2 from __future__
import unicode_literals
16 from ..compat
import (
17 compat_cookiejar_Cookie
,
18 compat_cookies_SimpleCookie
,
20 compat_etree_fromstring
,
26 compat_urllib_parse_unquote
,
27 compat_urllib_parse_urlencode
,
28 compat_urllib_request
,
30 compat_xml_parse_error
,
32 from ..downloader
import FileDownloader
33 from ..downloader
.f4m
import (
35 remove_encrypted_media
,
65 parse_m3u8_attributes
,
89 class InfoExtractor(object):
90 """Information Extractor class.
92 Information extractors are the classes that, given a URL, extract
93 information about the video (or videos) the URL refers to. This
94 information includes the real video URL, the video title, author and
95 others. The information is stored in a dictionary which is then
96 passed to the YoutubeDL. The YoutubeDL processes this
97 information possibly downloading the video to the file system, among
98 other possible outcomes.
100 The type field determines the type of the result.
101 By far the most common value (and the default if _type is missing) is
102 "video", which indicates a single video.
104 For a video, the dictionaries must include the following fields:
106 id: Video identifier.
107 title: Video title, unescaped.
109 Additionally, it must contain either a formats entry or a url one:
111 formats: A list of dictionaries for each format available, ordered
112 from worst to best quality.
115 * url The mandatory URL representing the media:
116 for plain file media - HTTP URL of this file,
118 for HLS - URL of the M3U8 media playlist,
119 for HDS - URL of the F4M manifest,
121 - HTTP URL to plain file media (in case of
123 - URL of the MPD manifest or base URL
124 representing the media if MPD manifest
125 is parsed from a string (in case of
127 for MSS - URL of the ISM manifest.
129 The URL of the manifest file in case of
131 for HLS - URL of the M3U8 master playlist,
132 for HDS - URL of the F4M manifest,
133 for DASH - URL of the MPD manifest,
134 for MSS - URL of the ISM manifest.
135 * ext Will be calculated from URL if missing
136 * format A human-readable description of the format
137 ("mp4 container with h264/opus").
138 Calculated from the format_id, width, height.
139 and format_note fields if missing.
140 * format_id A short description of the format
141 ("mp4_h264_opus" or "19").
142 Technically optional, but strongly recommended.
143 * format_note Additional info about the format
144 ("3D" or "DASH video")
145 * width Width of the video, if known
146 * height Height of the video, if known
147 * resolution Textual description of width and height
148 * tbr Average bitrate of audio and video in KBit/s
149 * abr Average audio bitrate in KBit/s
150 * acodec Name of the audio codec in use
151 * asr Audio sampling rate in Hertz
152 * vbr Average video bitrate in KBit/s
154 * vcodec Name of the video codec in use
155 * container Name of the container format
156 * filesize The number of bytes, if known in advance
157 * filesize_approx An estimate for the number of bytes
158 * player_url SWF Player URL (used for rtmpdump).
159 * protocol The protocol that will be used for the actual
160 download, lower-case.
161 "http", "https", "rtsp", "rtmp", "rtmp_ffmpeg", "rtmpe",
162 "m3u8", "m3u8_native" or "http_dash_segments".
164 Base URL for fragments. Each fragment's path
165 value (if present) will be relative to
167 * fragments A list of fragments of a fragmented media.
168 Each fragment entry must contain either an url
169 or a path. If an url is present it should be
170 considered by a client. Otherwise both path and
171 fragment_base_url must be present. Here is
172 the list of all potential fields:
173 * "url" - fragment's URL
174 * "path" - fragment's path relative to
176 * "duration" (optional, int or float)
177 * "filesize" (optional, int)
178 * preference Order number of this format. If this field is
179 present and not None, the formats get sorted
180 by this field, regardless of all other values.
181 -1 for default (order by other properties),
182 -2 or smaller for less than default.
183 < -1000 to hide the format (if there is
184 another one which is strictly better)
185 * language Language code, e.g. "de" or "en-US".
186 * language_preference Is this in the language mentioned in
188 10 if it's what the URL is about,
189 -1 for default (don't know),
190 -10 otherwise, other values reserved for now.
191 * quality Order number of the video quality of this
192 format, irrespective of the file format.
193 -1 for default (order by other properties),
194 -2 or smaller for less than default.
195 * source_preference Order number for this video source
196 (quality takes higher priority)
197 -1 for default (order by other properties),
198 -2 or smaller for less than default.
199 * http_headers A dictionary of additional HTTP headers
200 to add to the request.
201 * stretched_ratio If given and not 1, indicates that the
202 video's pixels are not square.
203 width : height ratio as float.
204 * no_resume The server does not support resuming the
205 (HTTP or RTMP) download. Boolean.
206 * downloader_options A dictionary of downloader options as
207 described in FileDownloader
208 RTMP formats can also have the additional fields: page_url,
209 app, play_path, tc_url, flash_version, rtmp_live, rtmp_conn,
210 rtmp_protocol, rtmp_real_time
212 url: Final video URL.
213 ext: Video filename extension.
214 format: The video format, defaults to ext (used for --get-format)
215 player_url: SWF Player URL (used for rtmpdump).
217 The following fields are optional:
219 alt_title: A secondary title of the video.
220 display_id An alternative identifier for the video, not necessarily
221 unique, but available before title. Typically, id is
222 something like "4234987", title "Dancing naked mole rats",
223 and display_id "dancing-naked-mole-rats"
224 thumbnails: A list of dictionaries, with the following entries:
225 * "id" (optional, string) - Thumbnail format ID
227 * "preference" (optional, int) - quality of the image
228 * "width" (optional, int)
229 * "height" (optional, int)
230 * "resolution" (optional, string "{width}x{height}",
232 * "filesize" (optional, int)
233 * "_test_url" (optional, bool) - If true, test the URL
234 thumbnail: Full URL to a video thumbnail image.
235 description: Full video description.
236 uploader: Full name of the video uploader.
237 license: License name the video is licensed under.
238 creator: The creator of the video.
239 release_timestamp: UNIX timestamp of the moment the video was released.
240 release_date: The date (YYYYMMDD) when the video was released.
241 timestamp: UNIX timestamp of the moment the video was uploaded
242 upload_date: Video upload date (YYYYMMDD).
243 If not explicitly set, calculated from timestamp.
244 uploader_id: Nickname or id of the video uploader.
245 uploader_url: Full URL to a personal webpage of the video uploader.
246 channel: Full name of the channel the video is uploaded on.
247 Note that channel fields may or may not repeat uploader
248 fields. This depends on a particular extractor.
249 channel_id: Id of the channel.
250 channel_url: Full URL to a channel webpage.
251 location: Physical location where the video was filmed.
252 subtitles: The available subtitles as a dictionary in the format
253 {tag: subformats}. "tag" is usually a language code, and
254 "subformats" is a list sorted from lower to higher
255 preference, each element is a dictionary with the "ext"
257 * "data": The subtitles file contents
258 * "url": A URL pointing to the subtitles file
259 It can optionally also have:
260 * "name": Name or description of the subtitles
261 "ext" will be calculated from URL if missing
262 automatic_captions: Like 'subtitles'; contains automatically generated
263 captions instead of normal subtitles
264 duration: Length of the video in seconds, as an integer or float.
265 view_count: How many users have watched the video on the platform.
266 like_count: Number of positive ratings of the video
267 dislike_count: Number of negative ratings of the video
268 repost_count: Number of reposts of the video
269 average_rating: Average rating give by users, the scale used depends on the webpage
270 comment_count: Number of comments on the video
271 comments: A list of comments, each with one or more of the following
272 properties (all but one of text or html optional):
273 * "author" - human-readable name of the comment author
274 * "author_id" - user ID of the comment author
275 * "author_thumbnail" - The thumbnail of the comment author
277 * "html" - Comment as HTML
278 * "text" - Plain text of the comment
279 * "timestamp" - UNIX timestamp of comment
280 * "parent" - ID of the comment this one is replying to.
281 Set to "root" to indicate that this is a
282 comment to the original video.
283 * "like_count" - Number of positive ratings of the comment
284 * "dislike_count" - Number of negative ratings of the comment
285 * "is_favorited" - Whether the comment is marked as
286 favorite by the video uploader
287 * "author_is_uploader" - Whether the comment is made by
289 age_limit: Age restriction for the video, as an integer (years)
290 webpage_url: The URL to the video webpage, if given to yt-dlp it
291 should allow to get the same result again. (It will be set
292 by YoutubeDL if it's missing)
293 categories: A list of categories that the video falls in, for example
295 tags: A list of tags assigned to the video, e.g. ["sweden", "pop music"]
296 cast: A list of the video cast
297 is_live: True, False, or None (=unknown). Whether this video is a
298 live stream that goes on instead of a fixed-length video.
299 was_live: True, False, or None (=unknown). Whether this video was
300 originally a live stream.
301 live_status: 'is_live', 'is_upcoming', 'was_live', 'not_live' or None (=unknown)
302 If absent, automatically set from is_live, was_live
303 start_time: Time in seconds where the reproduction should start, as
304 specified in the URL.
305 end_time: Time in seconds where the reproduction should end, as
306 specified in the URL.
307 chapters: A list of dictionaries, with the following entries:
308 * "start_time" - The start time of the chapter in seconds
309 * "end_time" - The end time of the chapter in seconds
310 * "title" (optional, string)
311 playable_in_embed: Whether this video is allowed to play in embedded
312 players on other sites. Can be True (=always allowed),
313 False (=never allowed), None (=unknown), or a string
314 specifying the criteria for embedability (Eg: 'whitelist')
315 availability: Under what condition the video is available. One of
316 'private', 'premium_only', 'subscriber_only', 'needs_auth',
317 'unlisted' or 'public'. Use 'InfoExtractor._availability'
319 __post_extractor: A function to be called just before the metadata is
320 written to either disk, logger or console. The function
321 must return a dict which will be added to the info_dict.
322 This is usefull for additional information that is
323 time-consuming to extract. Note that the fields thus
324 extracted will not be available to output template and
325 match_filter. So, only "comments" and "comment_count" are
326 currently allowed to be extracted via this method.
328 The following fields should only be used when the video belongs to some logical
331 chapter: Name or title of the chapter the video belongs to.
332 chapter_number: Number of the chapter the video belongs to, as an integer.
333 chapter_id: Id of the chapter the video belongs to, as a unicode string.
335 The following fields should only be used when the video is an episode of some
336 series, programme or podcast:
338 series: Title of the series or programme the video episode belongs to.
339 season: Title of the season the video episode belongs to.
340 season_number: Number of the season the video episode belongs to, as an integer.
341 season_id: Id of the season the video episode belongs to, as a unicode string.
342 episode: Title of the video episode. Unlike mandatory video title field,
343 this field should denote the exact title of the video episode
344 without any kind of decoration.
345 episode_number: Number of the video episode within a season, as an integer.
346 episode_id: Id of the video episode, as a unicode string.
348 The following fields should only be used when the media is a track or a part of
351 track: Title of the track.
352 track_number: Number of the track within an album or a disc, as an integer.
353 track_id: Id of the track (useful in case of custom indexing, e.g. 6.iii),
355 artist: Artist(s) of the track.
356 genre: Genre(s) of the track.
357 album: Title of the album the track belongs to.
358 album_type: Type of the album (e.g. "Demo", "Full-length", "Split", "Compilation", etc).
359 album_artist: List of all artists appeared on the album (e.g.
360 "Ash Borer / Fell Voices" or "Various Artists", useful for splits
362 disc_number: Number of the disc or other physical medium the track belongs to,
364 release_year: Year (YYYY) when the album was released.
366 Unless mentioned otherwise, the fields should be Unicode strings.
368 Unless mentioned otherwise, None is equivalent to absence of information.
371 _type "playlist" indicates multiple videos.
372 There must be a key "entries", which is a list, an iterable, or a PagedList
373 object, each element of which is a valid dictionary by this specification.
375 Additionally, playlists can have "id", "title", and any other relevent
376 attributes with the same semantics as videos (see above).
379 _type "multi_video" indicates that there are multiple videos that
380 form a single show, for examples multiple acts of an opera or TV episode.
381 It must have an entries key like a playlist and contain all the keys
382 required for a video at the same time.
385 _type "url" indicates that the video must be extracted from another
386 location, possibly by a different extractor. Its only required key is:
387 "url" - the next URL to extract.
388 The key "ie_key" can be set to the class name (minus the trailing "IE",
389 e.g. "Youtube") if the extractor class is known in advance.
390 Additionally, the dictionary may have any properties of the resolved entity
391 known in advance, for example "title" if the title of the referred video is
395 _type "url_transparent" entities have the same specification as "url", but
396 indicate that the given additional information is more precise than the one
397 associated with the resolved URL.
398 This is useful when a site employs a video service that hosts the video and
399 its technical metadata, but that video service does not embed a useful
400 title, description etc.
403 Subclasses of this one should re-define the _real_initialize() and
404 _real_extract() methods and define a _VALID_URL regexp.
405 Probably, they should also be added to the list of extractors.
407 _GEO_BYPASS attribute may be set to False in order to disable
408 geo restriction bypass mechanisms for a particular extractor.
409 Though it won't disable explicit geo restriction bypass based on
410 country code provided with geo_bypass_country.
412 _GEO_COUNTRIES attribute may contain a list of presumably geo unrestricted
413 countries for this extractor. One of these countries will be used by
414 geo restriction bypass mechanism right away in order to bypass
415 geo restriction, of course, if the mechanism is not disabled.
417 _GEO_IP_BLOCKS attribute may contain a list of presumably geo unrestricted
418 IP blocks in CIDR notation for this extractor. One of these IP blocks
419 will be used by geo restriction bypass mechanism similarly
422 Finally, the _WORKING attribute should be set to False for broken IEs
423 in order to warn the users and skip the tests.
428 _x_forwarded_for_ip
= None
430 _GEO_COUNTRIES
= None
431 _GEO_IP_BLOCKS
= None
435 'any': 'Use --cookies, --username and --password or --netrc to provide account credentials',
437 'Use --cookies for the authentication. '
438 'See https://github.com/ytdl-org/youtube-dl#how-do-i-pass-cookies-to-youtube-dl for how to pass cookies'),
439 'password': 'Use --username and --password or --netrc to provide account credentials',
442 def __init__(self
, downloader
=None):
443 """Constructor. Receives an optional downloader."""
445 self
._x
_forwarded
_for
_ip
= None
446 self
._printed
_messages
= set()
447 self
.set_downloader(downloader
)
450 def _match_valid_url(cls
, url
):
451 # This does not use has/getattr intentionally - we want to know whether
452 # we have cached the regexp for *this* class, whereas getattr would also
453 # match the superclass
454 if '_VALID_URL_RE' not in cls
.__dict
__:
455 cls
._VALID
_URL
_RE
= re
.compile(cls
._VALID
_URL
)
456 return cls
._VALID
_URL
_RE
.match(url
)
459 def suitable(cls
, url
):
460 """Receives a URL and returns True if suitable for this IE."""
461 return cls
._match
_valid
_url
(url
) is not None
464 def _match_id(cls
, url
):
465 return cls
._match
_valid
_url
(url
).group('id')
469 """Getter method for _WORKING."""
472 def initialize(self
):
473 """Initializes an instance (authentication, etc)."""
474 self
._printed
_messages
= set()
475 self
._initialize
_geo
_bypass
({
476 'countries': self
._GEO
_COUNTRIES
,
477 'ip_blocks': self
._GEO
_IP
_BLOCKS
,
480 self
._real
_initialize
()
483 def _initialize_geo_bypass(self
, geo_bypass_context
):
485 Initialize geo restriction bypass mechanism.
487 This method is used to initialize geo bypass mechanism based on faking
488 X-Forwarded-For HTTP header. A random country from provided country list
489 is selected and a random IP belonging to this country is generated. This
490 IP will be passed as X-Forwarded-For HTTP header in all subsequent
493 This method will be used for initial geo bypass mechanism initialization
494 during the instance initialization with _GEO_COUNTRIES and
497 You may also manually call it from extractor's code if geo bypass
498 information is not available beforehand (e.g. obtained during
499 extraction) or due to some other reason. In this case you should pass
500 this information in geo bypass context passed as first argument. It may
501 contain following fields:
503 countries: List of geo unrestricted countries (similar
505 ip_blocks: List of geo unrestricted IP blocks in CIDR notation
506 (similar to _GEO_IP_BLOCKS)
509 if not self
._x
_forwarded
_for
_ip
:
511 # Geo bypass mechanism is explicitly disabled by user
512 if not self
.get_param('geo_bypass', True):
515 if not geo_bypass_context
:
516 geo_bypass_context
= {}
518 # Backward compatibility: previously _initialize_geo_bypass
519 # expected a list of countries, some 3rd party code may still use
521 if isinstance(geo_bypass_context
, (list, tuple)):
522 geo_bypass_context
= {
523 'countries': geo_bypass_context
,
526 # The whole point of geo bypass mechanism is to fake IP
527 # as X-Forwarded-For HTTP header based on some IP block or
530 # Path 1: bypassing based on IP block in CIDR notation
532 # Explicit IP block specified by user, use it right away
533 # regardless of whether extractor is geo bypassable or not
534 ip_block
= self
.get_param('geo_bypass_ip_block', None)
536 # Otherwise use random IP block from geo bypass context but only
537 # if extractor is known as geo bypassable
539 ip_blocks
= geo_bypass_context
.get('ip_blocks')
540 if self
._GEO
_BYPASS
and ip_blocks
:
541 ip_block
= random
.choice(ip_blocks
)
544 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(ip_block
)
545 self
._downloader
.write_debug(
546 '[debug] Using fake IP %s as X-Forwarded-For' % self
._x
_forwarded
_for
_ip
)
549 # Path 2: bypassing based on country code
551 # Explicit country code specified by user, use it right away
552 # regardless of whether extractor is geo bypassable or not
553 country
= self
.get_param('geo_bypass_country', None)
555 # Otherwise use random country code from geo bypass context but
556 # only if extractor is known as geo bypassable
558 countries
= geo_bypass_context
.get('countries')
559 if self
._GEO
_BYPASS
and countries
:
560 country
= random
.choice(countries
)
563 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(country
)
564 self
._downloader
.write_debug(
565 'Using fake IP %s (%s) as X-Forwarded-For' % (self
._x
_forwarded
_for
_ip
, country
.upper()))
567 def extract(self
, url
):
568 """Extracts URL information and returns it in list of dicts."""
573 self
.write_debug('Extracting URL: %s' % url
)
574 ie_result
= self
._real
_extract
(url
)
575 if ie_result
is None:
577 if self
._x
_forwarded
_for
_ip
:
578 ie_result
['__x_forwarded_for_ip'] = self
._x
_forwarded
_for
_ip
579 subtitles
= ie_result
.get('subtitles')
580 if (subtitles
and 'live_chat' in subtitles
581 and 'no-live-chat' in self
.get_param('compat_opts', [])):
582 del subtitles
['live_chat']
584 except GeoRestrictedError
as e
:
585 if self
.__maybe
_fake
_ip
_and
_retry
(e
.countries
):
588 except ExtractorError
:
590 except compat_http_client
.IncompleteRead
as e
:
591 raise ExtractorError('A network error has occurred.', cause
=e
, expected
=True)
592 except (KeyError, StopIteration) as e
:
593 raise ExtractorError('An extractor error has occurred.', cause
=e
)
595 def __maybe_fake_ip_and_retry(self
, countries
):
596 if (not self
.get_param('geo_bypass_country', None)
598 and self
.get_param('geo_bypass', True)
599 and not self
._x
_forwarded
_for
_ip
601 country_code
= random
.choice(countries
)
602 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(country_code
)
603 if self
._x
_forwarded
_for
_ip
:
605 'Video is geo restricted. Retrying extraction with fake IP %s (%s) as X-Forwarded-For.'
606 % (self
._x
_forwarded
_for
_ip
, country_code
.upper()))
610 def set_downloader(self
, downloader
):
611 """Sets the downloader for this IE."""
612 self
._downloader
= downloader
614 def _real_initialize(self
):
615 """Real initialization process. Redefine in subclasses."""
618 def _real_extract(self
, url
):
619 """Real extraction process. Redefine in subclasses."""
624 """A string for getting the InfoExtractor with get_info_extractor"""
625 return compat_str(cls
.__name
__[:-2])
629 return compat_str(type(self
).__name
__[:-2])
632 def __can_accept_status_code(err
, expected_status
):
633 assert isinstance(err
, compat_urllib_error
.HTTPError
)
634 if expected_status
is None:
636 elif callable(expected_status
):
637 return expected_status(err
.code
) is True
639 return err
.code
in variadic(expected_status
)
641 def _request_webpage(self
, url_or_request
, video_id
, note
=None, errnote
=None, fatal
=True, data
=None, headers
={}, query={}
, expected_status
=None):
643 Return the response handle.
645 See _download_webpage docstring for arguments specification.
647 if not self
._downloader
._first
_webpage
_request
:
648 sleep_interval
= float_or_none(self
.get_param('sleep_interval_requests')) or 0
649 if sleep_interval
> 0:
650 self
.to_screen('Sleeping %s seconds ...' % sleep_interval
)
651 time
.sleep(sleep_interval
)
653 self
._downloader
._first
_webpage
_request
= False
656 self
.report_download_webpage(video_id
)
657 elif note
is not False:
659 self
.to_screen('%s' % (note
,))
661 self
.to_screen('%s: %s' % (video_id
, note
))
663 # Some sites check X-Forwarded-For HTTP header in order to figure out
664 # the origin of the client behind proxy. This allows bypassing geo
665 # restriction by faking this header's value to IP that belongs to some
666 # geo unrestricted country. We will do so once we encounter any
667 # geo restriction error.
668 if self
._x
_forwarded
_for
_ip
:
669 if 'X-Forwarded-For' not in headers
:
670 headers
['X-Forwarded-For'] = self
._x
_forwarded
_for
_ip
672 if isinstance(url_or_request
, compat_urllib_request
.Request
):
673 url_or_request
= update_Request(
674 url_or_request
, data
=data
, headers
=headers
, query
=query
)
677 url_or_request
= update_url_query(url_or_request
, query
)
678 if data
is not None or headers
:
679 url_or_request
= sanitized_Request(url_or_request
, data
, headers
)
681 return self
._downloader
.urlopen(url_or_request
)
682 except network_exceptions
as err
:
683 if isinstance(err
, compat_urllib_error
.HTTPError
):
684 if self
.__can
_accept
_status
_code
(err
, expected_status
):
685 # Retain reference to error to prevent file object from
686 # being closed before it can be read. Works around the
687 # effects of <https://bugs.python.org/issue15002>
688 # introduced in Python 3.4.1.
695 errnote
= 'Unable to download webpage'
697 errmsg
= '%s: %s' % (errnote
, error_to_compat_str(err
))
699 raise ExtractorError(errmsg
, sys
.exc_info()[2], cause
=err
)
701 self
.report_warning(errmsg
)
704 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):
706 Return a tuple (page content as string, URL handle).
708 See _download_webpage docstring for arguments specification.
710 # Strip hashes from the URL (#1038)
711 if isinstance(url_or_request
, (compat_str
, str)):
712 url_or_request
= url_or_request
.partition('#')[0]
714 urlh
= self
._request
_webpage
(url_or_request
, video_id
, note
, errnote
, fatal
, data
=data
, headers
=headers
, query
=query
, expected_status
=expected_status
)
718 content
= self
._webpage
_read
_content
(urlh
, url_or_request
, video_id
, note
, errnote
, fatal
, encoding
=encoding
)
719 return (content
, urlh
)
722 def _guess_encoding_from_content(content_type
, webpage_bytes
):
723 m
= re
.match(r
'[a-zA-Z0-9_.-]+/[a-zA-Z0-9_.-]+\s*;\s*charset=(.+)', content_type
)
725 encoding
= m
.group(1)
727 m
= re
.search(br
'<meta[^>]+charset=[\'"]?([^\'")]+)[ /\'">]',
728 webpage_bytes[:1024])
730 encoding = m.group(1).decode('ascii')
731 elif webpage_bytes.startswith(b'\xff\xfe'):
738 def __check_blocked(self, content):
739 first_block = content[:512]
740 if ('<title>Access to this site is blocked</title>' in content
741 and 'Websense' in first_block):
742 msg = 'Access to this webpage has been blocked by Websense filtering software in your network.'
743 blocked_iframe = self._html_search_regex(
744 r'<iframe src="([^
"]+)"', content,
745 'Websense information URL
', default=None)
747 msg += ' Visit
%s for more details
' % blocked_iframe
748 raise ExtractorError(msg, expected=True)
749 if '<title
>The URL you requested has been blocked
</title
>' in first_block:
751 'Access to this webpage has been blocked by Indian censorship
. '
752 'Use a VPN
or proxy
server (with --proxy
) to route around it
.')
753 block_msg = self._html_search_regex(
754 r'</h1
><p
>(.*?
)</p
>',
755 content, 'block message
', default=None)
757 msg += ' (Message
: "%s")' % block_msg.replace('\n', ' ')
758 raise ExtractorError(msg, expected=True)
759 if ('<title
>TTK
:: Доступ к ресурсу ограничен
</title
>' in content
760 and 'blocklist
.rkn
.gov
.ru
' in content):
761 raise ExtractorError(
762 'Access to this webpage has been blocked by decision of the Russian government
. '
763 'Visit http
://blocklist
.rkn
.gov
.ru
/ for a block reason
.',
766 def _webpage_read_content(self, urlh, url_or_request, video_id, note=None, errnote=None, fatal=True, prefix=None, encoding=None):
767 content_type = urlh.headers.get('Content
-Type
', '')
768 webpage_bytes = urlh.read()
769 if prefix is not None:
770 webpage_bytes = prefix + webpage_bytes
772 encoding = self._guess_encoding_from_content(content_type, webpage_bytes)
773 if self.get_param('dump_intermediate_pages
', False):
774 self.to_screen('Dumping request to
' + urlh.geturl())
775 dump = base64.b64encode(webpage_bytes).decode('ascii
')
776 self._downloader.to_screen(dump)
777 if self.get_param('write_pages
', False):
778 basen = '%s_%s' % (video_id, urlh.geturl())
780 h = '___
' + hashlib.md5(basen.encode('utf
-8')).hexdigest()
781 basen = basen[:240 - len(h)] + h
782 raw_filename = basen + '.dump
'
783 filename = sanitize_filename(raw_filename, restricted=True)
784 self.to_screen('Saving request to
' + filename)
785 # Working around MAX_PATH limitation on Windows (see
786 # http://msdn.microsoft.com/en-us/library/windows/desktop/aa365247(v=vs.85).aspx)
787 if compat_os_name == 'nt
':
788 absfilepath = os.path.abspath(filename)
789 if len(absfilepath) > 259:
790 filename = '\\\\?
\\' + absfilepath
791 with open(filename, 'wb
') as outf:
792 outf.write(webpage_bytes)
795 content = webpage_bytes.decode(encoding, 'replace
')
797 content = webpage_bytes.decode('utf
-8', 'replace
')
799 self.__check_blocked(content)
803 def _download_webpage(
804 self, url_or_request, video_id, note=None, errnote=None,
805 fatal=True, tries=1, timeout=5, encoding=None, data=None,
806 headers={}, query={}, expected_status=None):
808 Return the data of the page as a string.
811 url_or_request -- plain text URL as a string or
812 a compat_urllib_request.Requestobject
813 video_id -- Video/playlist/item identifier (string)
816 note -- note printed before downloading (string)
817 errnote -- note printed in case of an error (string)
818 fatal -- flag denoting whether error should be considered fatal,
819 i.e. whether it should cause ExtractionError to be raised,
820 otherwise a warning will be reported and extraction continued
821 tries -- number of tries
822 timeout -- sleep interval between tries
823 encoding -- encoding for a page content decoding, guessed automatically
824 when not explicitly specified
825 data -- POST data (bytes)
826 headers -- HTTP headers (dict)
827 query -- URL query (dict)
828 expected_status -- allows to accept failed HTTP requests (non 2xx
829 status code) by explicitly specifying a set of accepted status
830 codes. Can be any of the following entities:
831 - an integer type specifying an exact failed status code to
833 - a list or a tuple of integer types specifying a list of
834 failed status codes to accept
835 - a callable accepting an actual failed status code and
836 returning True if it should be accepted
837 Note that this argument does not affect success status codes (2xx)
838 which are always accepted.
843 while success is False:
845 res = self._download_webpage_handle(
846 url_or_request, video_id, note, errnote, fatal,
847 encoding=encoding, data=data, headers=headers, query=query,
848 expected_status=expected_status)
850 except compat_http_client.IncompleteRead as e:
852 if try_count >= tries:
854 self._sleep(timeout, video_id)
861 def _download_xml_handle(
862 self, url_or_request, video_id, note='Downloading XML
',
863 errnote='Unable to download XML
', transform_source=None,
864 fatal=True, encoding=None, data=None, headers={}, query={},
865 expected_status=None):
867 Return a tuple (xml as an compat_etree_Element, URL handle).
869 See _download_webpage docstring for arguments specification.
871 res = self._download_webpage_handle(
872 url_or_request, video_id, note, errnote, fatal=fatal,
873 encoding=encoding, data=data, headers=headers, query=query,
874 expected_status=expected_status)
877 xml_string, urlh = res
878 return self._parse_xml(
879 xml_string, video_id, transform_source=transform_source,
883 self, url_or_request, video_id,
884 note='Downloading XML
', errnote='Unable to download XML
',
885 transform_source=None, fatal=True, encoding=None,
886 data=None, headers={}, query={}, expected_status=None):
888 Return the xml as an compat_etree_Element.
890 See _download_webpage docstring for arguments specification.
892 res = self._download_xml_handle(
893 url_or_request, video_id, note=note, errnote=errnote,
894 transform_source=transform_source, fatal=fatal, encoding=encoding,
895 data=data, headers=headers, query=query,
896 expected_status=expected_status)
897 return res if res is False else res[0]
899 def _parse_xml(self, xml_string, video_id, transform_source=None, fatal=True):
901 xml_string = transform_source(xml_string)
903 return compat_etree_fromstring(xml_string.encode('utf
-8'))
904 except compat_xml_parse_error as ve:
905 errmsg = '%s: Failed to parse XML
' % video_id
907 raise ExtractorError(errmsg, cause=ve)
909 self.report_warning(errmsg + str(ve))
911 def _download_json_handle(
912 self, url_or_request, video_id, note='Downloading JSON metadata
',
913 errnote='Unable to download JSON metadata
', transform_source=None,
914 fatal=True, encoding=None, data=None, headers={}, query={},
915 expected_status=None):
917 Return a tuple (JSON object, URL handle).
919 See _download_webpage docstring for arguments specification.
921 res = self._download_webpage_handle(
922 url_or_request, video_id, note, errnote, fatal=fatal,
923 encoding=encoding, data=data, headers=headers, query=query,
924 expected_status=expected_status)
927 json_string, urlh = res
928 return self._parse_json(
929 json_string, video_id, transform_source=transform_source,
933 self, url_or_request, video_id, note='Downloading JSON metadata
',
934 errnote='Unable to download JSON metadata
', transform_source=None,
935 fatal=True, encoding=None, data=None, headers={}, query={},
936 expected_status=None):
938 Return the JSON object as a dict.
940 See _download_webpage docstring for arguments specification.
942 res = self._download_json_handle(
943 url_or_request, video_id, note=note, errnote=errnote,
944 transform_source=transform_source, fatal=fatal, encoding=encoding,
945 data=data, headers=headers, query=query,
946 expected_status=expected_status)
947 return res if res is False else res[0]
949 def _parse_json(self, json_string, video_id, transform_source=None, fatal=True):
951 json_string = transform_source(json_string)
953 return json.loads(json_string)
954 except ValueError as ve:
955 errmsg = '%s: Failed to parse JSON
' % video_id
957 raise ExtractorError(errmsg, cause=ve)
959 self.report_warning(errmsg + str(ve))
961 def _parse_socket_response_as_json(self, data, video_id, transform_source=None, fatal=True):
962 return self._parse_json(
963 data[data.find('{'):data.rfind('}
') + 1],
964 video_id, transform_source, fatal)
966 def _download_socket_json_handle(
967 self, url_or_request, video_id, note='Polling socket
',
968 errnote='Unable to poll socket
', transform_source=None,
969 fatal=True, encoding=None, data=None, headers={}, query={},
970 expected_status=None):
972 Return a tuple (JSON object, URL handle).
974 See _download_webpage docstring for arguments specification.
976 res = self._download_webpage_handle(
977 url_or_request, video_id, note, errnote, fatal=fatal,
978 encoding=encoding, data=data, headers=headers, query=query,
979 expected_status=expected_status)
983 return self._parse_socket_response_as_json(
984 webpage, video_id, transform_source=transform_source,
987 def _download_socket_json(
988 self, url_or_request, video_id, note='Polling socket
',
989 errnote='Unable to poll socket
', transform_source=None,
990 fatal=True, encoding=None, data=None, headers={}, query={},
991 expected_status=None):
993 Return the JSON object as a dict.
995 See _download_webpage docstring for arguments specification.
997 res = self._download_socket_json_handle(
998 url_or_request, video_id, note=note, errnote=errnote,
999 transform_source=transform_source, fatal=fatal, encoding=encoding,
1000 data=data, headers=headers, query=query,
1001 expected_status=expected_status)
1002 return res if res is False else res[0]
1004 def report_warning(self, msg, video_id=None, *args, only_once=False, **kwargs):
1005 idstr = format_field(video_id, template='%s: ')
1006 msg = f'[{self.IE_NAME}
] {idstr}{msg}
'
1008 if f'WARNING
: {msg}
' in self._printed_messages:
1010 self._printed_messages.add(f'WARNING
: {msg}
')
1011 self._downloader.report_warning(msg, *args, **kwargs)
1013 def to_screen(self, msg, *args, **kwargs):
1014 """Print msg to screen, prefixing it with '[ie_name
]'"""
1015 self._downloader.to_screen('[%s] %s' % (self.IE_NAME, msg), *args, **kwargs)
1017 def write_debug(self, msg, *args, **kwargs):
1018 self._downloader.write_debug('[%s] %s' % (self.IE_NAME, msg), *args, **kwargs)
1020 def get_param(self, name, default=None, *args, **kwargs):
1021 if self._downloader:
1022 return self._downloader.params.get(name, default, *args, **kwargs)
1025 def report_extraction(self, id_or_name):
1026 """Report information extraction."""
1027 self.to_screen('%s: Extracting information
' % id_or_name)
1029 def report_download_webpage(self, video_id):
1030 """Report webpage download."""
1031 self.to_screen('%s: Downloading webpage
' % video_id)
1033 def report_age_confirmation(self):
1034 """Report attempt to confirm age."""
1035 self.to_screen('Confirming age
')
1037 def report_login(self):
1038 """Report attempt to log in."""
1039 self.to_screen('Logging
in')
1041 def raise_login_required(
1042 self, msg='This video
is only available
for registered users
',
1043 metadata_available=False, method='any
'):
1044 if metadata_available and self.get_param('ignore_no_formats_error
'):
1045 self.report_warning(msg)
1046 if method is not None:
1047 msg = '%s. %s' % (msg, self._LOGIN_HINTS[method])
1048 raise ExtractorError(msg, expected=True)
1050 def raise_geo_restricted(
1051 self, msg='This video
is not available
from your location due to geo restriction
',
1052 countries=None, metadata_available=False):
1053 if metadata_available and self.get_param('ignore_no_formats_error
'):
1054 self.report_warning(msg)
1056 raise GeoRestrictedError(msg, countries=countries)
1058 def raise_no_formats(self, msg, expected=False, video_id=None):
1059 if expected and self.get_param('ignore_no_formats_error
'):
1060 self.report_warning(msg, video_id)
1061 elif isinstance(msg, ExtractorError):
1064 raise ExtractorError(msg, expected=expected, video_id=video_id)
1066 # Methods for following #608
1068 def url_result(url, ie=None, video_id=None, video_title=None):
1069 """Returns a URL that points to a page that should be processed"""
1070 # TODO: ie should be the class used for getting the info
1071 video_info = {'_type
': 'url
',
1074 if video_id is not None:
1075 video_info['id'] = video_id
1076 if video_title is not None:
1077 video_info['title
'] = video_title
1080 def playlist_from_matches(self, matches, playlist_id=None, playlist_title=None, getter=None, ie=None):
1082 self.url_result(self._proto_relative_url(getter(m) if getter else m), ie)
1084 return self.playlist_result(
1085 urls, playlist_id=playlist_id, playlist_title=playlist_title)
1088 def playlist_result(entries, playlist_id=None, playlist_title=None, playlist_description=None, **kwargs):
1089 """Returns a playlist"""
1090 video_info = {'_type
': 'playlist
',
1092 video_info.update(kwargs)
1094 video_info['id'] = playlist_id
1096 video_info['title
'] = playlist_title
1097 if playlist_description is not None:
1098 video_info['description
'] = playlist_description
1101 def _search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
1103 Perform a regex search on the given string, using a single or a list of
1104 patterns returning the first matching group.
1105 In case of failure return a default value or raise a WARNING or a
1106 RegexNotFoundError, depending on fatal, specifying the field name.
1108 if isinstance(pattern, (str, compat_str, compiled_regex_type)):
1109 mobj = re.search(pattern, string, flags)
1112 mobj = re.search(p, string, flags)
1116 if not self.get_param('no_color
') and compat_os_name != 'nt
' and sys.stderr.isatty():
1117 _name = '\033[0;34m
%s\033[0m
' % name
1123 # return the first matching group
1124 return next(g for g in mobj.groups() if g is not None)
1125 elif isinstance(group, (list, tuple)):
1126 return tuple(mobj.group(g) for g in group)
1128 return mobj.group(group)
1129 elif default is not NO_DEFAULT:
1132 raise RegexNotFoundError('Unable to extract
%s' % _name)
1134 self.report_warning('unable to extract
%s' % _name + bug_reports_message())
1137 def _html_search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
1139 Like _search_regex, but strips HTML tags and unescapes entities.
1141 res = self._search_regex(pattern, string, name, default, fatal, flags, group)
1143 return clean_html(res).strip()
1147 def _get_netrc_login_info(self, netrc_machine=None):
1150 netrc_machine = netrc_machine or self._NETRC_MACHINE
1152 if self.get_param('usenetrc
', False):
1154 info = netrc.netrc().authenticators(netrc_machine)
1155 if info is not None:
1159 raise netrc.NetrcParseError(
1160 'No authenticators
for %s' % netrc_machine)
1161 except (IOError, netrc.NetrcParseError) as err:
1162 self.report_warning(
1163 'parsing
.netrc
: %s' % error_to_compat_str(err))
1165 return username, password
1167 def _get_login_info(self, username_option='username
', password_option='password
', netrc_machine=None):
1169 Get the login info as (username, password)
1170 First look for the manually specified credentials using username_option
1171 and password_option as keys in params dictionary. If no such credentials
1172 available look in the netrc file using the netrc_machine or _NETRC_MACHINE
1174 If there's no info available
, return (None, None)
1177 # Attempt to use provided username and password or .netrc data
1178 username = self.get_param(username_option)
1179 if username is not None:
1180 password = self.get_param(password_option)
1182 username, password = self._get_netrc_login_info(netrc_machine)
1184 return username, password
1186 def _get_tfa_info(self, note='two-factor verification code'):
1188 Get the two
-factor authentication info
1189 TODO
- asking the user will be required
for sms
/phone verify
1190 currently just uses the command line option
1191 If there
's no info available, return None
1194 tfa = self.get_param('twofactor
')
1198 return compat_getpass('Type
%s and press
[Return
]: ' % note)
1200 # Helper functions for extracting OpenGraph info
1202 def _og_regexes(prop):
1203 content_re = r'content
=(?
:"([^"]+?
)"|\'([^\']+?)\'|\s*([^\s"\'=<>`
]+?
))'
1204 property_re = (r'(?
:name|
property)=(?
:\'og
[:-]%(prop)s\'|
"og[:-]%(prop)s"|\s
*og
[:-]%(prop)s\b)'
1205 % {'prop': re.escape(prop)})
1206 template = r'<meta
[^
>]+?
%s[^
>]+?
%s'
1208 template % (property_re, content_re),
1209 template % (content_re, property_re),
1213 def _meta_regex(prop):
1214 return r'''(?isx)<meta
1215 (?=[^>]+(?:itemprop|name|property|id|http-equiv)=(["\']?)%s\1)
1216 [^>]+?content=(["\'])(?P<content>.*?)\2''' % re.escape(prop)
1218 def _og_search_property(self, prop, html, name=None, **kargs):
1219 prop = variadic(prop)
1221 name = 'OpenGraph
%s' % prop[0]
1224 og_regexes.extend(self._og_regexes(p))
1225 escaped = self._search_regex(og_regexes, html, name, flags=re.DOTALL, **kargs)
1228 return unescapeHTML(escaped)
1230 def _og_search_thumbnail(self, html, **kargs):
1231 return self._og_search_property('image
', html, 'thumbnail URL
', fatal=False, **kargs)
1233 def _og_search_description(self, html, **kargs):
1234 return self._og_search_property('description
', html, fatal=False, **kargs)
1236 def _og_search_title(self, html, **kargs):
1237 return self._og_search_property('title
', html, **kargs)
1239 def _og_search_video_url(self, html, name='video url
', secure=True, **kargs):
1240 regexes = self._og_regexes('video
') + self._og_regexes('video
:url
')
1242 regexes = self._og_regexes('video
:secure_url
') + regexes
1243 return self._html_search_regex(regexes, html, name, **kargs)
1245 def _og_search_url(self, html, **kargs):
1246 return self._og_search_property('url
', html, **kargs)
1248 def _html_search_meta(self, name, html, display_name=None, fatal=False, **kwargs):
1249 name = variadic(name)
1250 if display_name is None:
1251 display_name = name[0]
1252 return self._html_search_regex(
1253 [self._meta_regex(n) for n in name],
1254 html, display_name, fatal=fatal, group='content
', **kwargs)
1256 def _dc_search_uploader(self, html):
1257 return self._html_search_meta('dc
.creator
', html, 'uploader
')
1259 def _rta_search(self, html):
1260 # See http://www.rtalabel.org/index.php?content=howtofaq#single
1261 if re.search(r'(?ix
)<meta\s
+name
="rating"\s
+'
1262 r' content
="RTA-5042-1996-1400-1577-RTA"',
1267 def _media_rating_search(self, html):
1268 # See http://www.tjg-designs.com/WP/metadata-code-examples-adding-metadata-to-your-web-pages/
1269 rating = self._html_search_meta('rating
', html)
1281 return RATING_TABLE.get(rating.lower())
1283 def _family_friendly_search(self, html):
1284 # See http://schema.org/VideoObject
1285 family_friendly = self._html_search_meta(
1286 'isFamilyFriendly
', html, default=None)
1288 if not family_friendly:
1297 return RATING_TABLE.get(family_friendly.lower())
1299 def _twitter_search_player(self, html):
1300 return self._html_search_meta('twitter
:player
', html,
1301 'twitter card player
')
1303 def _search_json_ld(self, html, video_id, expected_type=None, **kwargs):
1304 json_ld_list = list(re.finditer(JSON_LD_RE, html))
1305 default = kwargs.get('default
', NO_DEFAULT)
1306 # JSON-LD may be malformed and thus `fatal` should be respected.
1307 # At the same time `default` may be passed that assumes `fatal=False`
1308 # for _search_regex. Let's simulate the same behavior here
as well
.
1309 fatal
= kwargs
.get('fatal', True) if default
is NO_DEFAULT
else False
1311 for mobj
in json_ld_list
:
1312 json_ld_item
= self
._parse
_json
(
1313 mobj
.group('json_ld'), video_id
, fatal
=fatal
)
1314 if not json_ld_item
:
1316 if isinstance(json_ld_item
, dict):
1317 json_ld
.append(json_ld_item
)
1318 elif isinstance(json_ld_item
, (list, tuple)):
1319 json_ld
.extend(json_ld_item
)
1321 json_ld
= self
._json
_ld
(json_ld
, video_id
, fatal
=fatal
, expected_type
=expected_type
)
1324 if default
is not NO_DEFAULT
:
1327 raise RegexNotFoundError('Unable to extract JSON-LD')
1329 self
.report_warning('unable to extract JSON-LD %s' % bug_reports_message())
1332 def _json_ld(self
, json_ld
, video_id
, fatal
=True, expected_type
=None):
1333 if isinstance(json_ld
, compat_str
):
1334 json_ld
= self
._parse
_json
(json_ld
, video_id
, fatal
=fatal
)
1338 if not isinstance(json_ld
, (list, tuple, dict)):
1340 if isinstance(json_ld
, dict):
1343 INTERACTION_TYPE_MAP
= {
1344 'CommentAction': 'comment',
1345 'AgreeAction': 'like',
1346 'DisagreeAction': 'dislike',
1347 'LikeAction': 'like',
1348 'DislikeAction': 'dislike',
1349 'ListenAction': 'view',
1350 'WatchAction': 'view',
1351 'ViewAction': 'view',
1354 def extract_interaction_type(e
):
1355 interaction_type
= e
.get('interactionType')
1356 if isinstance(interaction_type
, dict):
1357 interaction_type
= interaction_type
.get('@type')
1358 return str_or_none(interaction_type
)
1360 def extract_interaction_statistic(e
):
1361 interaction_statistic
= e
.get('interactionStatistic')
1362 if isinstance(interaction_statistic
, dict):
1363 interaction_statistic
= [interaction_statistic
]
1364 if not isinstance(interaction_statistic
, list):
1366 for is_e
in interaction_statistic
:
1367 if not isinstance(is_e
, dict):
1369 if is_e
.get('@type') != 'InteractionCounter':
1371 interaction_type
= extract_interaction_type(is_e
)
1372 if not interaction_type
:
1374 # For interaction count some sites provide string instead of
1375 # an integer (as per spec) with non digit characters (e.g. ",")
1376 # so extracting count with more relaxed str_to_int
1377 interaction_count
= str_to_int(is_e
.get('userInteractionCount'))
1378 if interaction_count
is None:
1380 count_kind
= INTERACTION_TYPE_MAP
.get(interaction_type
.split('/')[-1])
1383 count_key
= '%s_count' % count_kind
1384 if info
.get(count_key
) is not None:
1386 info
[count_key
] = interaction_count
1388 def extract_video_object(e
):
1389 assert e
['@type'] == 'VideoObject'
1390 author
= e
.get('author')
1392 'url': url_or_none(e
.get('contentUrl')),
1393 'title': unescapeHTML(e
.get('name')),
1394 'description': unescapeHTML(e
.get('description')),
1395 'thumbnail': url_or_none(e
.get('thumbnailUrl') or e
.get('thumbnailURL')),
1396 'duration': parse_duration(e
.get('duration')),
1397 'timestamp': unified_timestamp(e
.get('uploadDate')),
1398 # author can be an instance of 'Organization' or 'Person' types.
1399 # both types can have 'name' property(inherited from 'Thing' type). [1]
1400 # however some websites are using 'Text' type instead.
1401 # 1. https://schema.org/VideoObject
1402 'uploader': author
.get('name') if isinstance(author
, dict) else author
if isinstance(author
, compat_str
) else None,
1403 'filesize': float_or_none(e
.get('contentSize')),
1404 'tbr': int_or_none(e
.get('bitrate')),
1405 'width': int_or_none(e
.get('width')),
1406 'height': int_or_none(e
.get('height')),
1407 'view_count': int_or_none(e
.get('interactionCount')),
1409 extract_interaction_statistic(e
)
1413 item_type
= e
.get('@type')
1414 if expected_type
is not None and expected_type
!= item_type
:
1416 if item_type
in ('TVEpisode', 'Episode'):
1417 episode_name
= unescapeHTML(e
.get('name'))
1419 'episode': episode_name
,
1420 'episode_number': int_or_none(e
.get('episodeNumber')),
1421 'description': unescapeHTML(e
.get('description')),
1423 if not info
.get('title') and episode_name
:
1424 info
['title'] = episode_name
1425 part_of_season
= e
.get('partOfSeason')
1426 if isinstance(part_of_season
, dict) and part_of_season
.get('@type') in ('TVSeason', 'Season', 'CreativeWorkSeason'):
1428 'season': unescapeHTML(part_of_season
.get('name')),
1429 'season_number': int_or_none(part_of_season
.get('seasonNumber')),
1431 part_of_series
= e
.get('partOfSeries') or e
.get('partOfTVSeries')
1432 if isinstance(part_of_series
, dict) and part_of_series
.get('@type') in ('TVSeries', 'Series', 'CreativeWorkSeries'):
1433 info
['series'] = unescapeHTML(part_of_series
.get('name'))
1434 elif item_type
== 'Movie':
1436 'title': unescapeHTML(e
.get('name')),
1437 'description': unescapeHTML(e
.get('description')),
1438 'duration': parse_duration(e
.get('duration')),
1439 'timestamp': unified_timestamp(e
.get('dateCreated')),
1441 elif item_type
in ('Article', 'NewsArticle'):
1443 'timestamp': parse_iso8601(e
.get('datePublished')),
1444 'title': unescapeHTML(e
.get('headline')),
1445 'description': unescapeHTML(e
.get('articleBody')),
1447 elif item_type
== 'VideoObject':
1448 extract_video_object(e
)
1449 if expected_type
is None:
1453 video
= e
.get('video')
1454 if isinstance(video
, dict) and video
.get('@type') == 'VideoObject':
1455 extract_video_object(video
)
1456 if expected_type
is None:
1460 return dict((k
, v
) for k
, v
in info
.items() if v
is not None)
1463 def _hidden_inputs(html
):
1464 html
= re
.sub(r
'<!--(?:(?!<!--).)*-->', '', html
)
1466 for input in re
.findall(r
'(?i)(<input[^>]+>)', html
):
1467 attrs
= extract_attributes(input)
1470 if attrs
.get('type') not in ('hidden', 'submit'):
1472 name
= attrs
.get('name') or attrs
.get('id')
1473 value
= attrs
.get('value')
1474 if name
and value
is not None:
1475 hidden_inputs
[name
] = value
1476 return hidden_inputs
1478 def _form_hidden_inputs(self
, form_id
, html
):
1479 form
= self
._search
_regex
(
1480 r
'(?is)<form[^>]+?id=(["\'])%s\
1[^
>]*>(?P
<form
>.+?
)</form
>' % form_id,
1481 html, '%s form
' % form_id, group='form
')
1482 return self._hidden_inputs(form)
1485 regex = r' *((?P
<reverse
>\
+)?
(?P
<field
>[a
-zA
-Z0
-9_]+)((?P
<separator
>[~
:])(?P
<limit
>.*?
))?
)?
*$
'
1487 default = ('hidden
', 'aud_or_vid
', 'hasvid
', 'ie_pref
', 'lang
', 'quality
',
1488 'res
', 'fps
', 'codec
:vp9
.2
', 'size
', 'br
', 'asr
',
1489 'proto
', 'ext
', 'hasaud
', 'source
', 'format_id
') # These must not be aliases
1490 ytdl_default = ('hasaud
', 'quality
', 'tbr
', 'filesize
', 'vbr
',
1491 'height
', 'width
', 'proto
', 'vext
', 'abr
', 'aext
',
1492 'fps
', 'fs_approx
', 'source
', 'format_id
')
1495 'vcodec
': {'type': 'ordered
', 'regex
': True,
1496 'order
': ['av0?
1', 'vp0?
9.2', 'vp0?
9', '[hx
]265|he?vc?
', '[hx
]264|avc
', 'vp0?
8', 'mp4v|h263
', 'theora
', '', None, 'none
']},
1497 'acodec
': {'type': 'ordered
', 'regex
': True,
1498 'order
': ['opus
', 'vorbis
', 'aac
', 'mp?
4a?
', 'mp3
', 'e?a?c
-?
3', 'dts
', '', None, 'none
']},
1499 'proto
': {'type': 'ordered
', 'regex
': True, 'field
': 'protocol
',
1500 'order
': ['(ht|f
)tps
', '(ht|f
)tp$
', 'm3u8
.+', '.*dash
', 'ws|websocket
', '', 'mms|rtsp
', 'none
', 'f4
']},
1501 'vext
': {'type': 'ordered
', 'field
': 'video_ext
',
1502 'order
': ('mp4
', 'webm
', 'flv
', '', 'none
'),
1503 'order_free
': ('webm
', 'mp4
', 'flv
', '', 'none
')},
1504 'aext
': {'type': 'ordered
', 'field
': 'audio_ext
',
1505 'order
': ('m4a
', 'aac
', 'mp3
', 'ogg
', 'opus
', 'webm
', '', 'none
'),
1506 'order_free
': ('opus
', 'ogg
', 'webm
', 'm4a
', 'mp3
', 'aac
', '', 'none
')},
1507 'hidden
': {'visible': False, 'forced': True, 'type': 'extractor', 'max': -1000},
1508 'aud_or_vid
': {'visible
': False, 'forced
': True, 'type': 'multiple
',
1509 'field
': ('vcodec
', 'acodec
'),
1510 'function
': lambda it: int(any(v != 'none
' for v in it))},
1511 'ie_pref
': {'priority': True, 'type': 'extractor'},
1512 'hasvid
': {'priority': True, 'field': 'vcodec', 'type': 'boolean', 'not_in_list': ('none',)},
1513 'hasaud
': {'field': 'acodec', 'type': 'boolean', 'not_in_list': ('none',)},
1514 'lang
': {'priority': True, 'convert': 'ignore', 'field': 'language_preference'},
1515 'quality
': {'convert': 'float_none', 'default': -1},
1516 'filesize
': {'convert': 'bytes'},
1517 'fs_approx
': {'convert': 'bytes', 'field': 'filesize_approx'},
1518 'id': {'convert': 'string', 'field': 'format_id'},
1519 'height
': {'convert': 'float_none'},
1520 'width
': {'convert': 'float_none'},
1521 'fps
': {'convert': 'float_none'},
1522 'tbr
': {'convert': 'float_none'},
1523 'vbr
': {'convert': 'float_none'},
1524 'abr
': {'convert': 'float_none'},
1525 'asr
': {'convert': 'float_none'},
1526 'source
': {'convert': 'ignore', 'field': 'source_preference'},
1528 'codec
': {'type': 'combined', 'field': ('vcodec', 'acodec')},
1529 'br
': {'type': 'combined', 'field': ('tbr', 'vbr', 'abr'), 'same_limit': True},
1530 'size
': {'type': 'combined', 'same_limit': True, 'field': ('filesize', 'fs_approx')},
1531 'ext
': {'type': 'combined', 'field': ('vext', 'aext')},
1532 'res
': {'type': 'multiple
', 'field
': ('height
', 'width
'),
1533 'function
': lambda it: (lambda l: min(l) if l else 0)(tuple(filter(None, it)))},
1535 # Most of these exist only for compatibility reasons
1536 'dimension
': {'type': 'alias', 'field': 'res'},
1537 'resolution
': {'type': 'alias', 'field': 'res'},
1538 'extension
': {'type': 'alias', 'field': 'ext'},
1539 'bitrate
': {'type': 'alias', 'field': 'br'},
1540 'total_bitrate
': {'type': 'alias', 'field': 'tbr'},
1541 'video_bitrate
': {'type': 'alias', 'field': 'vbr'},
1542 'audio_bitrate
': {'type': 'alias', 'field': 'abr'},
1543 'framerate
': {'type': 'alias', 'field': 'fps'},
1544 'language_preference
': {'type': 'alias', 'field': 'lang'}, # not named as 'language
' because such a field exists
1545 'protocol
': {'type': 'alias', 'field': 'proto'},
1546 'source_preference
': {'type': 'alias', 'field': 'source'},
1547 'filesize_approx
': {'type': 'alias', 'field': 'fs_approx'},
1548 'filesize_estimate
': {'type': 'alias', 'field': 'size'},
1549 'samplerate
': {'type': 'alias', 'field': 'asr'},
1550 'video_ext
': {'type': 'alias', 'field': 'vext'},
1551 'audio_ext
': {'type': 'alias', 'field': 'aext'},
1552 'video_codec
': {'type': 'alias', 'field': 'vcodec'},
1553 'audio_codec
': {'type': 'alias', 'field': 'acodec'},
1554 'video
': {'type': 'alias', 'field': 'hasvid'},
1555 'has_video
': {'type': 'alias', 'field': 'hasvid'},
1556 'audio
': {'type': 'alias', 'field': 'hasaud'},
1557 'has_audio
': {'type': 'alias', 'field': 'hasaud'},
1558 'extractor
': {'type': 'alias', 'field': 'ie_pref'},
1559 'preference
': {'type': 'alias', 'field': 'ie_pref'},
1560 'extractor_preference
': {'type': 'alias', 'field': 'ie_pref'},
1561 'format_id
': {'type': 'alias', 'field': 'id'},
1566 def _get_field_setting(self, field, key):
1567 if field not in self.settings:
1568 self.settings[field] = {}
1569 propObj = self.settings[field]
1570 if key not in propObj:
1571 type = propObj.get('type')
1573 default = 'preference
' if type == 'extractor
' else (field,) if type in ('combined
', 'multiple
') else field
1574 elif key == 'convert
':
1575 default = 'order
' if type == 'ordered
' else 'float_string
' if field else 'ignore
'
1577 default = {'type': 'field', 'visible': True, 'order': [], 'not_in_list': (None,)}.get(key, None)
1578 propObj[key] = default
1581 def _resolve_field_value(self, field, value, convertNone=False):
1586 value = value.lower()
1587 conversion = self._get_field_setting(field, 'convert
')
1588 if conversion == 'ignore
':
1590 if conversion == 'string
':
1592 elif conversion == 'float_none
':
1593 return float_or_none(value)
1594 elif conversion == 'bytes':
1595 return FileDownloader.parse_bytes(value)
1596 elif conversion == 'order
':
1597 order_list = (self._use_free_order and self._get_field_setting(field, 'order_free
')) or self._get_field_setting(field, 'order
')
1598 use_regex = self._get_field_setting(field, 'regex
')
1599 list_length = len(order_list)
1600 empty_pos = order_list.index('') if '' in order_list else list_length + 1
1601 if use_regex and value is not None:
1602 for i, regex in enumerate(order_list):
1603 if regex and re.match(regex, value):
1604 return list_length - i
1605 return list_length - empty_pos # not in list
1606 else: # not regex or value = None
1607 return list_length - (order_list.index(value) if value in order_list else empty_pos)
1609 if value.isnumeric():
1612 self.settings[field]['convert
'] = 'string
'
1615 def evaluate_params(self, params, sort_extractor):
1616 self._use_free_order = params.get('prefer_free_formats
', False)
1617 self._sort_user = params.get('format_sort
', [])
1618 self._sort_extractor = sort_extractor
1620 def add_item(field, reverse, closest, limit_text):
1621 field = field.lower()
1622 if field in self._order:
1624 self._order.append(field)
1625 limit = self._resolve_field_value(field, limit_text)
1628 'closest
': False if limit is None else closest,
1629 'limit_text
': limit_text,
1631 if field in self.settings:
1632 self.settings[field].update(data)
1634 self.settings[field] = data
1637 tuple(field for field in self.default if self._get_field_setting(field, 'forced
'))
1638 + (tuple() if params.get('format_sort_force
', False)
1639 else tuple(field for field in self.default if self._get_field_setting(field, 'priority
')))
1640 + tuple(self._sort_user) + tuple(sort_extractor) + self.default)
1642 for item in sort_list:
1643 match = re.match(self.regex, item)
1645 raise ExtractorError('Invalid format sort string
"%s" given by extractor
' % item)
1646 field = match.group('field
')
1649 if self._get_field_setting(field, 'type') == 'alias
':
1650 field = self._get_field_setting(field, 'field
')
1651 reverse = match.group('reverse
') is not None
1652 closest = match.group('separator
') == '~
'
1653 limit_text = match.group('limit
')
1655 has_limit = limit_text is not None
1656 has_multiple_fields = self._get_field_setting(field, 'type') == 'combined
'
1657 has_multiple_limits = has_limit and has_multiple_fields and not self._get_field_setting(field, 'same_limit
')
1659 fields = self._get_field_setting(field, 'field
') if has_multiple_fields else (field,)
1660 limits = limit_text.split(":") if has_multiple_limits else (limit_text,) if has_limit else tuple()
1661 limit_count = len(limits)
1662 for (i, f) in enumerate(fields):
1663 add_item(f, reverse, closest,
1664 limits[i] if i < limit_count
1665 else limits[0] if has_limit and not has_multiple_limits
1668 def print_verbose_info(self, write_debug):
1670 write_debug('Sort order given by user
: %s' % ', '.join(self._sort_user))
1671 if self._sort_extractor:
1672 write_debug('Sort order given by extractor
: %s' % ', '.join(self._sort_extractor))
1673 write_debug('Formats
sorted by
: %s' % ', '.join(['%s%s%s' % (
1674 '+' if self._get_field_setting(field, 'reverse
') else '', field,
1675 '%s%s(%s)' % ('~
' if self._get_field_setting(field, 'closest
') else ':',
1676 self._get_field_setting(field, 'limit_text
'),
1677 self._get_field_setting(field, 'limit
'))
1678 if self._get_field_setting(field, 'limit_text
') is not None else '')
1679 for field in self._order if self._get_field_setting(field, 'visible
')]))
1681 def _calculate_field_preference_from_value(self, format, field, type, value):
1682 reverse = self._get_field_setting(field, 'reverse
')
1683 closest = self._get_field_setting(field, 'closest
')
1684 limit = self._get_field_setting(field, 'limit
')
1686 if type == 'extractor
':
1687 maximum = self._get_field_setting(field, 'max')
1688 if value is None or (maximum is not None and value >= maximum):
1690 elif type == 'boolean
':
1691 in_list = self._get_field_setting(field, 'in_list
')
1692 not_in_list = self._get_field_setting(field, 'not_in_list
')
1693 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
1694 elif type == 'ordered
':
1695 value = self._resolve_field_value(field, value, True)
1697 # try to convert to number
1698 val_num = float_or_none(value, default=self._get_field_setting(field, 'default
'))
1699 is_num = self._get_field_setting(field, 'convert
') != 'string
' and val_num is not None
1703 return ((-10, 0) if value is None
1704 else (1, value, 0) if not is_num # if a field has mixed strings and numbers, strings are sorted higher
1705 else (0, -abs(value - limit), value - limit if reverse else limit - value) if closest
1706 else (0, value, 0) if not reverse and (limit is None or value <= limit)
1707 else (0, -value, 0) if limit is None or (reverse and value == limit) or value > limit
1708 else (-1, value, 0))
1710 def _calculate_field_preference(self, format, field):
1711 type = self._get_field_setting(field, 'type') # extractor, boolean, ordered, field, multiple
1712 get_value = lambda f: format.get(self._get_field_setting(f, 'field
'))
1713 if type == 'multiple
':
1714 type = 'field
' # Only 'field
' is allowed in multiple for now
1715 actual_fields = self._get_field_setting(field, 'field
')
1717 value = self._get_field_setting(field, 'function
')(get_value(f) for f in actual_fields)
1719 value = get_value(field)
1720 return self._calculate_field_preference_from_value(format, field, type, value)
1722 def calculate_preference(self, format):
1723 # Determine missing protocol
1724 if not format.get('protocol
'):
1725 format['protocol
'] = determine_protocol(format)
1727 # Determine missing ext
1728 if not format.get('ext
') and 'url
' in format:
1729 format['ext
'] = determine_ext(format['url
'])
1730 if format.get('vcodec
') == 'none
':
1731 format['audio_ext
'] = format['ext
'] if format.get('acodec
') != 'none
' else 'none
'
1732 format['video_ext
'] = 'none
'
1734 format['video_ext
'] = format['ext
']
1735 format['audio_ext
'] = 'none
'
1736 # if format.get('preference
') is None and format.get('ext
') in ('f4f
', 'f4m
'): # Not supported?
1737 # format['preference
'] = -1000
1739 # Determine missing bitrates
1740 if format.get('tbr
') is None:
1741 if format.get('vbr
') is not None and format.get('abr
') is not None:
1742 format['tbr
'] = format.get('vbr
', 0) + format.get('abr
', 0)
1744 if format.get('vcodec
') != "none" and format.get('vbr
') is None:
1745 format['vbr
'] = format.get('tbr
') - format.get('abr
', 0)
1746 if format.get('acodec
') != "none" and format.get('abr
') is None:
1747 format['abr
'] = format.get('tbr
') - format.get('vbr
', 0)
1749 return tuple(self._calculate_field_preference(format, field) for field in self._order)
1751 def _sort_formats(self, formats, field_preference=[]):
1753 if self.get_param('ignore_no_formats_error
'):
1755 raise ExtractorError('No video formats found
')
1756 format_sort = self.FormatSort() # params and to_screen are taken from the downloader
1757 format_sort.evaluate_params(self._downloader.params, field_preference)
1758 if self.get_param('verbose
', False):
1759 format_sort.print_verbose_info(self._downloader.write_debug)
1760 formats.sort(key=lambda f: format_sort.calculate_preference(f))
1762 def _check_formats(self, formats, video_id):
1764 formats[:] = filter(
1765 lambda f: self._is_valid_url(
1767 item='%s video format
' % f.get('format_id
') if f.get('format_id
') else 'video
'),
1771 def _remove_duplicate_formats(formats):
1775 if f['url
'] not in format_urls:
1776 format_urls.add(f['url
'])
1777 unique_formats.append(f)
1778 formats[:] = unique_formats
1780 def _is_valid_url(self, url, video_id, item='video
', headers={}):
1781 url = self._proto_relative_url(url, scheme='http
:')
1782 # For now assume non HTTP(S) URLs always valid
1783 if not (url.startswith('http
://') or url.startswith('https
://')):
1786 self._request_webpage(url, video_id, 'Checking
%s URL
' % item, headers=headers)
1788 except ExtractorError as e:
1790 '%s: %s URL
is invalid
, skipping
: %s'
1791 % (video_id, item, error_to_compat_str(e.cause)))
1794 def http_scheme(self):
1795 """ Either "http:" or "https:", depending on the user's preferences
"""
1798 if self.get_param('prefer_insecure', False)
1801 def _proto_relative_url(self, url, scheme=None):
1804 if url.startswith('//'):
1806 scheme = self.http_scheme()
1811 def _sleep(self, timeout, video_id, msg_template=None):
1812 if msg_template is None:
1813 msg_template = '%(video_id)s: Waiting for %(timeout)s seconds'
1814 msg = msg_template % {'video_id': video_id, 'timeout': timeout}
1818 def _extract_f4m_formats(self, manifest_url, video_id, preference=None, quality=None, f4m_id=None,
1819 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1820 fatal=True, m3u8_id=None, data=None, headers={}, query={}):
1821 manifest = self._download_xml(
1822 manifest_url, video_id, 'Downloading f4m manifest',
1823 'Unable to download f4m manifest',
1824 # Some manifests may be malformed, e.g. prosiebensat1 generated manifests
1825 # (see https://github.com/ytdl-org/youtube-dl/issues/6215#issuecomment-121704244)
1826 transform_source=transform_source,
1827 fatal=fatal, data=data, headers=headers, query=query)
1829 if manifest is False:
1832 return self._parse_f4m_formats(
1833 manifest, manifest_url, video_id, preference=preference, quality=quality, f4m_id=f4m_id,
1834 transform_source=transform_source, fatal=fatal, m3u8_id=m3u8_id)
1836 def _parse_f4m_formats(self, manifest, manifest_url, video_id, preference=None, quality=None, f4m_id=None,
1837 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1838 fatal=True, m3u8_id=None):
1839 if not isinstance(manifest, compat_etree_Element) and not fatal:
1842 # currently yt-dlp cannot decode the playerVerificationChallenge as Akamai uses Adobe Alchemy
1843 akamai_pv = manifest.find('{http://ns.adobe.com/f4m/1.0}pv-2.0')
1844 if akamai_pv is not None and ';' in akamai_pv.text:
1845 playerVerificationChallenge = akamai_pv.text.split(';')[0]
1846 if playerVerificationChallenge.strip() != '':
1850 manifest_version = '1.0'
1851 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/1.0}media')
1853 manifest_version = '2.0'
1854 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/2.0}media')
1855 # Remove unsupported DRM protected media from final formats
1856 # rendition (see https://github.com/ytdl-org/youtube-dl/issues/8573).
1857 media_nodes = remove_encrypted_media(media_nodes)
1861 manifest_base_url = get_base_url(manifest)
1863 bootstrap_info = xpath_element(
1864 manifest, ['{http://ns.adobe.com/f4m/1.0}bootstrapInfo', '{http://ns.adobe.com/f4m/2.0}bootstrapInfo'],
1865 'bootstrap info', default=None)
1868 mime_type = xpath_text(
1869 manifest, ['{http://ns.adobe.com/f4m/1.0}mimeType', '{http://ns.adobe.com/f4m/2.0}mimeType'],
1870 'base URL', default=None)
1871 if mime_type and mime_type.startswith('audio/'):
1874 for i, media_el in enumerate(media_nodes):
1875 tbr = int_or_none(media_el.attrib.get('bitrate'))
1876 width = int_or_none(media_el.attrib.get('width'))
1877 height = int_or_none(media_el.attrib.get('height'))
1878 format_id = '-'.join(filter(None, [f4m_id, compat_str(i if tbr is None else tbr)]))
1879 # If <bootstrapInfo> is present, the specified f4m is a
1880 # stream-level manifest, and only set-level manifests may refer to
1881 # external resources. See section 11.4 and section 4 of F4M spec
1882 if bootstrap_info is None:
1884 # @href is introduced in 2.0, see section 11.6 of F4M spec
1885 if manifest_version == '2.0':
1886 media_url = media_el.attrib.get('href')
1887 if media_url is None:
1888 media_url = media_el.attrib.get('url')
1892 media_url if media_url.startswith('http://') or media_url.startswith('https://')
1893 else ((manifest_base_url or '/'.join(manifest_url.split('/')[:-1])) + '/' + media_url))
1894 # If media_url is itself a f4m manifest do the recursive extraction
1895 # since bitrates in parent manifest (this one) and media_url manifest
1896 # may differ leading to inability to resolve the format by requested
1897 # bitrate in f4m downloader
1898 ext = determine_ext(manifest_url)
1900 f4m_formats = self._extract_f4m_formats(
1901 manifest_url, video_id, preference=preference, quality=quality, f4m_id=f4m_id,
1902 transform_source=transform_source, fatal=fatal)
1903 # Sometimes stream-level manifest contains single media entry that
1904 # does not contain any quality metadata (e.g. http://matchtv.ru/#live-player).
1905 # At the same time parent's media entry in set-level manifest may
1906 # contain it. We will copy it from parent in such cases.
1907 if len(f4m_formats) == 1:
1910 'tbr': f.get('tbr') or tbr,
1911 'width': f.get('width') or width,
1912 'height': f.get('height') or height,
1913 'format_id': f.get('format_id') if not tbr else format_id,
1916 formats.extend(f4m_formats)
1919 formats.extend(self._extract_m3u8_formats(
1920 manifest_url, video_id, 'mp4', preference=preference,
1921 quality=quality, m3u8_id=m3u8_id, fatal=fatal))
1924 'format_id': format_id,
1925 'url': manifest_url,
1926 'manifest_url': manifest_url,
1927 'ext': 'flv' if bootstrap_info is not None else None,
1933 'preference': preference,
1938 def _m3u8_meta_format(self, m3u8_url, ext=None, preference=None, quality=None, m3u8_id=None):
1940 'format_id': '-'.join(filter(None, [m3u8_id, 'meta'])),
1944 'preference': preference - 100 if preference else -100,
1946 'resolution': 'multiple',
1947 'format_note': 'Quality selection URL',
1950 def _extract_m3u8_formats(self, *args, **kwargs):
1951 fmts, subs = self._extract_m3u8_formats_and_subtitles(*args, **kwargs)
1953 self.report_warning(bug_reports_message(
1954 "Ignoring subtitle tracks found in the HLS manifest; "
1955 "if any subtitle tracks are missing,"
1959 def _extract_m3u8_formats_and_subtitles(
1960 self, m3u8_url, video_id, ext=None, entry_protocol='m3u8_native',
1961 preference=None, quality=None, m3u8_id=None, note=None,
1962 errnote=None, fatal=True, live=False, data=None, headers={},
1965 res = self._download_webpage_handle(
1967 note='Downloading m3u8 information' if note is None else note,
1968 errnote='Failed to download m3u8 information' if errnote is None else errnote,
1969 fatal=fatal, data=data, headers=headers, query=query)
1974 m3u8_doc, urlh = res
1975 m3u8_url = urlh.geturl()
1977 return self._parse_m3u8_formats_and_subtitles(
1978 m3u8_doc, m3u8_url, ext=ext, entry_protocol=entry_protocol,
1979 preference=preference, quality=quality, m3u8_id=m3u8_id,
1980 note=note, errnote=errnote, fatal=fatal, live=live, data=data,
1981 headers=headers, query=query, video_id=video_id)
1983 def _parse_m3u8_formats_and_subtitles(
1984 self, m3u8_doc, m3u8_url, ext=None, entry_protocol='m3u8_native',
1985 preference=None, quality=None, m3u8_id=None, live=False, note=None,
1986 errnote=None, fatal=True, data=None, headers={}, query={},
1988 formats, subtitles = [], {}
1990 if '#EXT-X-FAXS-CM:' in m3u8_doc: # Adobe Flash Access
1991 return formats, subtitles
1993 if (not self.get_param('allow_unplayable_formats')
1994 and re.search(r'#EXT-X-SESSION-KEY:.*?URI="skd://', m3u8_doc)): # Apple FairPlay
1995 return formats, subtitles
1997 def format_url(url):
1998 return url if re.match(r'^https?://', url) else compat_urlparse.urljoin(m3u8_url, url)
2000 if self.get_param('hls_split_discontinuity', False):
2001 def _extract_m3u8_playlist_indices(manifest_url=None, m3u8_doc=None):
2003 if not manifest_url:
2005 m3u8_doc = self._download_webpage(
2006 manifest_url, video_id, fatal=fatal, data=data, headers=headers,
2007 note=False, errnote='Failed to download m3u8 playlist information')
2008 if m3u8_doc is False:
2010 return range(1 + sum(line.startswith('#EXT-X-DISCONTINUITY') for line in m3u8_doc.splitlines()))
2013 def _extract_m3u8_playlist_indices(*args, **kwargs):
2017 # 1. https://tools.ietf.org/html/draft-pantos-http-live-streaming-21
2018 # 2. https://github.com/ytdl-org/youtube-dl/issues/12211
2019 # 3. https://github.com/ytdl-org/youtube-dl/issues/18923
2021 # We should try extracting formats only from master playlists [1, 4.3.4],
2022 # i.e. playlists that describe available qualities. On the other hand
2023 # media playlists [1, 4.3.3] should be returned as is since they contain
2024 # just the media without qualities renditions.
2025 # Fortunately, master playlist can be easily distinguished from media
2026 # playlist based on particular tags availability. As of [1, 4.3.3, 4.3.4]
2027 # master playlist tags MUST NOT appear in a media playlist and vice versa.
2028 # As of [1, 4.3.3.1] #EXT-X-TARGETDURATION tag is REQUIRED for every
2029 # media playlist and MUST NOT appear in master playlist thus we can
2030 # clearly detect media playlist with this criterion.
2032 if '#EXT-X-TARGETDURATION' in m3u8_doc: # media playlist, return as is
2034 'format_id': '-'.join(map(str, filter(None, [m3u8_id, idx]))),
2035 'format_index': idx,
2038 'protocol': entry_protocol,
2039 'preference': preference,
2041 } for idx in _extract_m3u8_playlist_indices(m3u8_doc=m3u8_doc)]
2043 return formats, subtitles
2046 last_stream_inf = {}
2048 def extract_media(x_media_line):
2049 media = parse_m3u8_attributes(x_media_line)
2050 # As per [1, 4.3.4.1] TYPE, GROUP-ID and NAME are REQUIRED
2051 media_type, group_id, name = media.get('TYPE'), media.get('GROUP-ID'), media.get('NAME')
2052 if not (media_type and group_id and name):
2054 groups.setdefault(group_id, []).append(media)
2055 # <https://tools.ietf.org/html/rfc8216#section-4.3.4.1>
2056 if media_type == 'SUBTITLES':
2057 # According to RFC 8216 §4.3.4.2.1, URI is REQUIRED in the
2058 # EXT-X-MEDIA tag if the media type is SUBTITLES.
2059 # However, lack of URI has been spotted in the wild.
2060 # e.g. NebulaIE; see https://github.com/yt-dlp/yt-dlp/issues/339
2061 if not media.get('URI'):
2063 url = format_url(media['URI'])
2066 'ext': determine_ext(url),
2068 if sub_info['ext'] == 'm3u8':
2069 # Per RFC 8216 §3.1, the only possible subtitle format m3u8
2070 # files may contain is WebVTT:
2071 # <https://tools.ietf.org/html/rfc8216#section-3.1>
2072 sub_info['ext'] = 'vtt'
2073 sub_info['protocol'] = 'm3u8_native'
2074 lang = media.get('LANGUAGE') or 'und'
2075 subtitles.setdefault(lang, []).append(sub_info)
2076 if media_type not in ('VIDEO', 'AUDIO'):
2078 media_url = media.get('URI')
2080 manifest_url = format_url(media_url)
2082 'format_id': '-'.join(map(str, filter(None, (m3u8_id, group_id, name, idx)))),
2083 'format_note': name,
2084 'format_index': idx,
2085 'url': manifest_url,
2086 'manifest_url': m3u8_url,
2087 'language': media.get('LANGUAGE'),
2089 'protocol': entry_protocol,
2090 'preference': preference,
2092 'vcodec': 'none' if media_type == 'AUDIO' else None,
2093 } for idx in _extract_m3u8_playlist_indices(manifest_url))
2095 def build_stream_name():
2096 # Despite specification does not mention NAME attribute for
2097 # EXT-X-STREAM-INF tag it still sometimes may be present (see [1]
2098 # or vidio test in TestInfoExtractor.test_parse_m3u8_formats)
2099 # 1. http://www.vidio.com/watch/165683-dj_ambred-booyah-live-2015
2100 stream_name = last_stream_inf.get('NAME')
2103 # If there is no NAME in EXT-X-STREAM-INF it will be obtained
2104 # from corresponding rendition group
2105 stream_group_id = last_stream_inf.get('VIDEO')
2106 if not stream_group_id:
2108 stream_group = groups.get(stream_group_id)
2109 if not stream_group:
2110 return stream_group_id
2111 rendition = stream_group[0]
2112 return rendition.get('NAME') or stream_group_id
2114 # parse EXT-X-MEDIA tags before EXT-X-STREAM-INF in order to have the
2115 # chance to detect video only formats when EXT-X-STREAM-INF tags
2116 # precede EXT-X-MEDIA tags in HLS manifest such as [3].
2117 for line in m3u8_doc.splitlines():
2118 if line.startswith('#EXT-X-MEDIA:'):
2121 for line in m3u8_doc.splitlines():
2122 if line.startswith('#EXT-X-STREAM-INF:'):
2123 last_stream_inf = parse_m3u8_attributes(line)
2124 elif line.startswith('#') or not line.strip():
2127 tbr = float_or_none(
2128 last_stream_inf.get('AVERAGE-BANDWIDTH')
2129 or last_stream_inf.get('BANDWIDTH'), scale=1000)
2130 manifest_url = format_url(line.strip())
2132 for idx in _extract_m3u8_playlist_indices(manifest_url):
2133 format_id = [m3u8_id, None, idx]
2134 # Bandwidth of live streams may differ over time thus making
2135 # format_id unpredictable. So it's better to keep provided
2138 stream_name = build_stream_name()
2139 format_id[1] = stream_name if stream_name else '%d' % (tbr if tbr else len(formats))
2141 'format_id': '-'.join(map(str, filter(None, format_id))),
2142 'format_index': idx,
2143 'url': manifest_url,
2144 'manifest_url': m3u8_url,
2147 'fps': float_or_none(last_stream_inf.get('FRAME-RATE')),
2148 'protocol': entry_protocol,
2149 'preference': preference,
2152 resolution = last_stream_inf.get('RESOLUTION')
2154 mobj = re.search(r'(?P<width>\d+)[xX](?P<height>\d+)', resolution)
2156 f['width'] = int(mobj.group('width'))
2157 f['height'] = int(mobj.group('height'))
2158 # Unified Streaming Platform
2160 r'audio.*?(?:%3D|=)(\d+)(?:-video.*?(?:%3D|=)(\d+))?', f['url'])
2162 abr, vbr = mobj.groups()
2163 abr, vbr = float_or_none(abr, 1000), float_or_none(vbr, 1000)
2168 codecs = parse_codecs(last_stream_inf.get('CODECS'))
2170 audio_group_id = last_stream_inf.get('AUDIO')
2171 # As per [1, 4.3.4.1.1] any EXT-X-STREAM-INF tag which
2172 # references a rendition group MUST have a CODECS attribute.
2173 # However, this is not always respected, for example, [2]
2174 # contains EXT-X-STREAM-INF tag which references AUDIO
2175 # rendition group but does not have CODECS and despite
2176 # referencing an audio group it represents a complete
2177 # (with audio and video) format. So, for such cases we will
2178 # ignore references to rendition groups and treat them
2179 # as complete formats.
2180 if audio_group_id and codecs and f.get('vcodec') != 'none':
2181 audio_group = groups.get(audio_group_id)
2182 if audio_group and audio_group[0].get('URI'):
2183 # TODO: update acodec for audio only formats with
2185 f['acodec'] = 'none'
2186 if not f.get('ext'):
2187 f['ext'] = 'm4a' if f.get('vcodec') == 'none' else 'mp4'
2191 progressive_uri = last_stream_inf.get('PROGRESSIVE-URI')
2194 del http_f['manifest_url']
2196 'format_id': f['format_id'].replace('hls-', 'http-'),
2198 'url': progressive_uri,
2200 formats.append(http_f)
2202 last_stream_inf = {}
2203 return formats, subtitles
2206 def _xpath_ns(path, namespace=None):
2210 for c in path.split('/'):
2211 if not c or c == '.':
2214 out.append('{%s}%s' % (namespace, c))
2215 return '/'.join(out)
2217 def _extract_smil_formats_and_subtitles(self, smil_url, video_id, fatal=True, f4m_params=None, transform_source=None):
2218 smil = self._download_smil(smil_url, video_id, fatal=fatal, transform_source=transform_source)
2224 namespace = self._parse_smil_namespace(smil)
2226 fmts = self._parse_smil_formats(
2227 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
2228 subs = self._parse_smil_subtitles(
2229 smil, namespace=namespace)
2233 def _extract_smil_formats(self, *args, **kwargs):
2234 fmts, subs = self._extract_smil_formats_and_subtitles(*args, **kwargs)
2236 self.report_warning(bug_reports_message(
2237 "Ignoring subtitle tracks found in the SMIL manifest; "
2238 "if any subtitle tracks are missing,"
2242 def _extract_smil_info(self, smil_url, video_id, fatal=True, f4m_params=None):
2243 smil = self._download_smil(smil_url, video_id, fatal=fatal)
2246 return self._parse_smil(smil, smil_url, video_id, f4m_params=f4m_params)
2248 def _download_smil(self, smil_url, video_id, fatal=True, transform_source=None):
2249 return self._download_xml(
2250 smil_url, video_id, 'Downloading SMIL file',
2251 'Unable to download SMIL file', fatal=fatal, transform_source=transform_source)
2253 def _parse_smil(self, smil, smil_url, video_id, f4m_params=None):
2254 namespace = self._parse_smil_namespace(smil)
2256 formats = self._parse_smil_formats(
2257 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
2258 subtitles = self._parse_smil_subtitles(smil, namespace=namespace)
2260 video_id = os.path.splitext(url_basename(smil_url))[0]
2264 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
2265 name = meta.attrib.get('name')
2266 content = meta.attrib.get('content')
2267 if not name or not content:
2269 if not title and name == 'title':
2271 elif not description and name in ('description', 'abstract'):
2272 description = content
2273 elif not upload_date and name == 'date':
2274 upload_date = unified_strdate(content)
2277 'id': image.get('type'),
2278 'url': image.get('src'),
2279 'width': int_or_none(image.get('width')),
2280 'height': int_or_none(image.get('height')),
2281 } for image in smil.findall(self._xpath_ns('.//image', namespace)) if image.get('src')]
2285 'title': title or video_id,
2286 'description': description,
2287 'upload_date': upload_date,
2288 'thumbnails': thumbnails,
2290 'subtitles': subtitles,
2293 def _parse_smil_namespace(self, smil):
2294 return self._search_regex(
2295 r'(?i)^{([^}]+)?}smil$', smil.tag, 'namespace', default=None)
2297 def _parse_smil_formats(self, smil, smil_url, video_id, namespace=None, f4m_params=None, transform_rtmp_url=None):
2299 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
2300 b = meta.get('base') or meta.get('httpBase')
2311 media = smil.findall(self._xpath_ns('.//video', namespace)) + smil.findall(self._xpath_ns('.//audio', namespace))
2312 for medium in media:
2313 src = medium.get('src')
2314 if not src or src in srcs:
2318 bitrate = float_or_none(medium.get('system-bitrate') or medium.get('systemBitrate'), 1000)
2319 filesize = int_or_none(medium.get('size') or medium.get('fileSize'))
2320 width = int_or_none(medium.get('width'))
2321 height = int_or_none(medium.get('height'))
2322 proto = medium.get('proto')
2323 ext = medium.get('ext')
2324 src_ext = determine_ext(src)
2325 streamer = medium.get('streamer') or base
2327 if proto == 'rtmp' or streamer.startswith('rtmp'):
2333 'format_id': 'rtmp-%d' % (rtmp_count if bitrate is None else bitrate),
2335 'filesize': filesize,
2339 if transform_rtmp_url:
2340 streamer, src = transform_rtmp_url(streamer, src)
2341 formats[-1].update({
2347 src_url = src if src.startswith('http') else compat_urlparse.urljoin(base, src)
2348 src_url = src_url.strip()
2350 if proto == 'm3u8' or src_ext == 'm3u8':
2351 m3u8_formats = self._extract_m3u8_formats(
2352 src_url, video_id, ext or 'mp4', m3u8_id='hls', fatal=False)
2353 if len(m3u8_formats) == 1:
2355 m3u8_formats[0].update({
2356 'format_id': 'hls-%d' % (m3u8_count if bitrate is None else bitrate),
2361 formats.extend(m3u8_formats)
2362 elif src_ext == 'f4m':
2367 'plugin': 'flowplayer-3.2.0.1',
2369 f4m_url += '&' if '?' in f4m_url else '?'
2370 f4m_url += compat_urllib_parse_urlencode(f4m_params)
2371 formats.extend(self._extract_f4m_formats(f4m_url, video_id, f4m_id='hds', fatal=False))
2372 elif src_ext == 'mpd':
2373 formats.extend(self._extract_mpd_formats(
2374 src_url, video_id, mpd_id='dash', fatal=False))
2375 elif re.search(r'\.ism/[Mm]anifest', src_url):
2376 formats.extend(self._extract_ism_formats(
2377 src_url, video_id, ism_id='mss', fatal=False))
2378 elif src_url.startswith('http') and self._is_valid_url(src, video_id):
2382 'ext': ext or src_ext or 'flv',
2383 'format_id': 'http-%d' % (bitrate or http_count),
2385 'filesize': filesize,
2392 def _parse_smil_subtitles(self, smil, namespace=None, subtitles_lang='en'):
2395 for num, textstream in enumerate(smil.findall(self._xpath_ns('.//textstream', namespace))):
2396 src = textstream.get('src')
2397 if not src or src in urls:
2400 ext = textstream.get('ext') or mimetype2ext(textstream.get('type')) or determine_ext(src)
2401 lang = textstream.get('systemLanguage') or textstream.get('systemLanguageName') or textstream.get('lang') or subtitles_lang
2402 subtitles.setdefault(lang, []).append({
2408 def _extract_xspf_playlist(self, xspf_url, playlist_id, fatal=True):
2409 xspf = self._download_xml(
2410 xspf_url, playlist_id, 'Downloading xpsf playlist',
2411 'Unable to download xspf manifest', fatal=fatal)
2414 return self._parse_xspf(
2415 xspf, playlist_id, xspf_url=xspf_url,
2416 xspf_base_url=base_url(xspf_url))
2418 def _parse_xspf(self, xspf_doc, playlist_id, xspf_url=None, xspf_base_url=None):
2420 'xspf': 'http://xspf.org/ns/0/',
2421 's1': 'http://static.streamone.nl/player/ns/0',
2425 for track in xspf_doc.findall(xpath_with_ns('./xspf:trackList/xspf:track', NS_MAP)):
2427 track, xpath_with_ns('./xspf:title', NS_MAP), 'title', default=playlist_id)
2428 description = xpath_text(
2429 track, xpath_with_ns('./xspf:annotation', NS_MAP), 'description')
2430 thumbnail = xpath_text(
2431 track, xpath_with_ns('./xspf:image', NS_MAP), 'thumbnail')
2432 duration = float_or_none(
2433 xpath_text(track, xpath_with_ns('./xspf:duration', NS_MAP), 'duration'), 1000)
2436 for location in track.findall(xpath_with_ns('./xspf:location', NS_MAP)):
2437 format_url = urljoin(xspf_base_url, location.text)
2442 'manifest_url': xspf_url,
2443 'format_id': location.get(xpath_with_ns('s1:label', NS_MAP)),
2444 'width': int_or_none(location.get(xpath_with_ns('s1:width', NS_MAP))),
2445 'height': int_or_none(location.get(xpath_with_ns('s1:height', NS_MAP))),
2447 self._sort_formats(formats)
2452 'description': description,
2453 'thumbnail': thumbnail,
2454 'duration': duration,
2459 def _extract_mpd_formats(self, *args, **kwargs):
2460 fmts, subs = self._extract_mpd_formats_and_subtitles(*args, **kwargs)
2462 self.report_warning(bug_reports_message(
2463 "Ignoring subtitle tracks found in the DASH manifest; "
2464 "if any subtitle tracks are missing,"
2468 def _extract_mpd_formats_and_subtitles(
2469 self, mpd_url, video_id, mpd_id=None, note=None, errnote=None,
2470 fatal=True, data=None, headers={}, query={}):
2471 res = self._download_xml_handle(
2473 note='Downloading MPD manifest' if note is None else note,
2474 errnote='Failed to download MPD manifest' if errnote is None else errnote,
2475 fatal=fatal, data=data, headers=headers, query=query)
2481 mpd_base_url = base_url(urlh.geturl())
2483 return self._parse_mpd_formats_and_subtitles(
2484 mpd_doc, mpd_id, mpd_base_url, mpd_url)
2486 def _parse_mpd_formats(self, *args, **kwargs):
2487 fmts, subs = self._parse_mpd_formats_and_subtitles(*args, **kwargs)
2489 self.report_warning(bug_reports_message(
2490 "Ignoring subtitle tracks found in the DASH manifest; "
2491 "if any subtitle tracks are missing,"
2495 def _parse_mpd_formats_and_subtitles(
2496 self, mpd_doc, mpd_id=None, mpd_base_url='', mpd_url=None):
2498 Parse formats
from MPD manifest
.
2500 1. MPEG
-DASH Standard
, ISO
/IEC
23009-1:2014(E
),
2501 http
://standards
.iso
.org
/ittf
/PubliclyAvailableStandards
/c065274_ISO_IEC_23009
-1_2014.zip
2502 2. https
://en
.wikipedia
.org
/wiki
/Dynamic_Adaptive_Streaming_over_HTTP
2504 if not self.get_param('dynamic_mpd', True):
2505 if mpd_doc.get('type') == 'dynamic':
2508 namespace = self._search_regex(r'(?i)^{([^}]+)?}MPD$', mpd_doc.tag, 'namespace', default=None)
2511 return self._xpath_ns(path, namespace)
2513 def is_drm_protected(element):
2514 return element.find(_add_ns('ContentProtection')) is not None
2516 def extract_multisegment_info(element, ms_parent_info):
2517 ms_info = ms_parent_info.copy()
2519 # As per [1, 5.3.9.2.2] SegmentList and SegmentTemplate share some
2520 # common attributes and elements. We will only extract relevant
2522 def extract_common(source):
2523 segment_timeline = source.find(_add_ns('SegmentTimeline'))
2524 if segment_timeline is not None:
2525 s_e = segment_timeline.findall(_add_ns('S'))
2527 ms_info['total_number'] = 0
2530 r = int(s.get('r', 0))
2531 ms_info['total_number'] += 1 + r
2532 ms_info['s'].append({
2533 't': int(s.get('t', 0)),
2534 # @d is mandatory (see [1, 5.3.9.6.2, Table 17, page 60])
2535 'd': int(s.attrib['d']),
2538 start_number = source.get('startNumber')
2540 ms_info['start_number'] = int(start_number)
2541 timescale = source.get('timescale')
2543 ms_info['timescale'] = int(timescale)
2544 segment_duration = source.get('duration')
2545 if segment_duration:
2546 ms_info['segment_duration'] = float(segment_duration)
2548 def extract_Initialization(source):
2549 initialization = source.find(_add_ns('Initialization'))
2550 if initialization is not None:
2551 ms_info['initialization_url'] = initialization.attrib['sourceURL']
2553 segment_list = element.find(_add_ns('SegmentList'))
2554 if segment_list is not None:
2555 extract_common(segment_list)
2556 extract_Initialization(segment_list)
2557 segment_urls_e = segment_list.findall(_add_ns('SegmentURL'))
2559 ms_info['segment_urls'] = [segment.attrib['media'] for segment in segment_urls_e]
2561 segment_template = element.find(_add_ns('SegmentTemplate'))
2562 if segment_template is not None:
2563 extract_common(segment_template)
2564 media = segment_template.get('media')
2566 ms_info['media'] = media
2567 initialization = segment_template.get('initialization')
2569 ms_info['initialization'] = initialization
2571 extract_Initialization(segment_template)
2574 skip_unplayable = not self.get_param('allow_unplayable_formats')
2576 mpd_duration = parse_duration(mpd_doc.get('mediaPresentationDuration'))
2579 for period in mpd_doc.findall(_add_ns('Period')):
2580 period_duration = parse_duration(period.get('duration')) or mpd_duration
2581 period_ms_info = extract_multisegment_info(period, {
2585 for adaptation_set in period.findall(_add_ns('AdaptationSet')):
2586 if skip_unplayable and is_drm_protected(adaptation_set):
2588 adaption_set_ms_info = extract_multisegment_info(adaptation_set, period_ms_info)
2589 for representation in adaptation_set.findall(_add_ns('Representation')):
2590 if skip_unplayable and is_drm_protected(representation):
2592 representation_attrib = adaptation_set.attrib.copy()
2593 representation_attrib.update(representation.attrib)
2594 # According to [1, 5.3.7.2, Table 9, page 41], @mimeType is mandatory
2595 mime_type = representation_attrib['mimeType']
2596 content_type = representation_attrib.get('contentType', mime_type.split('/')[0])
2598 codecs = representation_attrib.get('codecs', '')
2599 if content_type not in ('video', 'audio', 'text'):
2600 if mime_type == 'image/jpeg':
2601 content_type = mime_type
2602 elif codecs.split('.')[0] == 'stpp':
2603 content_type = 'text'
2605 self.report_warning('Unknown MIME type %s in DASH manifest' % mime_type)
2609 for element in (representation, adaptation_set, period, mpd_doc):
2610 base_url_e = element.find(_add_ns('BaseURL'))
2611 if base_url_e is not None:
2612 base_url = base_url_e.text + base_url
2613 if re.match(r'^https?://', base_url):
2615 if mpd_base_url and not re.match(r'^https?://', base_url):
2616 if not mpd_base_url.endswith('/') and not base_url.startswith('/'):
2618 base_url = mpd_base_url + base_url
2619 representation_id = representation_attrib.get('id')
2620 lang = representation_attrib.get('lang')
2621 url_el = representation.find(_add_ns('BaseURL'))
2622 filesize = int_or_none(url_el.attrib.get('{http://youtube.com/yt/2012/10/10}contentLength') if url_el is not None else None)
2623 bandwidth = int_or_none(representation_attrib.get('bandwidth'))
2624 if representation_id is not None:
2625 format_id = representation_id
2627 format_id = content_type
2629 format_id = mpd_id + '-' + format_id
2630 if content_type in ('video', 'audio'):
2632 'format_id': format_id,
2633 'manifest_url': mpd_url,
2634 'ext': mimetype2ext(mime_type),
2635 'width': int_or_none(representation_attrib.get('width')),
2636 'height': int_or_none(representation_attrib.get('height')),
2637 'tbr': float_or_none(bandwidth, 1000),
2638 'asr': int_or_none(representation_attrib.get('audioSamplingRate')),
2639 'fps': int_or_none(representation_attrib.get('frameRate')),
2640 'language': lang if lang not in ('mul', 'und', 'zxx', 'mis') else None,
2641 'format_note': 'DASH %s' % content_type,
2642 'filesize': filesize,
2643 'container': mimetype2ext(mime_type) + '_dash',
2645 f.update(parse_codecs(codecs))
2646 elif content_type == 'text':
2648 'ext': mimetype2ext(mime_type),
2649 'manifest_url': mpd_url,
2650 'filesize': filesize,
2652 elif content_type == 'image/jpeg':
2653 # See test case in VikiIE
2654 # https://www.viki.com/videos/1175236v-choosing-spouse-by-lottery-episode-1
2656 'format_id': format_id,
2658 'manifest_url': mpd_url,
2659 'format_note': 'DASH storyboards (jpeg)',
2663 representation_ms_info = extract_multisegment_info(representation, adaption_set_ms_info)
2665 def prepare_template(template_name, identifiers):
2666 tmpl = representation_ms_info[template_name]
2667 # First of, % characters outside $...$ templates
2668 # must be escaped by doubling for proper processing
2669 # by % operator string formatting used further (see
2670 # https://github.com/ytdl-org/youtube-dl/issues/16867).
2676 in_template = not in_template
2677 elif c == '%' and not in_template:
2679 # Next, $...$ templates are translated to their
2680 # %(...) counterparts to be used with % operator
2681 if representation_id is not None:
2682 t = t.replace('$RepresentationID$', representation_id)
2683 t = re.sub(r'\$(%s)\$' % '|'.join(identifiers), r'%(\1)d', t)
2684 t = re.sub(r'\$(%s)%%([^$]+)\$' % '|'.join(identifiers), r'%(\1)\2', t)
2685 t.replace('$$', '$')
2688 # @initialization is a regular template like @media one
2689 # so it should be handled just the same way (see
2690 # https://github.com/ytdl-org/youtube-dl/issues/11605)
2691 if 'initialization' in representation_ms_info:
2692 initialization_template = prepare_template(
2694 # As per [1, 5.3.9.4.2, Table 15, page 54] $Number$ and
2695 # $Time$ shall not be included for @initialization thus
2696 # only $Bandwidth$ remains
2698 representation_ms_info['initialization_url'] = initialization_template % {
2699 'Bandwidth': bandwidth,
2702 def location_key(location):
2703 return 'url' if re.match(r'^https?://', location) else 'path'
2705 if 'segment_urls' not in representation_ms_info and 'media' in representation_ms_info:
2707 media_template = prepare_template('media', ('Number', 'Bandwidth', 'Time'))
2708 media_location_key = location_key(media_template)
2710 # As per [1, 5.3.9.4.4, Table 16, page 55] $Number$ and $Time$
2711 # can't be used at the same time
2712 if '%(Number' in media_template and 's' not in representation_ms_info:
2713 segment_duration = None
2714 if 'total_number' not in representation_ms_info and 'segment_duration' in representation_ms_info:
2715 segment_duration = float_or_none(representation_ms_info['segment_duration'], representation_ms_info['timescale'])
2716 representation_ms_info['total_number'] = int(math.ceil(float(period_duration) / segment_duration))
2717 representation_ms_info['fragments'] = [{
2718 media_location_key: media_template % {
2719 'Number': segment_number,
2720 'Bandwidth': bandwidth,
2722 'duration': segment_duration,
2723 } for segment_number in range(
2724 representation_ms_info['start_number'],
2725 representation_ms_info['total_number'] + representation_ms_info['start_number'])]
2727 # $Number*$ or $Time$ in media template with S list available
2728 # Example $Number*$: http://www.svtplay.se/klipp/9023742/stopptid-om-bjorn-borg
2729 # Example $Time$: https://play.arkena.com/embed/avp/v2/player/media/b41dda37-d8e7-4d3f-b1b5-9a9db578bdfe/1/129411
2730 representation_ms_info['fragments'] = []
2733 segment_number = representation_ms_info['start_number']
2735 def add_segment_url():
2736 segment_url = media_template % {
2737 'Time': segment_time,
2738 'Bandwidth': bandwidth,
2739 'Number': segment_number,
2741 representation_ms_info['fragments'].append({
2742 media_location_key: segment_url,
2743 'duration': float_or_none(segment_d, representation_ms_info['timescale']),
2746 for num, s in enumerate(representation_ms_info['s']):
2747 segment_time = s.get('t') or segment_time
2751 for r in range(s.get('r', 0)):
2752 segment_time += segment_d
2755 segment_time += segment_d
2756 elif 'segment_urls' in representation_ms_info and 's' in representation_ms_info:
2758 # Example: https://www.youtube.com/watch?v=iXZV5uAYMJI
2759 # or any YouTube dashsegments video
2762 timescale = representation_ms_info['timescale']
2763 for s in representation_ms_info['s']:
2764 duration = float_or_none(s['d'], timescale)
2765 for r in range(s.get('r', 0) + 1):
2766 segment_uri = representation_ms_info['segment_urls'][segment_index]
2768 location_key(segment_uri): segment_uri,
2769 'duration': duration,
2772 representation_ms_info['fragments'] = fragments
2773 elif 'segment_urls' in representation_ms_info:
2774 # Segment URLs with no SegmentTimeline
2775 # Example: https://www.seznam.cz/zpravy/clanek/cesko-zasahne-vitr-o-sile-vichrice-muze-byt-i-zivotu-nebezpecny-39091
2776 # https://github.com/ytdl-org/youtube-dl/pull/14844
2778 segment_duration = float_or_none(
2779 representation_ms_info['segment_duration'],
2780 representation_ms_info['timescale']) if 'segment_duration' in representation_ms_info else None
2781 for segment_url in representation_ms_info['segment_urls']:
2783 location_key(segment_url): segment_url,
2785 if segment_duration:
2786 fragment['duration'] = segment_duration
2787 fragments.append(fragment)
2788 representation_ms_info['fragments'] = fragments
2789 # If there is a fragments key available then we correctly recognized fragmented media.
2790 # Otherwise we will assume unfragmented media with direct access. Technically, such
2791 # assumption is not necessarily correct since we may simply have no support for
2792 # some forms of fragmented media renditions yet, but for now we'll use this fallback.
2793 if 'fragments' in representation_ms_info:
2795 # NB: mpd_url may be empty when MPD manifest is parsed from a string
2796 'url': mpd_url or base_url,
2797 'fragment_base_url': base_url,
2799 'protocol': 'http_dash_segments' if mime_type != 'image/jpeg' else 'mhtml',
2801 if 'initialization_url' in representation_ms_info:
2802 initialization_url = representation_ms_info['initialization_url']
2803 if not f.get('url'):
2804 f['url'] = initialization_url
2805 f['fragments'].append({location_key(initialization_url): initialization_url})
2806 f['fragments'].extend(representation_ms_info['fragments'])
2808 # Assuming direct URL to unfragmented media.
2810 if content_type in ('video', 'audio') or mime_type == 'image/jpeg':
2812 elif content_type == 'text':
2813 subtitles.setdefault(lang or 'und', []).append(f)
2815 return formats, subtitles
2817 def _extract_ism_formats(self, *args, **kwargs):
2818 fmts, subs = self._extract_ism_formats_and_subtitles(*args, **kwargs)
2820 self.report_warning(bug_reports_message(
2821 "Ignoring subtitle tracks found in the ISM manifest; "
2822 "if any subtitle tracks are missing,"
2826 def _extract_ism_formats_and_subtitles(self, ism_url, video_id, ism_id=None, note=None, errnote=None, fatal=True, data=None, headers={}, query={}):
2827 res = self._download_xml_handle(
2829 note='Downloading ISM manifest' if note is None else note,
2830 errnote='Failed to download ISM manifest' if errnote is None else errnote,
2831 fatal=fatal, data=data, headers=headers, query=query)
2838 return self._parse_ism_formats_and_subtitles(ism_doc, urlh.geturl(), ism_id)
2840 def _parse_ism_formats_and_subtitles(self, ism_doc, ism_url, ism_id=None):
2842 Parse formats
from ISM manifest
.
2844 1. [MS
-SSTR
]: Smooth Streaming Protocol
,
2845 https
://msdn
.microsoft
.com
/en
-us
/library
/ff469518
.aspx
2847 if ism_doc.get('IsLive') == 'TRUE':
2849 if (not self.get_param('allow_unplayable_formats')
2850 and ism_doc.find('Protection') is not None):
2853 duration = int(ism_doc.attrib['Duration'])
2854 timescale = int_or_none(ism_doc.get('TimeScale')) or 10000000
2858 for stream in ism_doc.findall('StreamIndex'):
2859 stream_type = stream.get('Type')
2860 if stream_type not in ('video', 'audio', 'text'):
2862 url_pattern = stream.attrib['Url']
2863 stream_timescale = int_or_none(stream.get('TimeScale')) or timescale
2864 stream_name = stream.get('Name')
2865 stream_language = stream.get('Language', 'und')
2866 for track in stream.findall('QualityLevel'):
2867 fourcc = track.get('FourCC') or ('AACL' if track.get('AudioTag') == '255' else None)
2868 # TODO: add support for WVC1 and WMAP
2869 if fourcc not in ('H264', 'AVC1', 'AACL', 'TTML'):
2870 self.report_warning('%s is not a supported codec' % fourcc)
2872 tbr = int(track.attrib['Bitrate']) // 1000
2873 # [1] does not mention Width and Height attributes. However,
2874 # they're often present while MaxWidth and MaxHeight are
2875 # missing, so should be used as fallbacks
2876 width = int_or_none(track.get('MaxWidth') or track.get('Width'))
2877 height = int_or_none(track.get('MaxHeight') or track.get('Height'))
2878 sampling_rate = int_or_none(track.get('SamplingRate'))
2880 track_url_pattern = re.sub(r'{[Bb]itrate}', track.attrib['Bitrate'], url_pattern)
2881 track_url_pattern = compat_urlparse.urljoin(ism_url, track_url_pattern)
2887 stream_fragments = stream.findall('c')
2888 for stream_fragment_index, stream_fragment in enumerate(stream_fragments):
2889 fragment_ctx['time'] = int_or_none(stream_fragment.get('t')) or fragment_ctx['time']
2890 fragment_repeat = int_or_none(stream_fragment.get('r')) or 1
2891 fragment_ctx['duration'] = int_or_none(stream_fragment.get('d'))
2892 if not fragment_ctx['duration']:
2894 next_fragment_time = int(stream_fragment[stream_fragment_index + 1].attrib['t'])
2896 next_fragment_time = duration
2897 fragment_ctx['duration'] = (next_fragment_time - fragment_ctx['time']) / fragment_repeat
2898 for _ in range(fragment_repeat):
2900 'url': re.sub(r'{start[ _]time}', compat_str(fragment_ctx['time']), track_url_pattern),
2901 'duration': fragment_ctx['duration'] / stream_timescale,
2903 fragment_ctx['time'] += fragment_ctx['duration']
2907 format_id.append(ism_id)
2909 format_id.append(stream_name)
2910 format_id.append(compat_str(tbr))
2912 if stream_type == 'text':
2913 subtitles.setdefault(stream_language, []).append({
2917 'manifest_url': ism_url,
2918 'fragments': fragments,
2919 '_download_params': {
2920 'stream_type': stream_type,
2921 'duration': duration,
2922 'timescale': stream_timescale,
2924 'language': stream_language,
2925 'codec_private_data': track.get('CodecPrivateData'),
2928 elif stream_type in ('video', 'audio'):
2930 'format_id': '-'.join(format_id),
2932 'manifest_url': ism_url,
2933 'ext': 'ismv' if stream_type == 'video' else 'isma',
2937 'asr': sampling_rate,
2938 'vcodec': 'none' if stream_type == 'audio' else fourcc,
2939 'acodec': 'none' if stream_type == 'video' else fourcc,
2941 'fragments': fragments,
2942 '_download_params': {
2943 'stream_type': stream_type,
2944 'duration': duration,
2945 'timescale': stream_timescale,
2946 'width': width or 0,
2947 'height': height or 0,
2949 'language': stream_language,
2950 'codec_private_data': track.get('CodecPrivateData'),
2951 'sampling_rate': sampling_rate,
2952 'channels': int_or_none(track.get('Channels', 2)),
2953 'bits_per_sample': int_or_none(track.get('BitsPerSample', 16)),
2954 'nal_unit_length_field': int_or_none(track.get('NALUnitLengthField', 4)),
2957 return formats, subtitles
2959 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):
2960 def absolute_url(item_url):
2961 return urljoin(base_url, item_url)
2963 def parse_content_type(content_type):
2964 if not content_type:
2966 ctr = re.search(r'(?P<mimetype>[^/]+/[^;]+)(?:;\s*codecs="?(?P<codecs>[^"]+))?', content_type)
2968 mimetype, codecs = ctr.groups()
2969 f = parse_codecs(codecs)
2970 f['ext'] = mimetype2ext(mimetype)
2974 def _media_formats(src, cur_media_type, type_info={}):
2975 full_url = absolute_url(src)
2976 ext = type_info.get('ext') or determine_ext(full_url)
2978 is_plain_url = False
2979 formats = self._extract_m3u8_formats(
2980 full_url, video_id, ext='mp4',
2981 entry_protocol=m3u8_entry_protocol, m3u8_id=m3u8_id,
2982 preference=preference, quality=quality, fatal=False)
2984 is_plain_url = False
2985 formats = self._extract_mpd_formats(
2986 full_url, video_id, mpd_id=mpd_id, fatal=False)
2991 'vcodec': 'none' if cur_media_type == 'audio' else None,
2993 return is_plain_url, formats
2996 # amp-video and amp-audio are very similar to their HTML5 counterparts
2997 # so we wll include them right here (see
2998 # https://www.ampproject.org/docs/reference/components/amp-video)
2999 # For dl8-* tags see https://delight-vr.com/documentation/dl8-video/
3000 _MEDIA_TAG_NAME_RE = r'(?:(?:amp|dl8(?:-live)?)-)?(video|audio)'
3001 media_tags = [(media_tag, media_tag_name, media_type, '')
3002 for media_tag, media_tag_name, media_type
3003 in re.findall(r'(?s)(<(%s)[^>]*/>)' % _MEDIA_TAG_NAME_RE, webpage)]
3004 media_tags.extend(re.findall(
3005 # We only allow video|audio followed by a whitespace or '>'.
3006 # Allowing more characters may end up in significant slow down (see
3007 # https://github.com/ytdl-org/youtube-dl/issues/11979, example URL:
3008 # http://www.porntrex.com/maps/videositemap.xml).
3009 r'(?s)(<(?P<tag>%s)(?:\s+[^>]*)?>)(.*?)</(?P=tag)>' % _MEDIA_TAG_NAME_RE, webpage))
3010 for media_tag, _, media_type, media_content in media_tags:
3015 media_attributes = extract_attributes(media_tag)
3016 src = strip_or_none(media_attributes.get('src'))
3018 _, formats = _media_formats(src, media_type)
3019 media_info['formats'].extend(formats)
3020 media_info['thumbnail'] = absolute_url(media_attributes.get('poster'))
3022 for source_tag in re.findall(r'<source[^>]+>', media_content):
3023 s_attr = extract_attributes(source_tag)
3024 # data-video-src and data-src are non standard but seen
3025 # several times in the wild
3026 src = strip_or_none(dict_get(s_attr, ('src', 'data-video-src', 'data-src')))
3029 f = parse_content_type(s_attr.get('type'))
3030 is_plain_url, formats = _media_formats(src, media_type, f)
3032 # width, height, res, label and title attributes are
3033 # all not standard but seen several times in the wild
3036 for lbl in ('label', 'title')
3037 if str_or_none(s_attr.get(lbl))
3039 width = int_or_none(s_attr.get('width'))
3040 height = (int_or_none(s_attr.get('height'))
3041 or int_or_none(s_attr.get('res')))
3042 if not width or not height:
3044 resolution = parse_resolution(lbl)
3047 width = width or resolution.get('width')
3048 height = height or resolution.get('height')
3050 tbr = parse_bitrate(lbl)
3059 'format_id': s_attr.get('label') or s_attr.get('title'),
3061 f.update(formats[0])
3062 media_info['formats'].append(f)
3064 media_info['formats'].extend(formats)
3065 for track_tag in re.findall(r'<track[^>]+>', media_content):
3066 track_attributes = extract_attributes(track_tag)
3067 kind = track_attributes.get('kind')
3068 if not kind or kind in ('subtitles', 'captions'):
3069 src = strip_or_none(track_attributes.get('src'))
3072 lang = track_attributes.get('srclang') or track_attributes.get('lang') or track_attributes.get('label')
3073 media_info['subtitles'].setdefault(lang, []).append({
3074 'url': absolute_url(src),
3076 for f in media_info['formats']:
3077 f.setdefault('http_headers', {})['Referer'] = base_url
3078 if media_info['formats'] or media_info['subtitles']:
3079 entries.append(media_info)
3082 def _extract_akamai_formats(self, *args, **kwargs):
3083 fmts, subs = self._extract_akamai_formats_and_subtitles(*args, **kwargs)
3085 self.report_warning(bug_reports_message(
3086 "Ignoring subtitle tracks found in the manifests; "
3087 "if any subtitle tracks are missing,"
3091 def _extract_akamai_formats_and_subtitles(self, manifest_url, video_id, hosts={}):
3092 signed = 'hdnea=' in manifest_url
3094 # https://learn.akamai.com/en-us/webhelp/media-services-on-demand/stream-packaging-user-guide/GUID-BE6C0F73-1E06-483B-B0EA-57984B91B7F9.html
3095 manifest_url = re.sub(
3096 r'(?:b=[\d,-]+|(?:__a__|attributes)=off|__b__=\d+)&?',
3097 '', manifest_url).strip('?')
3102 hdcore_sign = 'hdcore=3.7.0'
3103 f4m_url = re.sub(r'(https?://[^/]+)/i/', r'\1/z/', manifest_url).replace('/master.m3u8', '/manifest.f4m')
3104 hds_host = hosts.get('hds')
3106 f4m_url = re.sub(r'(https?://)[^/]+', r'\1' + hds_host, f4m_url)
3107 if 'hdcore=' not in f4m_url:
3108 f4m_url += ('&' if '?' in f4m_url else '?') + hdcore_sign
3109 f4m_formats = self._extract_f4m_formats(
3110 f4m_url, video_id, f4m_id='hds', fatal=False)
3111 for entry in f4m_formats:
3112 entry.update({'extra_param_to_segment_url': hdcore_sign})
3113 formats.extend(f4m_formats)
3115 m3u8_url = re.sub(r'(https?://[^/]+)/z/', r'\1/i/', manifest_url).replace('/manifest.f4m', '/master.m3u8')
3116 hls_host = hosts.get('hls')
3118 m3u8_url = re.sub(r'(https?://)[^/]+', r'\1' + hls_host, m3u8_url)
3119 m3u8_formats, m3u8_subtitles = self._extract_m3u8_formats_and_subtitles(
3120 m3u8_url, video_id, 'mp4', 'm3u8_native',
3121 m3u8_id='hls', fatal=False)
3122 formats.extend(m3u8_formats)
3123 subtitles = self._merge_subtitles(subtitles, m3u8_subtitles)
3125 http_host = hosts.get('http')
3126 if http_host and m3u8_formats and not signed:
3127 REPL_REGEX = r'https?://[^/]+/i/([^,]+),([^/]+),([^/]+)\.csmil/.+'
3128 qualities = re.match(REPL_REGEX, m3u8_url).group(2).split(',')
3129 qualities_length = len(qualities)
3130 if len(m3u8_formats) in (qualities_length, qualities_length + 1):
3132 for f in m3u8_formats:
3133 if f['vcodec'] != 'none':
3134 for protocol in ('http', 'https'):
3136 del http_f['manifest_url']
3138 REPL_REGEX, protocol + r'://%s/\g<1>%s\3' % (http_host, qualities[i]), f['url'])
3140 'format_id': http_f['format_id'].replace('hls-', protocol + '-'),
3142 'protocol': protocol,
3144 formats.append(http_f)
3147 return formats, subtitles
3149 def _extract_wowza_formats(self, url, video_id, m3u8_entry_protocol='m3u8_native', skip_protocols=[]):
3150 query = compat_urlparse.urlparse(url).query
3151 url = re.sub(r'/(?:manifest|playlist|jwplayer)\.(?:m3u8|f4m|mpd|smil)', '', url)
3153 r'(?:(?:http|rtmp|rtsp)(?P<s>s)?:)?(?P<url>//[^?]+)', url)
3154 url_base = mobj.group('url')
3155 http_base_url = '%s%s:%s' % ('http', mobj.group('s') or '', url_base)
3158 def manifest_url(manifest):
3159 m_url = '%s/%s' % (http_base_url, manifest)
3161 m_url += '?%s' % query
3164 if 'm3u8' not in skip_protocols:
3165 formats.extend(self._extract_m3u8_formats(
3166 manifest_url('playlist.m3u8'), video_id, 'mp4',
3167 m3u8_entry_protocol, m3u8_id='hls', fatal=False))
3168 if 'f4m' not in skip_protocols:
3169 formats.extend(self._extract_f4m_formats(
3170 manifest_url('manifest.f4m'),
3171 video_id, f4m_id='hds', fatal=False))
3172 if 'dash' not in skip_protocols:
3173 formats.extend(self._extract_mpd_formats(
3174 manifest_url('manifest.mpd'),
3175 video_id, mpd_id='dash', fatal=False))
3176 if re.search(r'(?:/smil:|\.smil)', url_base):
3177 if 'smil' not in skip_protocols:
3178 rtmp_formats = self._extract_smil_formats(
3179 manifest_url('jwplayer.smil'),
3180 video_id, fatal=False)
3181 for rtmp_format in rtmp_formats:
3182 rtsp_format = rtmp_format.copy()
3183 rtsp_format['url'] = '%s/%s' % (rtmp_format['url'], rtmp_format['play_path'])
3184 del rtsp_format['play_path']
3185 del rtsp_format['ext']
3186 rtsp_format.update({
3187 'url': rtsp_format['url'].replace('rtmp://', 'rtsp://'),
3188 'format_id': rtmp_format['format_id'].replace('rtmp', 'rtsp'),
3191 formats.extend([rtmp_format, rtsp_format])
3193 for protocol in ('rtmp', 'rtsp'):
3194 if protocol not in skip_protocols:
3196 'url': '%s:%s' % (protocol, url_base),
3197 'format_id': protocol,
3198 'protocol': protocol,
3202 def _find_jwplayer_data(self, webpage, video_id=None, transform_source=js_to_json):
3204 r'(?s)jwplayer\((?P<quote>[\'"])[^\'" ]+(?P=quote)\)(?!</script>).*?\.setup\s*\((?P<options>[^)]+)\)',
3208 jwplayer_data = self._parse_json(mobj.group('options'),
3210 transform_source=transform_source)
3211 except ExtractorError:
3214 if isinstance(jwplayer_data, dict):
3215 return jwplayer_data
3217 def _extract_jwplayer_data(self, webpage, video_id, *args, **kwargs):
3218 jwplayer_data = self._find_jwplayer_data(
3219 webpage, video_id, transform_source=js_to_json)
3220 return self._parse_jwplayer_data(
3221 jwplayer_data, video_id, *args, **kwargs)
3223 def _parse_jwplayer_data(self, jwplayer_data, video_id=None, require_title=True,
3224 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
3225 # JWPlayer backward compatibility: flattened playlists
3226 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/api/config.js#L81-L96
3227 if 'playlist' not in jwplayer_data:
3228 jwplayer_data = {'playlist': [jwplayer_data]}
3232 # JWPlayer backward compatibility: single playlist item
3233 # https://github.com/jwplayer/jwplayer/blob/v7.7.0/src/js/playlist/playlist.js#L10
3234 if not isinstance(jwplayer_data['playlist'], list):
3235 jwplayer_data['playlist'] = [jwplayer_data['playlist']]
3237 for video_data in jwplayer_data['playlist']:
3238 # JWPlayer backward compatibility: flattened sources
3239 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/playlist/item.js#L29-L35
3240 if 'sources' not in video_data:
3241 video_data['sources'] = [video_data]
3243 this_video_id = video_id or video_data['mediaid']
3245 formats = self._parse_jwplayer_formats(
3246 video_data['sources'], video_id=this_video_id, m3u8_id=m3u8_id,
3247 mpd_id=mpd_id, rtmp_params=rtmp_params, base_url=base_url)
3250 tracks = video_data.get('tracks')
3251 if tracks and isinstance(tracks, list):
3252 for track in tracks:
3253 if not isinstance(track, dict):
3255 track_kind = track.get('kind')
3256 if not track_kind or not isinstance(track_kind, compat_str):
3258 if track_kind.lower() not in ('captions', 'subtitles'):
3260 track_url = urljoin(base_url, track.get('file'))
3263 subtitles.setdefault(track.get('label') or 'en', []).append({
3264 'url': self._proto_relative_url(track_url)
3268 'id': this_video_id,
3269 'title': unescapeHTML(video_data['title'] if require_title else video_data.get('title')),
3270 'description': clean_html(video_data.get('description')),
3271 'thumbnail': urljoin(base_url, self._proto_relative_url(video_data.get('image'))),
3272 'timestamp': int_or_none(video_data.get('pubdate')),
3273 'duration': float_or_none(jwplayer_data.get('duration') or video_data.get('duration')),
3274 'subtitles': subtitles,
3276 # https://github.com/jwplayer/jwplayer/blob/master/src/js/utils/validator.js#L32
3277 if len(formats) == 1 and re.search(r'^(?:http|//).*(?:youtube\.com|youtu\.be)/.+', formats[0]['url']):
3279 '_type': 'url_transparent',
3280 'url': formats[0]['url'],
3283 self._sort_formats(formats)
3284 entry['formats'] = formats
3285 entries.append(entry)
3286 if len(entries) == 1:
3289 return self.playlist_result(entries)
3291 def _parse_jwplayer_formats(self, jwplayer_sources_data, video_id=None,
3292 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
3295 for source in jwplayer_sources_data:
3296 if not isinstance(source, dict):
3298 source_url = urljoin(
3299 base_url, self._proto_relative_url(source.get('file')))
3300 if not source_url or source_url in urls:
3302 urls.append(source_url)
3303 source_type = source.get('type') or ''
3304 ext = mimetype2ext(source_type) or determine_ext(source_url)
3305 if source_type == 'hls' or ext == 'm3u8':
3306 formats.extend(self._extract_m3u8_formats(
3307 source_url, video_id, 'mp4', entry_protocol='m3u8_native',
3308 m3u8_id=m3u8_id, fatal=False))
3309 elif source_type == 'dash' or ext == 'mpd':
3310 formats.extend(self._extract_mpd_formats(
3311 source_url, video_id, mpd_id=mpd_id, fatal=False))
3313 formats.extend(self._extract_smil_formats(
3314 source_url, video_id, fatal=False))
3315 # https://github.com/jwplayer/jwplayer/blob/master/src/js/providers/default.js#L67
3316 elif source_type.startswith('audio') or ext in (
3317 'oga', 'aac', 'mp3', 'mpeg', 'vorbis'):
3324 height = int_or_none(source.get('height'))
3326 # Often no height is provided but there is a label in
3327 # format like "1080p", "720p SD", or 1080.
3328 height = int_or_none(self._search_regex(
3329 r'^(\d{3,4})[pP]?(?:\b|$)', compat_str(source.get('label') or ''),
3330 'height', default=None))
3333 'width': int_or_none(source.get('width')),
3335 'tbr': int_or_none(source.get('bitrate')),
3338 if source_url.startswith('rtmp'):
3339 a_format['ext'] = 'flv'
3340 # See com/longtailvideo/jwplayer/media/RTMPMediaProvider.as
3341 # of jwplayer.flash.swf
3342 rtmp_url_parts = re.split(
3343 r'((?:mp4|mp3|flv):)', source_url, 1)
3344 if len(rtmp_url_parts) == 3:
3345 rtmp_url, prefix, play_path = rtmp_url_parts
3348 'play_path': prefix + play_path,
3351 a_format.update(rtmp_params)
3352 formats.append(a_format)
3355 def _live_title(self, name):
3356 """ Generate the title
for a live video
"""
3357 now = datetime.datetime.now()
3358 now_str = now.strftime('%Y-%m-%d %H:%M')
3359 return name + ' ' + now_str
3361 def _int(self, v, name, fatal=False, **kwargs):
3362 res = int_or_none(v, **kwargs)
3363 if 'get_attr' in kwargs:
3364 print(getattr(v, kwargs['get_attr']))
3366 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
3368 raise ExtractorError(msg)
3370 self.report_warning(msg)
3373 def _float(self, v, name, fatal=False, **kwargs):
3374 res = float_or_none(v, **kwargs)
3376 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
3378 raise ExtractorError(msg)
3380 self.report_warning(msg)
3383 def _set_cookie(self, domain, name, value, expire_time=None, port=None,
3384 path='/', secure=False, discard=False, rest={}, **kwargs):
3385 cookie = compat_cookiejar_Cookie(
3386 0, name, value, port, port is not None, domain, True,
3387 domain.startswith('.'), path, True, secure, expire_time,
3388 discard, None, None, rest)
3389 self._downloader.cookiejar.set_cookie(cookie)
3391 def _get_cookies(self, url):
3392 """ Return a compat_cookies_SimpleCookie
with the cookies
for the url
"""
3393 req = sanitized_Request(url)
3394 self._downloader.cookiejar.add_cookie_header(req)
3395 return compat_cookies_SimpleCookie(req.get_header('Cookie'))
3397 def _apply_first_set_cookie_header(self, url_handle, cookie):
3399 Apply first Set
-Cookie header instead of the last
. Experimental
.
3401 Some
sites (e
.g
. [1-3]) may serve two cookies under the same name
3402 in Set
-Cookie header
and expect the
first (old
) one to be
set rather
3403 than
second (new
). However
, as of RFC6265 the newer one cookie
3404 should be
set into cookie store what actually happens
.
3405 We will workaround this issue by resetting the cookie to
3406 the first one manually
.
3407 1. https
://new
.vk
.com
/
3408 2. https
://github
.com
/ytdl
-org
/youtube
-dl
/issues
/9841#issuecomment-227871201
3409 3. https
://learning
.oreilly
.com
/
3411 for header, cookies in url_handle.headers.items():
3412 if header.lower() != 'set-cookie':
3414 if sys.version_info[0] >= 3:
3415 cookies = cookies.encode('iso-8859-1')
3416 cookies = cookies.decode('utf-8')
3417 cookie_value = re.search(
3418 r'%s=(.+?);.*?\b[Dd]omain=(.+?)(?:[,;]|$)' % cookie, cookies)
3420 value, domain = cookie_value.groups()
3421 self._set_cookie(domain, cookie, value)
3424 def get_testcases(self, include_onlymatching=False):
3425 t = getattr(self, '_TEST', None)
3427 assert not hasattr(self, '_TESTS'), \
3428 '%s has _TEST and _TESTS' % type(self).__name__
3431 tests = getattr(self, '_TESTS', [])
3433 if not include_onlymatching and t.get('only_matching', False):
3435 t['name'] = type(self).__name__[:-len('IE')]
3438 def is_suitable(self, age_limit):
3439 """ Test whether the extractor
is generally suitable
for the given
3440 age
limit (i
.e
. pornographic sites are
not, all others usually are
) """
3442 any_restricted = False
3443 for tc in self.get_testcases(include_onlymatching=False):
3444 if tc.get('playlist', []):
3445 tc = tc['playlist'][0]
3446 is_restricted = age_restricted(
3447 tc.get('info_dict', {}).get('age_limit'), age_limit)
3448 if not is_restricted:
3450 any_restricted = any_restricted or is_restricted
3451 return not any_restricted
3453 def extract_subtitles(self, *args, **kwargs):
3454 if (self.get_param('writesubtitles', False)
3455 or self.get_param('listsubtitles')):
3456 return self._get_subtitles(*args, **kwargs)
3459 def _get_subtitles(self, *args, **kwargs):
3460 raise NotImplementedError('This method must be implemented by subclasses')
3463 def _merge_subtitle_items(subtitle_list1, subtitle_list2):
3464 """ Merge subtitle items
for one language
. Items
with duplicated URLs
3465 will be dropped
. """
3466 list1_urls = set([item['url'] for item in subtitle_list1])
3467 ret = list(subtitle_list1)
3468 ret.extend([item for item in subtitle_list2 if item['url'] not in list1_urls])
3472 def _merge_subtitles(cls, *dicts, target=None):
3473 """ Merge subtitle dictionaries
, language by language
. """
3477 for lang, subs in d.items():
3478 target[lang] = cls._merge_subtitle_items(target.get(lang, []), subs)
3481 def extract_automatic_captions(self, *args, **kwargs):
3482 if (self.get_param('writeautomaticsub', False)
3483 or self.get_param('listsubtitles')):
3484 return self._get_automatic_captions(*args, **kwargs)
3487 def _get_automatic_captions(self, *args, **kwargs):
3488 raise NotImplementedError('This method must be implemented by subclasses')
3490 def mark_watched(self, *args, **kwargs):
3491 if (self.get_param('mark_watched', False)
3492 and (self._get_login_info()[0] is not None
3493 or self.get_param('cookiefile') is not None)):
3494 self._mark_watched(*args, **kwargs)
3496 def _mark_watched(self, *args, **kwargs):
3497 raise NotImplementedError('This method must be implemented by subclasses')
3499 def geo_verification_headers(self):
3501 geo_verification_proxy = self.get_param('geo_verification_proxy')
3502 if geo_verification_proxy:
3503 headers['Ytdl-request-proxy'] = geo_verification_proxy
3506 def _generic_id(self, url):
3507 return compat_urllib_parse_unquote(os.path.splitext(url.rstrip('/').split('/')[-1])[0])
3509 def _generic_title(self, url):
3510 return compat_urllib_parse_unquote(os.path.splitext(url_basename(url))[0])
3513 def _availability(is_private=None, needs_premium=None, needs_subscription=None, needs_auth=None, is_unlisted=None):
3514 all_known = all(map(
3515 lambda x: x is not None,
3516 (is_private, needs_premium, needs_subscription, needs_auth, is_unlisted)))
3518 'private' if is_private
3519 else 'premium_only' if needs_premium
3520 else 'subscriber_only' if needs_subscription
3521 else 'needs_auth' if needs_auth
3522 else 'unlisted' if is_unlisted
3523 else 'public' if all_known
3526 def _configuration_arg(self, key, default=NO_DEFAULT, casesense=False):
3528 @returns A list of values for the extractor argument given by "key"
3529 or "default" if no such key is present
3530 @param default The default value to return when the key is not present (default: [])
3531 @param casesense When false, the values are converted to lower case
3534 self._downloader.params, ('extractor_args', self.ie_key().lower(), key))
3536 return [] if default is NO_DEFAULT else default
3537 return list(val) if casesense else [x.lower() for x in val]
3540 class SearchInfoExtractor(InfoExtractor):
3542 Base
class for paged search queries extractors
.
3543 They accept URLs
in the format
_SEARCH_KEY(|all|
[0-9]):{query}
3544 Instances should define _SEARCH_KEY
and _MAX_RESULTS
.
3548 def _make_valid_url(cls):
3549 return r'%s(?P<prefix>|[1-9][0-9]*|all):(?P<query>[\s\S]+)' % cls._SEARCH_KEY
3552 def suitable(cls, url):
3553 return re.match(cls._make_valid_url(), url) is not None
3555 def _real_extract(self, query):
3556 mobj = re.match(self._make_valid_url(), query)
3558 raise ExtractorError('Invalid search query "%s"' % query)
3560 prefix = mobj.group('prefix')
3561 query = mobj.group('query')
3563 return self._get_n_results(query, 1)
3564 elif prefix == 'all':
3565 return self._get_n_results(query, self._MAX_RESULTS)
3569 raise ExtractorError('invalid download number %s for query "%s"' % (n, query))
3570 elif n > self._MAX_RESULTS:
3571 self.report_warning('%s returns max %i results (you requested %i)' % (self._SEARCH_KEY, self._MAX_RESULTS, n))
3572 n = self._MAX_RESULTS
3573 return self._get_n_results(query, n)
3575 def _get_n_results(self, query, n):
3576 """Get a specified number of results
for a query
"""
3577 raise NotImplementedError('This method must be implemented by subclasses')
3580 def SEARCH_KEY(self):
3581 return self._SEARCH_KEY