545 'order': ['dv', '(hdr)?12', r'(hdr)?10\+', '(hdr)?10', 'hlg', '', 'sdr', None]},
1546 'proto': {'type': 'ordered', 'regex': True, 'field': 'protocol',
1547 'order': ['(ht|f)tps', '(ht|f)tp$', 'm3u8.+', '.*dash', 'ws|websocket', '', 'mms|rtsp', 'none', 'f4']},
1548 'vext': {'type': 'ordered', 'field': 'video_ext',
1549 'order': ('mp4', 'webm', 'flv', '', 'none'),
1550 'order_free': ('webm', 'mp4', 'flv', '', 'none')},
1551 'aext': {'type': 'ordered', 'field': 'audio_ext',
1552 'order': ('m4a', 'aac', 'mp3', 'ogg', 'opus', 'webm', '', 'none'),
1553 'order_free': ('opus', 'ogg', 'webm', 'm4a', 'mp3', 'aac', '', 'none')},
1554 'hidden': {'visible': False, 'forced': True, 'type': 'extractor', 'max': -1000},
1555 'aud_or_vid': {'visible': False, 'forced': True, 'type': 'multiple',
1556 'field': ('vcodec', 'acodec'),
1557 'function': lambda it: int(any(v != 'none' for v in it))},
1558 'ie_pref': {'priority': True, 'type': 'extractor'},
1559 'hasvid': {'priority': True, 'field': 'vcodec', 'type': 'boolean', 'not_in_list': ('none',)},
1560 'hasaud': {'field': 'acodec', 'type': 'boolean', 'not_in_list': ('none',)},
1561 'lang': {'convert': 'float', 'field': 'language_preference', 'default': -1},
1562 'quality': {'convert': 'float', 'default': -1},
1563 'filesize': {'convert': 'bytes'},
1564 'fs_approx': {'convert': 'bytes', 'field': 'filesize_approx'},
1565 'id': {'convert': 'string', 'field': 'format_id'},
1566 'height': {'convert': 'float_none'},
1567 'width': {'convert': 'float_none'},
1568 'fps': {'convert': 'float_none'},
1569 'tbr': {'convert': 'float_none'},
1570 'vbr': {'convert': 'float_none'},
1571 'abr': {'convert': 'float_none'},
1572 'asr': {'convert': 'float_none'},
1573 'source': {'convert': 'float', 'field': 'source_preference', 'default': -1},
1575 'codec': {'type': 'combined', 'field': ('vcodec', 'acodec')},
1576 'br': {'type': 'combined', 'field': ('tbr', 'vbr', 'abr'), 'same_limit': True},
1577 'size': {'type': 'combined', 'same_limit': True, 'field': ('filesize', 'fs_approx')},
1578 'ext': {'type': 'combined', 'field': ('vext', 'aext')},
1579 'res': {'type': 'multiple', 'field': ('height', 'width'),
1580 'function': lambda it: (lambda l: min(l) if l else 0)(tuple(filter(None, it)))},
1582 # Most of these exist only for compatibility reasons
1583 'dimension': {'type': 'alias', 'field': 'res'},
1584 'resolution': {'type': 'alias', 'field': 'res'},
1585 'extension': {'type': 'alias', 'field': 'ext'},
1586 'bitrate': {'type': 'alias', 'field': 'br'},
1587 'total_bitrate': {'type': 'alias', 'field': 'tbr'},
1588 'video_bitrate': {'type': 'alias', 'field': 'vbr'},
1589 'audio_bitrate': {'type': 'alias', 'field': 'abr'},
1590 'framerate': {'type': 'alias', 'field': 'fps'},
1591 'language_preference': {'type': 'alias', 'field': 'lang'}, # not named as 'language' because such a field exists
1592 'protocol': {'type': 'alias', 'field': 'proto'},
1593 'source_preference': {'type': 'alias', 'field': 'source'},
1594 'filesize_approx': {'type': 'alias', 'field': 'fs_approx'},
1595 'filesize_estimate': {'type': 'alias', 'field': 'size'},
1596 'samplerate': {'type': 'alias', 'field': 'asr'},
1597 'video_ext': {'type': 'alias', 'field': 'vext'},
1598 'audio_ext': {'type': 'alias', 'field': 'aext'},
1599 'video_codec': {'type': 'alias', 'field': 'vcodec'},
1600 'audio_codec': {'type': 'alias', 'field': 'acodec'},
1601 'video': {'type': 'alias', 'field': 'hasvid'},
1602 'has_video': {'type': 'alias', 'field': 'hasvid'},
1603 'audio': {'type': 'alias', 'field': 'hasaud'},
1604 'has_audio': {'type': 'alias', 'field': 'hasaud'},
1605 'extractor': {'type': 'alias', 'field': 'ie_pref'},
1606 'preference': {'type': 'alias', 'field': 'ie_pref'},
1607 'extractor_preference': {'type': 'alias', 'field': 'ie_pref'},
1608 'format_id': {'type': 'alias', 'field': 'id'},
1613 def _get_field_setting(self, field, key):
1614 if field not in self.settings:
1615 self.settings[field] = {}
1616 propObj = self.settings[field]
1617 if key not in propObj:
1618 type = propObj.get('type')
1620 default = 'preference' if type == 'extractor' else (field,) if type in ('combined', 'multiple') else field
1621 elif key == 'convert':
1622 default = 'order' if type == 'ordered' else 'float_string' if field else 'ignore'
1624 default = {'type': 'field', 'visible': True, 'order': [], 'not_in_list': (None,)}.get(key, None)
1625 propObj[key] = default
1628 def _resolve_field_value(self, field, value, convertNone=False):
1633 value = value.lower()
1634 conversion = self._get_field_setting(field, 'convert')
1635 if conversion == 'ignore':
1637 if conversion == 'string':
1639 elif conversion == 'float_none':
1640 return float_or_none(value)
1641 elif conversion == 'bytes':
1642 return FileDownloader.parse_bytes(value)
1643 elif conversion == 'order':
1644 order_list = (self._use_free_order and self._get_field_setting(field, 'order_free')) or self._get_field_setting(field, 'order')
1645 use_regex = self._get_field_setting(field, 'regex')
1646 list_length = len(order_list)
1647 empty_pos = order_list.index('') if '' in order_list else list_length + 1
1648 if use_regex and value is not None:
1649 for i, regex in enumerate(order_list):
1650 if regex and re.match(regex, value):
1651 return list_length - i
1652 return list_length - empty_pos # not in list
1653 else: # not regex or value = None
1654 return list_length - (order_list.index(value) if value in order_list else empty_pos)
1656 if value.isnumeric():
1659 self.settings[field]['convert'] = 'string'
1662 def evaluate_params(self, params, sort_extractor):
1663 self._use_free_order = params.get('prefer_free_formats', False)
1664 self._sort_user = params.get('format_sort', [])
1665 self._sort_extractor = sort_extractor
1667 def add_item(field, reverse, closest, limit_text):
1668 field = field.lower()
1669 if field in self._order:
1671 self._order.append(field)
1672 limit = self._resolve_field_value(field, limit_text)
1675 'closest': False if limit is None else closest,
1676 'limit_text': limit_text,
1678 if field in self.settings:
1679 self.settings[field].update(data)
1681 self.settings[field] = data
1684 tuple(field for field in self.default if self._get_field_setting(field, 'forced'))
1685 + (tuple() if params.get('format_sort_force', False)
1686 else tuple(field for field in self.default if self._get_field_setting(field, 'priority')))
1687 + tuple(self._sort_user) + tuple(sort_extractor) + self.default)
1689 for item in sort_list:
1690 match = re.match(self.regex, item)
1692 raise ExtractorError('Invalid format sort string "%s" given by extractor' % item)
1693 field = match.group('field')
1696 if self._get_field_setting(field, 'type') == 'alias':
1697 field = self._get_field_setting(field, 'field')
1698 reverse = match.group('reverse') is not None
1699 closest = match.group('separator') == '~'
1700 limit_text = match.group('limit')
1702 has_limit = limit_text is not None
1703 has_multiple_fields = self._get_field_setting(field, 'type') == 'combined'
1704 has_multiple_limits = has_limit and has_multiple_fields and not self._get_field_setting(field, 'same_limit')
1706 fields = self._get_field_setting(field, 'field') if has_multiple_fields else (field,)
1707 limits = limit_text.split(':') if has_multiple_limits else (limit_text,) if has_limit else tuple()
1708 limit_count = len(limits)
1709 for (i, f) in enumerate(fields):
1710 add_item(f, reverse, closest,
1711 limits[i] if i < limit_count
1712 else limits[0] if has_limit and not has_multiple_limits
1715 def print_verbose_info(self, write_debug):
1717 write_debug('Sort order given by user: %s' % ', '.join(self._sort_user))
1718 if self._sort_extractor:
1719 write_debug('Sort order given by extractor: %s' % ', '.join(self._sort_extractor))
1720 write_debug('Formats sorted by: %s' % ', '.join(['%s%s%s' % (
1721 '+' if self._get_field_setting(field, 'reverse') else '', field,
1722 '%s%s(%s)' % ('~' if self._get_field_setting(field, 'closest') else ':',
1723 self._get_field_setting(field, 'limit_text'),
1724 self._get_field_setting(field, 'limit'))
1725 if self._get_field_setting(field, 'limit_text') is not None else '')
1726 for field in self._order if self._get_field_setting(field, 'visible')]))
1728 def _calculate_field_preference_from_value(self, format, field, type, value):
1729 reverse = self._get_field_setting(field, 'reverse')
1730 closest = self._get_field_setting(field, 'closest')
1731 limit = self._get_field_setting(field, 'limit')
1733 if type == 'extractor':
1734 maximum = self._get_field_setting(field, 'max')
1735 if value is None or (maximum is not None and value >= maximum):
1737 elif type == 'boolean':
1738 in_list = self._get_field_setting(field, 'in_list')
1739 not_in_list = self._get_field_setting(field, 'not_in_list')
1740 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
1741 elif type == 'ordered':
1742 value = self._resolve_field_value(field, value, True)
1744 # try to convert to number
1745 val_num = float_or_none(value, default=self._get_field_setting(field, 'default'))
1746 is_num = self._get_field_setting(field, 'convert') != 'string' and val_num is not None
1750 return ((-10, 0) if value is None
1751 else (1, value, 0) if not is_num # if a field has mixed strings and numbers, strings are sorted higher
1752 else (0, -abs(value - limit), value - limit if reverse else limit - value) if closest
1753 else (0, value, 0) if not reverse and (limit is None or value <= limit)
1754 else (0, -value, 0) if limit is None or (reverse and value == limit) or value > limit
1755 else (-1, value, 0))
1757 def _calculate_field_preference(self, format, field):
1758 type = self._get_field_setting(field, 'type') # extractor, boolean, ordered, field, multiple
1759 get_value = lambda f: format.get(self._get_field_setting(f, 'field'))
1760 if type == 'multiple':
1761 type = 'field' # Only 'field' is allowed in multiple for now
1762 actual_fields = self._get_field_setting(field, 'field')
1764 value = self._get_field_setting(field, 'function')(get_value(f) for f in actual_fields)
1766 value = get_value(field)
1767 return self._calculate_field_preference_from_value(format, field, type, value)
1769 def calculate_preference(self, format):
1770 # Determine missing protocol
1771 if not format.get('protocol'):
1772 format['protocol'] = determine_protocol(format)
1774 # Determine missing ext
1775 if not format.get('ext') and 'url' in format:
1776 format['ext'] = determine_ext(format['url'])
1777 if format.get('vcodec') == 'none':
1778 format['audio_ext'] = format['ext'] if format.get('acodec') != 'none' else 'none'
1779 format['video_ext'] = 'none'
1781 format['video_ext'] = format['ext']
1782 format['audio_ext'] = 'none'
1783 # if format.get('preference') is None and format.get('ext') in ('f4f', 'f4m'): # Not supported?
1784 # format['preference'] = -1000
1786 # Determine missing bitrates
1787 if format.get('tbr') is None:
1788 if format.get('vbr') is not None and format.get('abr') is not None:
1789 format['tbr'] = format.get('vbr', 0) + format.get('abr', 0)
1791 if format.get('vcodec') != 'none' and format.get('vbr') is None:
1792 format['vbr'] = format.get('tbr') - format.get('abr', 0)
1793 if format.get('acodec') != 'none' and format.get('abr') is None:
1794 format['abr'] = format.get('tbr') - format.get('vbr', 0)
1796 return tuple(self._calculate_field_preference(format, field) for field in self._order)
1798 def _sort_formats(self, formats, field_preference=[]):
1801 format_sort = self.FormatSort() # params and to_screen are taken from the downloader
1802 format_sort.evaluate_params(self._downloader.params, field_preference)
1803 if self.get_param('verbose', False):
1804 format_sort.print_verbose_info(self._downloader.write_debug)
1805 formats.sort(key=lambda f: format_sort.calculate_preference(f))
1807 def _check_formats(self, formats, video_id):
1809 formats[:] = filter(
1810 lambda f: self._is_valid_url(
1812 item='%s video format' % f.get('format_id') if f.get('format_id') else 'video'),
1816 def _remove_duplicate_formats(formats):
1820 if f['url'] not in format_urls:
1821 format_urls.add(f['url'])
1822 unique_formats.append(f)
1823 formats[:] = unique_formats
1825 def _is_valid_url(self, url, video_id, item='video', headers={}):
1826 url = self._proto_relative_url(url, scheme='http:')
1827 # For now assume non HTTP(S) URLs always valid
1828 if not (url.startswith('http://') or url.startswith('https://')):
1831 self._request_webpage(url, video_id, 'Checking %s URL' % item, headers=headers)
1833 except ExtractorError as e:
1835 '%s: %s URL is invalid, skipping: %s'
1836 % (video_id, item, error_to_compat_str(e.cause)))
1839 def http_scheme(self):
1840 """ Either "http
:" or "https
:", depending on the user's preferences """
1843 if self.get_param('prefer_insecure', False)
1846 def _proto_relative_url(self, url, scheme=None):
1849 if url.startswith('//'):
1851 scheme = self.http_scheme()
1856 def _sleep(self, timeout, video_id, msg_template=None):
1857 if msg_template is None:
1858 msg_template = '%(video_id)s: Waiting for %(timeout)s seconds'
1859 msg = msg_template % {'video_id': video_id, 'timeout': timeout}
1863 def _extract_f4m_formats(self, manifest_url, video_id, preference=None, quality=None, f4m_id=None,
1864 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1865 fatal=True, m3u8_id=None, data=None, headers={}, query={}):
1866 manifest = self._download_xml(
1867 manifest_url, video_id, 'Downloading f4m manifest',
1868 'Unable to download f4m manifest',
1869 # Some manifests may be malformed, e.g. prosiebensat1 generated manifests
1870 # (see https://github.com/ytdl-org/youtube-dl/issues/6215#issuecomment-121704244)
1871 transform_source=transform_source,
1872 fatal=fatal, data=data, headers=headers, query=query)
1874 if manifest is False:
1877 return self._parse_f4m_formats(
1878 manifest, manifest_url, video_id, preference=preference, quality=quality, f4m_id=f4m_id,
1879 transform_source=transform_source, fatal=fatal, m3u8_id=m3u8_id)
1881 def _parse_f4m_formats(self, manifest, manifest_url, video_id, preference=None, quality=None, f4m_id=None,
1882 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1883 fatal=True, m3u8_id=None):
1884 if not isinstance(manifest, compat_etree_Element) and not fatal:
1887 # currently yt-dlp cannot decode the playerVerificationChallenge as Akamai uses Adobe Alchemy
1888 akamai_pv = manifest.find('{http://ns.adobe.com/f4m/1.0}pv-2.0')
1889 if akamai_pv is not None and ';' in akamai_pv.text:
1890 playerVerificationChallenge = akamai_pv.text.split(';')[0]
1891 if playerVerificationChallenge.strip() != '':
1895 manifest_version = '1.0'
1896 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/1.0}media')
1898 manifest_version = '2.0'
1899 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/2.0}media')
1900 # Remove unsupported DRM protected media from final formats
1901 # rendition (see https://github.com/ytdl-org/youtube-dl/issues/8573).
1902 media_nodes = remove_encrypted_media(media_nodes)
1906 manifest_base_url = get_base_url(manifest)
1908 bootstrap_info = xpath_element(
1909 manifest, ['{http://ns.adobe.com/f4m/1.0}bootstrapInfo', '{http://ns.adobe.com/f4m/2.0}bootstrapInfo'],
1910 'bootstrap info', default=None)
1913 mime_type = xpath_text(
1914 manifest, ['{http://ns.adobe.com/f4m/1.0}mimeType', '{http://ns.adobe.com/f4m/2.0}mimeType'],
1915 'base URL', default=None)
1916 if mime_type and mime_type.startswith('audio/'):
1919 for i, media_el in enumerate(media_nodes):
1920 tbr = int_or_none(media_el.attrib.get('bitrate'))
1921 width = int_or_none(media_el.attrib.get('width'))
1922 height = int_or_none(media_el.attrib.get('height'))
1923 format_id = join_nonempty(f4m_id, tbr or i)
1924 # If <bootstrapInfo> is present, the specified f4m is a
1925 # stream-level manifest, and only set-level manifests may refer to
1926 # external resources. See section 11.4 and section 4 of F4M spec
1927 if bootstrap_info is None:
1929 # @href is introduced in 2.0, see section 11.6 of F4M spec
1930 if manifest_version == '2.0':
1931 media_url = media_el.attrib.get('href')
1932 if media_url is None:
1933 media_url = media_el.attrib.get('url')
1937 media_url if media_url.startswith('http://') or media_url.startswith('https://')
1938 else ((manifest_base_url or '/'.join(manifest_url.split('/')[:-1])) + '/' + media_url))
1939 # If media_url is itself a f4m manifest do the recursive extraction
1940 # since bitrates in parent manifest (this one) and media_url manifest
1941 # may differ leading to inability to resolve the format by requested
1942 # bitrate in f4m downloader
1943 ext = determine_ext(manifest_url)
1945 f4m_formats = self._extract_f4m_formats(
1946 manifest_url, video_id, preference=preference, quality=quality, f4m_id=f4m_id,
1947 transform_source=transform_source, fatal=fatal)
1948 # Sometimes stream-level manifest contains single media entry that
1949 # does not contain any quality metadata (e.g. http://matchtv.ru/#live-player).
1950 # At the same time parent's media entry in set-level manifest may
1951 # contain it. We will copy it from parent in such cases.
1952 if len(f4m_formats) == 1:
1955 'tbr': f.get('tbr') or tbr,
1956 'width': f.get('width') or width,
1957 'height': f.get('height') or height,
1958 'format_id': f.get('format_id') if not tbr else format_id,
1961 formats.extend(f4m_formats)
1964 formats.extend(self._extract_m3u8_formats(
1965 manifest_url, video_id, 'mp4', preference=preference,
1966 quality=quality, m3u8_id=m3u8_id, fatal=fatal))
1969 'format_id': format_id,
1970 'url': manifest_url,
1971 'manifest_url': manifest_url,
1972 'ext': 'flv' if bootstrap_info is not None else None,
1978 'preference': preference,
1983 def _m3u8_meta_format(self, m3u8_url, ext=None, preference=None, quality=None, m3u8_id=None):
1985 'format_id': join_nonempty(m3u8_id, 'meta'),
1989 'preference': preference - 100 if preference else -100,
1991 'resolution': 'multiple',
1992 'format_note': 'Quality selection URL',
1995 def _report_ignoring_subs(self, name):
1996 self.report_warning(bug_reports_message(
1997 f'Ignoring subtitle tracks found in the {name} manifest; '
1998 'if any subtitle tracks are missing,'
2001 def _extract_m3u8_formats(self, *args, **kwargs):
2002 fmts, subs = self._extract_m3u8_formats_and_subtitles(*args, **kwargs)
2004 self._report_ignoring_subs('HLS')
2007 def _extract_m3u8_formats_and_subtitles(
2008 self, m3u8_url, video_id, ext=None, entry_protocol='m3u8_native',
2009 preference=None, quality=None, m3u8_id=None, note=None,
2010 errnote=None, fatal=True, live=False, data=None, headers={},
2013 res = self._download_webpage_handle(
2015 note='Downloading m3u8 information' if note is None else note,
2016 errnote='Failed to download m3u8 information' if errnote is None else errnote,
2017 fatal=fatal, data=data, headers=headers, query=query)
2022 m3u8_doc, urlh = res
2023 m3u8_url = urlh.geturl()
2025 return self._parse_m3u8_formats_and_subtitles(
2026 m3u8_doc, m3u8_url, ext=ext, entry_protocol=entry_protocol,
2027 preference=preference, quality=quality, m3u8_id=m3u8_id,
2028 note=note, errnote=errnote, fatal=fatal, live=live, data=data,
2029 headers=headers, query=query, video_id=video_id)
2031 def _parse_m3u8_formats_and_subtitles(
2032 self, m3u8_doc, m3u8_url, ext=None, entry_protocol='m3u8_native',
2033 preference=None, quality=None, m3u8_id=None, live=False, note=None,
2034 errnote=None, fatal=True, data=None, headers={}, query={},
2036 formats, subtitles = [], {}
2038 if '#EXT-X-FAXS-CM:' in m3u8_doc: # Adobe Flash Access
2039 return formats, subtitles
2041 has_drm = re.search(r'#EXT-X-(?:SESSION-)?KEY:.*?URI="skd
://', m3u8_doc)
2043 def format_url(url):
2044 return url if re.match(r'^https?
://', url) else compat_urlparse.urljoin(m3u8_url, url)
2046 if self.get_param('hls_split_discontinuity
', False):
2047 def _extract_m3u8_playlist_indices(manifest_url=None, m3u8_doc=None):
2049 if not manifest_url:
2051 m3u8_doc = self._download_webpage(
2052 manifest_url, video_id, fatal=fatal, data=data, headers=headers,
2053 note=False, errnote='Failed to download m3u8 playlist information
')
2054 if m3u8_doc is False:
2056 return range(1 + sum(line.startswith('#EXT-X-DISCONTINUITY') for line in m3u8_doc.splitlines()))
2059 def _extract_m3u8_playlist_indices(*args
, **kwargs
):
2063 # 1. https://tools.ietf.org/html/draft-pantos-http-live-streaming-21
2064 # 2. https://github.com/ytdl-org/youtube-dl/issues/12211
2065 # 3. https://github.com/ytdl-org/youtube-dl/issues/18923
2067 # We should try extracting formats only from master playlists [1, 4.3.4],
2068 # i.e. playlists that describe available qualities. On the other hand
2069 # media playlists [1, 4.3.3] should be returned as is since they contain
2070 # just the media without qualities renditions.
2071 # Fortunately, master playlist can be easily distinguished from media
2072 # playlist based on particular tags availability. As of [1, 4.3.3, 4.3.4]
2073 # master playlist tags MUST NOT appear in a media playlist and vice versa.
2074 # As of [1, 4.3.3.1] #EXT-X-TARGETDURATION tag is REQUIRED for every
2075 # media playlist and MUST NOT appear in master playlist thus we can
2076 # clearly detect media playlist with this criterion.
2078 if '#EXT-X-TARGETDURATION' in m3u8_doc
: # media playlist, return as is
2080 'format_id': join_nonempty(m3u8_id
, idx
),
2081 'format_index': idx
,
2084 'protocol': entry_protocol
,
2085 'preference': preference
,
2088 } for idx
in _extract_m3u8_playlist_indices(m3u8_doc
=m3u8_doc
)]
2090 return formats
, subtitles
2093 last_stream_inf
= {}
2095 def extract_media(x_media_line
):
2096 media
= parse_m3u8_attributes(x_media_line
)
2097 # As per [1, 4.3.4.1] TYPE, GROUP-ID and NAME are REQUIRED
2098 media_type
, group_id
, name
= media
.get('TYPE'), media
.get('GROUP-ID'), media
.get('NAME')
2099 if not (media_type
and group_id
and name
):
2101 groups
.setdefault(group_id
, []).append(media
)
2102 # <https://tools.ietf.org/html/rfc8216#section-4.3.4.1>
2103 if media_type
== 'SUBTITLES':
2104 # According to RFC 8216 §4.3.4.2.1, URI is REQUIRED in the
2105 # EXT-X-MEDIA tag if the media type is SUBTITLES.
2106 # However, lack of URI has been spotted in the wild.
2107 # e.g. NebulaIE; see https://github.com/yt-dlp/yt-dlp/issues/339
2108 if not media
.get('URI'):
2110 url
= format_url(media
['URI'])
2113 'ext': determine_ext(url
),
2115 if sub_info
['ext'] == 'm3u8':
2116 # Per RFC 8216 §3.1, the only possible subtitle format m3u8
2117 # files may contain is WebVTT:
2118 # <https://tools.ietf.org/html/rfc8216#section-3.1>
2119 sub_info
['ext'] = 'vtt'
2120 sub_info
['protocol'] = 'm3u8_native'
2121 lang
= media
.get('LANGUAGE') or 'und'
2122 subtitles
.setdefault(lang
, []).append(sub_info
)
2123 if media_type
not in ('VIDEO', 'AUDIO'):
2125 media_url
= media
.get('URI')
2127 manifest_url
= format_url(media_url
)
2129 'format_id': join_nonempty(m3u8_id
, group_id
, name
, idx
),
2130 'format_note': name
,
2131 'format_index': idx
,
2132 'url': manifest_url
,
2133 'manifest_url': m3u8_url
,
2134 'language': media
.get('LANGUAGE'),
2136 'protocol': entry_protocol
,
2137 'preference': preference
,
2139 'vcodec': 'none' if media_type
== 'AUDIO' else None,
2140 } for idx
in _extract_m3u8_playlist_indices(manifest_url
))
2142 def build_stream_name():
2143 # Despite specification does not mention NAME attribute for
2144 # EXT-X-STREAM-INF tag it still sometimes may be present (see [1]
2145 # or vidio test in TestInfoExtractor.test_parse_m3u8_formats)
2146 # 1. http://www.vidio.com/watch/165683-dj_ambred-booyah-live-2015
2147 stream_name
= last_stream_inf
.get('NAME')
2150 # If there is no NAME in EXT-X-STREAM-INF it will be obtained
2151 # from corresponding rendition group
2152 stream_group_id
= last_stream_inf
.get('VIDEO')
2153 if not stream_group_id
:
2155 stream_group
= groups
.get(stream_group_id
)
2156 if not stream_group
:
2157 return stream_group_id
2158 rendition
= stream_group
[0]
2159 return rendition
.get('NAME') or stream_group_id
2161 # parse EXT-X-MEDIA tags before EXT-X-STREAM-INF in order to have the
2162 # chance to detect video only formats when EXT-X-STREAM-INF tags
2163 # precede EXT-X-MEDIA tags in HLS manifest such as [3].
2164 for line
in m3u8_doc
.splitlines():
2165 if line
.startswith('#EXT-X-MEDIA:'):
2168 for line
in m3u8_doc
.splitlines():
2169 if line
.startswith('#EXT-X-STREAM-INF:'):
2170 last_stream_inf
= parse_m3u8_attributes(line
)
2171 elif line
.startswith('#') or not line
.strip():
2174 tbr
= float_or_none(
2175 last_stream_inf
.get('AVERAGE-BANDWIDTH')
2176 or last_stream_inf
.get('BANDWIDTH'), scale
=1000)
2177 manifest_url
= format_url(line
.strip())
2179 for idx
in _extract_m3u8_playlist_indices(manifest_url
):
2180 format_id
= [m3u8_id
, None, idx
]
2181 # Bandwidth of live streams may differ over time thus making
2182 # format_id unpredictable. So it's better to keep provided
2185 stream_name
= build_stream_name()
2186 format_id
[1] = stream_name
or '%d' % (tbr
or len(formats
))
2188 'format_id': join_nonempty(*format_id
),
2189 'format_index': idx
,
2190 'url': manifest_url
,
2191 'manifest_url': m3u8_url
,
2194 'fps': float_or_none(last_stream_inf
.get('FRAME-RATE')),
2195 'protocol': entry_protocol
,
2196 'preference': preference
,
2199 resolution
= last_stream_inf
.get('RESOLUTION')
2201 mobj
= re
.search(r
'(?P<width>\d+)[xX](?P<height>\d+)', resolution
)
2203 f
['width'] = int(mobj
.group('width'))
2204 f
['height'] = int(mobj
.group('height'))
2205 # Unified Streaming Platform
2207 r
'audio.*?(?:%3D|=)(\d+)(?:-video.*?(?:%3D|=)(\d+))?', f
['url'])
2209 abr
, vbr
= mobj
.groups()
2210 abr
, vbr
= float_or_none(abr
, 1000), float_or_none(vbr
, 1000)
2215 codecs
= parse_codecs(last_stream_inf
.get('CODECS'))
2217 audio_group_id
= last_stream_inf
.get('AUDIO')
2218 # As per [1, 4.3.4.1.1] any EXT-X-STREAM-INF tag which
2219 # references a rendition group MUST have a CODECS attribute.
2220 # However, this is not always respected, for example, [2]
2221 # contains EXT-X-STREAM-INF tag which references AUDIO
2222 # rendition group but does not have CODECS and despite
2223 # referencing an audio group it represents a complete
2224 # (with audio and video) format. So, for such cases we will
2225 # ignore references to rendition groups and treat them
2226 # as complete formats.
2227 if audio_group_id
and codecs
and f
.get('vcodec') != 'none':
2228 audio_group
= groups
.get(audio_group_id
)
2229 if audio_group
and audio_group
[0].get('URI'):
2230 # TODO: update acodec for audio only formats with
2232 f
['acodec'] = 'none'
2233 if not f
.get('ext'):
2234 f
['ext'] = 'm4a' if f
.get('vcodec') == 'none' else 'mp4'
2238 progressive_uri
= last_stream_inf
.get('PROGRESSIVE-URI')
2241 del http_f
['manifest_url']
2243 'format_id': f
['format_id'].replace('hls-', 'http-'),
2245 'url': progressive_uri
,
2247 formats
.append(http_f
)
2249 last_stream_inf
= {}
2250 return formats
, subtitles
2252 def _extract_m3u8_vod_duration(
2253 self
, m3u8_vod_url
, video_id
, note
=None, errnote
=None, data
=None, headers
={}, query={}
):
2255 m3u8_vod
= self
._download
_webpage
(
2256 m3u8_vod_url
, video_id
,
2257 note
='Downloading m3u8 VOD manifest' if note
is None else note
,
2258 errnote
='Failed to download VOD manifest' if errnote
is None else errnote
,
2259 fatal
=False, data
=data
, headers
=headers
, query
=query
)
2261 return self
._parse
_m
3u8_vod
_duration
(m3u8_vod
or '', video_id
)
2263 def _parse_m3u8_vod_duration(self
, m3u8_vod
, video_id
):
2264 if '#EXT-X-PLAYLIST-TYPE:VOD' not in m3u8_vod
:
2268 float(line
[len('#EXTINF:'):].split(',')[0])
2269 for line
in m3u8_vod
.splitlines() if line
.startswith('#EXTINF:'))) or None
2272 def _xpath_ns(path
, namespace
=None):
2276 for c
in path
.split('/'):
2277 if not c
or c
== '.':
2280 out
.append('{%s}%s' % (namespace
, c
))
2281 return '/'.join(out
)
2283 def _extract_smil_formats_and_subtitles(self
, smil_url
, video_id
, fatal
=True, f4m_params
=None, transform_source
=None):
2284 smil
= self
._download
_smil
(smil_url
, video_id
, fatal
=fatal
, transform_source
=transform_source
)
2290 namespace
= self
._parse
_smil
_namespace
(smil
)
2292 fmts
= self
._parse
_smil
_formats
(
2293 smil
, smil_url
, video_id
, namespace
=namespace
, f4m_params
=f4m_params
)
2294 subs
= self
._parse
_smil
_subtitles
(
2295 smil
, namespace
=namespace
)
2299 def _extract_smil_formats(self
, *args
, **kwargs
):
2300 fmts
, subs
= self
._extract
_smil
_formats
_and
_subtitles
(*args
, **kwargs
)
2302 self
._report
_ignoring
_subs
('SMIL')
2305 def _extract_smil_info(self
, smil_url
, video_id
, fatal
=True, f4m_params
=None):
2306 smil
= self
._download
_smil
(smil_url
, video_id
, fatal
=fatal
)
2309 return self
._parse
_smil
(smil
, smil_url
, video_id
, f4m_params
=f4m_params
)
2311 def _download_smil(self
, smil_url
, video_id
, fatal
=True, transform_source
=None):
2312 return self
._download
_xml
(
2313 smil_url
, video_id
, 'Downloading SMIL file',
2314 'Unable to download SMIL file', fatal
=fatal
, transform_source
=transform_source
)
2316 def _parse_smil(self
, smil
, smil_url
, video_id
, f4m_params
=None):
2317 namespace
= self
._parse
_smil
_namespace
(smil
)
2319 formats
= self
._parse
_smil
_formats
(
2320 smil
, smil_url
, video_id
, namespace
=namespace
, f4m_params
=f4m_params
)
2321 subtitles
= self
._parse
_smil
_subtitles
(smil
, namespace
=namespace
)
2323 video_id
= os
.path
.splitext(url_basename(smil_url
))[0]
2327 for meta
in smil
.findall(self
._xpath
_ns
('./head/meta', namespace
)):
2328 name
= meta
.attrib
.get('name')
2329 content
= meta
.attrib
.get('content')
2330 if not name
or not content
:
2332 if not title
and name
== 'title':
2334 elif not description
and name
in ('description', 'abstract'):
2335 description
= content
2336 elif not upload_date
and name
== 'date':
2337 upload_date
= unified_strdate(content
)
2340 'id': image
.get('type'),
2341 'url': image
.get('src'),
2342 'width': int_or_none(image
.get('width')),
2343 'height': int_or_none(image
.get('height')),
2344 } for image
in smil
.findall(self
._xpath
_ns
('.//image', namespace
)) if image
.get('src')]
2348 'title': title
or video_id
,
2349 'description': description
,
2350 'upload_date': upload_date
,
2351 'thumbnails': thumbnails
,
2353 'subtitles': subtitles
,
2356 def _parse_smil_namespace(self
, smil
):
2357 return self
._search
_regex
(
2358 r
'(?i)^{([^}]+)?}smil$', smil
.tag
, 'namespace', default
=None)
2360 def _parse_smil_formats(self
, smil
, smil_url
, video_id
, namespace
=None, f4m_params
=None, transform_rtmp_url
=None):
2362 for meta
in smil
.findall(self
._xpath
_ns
('./head/meta', namespace
)):
2363 b
= meta
.get('base') or meta
.get('httpBase')
2375 media
= smil
.findall(self
._xpath
_ns
('.//video', namespace
)) + smil
.findall(self
._xpath
_ns
('.//audio', namespace
))
2376 for medium
in media
:
2377 src
= medium
.get('src')
2378 if not src
or src
in srcs
:
2382 bitrate
= float_or_none(medium
.get('system-bitrate') or medium
.get('systemBitrate'), 1000)
2383 filesize
= int_or_none(medium
.get('size') or medium
.get('fileSize'))
2384 width
= int_or_none(medium
.get('width'))
2385 height
= int_or_none(medium
.get('height'))
2386 proto
= medium
.get('proto')
2387 ext
= medium
.get('ext')
2388 src_ext
= determine_ext(src
)
2389 streamer
= medium
.get('streamer') or base
2391 if proto
== 'rtmp' or streamer
.startswith('rtmp'):
2397 'format_id': 'rtmp-%d' % (rtmp_count
if bitrate
is None else bitrate
),
2399 'filesize': filesize
,
2403 if transform_rtmp_url
:
2404 streamer
, src
= transform_rtmp_url(streamer
, src
)
2405 formats
[-1].update({
2411 src_url
= src
if src
.startswith('http') else compat_urlparse
.urljoin(base
, src
)
2412 src_url
= src_url
.strip()
2414 if proto
== 'm3u8' or src_ext
== 'm3u8':
2415 m3u8_formats
= self
._extract
_m
3u8_formats
(
2416 src_url
, video_id
, ext
or 'mp4', m3u8_id
='hls', fatal
=False)
2417 if len(m3u8_formats
) == 1:
2419 m3u8_formats
[0].update({
2420 'format_id': 'hls-%d' % (m3u8_count
if bitrate
is None else bitrate
),
2425 formats
.extend(m3u8_formats
)
2426 elif src_ext
== 'f4m':
2431 'plugin': 'flowplayer-3.2.0.1',
2433 f4m_url
+= '&' if '?' in f4m_url
else '?'
2434 f4m_url
+= compat_urllib_parse_urlencode(f4m_params
)
2435 formats
.extend(self
._extract
_f
4m
_formats
(f4m_url
, video_id
, f4m_id
='hds', fatal
=False))
2436 elif src_ext
== 'mpd':
2437 formats
.extend(self
._extract
_mpd
_formats
(
2438 src_url
, video_id
, mpd_id
='dash', fatal
=False))
2439 elif re
.search(r
'\.ism/[Mm]anifest', src_url
):
2440 formats
.extend(self
._extract
_ism
_formats
(
2441 src_url
, video_id
, ism_id
='mss', fatal
=False))
2442 elif src_url
.startswith('http') and self
._is
_valid
_url
(src
, video_id
):
2446 'ext': ext
or src_ext
or 'flv',
2447 'format_id': 'http-%d' % (bitrate
or http_count
),
2449 'filesize': filesize
,
2454 for medium
in smil
.findall(self
._xpath
_ns
('.//imagestream', namespace
)):
2455 src
= medium
.get('src')
2456 if not src
or src
in srcs
:
2462 'format_id': 'imagestream-%d' % (imgs_count
),
2464 'ext': mimetype2ext(medium
.get('type')),
2467 'width': int_or_none(medium
.get('width')),
2468 'height': int_or_none(medium
.get('height')),
2469 'format_note': 'SMIL storyboards',
2474 def _parse_smil_subtitles(self
, smil
, namespace
=None, subtitles_lang
='en'):
2477 for num
, textstream
in enumerate(smil
.findall(self
._xpath
_ns
('.//textstream', namespace
))):
2478 src
= textstream
.get('src')
2479 if not src
or src
in urls
:
2482 ext
= textstream
.get('ext') or mimetype2ext(textstream
.get('type')) or determine_ext(src
)
2483 lang
= textstream
.get('systemLanguage') or textstream
.get('systemLanguageName') or textstream
.get('lang') or subtitles_lang
2484 subtitles
.setdefault(lang
, []).append({
2490 def _extract_xspf_playlist(self
, xspf_url
, playlist_id
, fatal
=True):
2491 xspf
= self
._download
_xml
(
2492 xspf_url
, playlist_id
, 'Downloading xpsf playlist',
2493 'Unable to download xspf manifest', fatal
=fatal
)
2496 return self
._parse
_xspf
(
2497 xspf
, playlist_id
, xspf_url
=xspf_url
,
2498 xspf_base_url
=base_url(xspf_url
))
2500 def _parse_xspf(self
, xspf_doc
, playlist_id
, xspf_url
=None, xspf_base_url
=None):
2502 'xspf': 'http://xspf.org/ns/0/',
2503 's1': 'http://static.streamone.nl/player/ns/0',
2507 for track
in xspf_doc
.findall(xpath_with_ns('./xspf:trackList/xspf:track', NS_MAP
)):
2509 track
, xpath_with_ns('./xspf:title', NS_MAP
), 'title', default
=playlist_id
)
2510 description
= xpath_text(
2511 track
, xpath_with_ns('./xspf:annotation', NS_MAP
), 'description')
2512 thumbnail
= xpath_text(
2513 track
, xpath_with_ns('./xspf:image', NS_MAP
), 'thumbnail')
2514 duration
= float_or_none(
2515 xpath_text(track
, xpath_with_ns('./xspf:duration', NS_MAP
), 'duration'), 1000)
2518 for location
in track
.findall(xpath_with_ns('./xspf:location', NS_MAP
)):
2519 format_url
= urljoin(xspf_base_url
, location
.text
)
2524 'manifest_url': xspf_url
,
2525 'format_id': location
.get(xpath_with_ns('s1:label', NS_MAP
)),
2526 'width': int_or_none(location
.get(xpath_with_ns('s1:width', NS_MAP
))),
2527 'height': int_or_none(location
.get(xpath_with_ns('s1:height', NS_MAP
))),
2529 self
._sort
_formats
(formats
)
2534 'description': description
,
2535 'thumbnail': thumbnail
,
2536 'duration': duration
,
2541 def _extract_mpd_formats(self
, *args
, **kwargs
):
2542 fmts
, subs
= self
._extract
_mpd
_formats
_and
_subtitles
(*args
, **kwargs
)
2544 self
._report
_ignoring
_subs
('DASH')
2547 def _extract_mpd_formats_and_subtitles(
2548 self
, mpd_url
, video_id
, mpd_id
=None, note
=None, errnote
=None,
2549 fatal
=True, data
=None, headers
={}, query={}
):
2550 res
= self
._download
_xml
_handle
(
2552 note
='Downloading MPD manifest' if note
is None else note
,
2553 errnote
='Failed to download MPD manifest' if errnote
is None else errnote
,
2554 fatal
=fatal
, data
=data
, headers
=headers
, query
=query
)
2560 mpd_base_url
= base_url(urlh
.geturl())
2562 return self
._parse
_mpd
_formats
_and
_subtitles
(
2563 mpd_doc
, mpd_id
, mpd_base_url
, mpd_url
)
2565 def _parse_mpd_formats(self
, *args
, **kwargs
):
2566 fmts
, subs
= self
._parse
_mpd
_formats
_and
_subtitles
(*args
, **kwargs
)
2568 self
._report
_ignoring
_subs
('DASH')
2571 def _parse_mpd_formats_and_subtitles(
2572 self
, mpd_doc
, mpd_id
=None, mpd_base_url
='', mpd_url
=None):
2574 Parse formats from MPD manifest.
2576 1. MPEG-DASH Standard, ISO/IEC 23009-1:2014(E),
2577 http://standards.iso.org/ittf/PubliclyAvailableStandards/c065274_ISO_IEC_23009-1_2014.zip
2578 2. https://en.wikipedia.org/wiki/Dynamic_Adaptive_Streaming_over_HTTP
2580 if not self
.get_param('dynamic_mpd', True):
2581 if mpd_doc
.get('type') == 'dynamic':
2584 namespace
= self
._search
_regex
(r
'(?i)^{([^}]+)?}MPD$', mpd_doc
.tag
, 'namespace', default
=None)
2587 return self
._xpath
_ns
(path
, namespace
)
2589 def is_drm_protected(element
):
2590 return element
.find(_add_ns('ContentProtection')) is not None
2592 def extract_multisegment_info(element
, ms_parent_info
):
2593 ms_info
= ms_parent_info
.copy()
2595 # As per [1, 5.3.9.2.2] SegmentList and SegmentTemplate share some
2596 # common attributes and elements. We will only extract relevant
2598 def extract_common(source
):
2599 segment_timeline
= source
.find(_add_ns('SegmentTimeline'))
2600 if segment_timeline
is not None:
2601 s_e
= segment_timeline
.findall(_add_ns('S'))
2603 ms_info
['total_number'] = 0
2606 r
= int(s
.get('r', 0))
2607 ms_info
['total_number'] += 1 + r
2608 ms_info
['s'].append({
2609 't': int(s
.get('t', 0)),
2610 # @d is mandatory (see [1, 5.3.9.6.2, Table 17, page 60])
2611 'd': int(s
.attrib
['d']),
2614 start_number
= source
.get('startNumber')
2616 ms_info
['start_number'] = int(start_number
)
2617 timescale
= source
.get('timescale')
2619 ms_info
['timescale'] = int(timescale
)
2620 segment_duration
= source
.get('duration')
2621 if segment_duration
:
2622 ms_info
['segment_duration'] = float(segment_duration
)
2624 def extract_Initialization(source
):
2625 initialization
= source
.find(_add_ns('Initialization'))
2626 if initialization
is not None:
2627 ms_info
['initialization_url'] = initialization
.attrib
['sourceURL']
2629 segment_list
= element
.find(_add_ns('SegmentList'))
2630 if segment_list
is not None:
2631 extract_common(segment_list
)
2632 extract_Initialization(segment_list
)
2633 segment_urls_e
= segment_list
.findall(_add_ns('SegmentURL'))
2635 ms_info
['segment_urls'] = [segment
.attrib
['media'] for segment
in segment_urls_e
]
2637 segment_template
= element
.find(_add_ns('SegmentTemplate'))
2638 if segment_template
is not None:
2639 extract_common(segment_template
)
2640 media
= segment_template
.get('media')
2642 ms_info
['media'] = media
2643 initialization
= segment_template
.get('initialization')
2645 ms_info
['initialization'] = initialization
2647 extract_Initialization(segment_template
)
2650 mpd_duration
= parse_duration(mpd_doc
.get('mediaPresentationDuration'))
2651 formats
, subtitles
= [], {}
2652 stream_numbers
= {'audio': 0, 'video': 0}
2653 for period
in mpd_doc
.findall(_add_ns('Period')):
2654 period_duration
= parse_duration(period
.get('duration')) or mpd_duration
2655 period_ms_info
= extract_multisegment_info(period
, {
2659 for adaptation_set
in period
.findall(_add_ns('AdaptationSet')):
2660 adaption_set_ms_info
= extract_multisegment_info(adaptation_set
, period_ms_info
)
2661 for representation
in adaptation_set
.findall(_add_ns('Representation')):
2662 representation_attrib
= adaptation_set
.attrib
.copy()
2663 representation_attrib
.update(representation
.attrib
)
2664 # According to [1, 5.3.7.2, Table 9, page 41], @mimeType is mandatory
2665 mime_type
= representation_attrib
['mimeType']
2666 content_type
= representation_attrib
.get('contentType', mime_type
.split('/')[0])
2668 codecs
= representation_attrib
.get('codecs', '')
2669 if content_type
not in ('video', 'audio', 'text'):
2670 if mime_type
== 'image/jpeg':
2671 content_type
= mime_type
2672 elif codecs
.split('.')[0] == 'stpp':
2673 content_type
= 'text'
2674 elif mimetype2ext(mime_type
) in ('tt', 'dfxp', 'ttml', 'xml', 'json'):
2675 content_type
= 'text'
2677 self
.report_warning('Unknown MIME type %s in DASH manifest' % mime_type
)
2681 for element
in (representation
, adaptation_set
, period
, mpd_doc
):
2682 base_url_e
= element
.find(_add_ns('BaseURL'))
2683 if base_url_e
is not None:
2684 base_url
= base_url_e
.text
+ base_url
2685 if re
.match(r
'^https?://', base_url
):
2687 if mpd_base_url
and base_url
.startswith('/'):
2688 base_url
= compat_urlparse
.urljoin(mpd_base_url
, base_url
)
2689 elif mpd_base_url
and not re
.match(r
'^https?://', base_url
):
2690 if not mpd_base_url
.endswith('/'):
2692 base_url
= mpd_base_url
+ base_url
2693 representation_id
= representation_attrib
.get('id')
2694 lang
= representation_attrib
.get('lang')
2695 url_el
= representation
.find(_add_ns('BaseURL'))
2696 filesize
= int_or_none(url_el
.attrib
.get('{http://youtube.com/yt/2012/10/10}contentLength') if url_el
is not None else None)
2697 bandwidth
= int_or_none(representation_attrib
.get('bandwidth'))
2698 if representation_id
is not None:
2699 format_id
= representation_id
2701 format_id
= content_type
2703 format_id
= mpd_id
+ '-' + format_id
2704 if content_type
in ('video', 'audio'):
2706 'format_id': format_id
,
2707 'manifest_url': mpd_url
,
2708 'ext': mimetype2ext(mime_type
),
2709 'width': int_or_none(representation_attrib
.get('width')),
2710 'height': int_or_none(representation_attrib
.get('height')),
2711 'tbr': float_or_none(bandwidth
, 1000),
2712 'asr': int_or_none(representation_attrib
.get('audioSamplingRate')),
2713 'fps': int_or_none(representation_attrib
.get('frameRate')),
2714 'language': lang
if lang
not in ('mul', 'und', 'zxx', 'mis') else None,
2715 'format_note': 'DASH %s' % content_type
,
2716 'filesize': filesize
,
2717 'container': mimetype2ext(mime_type
) + '_dash',
2718 'manifest_stream_number': stream_numbers
[content_type
]
2720 f
.update(parse_codecs(codecs
))
2721 stream_numbers
[content_type
] += 1
2722 elif content_type
== 'text':
2724 'ext': mimetype2ext(mime_type
),
2725 'manifest_url': mpd_url
,
2726 'filesize': filesize
,
2728 elif content_type
== 'image/jpeg':
2729 # See test case in VikiIE
2730 # https://www.viki.com/videos/1175236v-choosing-spouse-by-lottery-episode-1
2732 'format_id': format_id
,
2734 'manifest_url': mpd_url
,
2735 'format_note': 'DASH storyboards (jpeg)',
2739 if is_drm_protected(adaptation_set
) or is_drm_protected(representation
):
2741 representation_ms_info
= extract_multisegment_info(representation
, adaption_set_ms_info
)
2743 def prepare_template(template_name
, identifiers
):
2744 tmpl
= representation_ms_info
[template_name
]
2745 # First of, % characters outside $...$ templates
2746 # must be escaped by doubling for proper processing
2747 # by % operator string formatting used further (see
2748 # https://github.com/ytdl-org/youtube-dl/issues/16867).
2754 in_template
= not in_template
2755 elif c
== '%' and not in_template
:
2757 # Next, $...$ templates are translated to their
2758 # %(...) counterparts to be used with % operator
2759 if representation_id
is not None:
2760 t
= t
.replace('$RepresentationID$', representation_id
)
2761 t
= re
.sub(r
'\$(%s)\$' % '|'.join(identifiers
), r
'%(\1)d', t
)
2762 t
= re
.sub(r
'\$(%s)%%([^$]+)\$' % '|'.join(identifiers
), r
'%(\1)\2', t
)
2763 t
.replace('$$', '$')
2766 # @initialization is a regular template like @media one
2767 # so it should be handled just the same way (see
2768 # https://github.com/ytdl-org/youtube-dl/issues/11605)
2769 if 'initialization' in representation_ms_info
:
2770 initialization_template
= prepare_template(
2772 # As per [1, 5.3.9.4.2, Table 15, page 54] $Number$ and
2773 # $Time$ shall not be included for @initialization thus
2774 # only $Bandwidth$ remains
2776 representation_ms_info
['initialization_url'] = initialization_template
% {
2777 'Bandwidth': bandwidth
,
2780 def location_key(location
):
2781 return 'url' if re
.match(r
'^https?://', location
) else 'path'
2783 if 'segment_urls' not in representation_ms_info
and 'media' in representation_ms_info
:
2785 media_template
= prepare_template('media', ('Number', 'Bandwidth', 'Time'))
2786 media_location_key
= location_key(media_template
)
2788 # As per [1, 5.3.9.4.4, Table 16, page 55] $Number$ and $Time$
2789 # can't be used at the same time
2790 if '%(Number' in media_template
and 's' not in representation_ms_info
:
2791 segment_duration
= None
2792 if 'total_number' not in representation_ms_info
and 'segment_duration' in representation_ms_info
:
2793 segment_duration
= float_or_none(representation_ms_info
['segment_duration'], representation_ms_info
['timescale'])
2794 representation_ms_info
['total_number'] = int(math
.ceil(float(period_duration
) / segment_duration
))
2795 representation_ms_info
['fragments'] = [{
2796 media_location_key
: media_template
% {
2797 'Number': segment_number
,
2798 'Bandwidth': bandwidth
,
2800 'duration': segment_duration
,
2801 } for segment_number
in range(
2802 representation_ms_info
['start_number'],
2803 representation_ms_info
['total_number'] + representation_ms_info
['start_number'])]
2805 # $Number*$ or $Time$ in media template with S list available
2806 # Example $Number*$: http://www.svtplay.se/klipp/9023742/stopptid-om-bjorn-borg
2807 # Example $Time$: https://play.arkena.com/embed/avp/v2/player/media/b41dda37-d8e7-4d3f-b1b5-9a9db578bdfe/1/129411
2808 representation_ms_info
['fragments'] = []
2811 segment_number
= representation_ms_info
['start_number']
2813 def add_segment_url():
2814 segment_url
= media_template
% {
2815 'Time': segment_time
,
2816 'Bandwidth': bandwidth
,
2817 'Number': segment_number
,
2819 representation_ms_info
['fragments'].append({
2820 media_location_key
: segment_url
,
2821 'duration': float_or_none(segment_d
, representation_ms_info
['timescale']),
2824 for num
, s
in enumerate(representation_ms_info
['s']):
2825 segment_time
= s
.get('t') or segment_time
2829 for r
in range(s
.get('r', 0)):
2830 segment_time
+= segment_d
2833 segment_time
+= segment_d
2834 elif 'segment_urls' in representation_ms_info
and 's' in representation_ms_info
:
2836 # Example: https://www.youtube.com/watch?v=iXZV5uAYMJI
2837 # or any YouTube dashsegments video
2840 timescale
= representation_ms_info
['timescale']
2841 for s
in representation_ms_info
['s']:
2842 duration
= float_or_none(s
['d'], timescale
)
2843 for r
in range(s
.get('r', 0) + 1):
2844 segment_uri
= representation_ms_info
['segment_urls'][segment_index
]
2846 location_key(segment_uri
): segment_uri
,
2847 'duration': duration
,
2850 representation_ms_info
['fragments'] = fragments
2851 elif 'segment_urls' in representation_ms_info
:
2852 # Segment URLs with no SegmentTimeline
2853 # Example: https://www.seznam.cz/zpravy/clanek/cesko-zasahne-vitr-o-sile-vichrice-muze-byt-i-zivotu-nebezpecny-39091
2854 # https://github.com/ytdl-org/youtube-dl/pull/14844
2856 segment_duration
= float_or_none(
2857 representation_ms_info
['segment_duration'],
2858 representation_ms_info
['timescale']) if 'segment_duration' in representation_ms_info
else None
2859 for segment_url
in representation_ms_info
['segment_urls']:
2861 location_key(segment_url
): segment_url
,
2863 if segment_duration
:
2864 fragment
['duration'] = segment_duration
2865 fragments
.append(fragment
)
2866 representation_ms_info
['fragments'] = fragments
2867 # If there is a fragments key available then we correctly recognized fragmented media.
2868 # Otherwise we will assume unfragmented media with direct access. Technically, such
2869 # assumption is not necessarily correct since we may simply have no support for
2870 # some forms of fragmented media renditions yet, but for now we'll use this fallback.
2871 if 'fragments' in representation_ms_info
:
2873 # NB: mpd_url may be empty when MPD manifest is parsed from a string
2874 'url': mpd_url
or base_url
,
2875 'fragment_base_url': base_url
,
2877 'protocol': 'http_dash_segments' if mime_type
!= 'image/jpeg' else 'mhtml',
2879 if 'initialization_url' in representation_ms_info
:
2880 initialization_url
= representation_ms_info
['initialization_url']
2881 if not f
.get('url'):
2882 f
['url'] = initialization_url
2883 f
['fragments'].append({location_key(initialization_url): initialization_url}
)
2884 f
['fragments'].extend(representation_ms_info
['fragments'])
2886 # Assuming direct URL to unfragmented media.
2888 if content_type
in ('video', 'audio') or mime_type
== 'image/jpeg':
2890 elif content_type
== 'text':
2891 subtitles
.setdefault(lang
or 'und', []).append(f
)
2893 return formats
, subtitles
2895 def _extract_ism_formats(self
, *args
, **kwargs
):
2896 fmts
, subs
= self
._extract
_ism
_formats
_and
_subtitles
(*args
, **kwargs
)
2898 self
._report
_ignoring
_subs
('ISM')
2901 def _extract_ism_formats_and_subtitles(self
, ism_url
, video_id
, ism_id
=None, note
=None, errnote
=None, fatal
=True, data
=None, headers
={}, query={}
):
2902 res
= self
._download
_xml
_handle
(
2904 note
='Downloading ISM manifest' if note
is None else note
,
2905 errnote
='Failed to download ISM manifest' if errnote
is None else errnote
,
2906 fatal
=fatal
, data
=data
, headers
=headers
, query
=query
)
2913 return self
._parse
_ism
_formats
_and
_subtitles
(ism_doc
, urlh
.geturl(), ism_id
)
2915 def _parse_ism_formats_and_subtitles(self
, ism_doc
, ism_url
, ism_id
=None):
2917 Parse formats from ISM manifest.
2919 1. [MS-SSTR]: Smooth Streaming Protocol,
2920 https://msdn.microsoft.com/en-us/library/ff469518.aspx
2922 if ism_doc
.get('IsLive') == 'TRUE':
2925 duration
= int(ism_doc
.attrib
['Duration'])
2926 timescale
= int_or_none(ism_doc
.get('TimeScale')) or 10000000
2930 for stream
in ism_doc
.findall('StreamIndex'):
2931 stream_type
= stream
.get('Type')
2932 if stream_type
not in ('video', 'audio', 'text'):
2934 url_pattern
= stream
.attrib
['Url']
2935 stream_timescale
= int_or_none(stream
.get('TimeScale')) or timescale
2936 stream_name
= stream
.get('Name')
2937 stream_language
= stream
.get('Language', 'und')
2938 for track
in stream
.findall('QualityLevel'):
2939 fourcc
= track
.get('FourCC') or ('AACL' if track
.get('AudioTag') == '255' else None)
2940 # TODO: add support for WVC1 and WMAP
2941 if fourcc
not in ('H264', 'AVC1', 'AACL', 'TTML'):
2942 self
.report_warning('%s is not a supported codec' % fourcc
)
2944 tbr
= int(track
.attrib
['Bitrate']) // 1000
2945 # [1] does not mention Width and Height attributes. However,
2946 # they're often present while MaxWidth and MaxHeight are
2947 # missing, so should be used as fallbacks
2948 width
= int_or_none(track
.get('MaxWidth') or track
.get('Width'))
2949 height
= int_or_none(track
.get('MaxHeight') or track
.get('Height'))
2950 sampling_rate
= int_or_none(track
.get('SamplingRate'))
2952 track_url_pattern
= re
.sub(r
'{[Bb]itrate}', track
.attrib
['Bitrate'], url_pattern
)
2953 track_url_pattern
= compat_urlparse
.urljoin(ism_url
, track_url_pattern
)
2959 stream_fragments
= stream
.findall('c')
2960 for stream_fragment_index
, stream_fragment
in enumerate(stream_fragments
):
2961 fragment_ctx
['time'] = int_or_none(stream_fragment
.get('t')) or fragment_ctx
['time']
2962 fragment_repeat
= int_or_none(stream_fragment
.get('r')) or 1
2963 fragment_ctx
['duration'] = int_or_none(stream_fragment
.get('d'))
2964 if not fragment_ctx
['duration']:
2966 next_fragment_time
= int(stream_fragment
[stream_fragment_index
+ 1].attrib
['t'])
2968 next_fragment_time
= duration
2969 fragment_ctx
['duration'] = (next_fragment_time
- fragment_ctx
['time']) / fragment_repeat
2970 for _
in range(fragment_repeat
):
2972 'url': re
.sub(r
'{start[ _]time}', compat_str(fragment_ctx
['time']), track_url_pattern
),
2973 'duration': fragment_ctx
['duration'] / stream_timescale
,
2975 fragment_ctx
['time'] += fragment_ctx
['duration']
2977 if stream_type
== 'text':
2978 subtitles
.setdefault(stream_language
, []).append({
2982 'manifest_url': ism_url
,
2983 'fragments': fragments
,
2984 '_download_params': {
2985 'stream_type': stream_type
,
2986 'duration': duration
,
2987 'timescale': stream_timescale
,
2989 'language': stream_language
,
2990 'codec_private_data': track
.get('CodecPrivateData'),
2993 elif stream_type
in ('video', 'audio'):
2995 'format_id': join_nonempty(ism_id
, stream_name
, tbr
),
2997 'manifest_url': ism_url
,
2998 'ext': 'ismv' if stream_type
== 'video' else 'isma',
3002 'asr': sampling_rate
,
3003 'vcodec': 'none' if stream_type
== 'audio' else fourcc
,
3004 'acodec': 'none' if stream_type
== 'video' else fourcc
,
3006 'fragments': fragments
,
3007 'has_drm': ism_doc
.find('Protection') is not None,
3008 '_download_params': {
3009 'stream_type': stream_type
,
3010 'duration': duration
,
3011 'timescale': stream_timescale
,
3012 'width': width
or 0,
3013 'height': height
or 0,
3015 'language': stream_language
,
3016 'codec_private_data': track
.get('CodecPrivateData'),
3017 'sampling_rate': sampling_rate
,
3018 'channels': int_or_none(track
.get('Channels', 2)),
3019 'bits_per_sample': int_or_none(track
.get('BitsPerSample', 16)),
3020 'nal_unit_length_field': int_or_none(track
.get('NALUnitLengthField', 4)),
3023 return formats
, subtitles
3025 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):
3026 def absolute_url(item_url
):
3027 return urljoin(base_url
, item_url
)
3029 def parse_content_type(content_type
):
3030 if not content_type
:
3032 ctr
= re
.search(r
'(?P<mimetype>[^/]+/[^;]+)(?:;\s*codecs="?(?P<codecs>[^"]+))?', content_type
)
3034 mimetype
, codecs
= ctr
.groups()
3035 f
= parse_codecs(codecs
)
3036 f
['ext'] = mimetype2ext(mimetype
)
3040 def _media_formats(src
, cur_media_type
, type_info
={}):
3041 full_url
= absolute_url(src
)
3042 ext
= type_info
.get('ext') or determine_ext(full_url
)
3044 is_plain_url
= False
3045 formats
= self
._extract
_m
3u8_formats
(
3046 full_url
, video_id
, ext
='mp4',
3047 entry_protocol
=m3u8_entry_protocol
, m3u8_id
=m3u8_id
,
3048 preference
=preference
, quality
=quality
, fatal
=False)
3050 is_plain_url
= False
3051 formats
= self
._extract
_mpd
_formats
(
3052 full_url
, video_id
, mpd_id
=mpd_id
, fatal
=False)
3057 'vcodec': 'none' if cur_media_type
== 'audio' else None,
3059 return is_plain_url
, formats
3062 # amp-video and amp-audio are very similar to their HTML5 counterparts
3063 # so we wll include them right here (see
3064 # https://www.ampproject.org/docs/reference/components/amp-video)
3065 # For dl8-* tags see https://delight-vr.com/documentation/dl8-video/
3066 _MEDIA_TAG_NAME_RE
= r
'(?:(?:amp|dl8(?:-live)?)-)?(video|audio)'
3067 media_tags
= [(media_tag
, media_tag_name
, media_type
, '')
3068 for media_tag
, media_tag_name
, media_type
3069 in re
.findall(r
'(?s)(<(%s)[^>]*/>)' % _MEDIA_TAG_NAME_RE
, webpage
)]
3070 media_tags
.extend(re
.findall(
3071 # We only allow video|audio followed by a whitespace or '>'.
3072 # Allowing more characters may end up in significant slow down (see
3073 # https://github.com/ytdl-org/youtube-dl/issues/11979, example URL:
3074 # http://www.porntrex.com/maps/videositemap.xml).
3075 r
'(?s)(<(?P<tag>%s)(?:\s+[^>]*)?>)(.*?)</(?P=tag)>' % _MEDIA_TAG_NAME_RE
, webpage
))
3076 for media_tag
, _
, media_type
, media_content
in media_tags
:
3081 media_attributes
= extract_attributes(media_tag
)
3082 src
= strip_or_none(media_attributes
.get('src'))
3084 _
, formats
= _media_formats(src
, media_type
)
3085 media_info
['formats'].extend(formats
)
3086 media_info
['thumbnail'] = absolute_url(media_attributes
.get('poster'))
3088 for source_tag
in re
.findall(r
'<source[^>]+>', media_content
):
3089 s_attr
= extract_attributes(source_tag
)
3090 # data-video-src and data-src are non standard but seen
3091 # several times in the wild
3092 src
= strip_or_none(dict_get(s_attr
, ('src', 'data-video-src', 'data-src')))
3095 f
= parse_content_type(s_attr
.get('type'))
3096 is_plain_url
, formats
= _media_formats(src
, media_type
, f
)
3098 # width, height, res, label and title attributes are
3099 # all not standard but seen several times in the wild
3102 for lbl
in ('label', 'title')
3103 if str_or_none(s_attr
.get(lbl
))
3105 width
= int_or_none(s_attr
.get('width'))
3106 height
= (int_or_none(s_attr
.get('height'))
3107 or int_or_none(s_attr
.get('res')))
3108 if not width
or not height
:
3110 resolution
= parse_resolution(lbl
)
3113 width
= width
or resolution
.get('width')
3114 height
= height
or resolution
.get('height')
3116 tbr
= parse_bitrate(lbl
)
3125 'format_id': s_attr
.get('label') or s_attr
.get('title'),
3127 f
.update(formats
[0])
3128 media_info
['formats'].append(f
)
3130 media_info
['formats'].extend(formats
)
3131 for track_tag
in re
.findall(r
'<track[^>]+>', media_content
):
3132 track_attributes
= extract_attributes(track_tag
)
3133 kind
= track_attributes
.get('kind')
3134 if not kind
or kind
in ('subtitles', 'captions'):
3135 src
= strip_or_none(track_attributes
.get('src'))
3138 lang
= track_attributes
.get('srclang') or track_attributes
.get('lang') or track_attributes
.get('label')
3139 media_info
['subtitles'].setdefault(lang
, []).append({
3140 'url': absolute_url(src
),
3142 for f
in media_info
['formats']:
3143 f
.setdefault('http_headers', {})['Referer'] = base_url
3144 if media_info
['formats'] or media_info
['subtitles']:
3145 entries
.append(media_info
)
3148 def _extract_akamai_formats(self
, *args
, **kwargs
):
3149 fmts
, subs
= self
._extract
_akamai
_formats
_and
_subtitles
(*args
, **kwargs
)
3151 self
._report
_ignoring
_subs
('akamai')
3154 def _extract_akamai_formats_and_subtitles(self
, manifest_url
, video_id
, hosts
={}):
3155 signed
= 'hdnea=' in manifest_url
3157 # https://learn.akamai.com/en-us/webhelp/media-services-on-demand/stream-packaging-user-guide/GUID-BE6C0F73-1E06-483B-B0EA-57984B91B7F9.html
3158 manifest_url
= re
.sub(
3159 r
'(?:b=[\d,-]+|(?:__a__|attributes)=off|__b__=\d+)&?',
3160 '', manifest_url
).strip('?')
3165 hdcore_sign
= 'hdcore=3.7.0'
3166 f4m_url
= re
.sub(r
'(https?://[^/]+)/i/', r
'\1/z/', manifest_url
).replace('/master.m3u8', '/manifest.f4m')
3167 hds_host
= hosts
.get('hds')
3169 f4m_url
= re
.sub(r
'(https?://)[^/]+', r
'\1' + hds_host
, f4m_url
)
3170 if 'hdcore=' not in f4m_url
:
3171 f4m_url
+= ('&' if '?' in f4m_url
else '?') + hdcore_sign
3172 f4m_formats
= self
._extract
_f
4m
_formats
(
3173 f4m_url
, video_id
, f4m_id
='hds', fatal
=False)
3174 for entry
in f4m_formats
:
3175 entry
.update({'extra_param_to_segment_url': hdcore_sign}
)
3176 formats
.extend(f4m_formats
)
3178 m3u8_url
= re
.sub(r
'(https?://[^/]+)/z/', r
'\1/i/', manifest_url
).replace('/manifest.f4m', '/master.m3u8')
3179 hls_host
= hosts
.get('hls')
3181 m3u8_url
= re
.sub(r
'(https?://)[^/]+', r
'\1' + hls_host
, m3u8_url
)
3182 m3u8_formats
, m3u8_subtitles
= self
._extract
_m
3u8_formats
_and
_subtitles
(
3183 m3u8_url
, video_id
, 'mp4', 'm3u8_native',
3184 m3u8_id
='hls', fatal
=False)
3185 formats
.extend(m3u8_formats
)
3186 subtitles
= self
._merge
_subtitles
(subtitles
, m3u8_subtitles
)
3188 http_host
= hosts
.get('http')
3189 if http_host
and m3u8_formats
and not signed
:
3190 REPL_REGEX
= r
'https?://[^/]+/i/([^,]+),([^/]+),([^/]+)\.csmil/.+'
3191 qualities
= re
.match(REPL_REGEX
, m3u8_url
).group(2).split(',')
3192 qualities_length
= len(qualities
)
3193 if len(m3u8_formats
) in (qualities_length
, qualities_length
+ 1):
3195 for f
in m3u8_formats
:
3196 if f
['vcodec'] != 'none':
3197 for protocol
in ('http', 'https'):
3199 del http_f
['manifest_url']
3201 REPL_REGEX
, protocol
+ r
'://%s/\g<1>%s\3' % (http_host
, qualities
[i
]), f
['url'])
3203 'format_id': http_f
['format_id'].replace('hls-', protocol
+ '-'),
3205 'protocol': protocol
,
3207 formats
.append(http_f
)
3210 return formats
, subtitles
3212 def _extract_wowza_formats(self
, url
, video_id
, m3u8_entry_protocol
='m3u8_native', skip_protocols
=[]):
3213 query
= compat_urlparse
.urlparse(url
).query
3214 url
= re
.sub(r
'/(?:manifest|playlist|jwplayer)\.(?:m3u8|f4m|mpd|smil)', '', url
)
3216 r
'(?:(?:http|rtmp|rtsp)(?P<s>s)?:)?(?P<url>//[^?]+)', url
)
3217 url_base
= mobj
.group('url')
3218 http_base_url
= '%s%s:%s' % ('http', mobj
.group('s') or '', url_base
)
3221 def manifest_url(manifest
):
3222 m_url
= '%s/%s' % (http_base_url
, manifest
)
3224 m_url
+= '?%s' % query
3227 if 'm3u8' not in skip_protocols
:
3228 formats
.extend(self
._extract
_m
3u8_formats
(
3229 manifest_url('playlist.m3u8'), video_id
, 'mp4',
3230 m3u8_entry_protocol
, m3u8_id
='hls', fatal
=False))
3231 if 'f4m' not in skip_protocols
:
3232 formats
.extend(self
._extract
_f
4m
_formats
(
3233 manifest_url('manifest.f4m'),
3234 video_id
, f4m_id
='hds', fatal
=False))
3235 if 'dash' not in skip_protocols
:
3236 formats
.extend(self
._extract
_mpd
_formats
(
3237 manifest_url('manifest.mpd'),
3238 video_id
, mpd_id
='dash', fatal
=False))
3239 if re
.search(r
'(?:/smil:|\.smil)', url_base
):
3240 if 'smil' not in skip_protocols
:
3241 rtmp_formats
= self
._extract
_smil
_formats
(
3242 manifest_url('jwplayer.smil'),
3243 video_id
, fatal
=False)
3244 for rtmp_format
in rtmp_formats
:
3245 rtsp_format
= rtmp_format
.copy()
3246 rtsp_format
['url'] = '%s/%s' % (rtmp_format
['url'], rtmp_format
['play_path'])
3247 del rtsp_format
['play_path']
3248 del rtsp_format
['ext']
3249 rtsp_format
.update({
3250 'url': rtsp_format
['url'].replace('rtmp://', 'rtsp://'),
3251 'format_id': rtmp_format
['format_id'].replace('rtmp', 'rtsp'),
3254 formats
.extend([rtmp_format
, rtsp_format
])
3256 for protocol
in ('rtmp', 'rtsp'):
3257 if protocol
not in skip_protocols
:
3259 'url': '%s:%s' % (protocol
, url_base
),
3260 'format_id': protocol
,
3261 'protocol': protocol
,
3265 def _find_jwplayer_data(self
, webpage
, video_id
=None, transform_source
=js_to_json
):
3267 r
'(?s)jwplayer\((?P<quote>[\'"])[^\'" ]+(?P
=quote
)\
)(?
!</script
>).*?\
.setup\s
*\
((?P
<options
>[^
)]+)\
)',
3271 jwplayer_data = self._parse_json(mobj.group('options
'),
3273 transform_source=transform_source)
3274 except ExtractorError:
3277 if isinstance(jwplayer_data, dict):
3278 return jwplayer_data
3280 def _extract_jwplayer_data(self, webpage, video_id, *args, **kwargs):
3281 jwplayer_data = self._find_jwplayer_data(
3282 webpage, video_id, transform_source=js_to_json)
3283 return self._parse_jwplayer_data(
3284 jwplayer_data, video_id, *args, **kwargs)
3286 def _parse_jwplayer_data(self, jwplayer_data, video_id=None, require_title=True,
3287 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
3288 # JWPlayer backward compatibility: flattened playlists
3289 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/api/config.js#L81-L96
3290 if 'playlist
' not in jwplayer_data:
3291 jwplayer_data = {'playlist': [jwplayer_data]}
3295 # JWPlayer backward compatibility: single playlist item
3296 # https://github.com/jwplayer/jwplayer/blob/v7.7.0/src/js/playlist/playlist.js#L10
3297 if not isinstance(jwplayer_data['playlist
'], list):
3298 jwplayer_data['playlist
'] = [jwplayer_data['playlist
']]
3300 for video_data in jwplayer_data['playlist
']:
3301 # JWPlayer backward compatibility: flattened sources
3302 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/playlist/item.js#L29-L35
3303 if 'sources
' not in video_data:
3304 video_data['sources
'] = [video_data]
3306 this_video_id = video_id or video_data['mediaid
']
3308 formats = self._parse_jwplayer_formats(
3309 video_data['sources
'], video_id=this_video_id, m3u8_id=m3u8_id,
3310 mpd_id=mpd_id, rtmp_params=rtmp_params, base_url=base_url)
3313 tracks = video_data.get('tracks
')
3314 if tracks and isinstance(tracks, list):
3315 for track in tracks:
3316 if not isinstance(track, dict):
3318 track_kind = track.get('kind
')
3319 if not track_kind or not isinstance(track_kind, compat_str):
3321 if track_kind.lower() not in ('captions
', 'subtitles
'):
3323 track_url = urljoin(base_url, track.get('file'))
3326 subtitles.setdefault(track.get('label
') or 'en
', []).append({
3327 'url
': self._proto_relative_url(track_url)
3331 'id': this_video_id,
3332 'title
': unescapeHTML(video_data['title
'] if require_title else video_data.get('title
')),
3333 'description
': clean_html(video_data.get('description
')),
3334 'thumbnail
': urljoin(base_url, self._proto_relative_url(video_data.get('image
'))),
3335 'timestamp
': int_or_none(video_data.get('pubdate
')),
3336 'duration
': float_or_none(jwplayer_data.get('duration
') or video_data.get('duration
')),
3337 'subtitles
': subtitles,
3339 # https://github.com/jwplayer/jwplayer/blob/master/src/js/utils/validator.js#L32
3340 if len(formats) == 1 and re.search(r'^
(?
:http|
//).*(?
:youtube\
.com|youtu\
.be
)/.+', formats[0]['url
']):
3342 '_type
': 'url_transparent
',
3343 'url
': formats[0]['url
'],
3346 self._sort_formats(formats)
3347 entry['formats
'] = formats
3348 entries.append(entry)
3349 if len(entries) == 1:
3352 return self.playlist_result(entries)
3354 def _parse_jwplayer_formats(self, jwplayer_sources_data, video_id=None,
3355 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
3358 for source in jwplayer_sources_data:
3359 if not isinstance(source, dict):
3361 source_url = urljoin(
3362 base_url, self._proto_relative_url(source.get('file')))
3363 if not source_url or source_url in urls:
3365 urls.append(source_url)
3366 source_type = source.get('type') or ''
3367 ext = mimetype2ext(source_type) or determine_ext(source_url)
3368 if source_type == 'hls
' or ext == 'm3u8
':
3369 formats.extend(self._extract_m3u8_formats(
3370 source_url, video_id, 'mp4
', entry_protocol='m3u8_native
',
3371 m3u8_id=m3u8_id, fatal=False))
3372 elif source_type == 'dash
' or ext == 'mpd
':
3373 formats.extend(self._extract_mpd_formats(
3374 source_url, video_id, mpd_id=mpd_id, fatal=False))
3376 formats.extend(self._extract_smil_formats(
3377 source_url, video_id, fatal=False))
3378 # https://github.com/jwplayer/jwplayer/blob/master/src/js/providers/default.js#L67
3379 elif source_type.startswith('audio
') or ext in (
3380 'oga
', 'aac
', 'mp3
', 'mpeg
', 'vorbis
'):
3387 height = int_or_none(source.get('height
'))
3389 # Often no height is provided but there is a label in
3390 # format like "1080p", "720p SD", or 1080.
3391 height = int_or_none(self._search_regex(
3392 r'^
(\d{3,4}
)[pP
]?
(?
:\b|$
)', compat_str(source.get('label
') or ''),
3393 'height
', default=None))
3396 'width
': int_or_none(source.get('width
')),
3398 'tbr
': int_or_none(source.get('bitrate
')),
3401 if source_url.startswith('rtmp
'):
3402 a_format['ext
'] = 'flv
'
3403 # See com/longtailvideo/jwplayer/media/RTMPMediaProvider.as
3404 # of jwplayer.flash.swf
3405 rtmp_url_parts = re.split(
3406 r'((?
:mp4|mp3|flv
):)', source_url, 1)
3407 if len(rtmp_url_parts) == 3:
3408 rtmp_url, prefix, play_path = rtmp_url_parts
3411 'play_path
': prefix + play_path,
3414 a_format.update(rtmp_params)
3415 formats.append(a_format)
3418 def _live_title(self, name):
3419 """ Generate the title for a live video """
3420 now = datetime.datetime.now()
3421 now_str = now.strftime('%Y
-%m
-%d %H
:%M
')
3422 return name + ' ' + now_str
3424 def _int(self, v, name, fatal=False, **kwargs):
3425 res = int_or_none(v, **kwargs)
3426 if 'get_attr
' in kwargs:
3427 print(getattr(v, kwargs['get_attr
']))
3429 msg = 'Failed to extract
%s: Could
not parse value
%r' % (name, v)
3431 raise ExtractorError(msg)
3433 self.report_warning(msg)
3436 def _float(self, v, name, fatal=False, **kwargs):
3437 res = float_or_none(v, **kwargs)
3439 msg = 'Failed to extract
%s: Could
not parse value
%r' % (name, v)
3441 raise ExtractorError(msg)
3443 self.report_warning(msg)
3446 def _set_cookie(self, domain, name, value, expire_time=None, port=None,
3447 path='/', secure=False, discard=False, rest={}, **kwargs):
3448 cookie = compat_cookiejar_Cookie(
3449 0, name, value, port, port is not None, domain, True,
3450 domain.startswith('.'), path, True, secure, expire_time,
3451 discard, None, None, rest)
3452 self._downloader.cookiejar.set_cookie(cookie)
3454 def _get_cookies(self, url):
3455 """ Return a compat_cookies_SimpleCookie with the cookies for the url """
3456 req = sanitized_Request(url)
3457 self._downloader.cookiejar.add_cookie_header(req)
3458 return compat_cookies_SimpleCookie(req.get_header('Cookie
'))
3460 def _apply_first_set_cookie_header(self, url_handle, cookie):
3462 Apply first Set-Cookie header instead of the last. Experimental.
3464 Some sites (e.g. [1-3]) may serve two cookies under the same name
3465 in Set-Cookie header and expect the first (old) one to be set rather
3466 than second (new). However, as of RFC6265 the newer one cookie
3467 should be set into cookie store what actually happens.
3468 We will workaround this issue by resetting the cookie to
3469 the first one manually.
3470 1. https://new.vk.com/
3471 2. https://github.com/ytdl-org/youtube-dl/issues/9841#issuecomment-227871201
3472 3. https://learning.oreilly.com/
3474 for header, cookies in url_handle.headers.items():
3475 if header.lower() != 'set-cookie
':
3477 if sys.version_info[0] >= 3:
3478 cookies = cookies.encode('iso
-8859-1')
3479 cookies = cookies.decode('utf
-8')
3480 cookie_value = re.search(
3481 r'%s=(.+?
);.*?
\b[Dd
]omain
=(.+?
)(?
:[,;]|$
)' % cookie, cookies)
3483 value, domain = cookie_value.groups()
3484 self._set_cookie(domain, cookie, value)
3487 def get_testcases(self, include_onlymatching=False):
3488 t = getattr(self, '_TEST
', None)
3490 assert not hasattr(self, '_TESTS
'), \
3491 '%s has _TEST
and _TESTS
' % type(self).__name__
3494 tests = getattr(self, '_TESTS
', [])
3496 if not include_onlymatching and t.get('only_matching
', False):
3498 t['name
'] = type(self).__name__[:-len('IE
')]
3501 def is_suitable(self, age_limit):
3502 """ Test whether the extractor is generally suitable for the given
3503 age limit (i.e. pornographic sites are not, all others usually are) """
3505 any_restricted = False
3506 for tc in self.get_testcases(include_onlymatching=False):
3507 if tc.get('playlist
', []):
3508 tc = tc['playlist
'][0]
3509 is_restricted = age_restricted(
3510 tc.get('info_dict
', {}).get('age_limit
'), age_limit)
3511 if not is_restricted:
3513 any_restricted = any_restricted or is_restricted
3514 return not any_restricted
3516 def extract_subtitles(self, *args, **kwargs):
3517 if (self.get_param('writesubtitles
', False)
3518 or self.get_param('listsubtitles
')):
3519 return self._get_subtitles(*args, **kwargs)
3522 def _get_subtitles(self, *args, **kwargs):
3523 raise NotImplementedError('This method must be implemented by subclasses
')
3525 def extract_comments(self, *args, **kwargs):
3526 if not self.get_param('getcomments
'):
3528 generator = self._get_comments(*args, **kwargs)
3534 comments.append(next(generator))
3535 except KeyboardInterrupt:
3537 self.to_screen('Interrupted by user
')
3538 except StopIteration:
3540 comment_count = len(comments)
3541 self.to_screen(f'Extracted {comment_count} comments
')
3543 'comments
': comments,
3544 'comment_count
': None if interrupted else comment_count
3548 def _get_comments(self, *args, **kwargs):
3549 raise NotImplementedError('This method must be implemented by subclasses
')
3552 def _merge_subtitle_items(subtitle_list1, subtitle_list2):
3553 """ Merge subtitle items for one language. Items with duplicated URLs
3554 will be dropped. """
3555 list1_urls = set([item['url
'] for item in subtitle_list1])
3556 ret = list(subtitle_list1)
3557 ret.extend([item for item in subtitle_list2 if item['url
'] not in list1_urls])
3561 def _merge_subtitles(cls, *dicts, target=None):
3562 """ Merge subtitle dictionaries, language by language. """
3566 for lang, subs in d.items():
3567 target[lang] = cls._merge_subtitle_items(target.get(lang, []), subs)
3570 def extract_automatic_captions(self, *args, **kwargs):
3571 if (self.get_param('writeautomaticsub
', False)
3572 or self.get_param('listsubtitles
')):
3573 return self._get_automatic_captions(*args, **kwargs)
3576 def _get_automatic_captions(self, *args, **kwargs):
3577 raise NotImplementedError('This method must be implemented by subclasses
')
3579 def mark_watched(self, *args, **kwargs):
3580 if not self.get_param('mark_watched
', False):
3582 if (self._get_login_info()[0] is not None
3583 or self.get_param('cookiefile
')
3584 or self.get_param('cookiesfrombrowser
')):
3585 self._mark_watched(*args, **kwargs)
3587 def _mark_watched(self, *args, **kwargs):
3588 raise NotImplementedError('This method must be implemented by subclasses
')
3590 def geo_verification_headers(self):
3592 geo_verification_proxy = self.get_param('geo_verification_proxy
')
3593 if geo_verification_proxy:
3594 headers['Ytdl
-request
-proxy
'] = geo_verification_proxy
3597 def _generic_id(self, url):
3598 return compat_urllib_parse_unquote(os.path.splitext(url.rstrip('/').split('/')[-1])[0])
3600 def _generic_title(self, url):
3601 return compat_urllib_parse_unquote(os.path.splitext(url_basename(url))[0])
3604 def _availability(is_private=None, needs_premium=None, needs_subscription=None, needs_auth=None, is_unlisted=None):
3605 all_known = all(map(
3606 lambda x: x is not None,
3607 (is_private, needs_premium, needs_subscription, needs_auth, is_unlisted)))
3609 'private
' if is_private
3610 else 'premium_only
' if needs_premium
3611 else 'subscriber_only
' if needs_subscription
3612 else 'needs_auth
' if needs_auth
3613 else 'unlisted
' if is_unlisted
3614 else 'public
' if all_known
3617 def _configuration_arg(self, key, default=NO_DEFAULT, casesense=False):
3619 @returns A list of values for the extractor argument given by "key"
3620 or "default" if no such key is present
3621 @param default The default value to return when the key is not present (default: [])
3622 @param casesense When false, the values are converted to lower case
3625 self._downloader.params, ('extractor_args
', self.ie_key().lower(), key))
3627 return [] if default is NO_DEFAULT else default
3628 return list(val) if casesense else [x.lower() for x in val]
3631 class SearchInfoExtractor(InfoExtractor):
3633 Base class for paged search queries extractors.
3634 They accept URLs in the format _SEARCH_KEY(|all|[0-9]):{query}
3635 Instances should define _SEARCH_KEY and optionally _MAX_RESULTS
3638 _MAX_RESULTS = float('inf
')
3641 def _make_valid_url(cls):
3642 return r'%s(?P
<prefix
>|
[1-9][0-9]*|all
):(?P
<query
>[\s\S
]+)' % cls._SEARCH_KEY
3645 def suitable(cls, url):
3646 return re.match(cls._make_valid_url(), url) is not None
3648 def _real_extract(self, query):
3649 mobj = re.match(self._make_valid_url(), query)
3651 raise ExtractorError('Invalid search query
"%s"' % query)
3653 prefix = mobj.group('prefix
')
3654 query = mobj.group('query
')
3656 return self._get_n_results(query, 1)
3657 elif prefix == 'all
':
3658 return self._get_n_results(query, self._MAX_RESULTS)
3662 raise ExtractorError('invalid download number
%s for query
"%s"' % (n, query))
3663 elif n > self._MAX_RESULTS:
3664 self.report_warning('%s returns
max %i results (you requested
%i)' % (self._SEARCH_KEY, self._MAX_RESULTS, n))
3665 n = self._MAX_RESULTS
3666 return self._get_n_results(query, n)
3668 def _get_n_results(self, query, n):
3669 """Get a specified number of results for a query.
3670 Either this function or _search_results must be overridden by subclasses """
3671 return self.playlist_result(
3672 itertools.islice(self._search_results(query), 0, None if n == float('inf
') else n),
3675 def _search_results(self, query):
3676 """Returns an iterator of search results"""
3677 raise NotImplementedError('This method must be implemented by subclasses
')
3680 def SEARCH_KEY(self):
3681 return self._SEARCH_KEY