2 from __future__ import unicode_literals
17 from ..compat import (
21 compat_etree_fromstring,
28 compat_urllib_parse_unquote,
29 compat_urllib_parse_urlencode,
30 compat_urllib_request,
32 compat_xml_parse_error,
34 from ..downloader.f4m import (
36 remove_encrypted_media,
64 parse_m3u8_attributes,
85 class InfoExtractor(object):
86 """Information Extractor class.
88 Information extractors are the classes that, given a URL, extract
89 information about the video (or videos) the URL refers to. This
90 information includes the real video URL, the video title, author and
91 others. The information is stored in a dictionary which is then
92 passed to the YoutubeDL. The YoutubeDL processes this
93 information possibly downloading the video to the file system, among
94 other possible outcomes.
96 The type field determines the type of the result.
97 By far the most common value (and the default if _type is missing) is
98 "video", which indicates a single video.
100 For a video, the dictionaries must include the following fields:
102 id: Video identifier.
103 title: Video title, unescaped.
105 Additionally, it must contain either a formats entry or a url one:
107 formats: A list of dictionaries for each format available, ordered
108 from worst to best quality.
111 * url The mandatory URL representing the media:
112 for plain file media - HTTP URL of this file,
114 for HLS - URL of the M3U8 media playlist,
115 for HDS - URL of the F4M manifest,
117 - HTTP URL to plain file media (in case of
119 - URL of the MPD manifest or base URL
120 representing the media if MPD manifest
121 is parsed from a string (in case of
123 for MSS - URL of the ISM manifest.
125 The URL of the manifest file in case of
127 for HLS - URL of the M3U8 master playlist,
128 for HDS - URL of the F4M manifest,
129 for DASH - URL of the MPD manifest,
130 for MSS - URL of the ISM manifest.
131 * ext Will be calculated from URL if missing
132 * format A human-readable description of the format
133 ("mp4 container with h264/opus").
134 Calculated from the format_id, width, height.
135 and format_note fields if missing.
136 * format_id A short description of the format
137 ("mp4_h264_opus" or "19").
138 Technically optional, but strongly recommended.
139 * format_note Additional info about the format
140 ("3D" or "DASH video")
141 * width Width of the video, if known
142 * height Height of the video, if known
143 * resolution Textual description of width and height
144 * tbr Average bitrate of audio and video in KBit/s
145 * abr Average audio bitrate in KBit/s
146 * acodec Name of the audio codec in use
147 * asr Audio sampling rate in Hertz
148 * vbr Average video bitrate in KBit/s
150 * vcodec Name of the video codec in use
151 * container Name of the container format
152 * filesize The number of bytes, if known in advance
153 * filesize_approx An estimate for the number of bytes
154 * player_url SWF Player URL (used for rtmpdump).
155 * protocol The protocol that will be used for the actual
156 download, lower-case.
157 "http", "https", "rtsp", "rtmp", "rtmpe",
158 "m3u8", "m3u8_native" or "http_dash_segments".
160 Base URL for fragments. Each fragment's path
161 value (if present) will be relative to
163 * fragments A list of fragments of a fragmented media.
164 Each fragment entry must contain either an url
165 or a path. If an url is present it should be
166 considered by a client. Otherwise both path and
167 fragment_base_url must be present. Here is
168 the list of all potential fields:
169 * "url" - fragment's URL
170 * "path" - fragment's path relative to
172 * "duration" (optional, int or float)
173 * "filesize" (optional, int)
174 * preference Order number of this format. If this field is
175 present and not None, the formats get sorted
176 by this field, regardless of all other values.
177 -1 for default (order by other properties),
178 -2 or smaller for less than default.
179 < -1000 to hide the format (if there is
180 another one which is strictly better)
181 * language Language code, e.g. "de" or "en-US".
182 * language_preference Is this in the language mentioned in
184 10 if it's what the URL is about,
185 -1 for default (don't know),
186 -10 otherwise, other values reserved for now.
187 * quality Order number of the video quality of this
188 format, irrespective of the file format.
189 -1 for default (order by other properties),
190 -2 or smaller for less than default.
191 * source_preference Order number for this video source
192 (quality takes higher priority)
193 -1 for default (order by other properties),
194 -2 or smaller for less than default.
195 * http_headers A dictionary of additional HTTP headers
196 to add to the request.
197 * stretched_ratio If given and not 1, indicates that the
198 video's pixels are not square.
199 width : height ratio as float.
200 * no_resume The server does not support resuming the
201 (HTTP or RTMP) download. Boolean.
202 * downloader_options A dictionary of downloader options as
203 described in FileDownloader
205 url: Final video URL.
206 ext: Video filename extension.
207 format: The video format, defaults to ext (used for --get-format)
208 player_url: SWF Player URL (used for rtmpdump).
210 The following fields are optional:
212 alt_title: A secondary title of the video.
213 display_id An alternative identifier for the video, not necessarily
214 unique, but available before title. Typically, id is
215 something like "4234987", title "Dancing naked mole rats",
216 and display_id "dancing-naked-mole-rats"
217 thumbnails: A list of dictionaries, with the following entries:
218 * "id" (optional, string) - Thumbnail format ID
220 * "preference" (optional, int) - quality of the image
221 * "width" (optional, int)
222 * "height" (optional, int)
223 * "resolution" (optional, string "{width}x{height}",
225 * "filesize" (optional, int)
226 thumbnail: Full URL to a video thumbnail image.
227 description: Full video description.
228 uploader: Full name of the video uploader.
229 license: License name the video is licensed under.
230 creator: The creator of the video.
231 release_date: The date (YYYYMMDD) when the video was released.
232 timestamp: UNIX timestamp of the moment the video became available.
233 upload_date: Video upload date (YYYYMMDD).
234 If not explicitly set, calculated from timestamp.
235 uploader_id: Nickname or id of the video uploader.
236 uploader_url: Full URL to a personal webpage of the video uploader.
237 channel: Full name of the channel the video is uploaded on.
238 Note that channel fields may or may not repeat uploader
239 fields. This depends on a particular extractor.
240 channel_id: Id of the channel.
241 channel_url: Full URL to a channel webpage.
242 location: Physical location where the video was filmed.
243 subtitles: The available subtitles as a dictionary in the format
244 {tag: subformats}. "tag" is usually a language code, and
245 "subformats" is a list sorted from lower to higher
246 preference, each element is a dictionary with the "ext"
248 * "data": The subtitles file contents
249 * "url": A URL pointing to the subtitles file
250 "ext" will be calculated from URL if missing
251 automatic_captions: Like 'subtitles', used by the YoutubeIE for
252 automatically generated captions
253 duration: Length of the video in seconds, as an integer or float.
254 view_count: How many users have watched the video on the platform.
255 like_count: Number of positive ratings of the video
256 dislike_count: Number of negative ratings of the video
257 repost_count: Number of reposts of the video
258 average_rating: Average rating give by users, the scale used depends on the webpage
259 comment_count: Number of comments on the video
260 comments: A list of comments, each with one or more of the following
261 properties (all but one of text or html optional):
262 * "author" - human-readable name of the comment author
263 * "author_id" - user ID of the comment author
265 * "html" - Comment as HTML
266 * "text" - Plain text of the comment
267 * "timestamp" - UNIX timestamp of comment
268 * "parent" - ID of the comment this one is replying to.
269 Set to "root" to indicate that this is a
270 comment to the original video.
271 age_limit: Age restriction for the video, as an integer (years)
272 webpage_url: The URL to the video webpage, if given to youtube-dl it
273 should allow to get the same result again. (It will be set
274 by YoutubeDL if it's missing)
275 categories: A list of categories that the video falls in, for example
277 tags: A list of tags assigned to the video, e.g. ["sweden", "pop music"]
278 is_live: True, False, or None (=unknown). Whether this video is a
279 live stream that goes on instead of a fixed-length video.
280 start_time: Time in seconds where the reproduction should start, as
281 specified in the URL.
282 end_time: Time in seconds where the reproduction should end, as
283 specified in the URL.
284 chapters: A list of dictionaries, with the following entries:
285 * "start_time" - The start time of the chapter in seconds
286 * "end_time" - The end time of the chapter in seconds
287 * "title" (optional, string)
289 The following fields should only be used when the video belongs to some logical
292 chapter: Name or title of the chapter the video belongs to.
293 chapter_number: Number of the chapter the video belongs to, as an integer.
294 chapter_id: Id of the chapter the video belongs to, as a unicode string.
296 The following fields should only be used when the video is an episode of some
297 series, programme or podcast:
299 series: Title of the series or programme the video episode belongs to.
300 season: Title of the season the video episode belongs to.
301 season_number: Number of the season the video episode belongs to, as an integer.
302 season_id: Id of the season the video episode belongs to, as a unicode string.
303 episode: Title of the video episode. Unlike mandatory video title field,
304 this field should denote the exact title of the video episode
305 without any kind of decoration.
306 episode_number: Number of the video episode within a season, as an integer.
307 episode_id: Id of the video episode, as a unicode string.
309 The following fields should only be used when the media is a track or a part of
312 track: Title of the track.
313 track_number: Number of the track within an album or a disc, as an integer.
314 track_id: Id of the track (useful in case of custom indexing, e.g. 6.iii),
316 artist: Artist(s) of the track.
317 genre: Genre(s) of the track.
318 album: Title of the album the track belongs to.
319 album_type: Type of the album (e.g. "Demo", "Full-length", "Split", "Compilation", etc).
320 album_artist: List of all artists appeared on the album (e.g.
321 "Ash Borer / Fell Voices" or "Various Artists", useful for splits
323 disc_number: Number of the disc or other physical medium the track belongs to,
325 release_year: Year (YYYY) when the album was released.
327 Unless mentioned otherwise, the fields should be Unicode strings.
329 Unless mentioned otherwise, None is equivalent to absence of information.
332 _type "playlist" indicates multiple videos.
333 There must be a key "entries", which is a list, an iterable, or a PagedList
334 object, each element of which is a valid dictionary by this specification.
336 Additionally, playlists can have "id", "title", "description", "uploader",
337 "uploader_id", "uploader_url" attributes with the same semantics as videos
341 _type "multi_video" indicates that there are multiple videos that
342 form a single show, for examples multiple acts of an opera or TV episode.
343 It must have an entries key like a playlist and contain all the keys
344 required for a video at the same time.
347 _type "url" indicates that the video must be extracted from another
348 location, possibly by a different extractor. Its only required key is:
349 "url" - the next URL to extract.
350 The key "ie_key" can be set to the class name (minus the trailing "IE",
351 e.g. "Youtube") if the extractor class is known in advance.
352 Additionally, the dictionary may have any properties of the resolved entity
353 known in advance, for example "title" if the title of the referred video is
357 _type "url_transparent" entities have the same specification as "url", but
358 indicate that the given additional information is more precise than the one
359 associated with the resolved URL.
360 This is useful when a site employs a video service that hosts the video and
361 its technical metadata, but that video service does not embed a useful
362 title, description etc.
365 Subclasses of this one should re-define the _real_initialize() and
366 _real_extract() methods and define a _VALID_URL regexp.
367 Probably, they should also be added to the list of extractors.
369 _GEO_BYPASS attribute may be set to False in order to disable
370 geo restriction bypass mechanisms for a particular extractor.
371 Though it won't disable explicit geo restriction bypass based on
372 country code provided with geo_bypass_country.
374 _GEO_COUNTRIES attribute may contain a list of presumably geo unrestricted
375 countries for this extractor. One of these countries will be used by
376 geo restriction bypass mechanism right away in order to bypass
377 geo restriction, of course, if the mechanism is not disabled.
379 _GEO_IP_BLOCKS attribute may contain a list of presumably geo unrestricted
380 IP blocks in CIDR notation for this extractor. One of these IP blocks
381 will be used by geo restriction bypass mechanism similarly
384 Finally, the _WORKING attribute should be set to False for broken IEs
385 in order to warn the users and skip the tests.
390 _x_forwarded_for_ip = None
392 _GEO_COUNTRIES = None
393 _GEO_IP_BLOCKS = None
396 def __init__(self, downloader=None):
397 """Constructor. Receives an optional downloader."""
399 self._x_forwarded_for_ip = None
400 self.set_downloader(downloader)
403 def suitable(cls, url):
404 """Receives a URL and returns True if suitable for this IE."""
406 # This does not use has/getattr intentionally - we want to know whether
407 # we have cached the regexp for *this* class, whereas getattr would also
408 # match the superclass
409 if '_VALID_URL_RE' not in cls.__dict__:
410 cls._VALID_URL_RE = re.compile(cls._VALID_URL)
411 return cls._VALID_URL_RE.match(url) is not None
414 def _match_id(cls, url):
415 if '_VALID_URL_RE' not in cls.__dict__:
416 cls._VALID_URL_RE = re.compile(cls._VALID_URL)
417 m = cls._VALID_URL_RE.match(url)
419 return compat_str(m.group('id'))
423 """Getter method for _WORKING."""
426 def initialize(self):
427 """Initializes an instance (authentication, etc)."""
428 self._initialize_geo_bypass({
429 'countries': self._GEO_COUNTRIES,
430 'ip_blocks': self._GEO_IP_BLOCKS,
433 self._real_initialize()
436 def _initialize_geo_bypass(self, geo_bypass_context):
438 Initialize geo restriction bypass mechanism.
440 This method is used to initialize geo bypass mechanism based on faking
441 X-Forwarded-For HTTP header. A random country from provided country list
442 is selected and a random IP belonging to this country is generated. This
443 IP will be passed as X-Forwarded-For HTTP header in all subsequent
446 This method will be used for initial geo bypass mechanism initialization
447 during the instance initialization with _GEO_COUNTRIES and
450 You may also manually call it from extractor's code if geo bypass
451 information is not available beforehand (e.g. obtained during
452 extraction) or due to some other reason. In this case you should pass
453 this information in geo bypass context passed as first argument. It may
454 contain following fields:
456 countries: List of geo unrestricted countries (similar
458 ip_blocks: List of geo unrestricted IP blocks in CIDR notation
459 (similar to _GEO_IP_BLOCKS)
462 if not self._x_forwarded_for_ip:
464 # Geo bypass mechanism is explicitly disabled by user
465 if not self._downloader.params.get('geo_bypass', True):
468 if not geo_bypass_context:
469 geo_bypass_context = {}
471 # Backward compatibility: previously _initialize_geo_bypass
472 # expected a list of countries, some 3rd party code may still use
474 if isinstance(geo_bypass_context, (list, tuple)):
475 geo_bypass_context = {
476 'countries': geo_bypass_context,
479 # The whole point of geo bypass mechanism is to fake IP
480 # as X-Forwarded-For HTTP header based on some IP block or
483 # Path 1: bypassing based on IP block in CIDR notation
485 # Explicit IP block specified by user, use it right away
486 # regardless of whether extractor is geo bypassable or not
487 ip_block = self._downloader.params.get('geo_bypass_ip_block', None)
489 # Otherwise use random IP block from geo bypass context but only
490 # if extractor is known as geo bypassable
492 ip_blocks = geo_bypass_context.get('ip_blocks')
493 if self._GEO_BYPASS and ip_blocks:
494 ip_block = random.choice(ip_blocks)
497 self._x_forwarded_for_ip = GeoUtils.random_ipv4(ip_block)
498 if self._downloader.params.get('verbose', False):
499 self._downloader.to_screen(
500 '[debug] Using fake IP %s as X-Forwarded-For.'
501 % self._x_forwarded_for_ip)
504 # Path 2: bypassing based on country code
506 # Explicit country code specified by user, use it right away
507 # regardless of whether extractor is geo bypassable or not
508 country = self._downloader.params.get('geo_bypass_country', None)
510 # Otherwise use random country code from geo bypass context but
511 # only if extractor is known as geo bypassable
513 countries = geo_bypass_context.get('countries')
514 if self._GEO_BYPASS and countries:
515 country = random.choice(countries)
518 self._x_forwarded_for_ip = GeoUtils.random_ipv4(country)
519 if self._downloader.params.get('verbose', False):
520 self._downloader.to_screen(
521 '[debug] Using fake IP %s (%s) as X-Forwarded-For.'
522 % (self._x_forwarded_for_ip, country.upper()))
524 def extract(self, url):
525 """Extracts URL information and returns it in list of dicts."""
530 ie_result = self._real_extract(url)
531 if self._x_forwarded_for_ip:
532 ie_result['__x_forwarded_for_ip'] = self._x_forwarded_for_ip
534 except GeoRestrictedError as e:
535 if self.__maybe_fake_ip_and_retry(e.countries):
538 except ExtractorError:
540 except compat_http_client.IncompleteRead as e:
541 raise ExtractorError('A network error has occurred.', cause=e, expected=True)
542 except (KeyError, StopIteration) as e:
543 raise ExtractorError('An extractor error has occurred.', cause=e)
545 def __maybe_fake_ip_and_retry(self, countries):
546 if (not self._downloader.params.get('geo_bypass_country', None)
548 and self._downloader.params.get('geo_bypass', True)
549 and not self._x_forwarded_for_ip
551 country_code = random.choice(countries)
552 self._x_forwarded_for_ip = GeoUtils.random_ipv4(country_code)
553 if self._x_forwarded_for_ip:
555 'Video is geo restricted. Retrying extraction with fake IP %s (%s) as X-Forwarded-For.'
556 % (self._x_forwarded_for_ip, country_code.upper()))
560 def set_downloader(self, downloader):
561 """Sets the downloader for this IE."""
562 self._downloader = downloader
564 def _real_initialize(self):
565 """Real initialization process. Redefine in subclasses."""
568 def _real_extract(self, url):
569 """Real extraction process. Redefine in subclasses."""
574 """A string for getting the InfoExtractor with get_info_extractor"""
575 return compat_str(cls.__name__[:-2])
579 return compat_str(type(self).__name__[:-2])
582 def __can_accept_status_code(err, expected_status):
583 assert isinstance(err, compat_urllib_error.HTTPError)
584 if expected_status is None:
586 if isinstance(expected_status, compat_integer_types):
587 return err.code == expected_status
588 elif isinstance(expected_status, (list, tuple)):
589 return err.code in expected_status
590 elif callable(expected_status):
591 return expected_status(err.code) is True
595 def _request_webpage(self, url_or_request, video_id, note=None, errnote=None, fatal=True, data=None, headers={}, query={}, expected_status=None):
597 Return the response handle.
599 See _download_webpage docstring for arguments specification.
602 self.report_download_webpage(video_id)
603 elif note is not False:
605 self.to_screen('%s' % (note,))
607 self.to_screen('%s: %s' % (video_id, note))
609 # Some sites check X-Forwarded-For HTTP header in order to figure out
610 # the origin of the client behind proxy. This allows bypassing geo
611 # restriction by faking this header's value to IP that belongs to some
612 # geo unrestricted country. We will do so once we encounter any
613 # geo restriction error.
614 if self._x_forwarded_for_ip:
615 if 'X-Forwarded-For' not in headers:
616 headers['X-Forwarded-For'] = self._x_forwarded_for_ip
618 if isinstance(url_or_request, compat_urllib_request.Request):
619 url_or_request = update_Request(
620 url_or_request, data=data, headers=headers, query=query)
623 url_or_request = update_url_query(url_or_request, query)
624 if data is not None or headers:
625 url_or_request = sanitized_Request(url_or_request, data, headers)
627 return self._downloader.urlopen(url_or_request)
628 except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err:
629 if isinstance(err, compat_urllib_error.HTTPError):
630 if self.__can_accept_status_code(err, expected_status):
631 # Retain reference to error to prevent file object from
632 # being closed before it can be read. Works around the
633 # effects of <https://bugs.python.org/issue15002>
634 # introduced in Python 3.4.1.
641 errnote = 'Unable to download webpage'
643 errmsg = '%s: %s' % (errnote, error_to_compat_str(err))
645 raise ExtractorError(errmsg, sys.exc_info()[2], cause=err)
647 self._downloader.report_warning(errmsg)
650 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):
652 Return a tuple (page content as string, URL handle).
654 See _download_webpage docstring for arguments specification.
656 # Strip hashes from the URL (#1038)
657 if isinstance(url_or_request, (compat_str, str)):
658 url_or_request = url_or_request.partition('#')[0]
660 urlh = self._request_webpage(url_or_request, video_id, note, errnote, fatal, data=data, headers=headers, query=query, expected_status=expected_status)
664 content = self._webpage_read_content(urlh, url_or_request, video_id, note, errnote, fatal, encoding=encoding)
665 return (content, urlh)
668 def _guess_encoding_from_content(content_type, webpage_bytes):
669 m = re.match(r'[a-zA-Z0-9_.-]+/[a-zA-Z0-9_.-]+\s*;\s*charset=(.+)', content_type)
671 encoding = m.group(1)
673 m = re.search(br'<meta[^>]+charset=[\'"]?([^\'")]+)[ /\'">]',
674 webpage_bytes[:1024])
676 encoding = m.group(1).decode('ascii')
677 elif webpage_bytes.startswith(b'\xff\xfe'):
684 def __check_blocked(self, content):
685 first_block = content[:512]
686 if ('<title>Access to this site is blocked</title>' in content
687 and 'Websense' in first_block):
688 msg = 'Access to this webpage has been blocked by Websense filtering software in your network.'
689 blocked_iframe = self._html_search_regex(
690 r'<iframe src="([^"]+)"', content,
691 'Websense information URL', default=None)
693 msg += ' Visit %s for more details' % blocked_iframe
694 raise ExtractorError(msg, expected=True)
695 if '<title>The URL you requested has been blocked</title>' in first_block:
697 'Access to this webpage has been blocked by Indian censorship. '
698 'Use a VPN or proxy server (with --proxy) to route around it.')
699 block_msg = self._html_search_regex(
700 r'</h1><p>(.*?)</p>',
701 content, 'block message', default=None)
703 msg += ' (Message: "%s")' % block_msg.replace('\n', ' ')
704 raise ExtractorError(msg, expected=True)
705 if ('<title>TTK :: Доступ к ресурсу ограничен</title>' in content
706 and 'blocklist.rkn.gov.ru' in content):
707 raise ExtractorError(
708 'Access to this webpage has been blocked by decision of the Russian government. '
709 'Visit http://blocklist.rkn.gov.ru/ for a block reason.',
712 def _webpage_read_content(self, urlh, url_or_request, video_id, note=None, errnote=None, fatal=True, prefix=None, encoding=None):
713 content_type = urlh.headers.get('Content-Type', '')
714 webpage_bytes = urlh.read()
715 if prefix is not None:
716 webpage_bytes = prefix + webpage_bytes
718 encoding = self._guess_encoding_from_content(content_type, webpage_bytes)
719 if self._downloader.params.get('dump_intermediate_pages', False):
720 self.to_screen('Dumping request to ' + urlh.geturl())
721 dump = base64.b64encode(webpage_bytes).decode('ascii')
722 self._downloader.to_screen(dump)
723 if self._downloader.params.get('write_pages', False):
724 basen = '%s_%s' % (video_id, urlh.geturl())
726 h = '___' + hashlib.md5(basen.encode('utf-8')).hexdigest()
727 basen = basen[:240 - len(h)] + h
728 raw_filename = basen + '.dump'
729 filename = sanitize_filename(raw_filename, restricted=True)
730 self.to_screen('Saving request to ' + filename)
731 # Working around MAX_PATH limitation on Windows (see
732 # http://msdn.microsoft.com/en-us/library/windows/desktop/aa365247(v=vs.85).aspx)
733 if compat_os_name == 'nt':
734 absfilepath = os.path.abspath(filename)
735 if len(absfilepath) > 259:
736 filename = '\\\\?\\' + absfilepath
737 with open(filename, 'wb') as outf:
738 outf.write(webpage_bytes)
741 content = webpage_bytes.decode(encoding, 'replace')
743 content = webpage_bytes.decode('utf-8', 'replace')
745 self.__check_blocked(content)
749 def _download_webpage(
750 self, url_or_request, video_id, note=None, errnote=None,
751 fatal=True, tries=1, timeout=5, encoding=None, data=None,
752 headers={}, query={}, expected_status=None):
754 Return the data of the page as a string.
757 url_or_request -- plain text URL as a string or
758 a compat_urllib_request.Requestobject
759 video_id -- Video/playlist/item identifier (string)
762 note -- note printed before downloading (string)
763 errnote -- note printed in case of an error (string)
764 fatal -- flag denoting whether error should be considered fatal,
765 i.e. whether it should cause ExtractionError to be raised,
766 otherwise a warning will be reported and extraction continued
767 tries -- number of tries
768 timeout -- sleep interval between tries
769 encoding -- encoding for a page content decoding, guessed automatically
770 when not explicitly specified
771 data -- POST data (bytes)
772 headers -- HTTP headers (dict)
773 query -- URL query (dict)
774 expected_status -- allows to accept failed HTTP requests (non 2xx
775 status code) by explicitly specifying a set of accepted status
776 codes. Can be any of the following entities:
777 - an integer type specifying an exact failed status code to
779 - a list or a tuple of integer types specifying a list of
780 failed status codes to accept
781 - a callable accepting an actual failed status code and
782 returning True if it should be accepted
783 Note that this argument does not affect success status codes (2xx)
784 which are always accepted.
789 while success is False:
791 res = self._download_webpage_handle(
792 url_or_request, video_id, note, errnote, fatal,
793 encoding=encoding, data=data, headers=headers, query=query,
794 expected_status=expected_status)
796 except compat_http_client.IncompleteRead as e:
798 if try_count >= tries:
800 self._sleep(timeout, video_id)
807 def _download_xml_handle(
808 self, url_or_request, video_id, note='Downloading XML',
809 errnote='Unable to download XML', transform_source=None,
810 fatal=True, encoding=None, data=None, headers={}, query={},
811 expected_status=None):
813 Return a tuple (xml as an compat_etree_Element, URL handle).
815 See _download_webpage docstring for arguments specification.
817 res = self._download_webpage_handle(
818 url_or_request, video_id, note, errnote, fatal=fatal,
819 encoding=encoding, data=data, headers=headers, query=query,
820 expected_status=expected_status)
823 xml_string, urlh = res
824 return self._parse_xml(
825 xml_string, video_id, transform_source=transform_source,
829 self, url_or_request, video_id,
830 note='Downloading XML', errnote='Unable to download XML',
831 transform_source=None, fatal=True, encoding=None,
832 data=None, headers={}, query={}, expected_status=None):
834 Return the xml as an compat_etree_Element.
836 See _download_webpage docstring for arguments specification.
838 res = self._download_xml_handle(
839 url_or_request, video_id, note=note, errnote=errnote,
840 transform_source=transform_source, fatal=fatal, encoding=encoding,
841 data=data, headers=headers, query=query,
842 expected_status=expected_status)
843 return res if res is False else res[0]
845 def _parse_xml(self, xml_string, video_id, transform_source=None, fatal=True):
847 xml_string = transform_source(xml_string)
849 return compat_etree_fromstring(xml_string.encode('utf-8'))
850 except compat_xml_parse_error as ve:
851 errmsg = '%s: Failed to parse XML ' % video_id
853 raise ExtractorError(errmsg, cause=ve)
855 self.report_warning(errmsg + str(ve))
857 def _download_json_handle(
858 self, url_or_request, video_id, note='Downloading JSON metadata',
859 errnote='Unable to download JSON metadata', transform_source=None,
860 fatal=True, encoding=None, data=None, headers={}, query={},
861 expected_status=None):
863 Return a tuple (JSON object, URL handle).
865 See _download_webpage docstring for arguments specification.
867 res = self._download_webpage_handle(
868 url_or_request, video_id, note, errnote, fatal=fatal,
869 encoding=encoding, data=data, headers=headers, query=query,
870 expected_status=expected_status)
873 json_string, urlh = res
874 return self._parse_json(
875 json_string, video_id, transform_source=transform_source,
879 self, url_or_request, video_id, note='Downloading JSON metadata',
880 errnote='Unable to download JSON metadata', transform_source=None,
881 fatal=True, encoding=None, data=None, headers={}, query={},
882 expected_status=None):
884 Return the JSON object as a dict.
886 See _download_webpage docstring for arguments specification.
888 res = self._download_json_handle(
889 url_or_request, video_id, note=note, errnote=errnote,
890 transform_source=transform_source, fatal=fatal, encoding=encoding,
891 data=data, headers=headers, query=query,
892 expected_status=expected_status)
893 return res if res is False else res[0]
895 def _parse_json(self, json_string, video_id, transform_source=None, fatal=True):
897 json_string = transform_source(json_string)
899 return json.loads(json_string)
900 except ValueError as ve:
901 errmsg = '%s: Failed to parse JSON ' % video_id
903 raise ExtractorError(errmsg, cause=ve)
905 self.report_warning(errmsg + str(ve))
907 def report_warning(self, msg, video_id=None):
908 idstr = '' if video_id is None else '%s: ' % video_id
909 self._downloader.report_warning(
910 '[%s] %s%s' % (self.IE_NAME, idstr, msg))
912 def to_screen(self, msg):
913 """Print msg to screen, prefixing it with '[ie_name]'"""
914 self._downloader.to_screen('[%s] %s' % (self.IE_NAME, msg))
916 def report_extraction(self, id_or_name):
917 """Report information extraction."""
918 self.to_screen('%s: Extracting information' % id_or_name)
920 def report_download_webpage(self, video_id):
921 """Report webpage download."""
922 self.to_screen('%s: Downloading webpage' % video_id)
924 def report_age_confirmation(self):
925 """Report attempt to confirm age."""
926 self.to_screen('Confirming age')
928 def report_login(self):
929 """Report attempt to log in."""
930 self.to_screen('Logging in')
933 def raise_login_required(msg='This video is only available for registered users'):
934 raise ExtractorError(
935 '%s. Use --username and --password or --netrc to provide account credentials.' % msg,
939 def raise_geo_restricted(msg='This video is not available from your location due to geo restriction', countries=None):
940 raise GeoRestrictedError(msg, countries=countries)
942 # Methods for following #608
944 def url_result(url, ie=None, video_id=None, video_title=None):
945 """Returns a URL that points to a page that should be processed"""
946 # TODO: ie should be the class used for getting the info
947 video_info = {'_type': 'url',
950 if video_id is not None:
951 video_info['id'] = video_id
952 if video_title is not None:
953 video_info['title'] = video_title
956 def playlist_from_matches(self, matches, playlist_id=None, playlist_title=None, getter=None, ie=None):
958 self.url_result(self._proto_relative_url(getter(m) if getter else m), ie)
960 return self.playlist_result(
961 urls, playlist_id=playlist_id, playlist_title=playlist_title)
964 def playlist_result(entries, playlist_id=None, playlist_title=None, playlist_description=None):
965 """Returns a playlist"""
966 video_info = {'_type': 'playlist',
969 video_info['id'] = playlist_id
971 video_info['title'] = playlist_title
972 if playlist_description:
973 video_info['description'] = playlist_description
976 def _search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
978 Perform a regex search on the given string, using a single or a list of
979 patterns returning the first matching group.
980 In case of failure return a default value or raise a WARNING or a
981 RegexNotFoundError, depending on fatal, specifying the field name.
983 if isinstance(pattern, (str, compat_str, compiled_regex_type)):
984 mobj = re.search(pattern, string, flags)
987 mobj = re.search(p, string, flags)
991 if not self._downloader.params.get('no_color') and compat_os_name != 'nt' and sys.stderr.isatty():
992 _name = '\033[0;34m%s\033[0m' % name
998 # return the first matching group
999 return next(g for g in mobj.groups() if g is not None)
1001 return mobj.group(group)
1002 elif default is not NO_DEFAULT:
1005 raise RegexNotFoundError('Unable to extract %s' % _name)
1007 self._downloader.report_warning('unable to extract %s' % _name + bug_reports_message())
1010 def _html_search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
1012 Like _search_regex, but strips HTML tags and unescapes entities.
1014 res = self._search_regex(pattern, string, name, default, fatal, flags, group)
1016 return clean_html(res).strip()
1020 def _get_netrc_login_info(self, netrc_machine=None):
1023 netrc_machine = netrc_machine or self._NETRC_MACHINE
1025 if self._downloader.params.get('usenetrc', False):
1027 info = netrc.netrc().authenticators(netrc_machine)
1028 if info is not None:
1032 raise netrc.NetrcParseError(
1033 'No authenticators for %s' % netrc_machine)
1034 except (IOError, netrc.NetrcParseError) as err:
1035 self._downloader.report_warning(
1036 'parsing .netrc: %s' % error_to_compat_str(err))
1038 return username, password
1040 def _get_login_info(self, username_option='username', password_option='password', netrc_machine=None):
1042 Get the login info as (username, password)
1043 First look for the manually specified credentials using username_option
1044 and password_option as keys in params dictionary. If no such credentials
1045 available look in the netrc file using the netrc_machine or _NETRC_MACHINE
1047 If there's no info available, return (None, None)
1049 if self._downloader is None:
1052 downloader_params = self._downloader.params
1054 # Attempt to use provided username and password or .netrc data
1055 if downloader_params.get(username_option) is not None:
1056 username = downloader_params[username_option]
1057 password = downloader_params[password_option]
1059 username, password = self._get_netrc_login_info(netrc_machine)
1061 return username, password
1063 def _get_tfa_info(self, note='two-factor verification code'):
1065 Get the two-factor authentication info
1066 TODO - asking the user will be required for sms/phone verify
1067 currently just uses the command line option
1068 If there's no info available, return None
1070 if self._downloader is None:
1072 downloader_params = self._downloader.params
1074 if downloader_params.get('twofactor') is not None:
1075 return downloader_params['twofactor']
1077 return compat_getpass('Type %s and press [Return]: ' % note)
1079 # Helper functions for extracting OpenGraph info
1081 def _og_regexes(prop):
1082 content_re = r'content=(?:"([^"]+?)"|\'([^\']+?)\'|\s*([^\s"\'=<>`]+?))'
1083 property_re = (r'(?:name|property)=(?:\'og[:-]%(prop)s\'|"og[:-]%(prop)s"|\s*og[:-]%(prop)s\b)'
1084 % {'prop': re.escape(prop)})
1085 template = r'<meta[^>]+?%s[^>]+?%s'
1087 template % (property_re, content_re),
1088 template % (content_re, property_re),
1092 def _meta_regex(prop):
1093 return r'''(?isx)<meta
1094 (?=[^>]+(?:itemprop|name|property|id|http-equiv)=(["\']?)%s\1)
1095 [^>]+?content=(["\'])(?P<content>.*?)\2''' % re.escape(prop)
1097 def _og_search_property(self, prop, html, name=None, **kargs):
1098 if not isinstance(prop, (list, tuple)):
1101 name = 'OpenGraph %s' % prop[0]
1104 og_regexes.extend(self._og_regexes(p))
1105 escaped = self._search_regex(og_regexes, html, name, flags=re.DOTALL, **kargs)
1108 return unescapeHTML(escaped)
1110 def _og_search_thumbnail(self, html, **kargs):
1111 return self._og_search_property('image', html, 'thumbnail URL', fatal=False, **kargs)
1113 def _og_search_description(self, html, **kargs):
1114 return self._og_search_property('description', html, fatal=False, **kargs)
1116 def _og_search_title(self, html, **kargs):
1117 return self._og_search_property('title', html, **kargs)
1119 def _og_search_video_url(self, html, name='video url', secure=True, **kargs):
1120 regexes = self._og_regexes('video') + self._og_regexes('video:url')
1122 regexes = self._og_regexes('video:secure_url') + regexes
1123 return self._html_search_regex(regexes, html, name, **kargs)
1125 def _og_search_url(self, html, **kargs):
1126 return self._og_search_property('url', html, **kargs)
1128 def _html_search_meta(self, name, html, display_name=None, fatal=False, **kwargs):
1129 if not isinstance(name, (list, tuple)):
1131 if display_name is None:
1132 display_name = name[0]
1133 return self._html_search_regex(
1134 [self._meta_regex(n) for n in name],
1135 html, display_name, fatal=fatal, group='content', **kwargs)
1137 def _dc_search_uploader(self, html):
1138 return self._html_search_meta('dc.creator', html, 'uploader')
1140 def _rta_search(self, html):
1141 # See http://www.rtalabel.org/index.php?content=howtofaq#single
1142 if re.search(r'(?ix)<meta\s+name="rating"\s+'
1143 r' content="RTA-5042-1996-1400-1577-RTA"',
1148 def _media_rating_search(self, html):
1149 # See http://www.tjg-designs.com/WP/metadata-code-examples-adding-metadata-to-your-web-pages/
1150 rating = self._html_search_meta('rating', html)
1162 return RATING_TABLE.get(rating.lower())
1164 def _family_friendly_search(self, html):
1165 # See http://schema.org/VideoObject
1166 family_friendly = self._html_search_meta(
1167 'isFamilyFriendly', html, default=None)
1169 if not family_friendly:
1178 return RATING_TABLE.get(family_friendly.lower())
1180 def _twitter_search_player(self, html):
1181 return self._html_search_meta('twitter:player', html,
1182 'twitter card player')
1184 def _search_json_ld(self, html, video_id, expected_type=None, **kwargs):
1185 json_ld = self._search_regex(
1186 JSON_LD_RE, html, 'JSON-LD', group='json_ld', **kwargs)
1187 default = kwargs.get('default', NO_DEFAULT)
1189 return default if default is not NO_DEFAULT else {}
1190 # JSON-LD may be malformed and thus `fatal` should be respected.
1191 # At the same time `default` may be passed that assumes `fatal=False`
1192 # for _search_regex. Let's simulate the same behavior here as well.
1193 fatal = kwargs.get('fatal', True) if default == NO_DEFAULT else False
1194 return self._json_ld(json_ld, video_id, fatal=fatal, expected_type=expected_type)
1196 def _json_ld(self, json_ld, video_id, fatal=True, expected_type=None):
1197 if isinstance(json_ld, compat_str):
1198 json_ld = self._parse_json(json_ld, video_id, fatal=fatal)
1202 if not isinstance(json_ld, (list, tuple, dict)):
1204 if isinstance(json_ld, dict):
1207 INTERACTION_TYPE_MAP = {
1208 'CommentAction': 'comment',
1209 'AgreeAction': 'like',
1210 'DisagreeAction': 'dislike',
1211 'LikeAction': 'like',
1212 'DislikeAction': 'dislike',
1213 'ListenAction': 'view',
1214 'WatchAction': 'view',
1215 'ViewAction': 'view',
1218 def extract_interaction_statistic(e):
1219 interaction_statistic = e.get('interactionStatistic')
1220 if not isinstance(interaction_statistic, list):
1222 for is_e in interaction_statistic:
1223 if not isinstance(is_e, dict):
1225 if is_e.get('@type') != 'InteractionCounter':
1227 interaction_type = is_e.get('interactionType')
1228 if not isinstance(interaction_type, compat_str):
1230 interaction_count = int_or_none(is_e.get('userInteractionCount'))
1231 if interaction_count is None:
1233 count_kind = INTERACTION_TYPE_MAP.get(interaction_type.split('/')[-1])
1236 count_key = '%s_count' % count_kind
1237 if info.get(count_key) is not None:
1239 info[count_key] = interaction_count
1241 def extract_video_object(e):
1242 assert e['@type'] == 'VideoObject'
1244 'url': url_or_none(e.get('contentUrl')),
1245 'title': unescapeHTML(e.get('name')),
1246 'description': unescapeHTML(e.get('description')),
1247 'thumbnail': url_or_none(e.get('thumbnailUrl') or e.get('thumbnailURL')),
1248 'duration': parse_duration(e.get('duration')),
1249 'timestamp': unified_timestamp(e.get('uploadDate')),
1250 'filesize': float_or_none(e.get('contentSize')),
1251 'tbr': int_or_none(e.get('bitrate')),
1252 'width': int_or_none(e.get('width')),
1253 'height': int_or_none(e.get('height')),
1254 'view_count': int_or_none(e.get('interactionCount')),
1256 extract_interaction_statistic(e)
1259 if isinstance(e.get('@context'), compat_str) and re.match(r'^https?://schema.org/?$', e.get('@context')):
1260 item_type = e.get('@type')
1261 if expected_type is not None and expected_type != item_type:
1263 if item_type in ('TVEpisode', 'Episode'):
1264 episode_name = unescapeHTML(e.get('name'))
1266 'episode': episode_name,
1267 'episode_number': int_or_none(e.get('episodeNumber')),
1268 'description': unescapeHTML(e.get('description')),
1270 if not info.get('title') and episode_name:
1271 info['title'] = episode_name
1272 part_of_season = e.get('partOfSeason')
1273 if isinstance(part_of_season, dict) and part_of_season.get('@type') in ('TVSeason', 'Season', 'CreativeWorkSeason'):
1275 'season': unescapeHTML(part_of_season.get('name')),
1276 'season_number': int_or_none(part_of_season.get('seasonNumber')),
1278 part_of_series = e.get('partOfSeries') or e.get('partOfTVSeries')
1279 if isinstance(part_of_series, dict) and part_of_series.get('@type') in ('TVSeries', 'Series', 'CreativeWorkSeries'):
1280 info['series'] = unescapeHTML(part_of_series.get('name'))
1281 elif item_type == 'Movie':
1283 'title': unescapeHTML(e.get('name')),
1284 'description': unescapeHTML(e.get('description')),
1285 'duration': parse_duration(e.get('duration')),
1286 'timestamp': unified_timestamp(e.get('dateCreated')),
1288 elif item_type in ('Article', 'NewsArticle'):
1290 'timestamp': parse_iso8601(e.get('datePublished')),
1291 'title': unescapeHTML(e.get('headline')),
1292 'description': unescapeHTML(e.get('articleBody')),
1294 elif item_type == 'VideoObject':
1295 extract_video_object(e)
1297 video = e.get('video')
1298 if isinstance(video, dict) and video.get('@type') == 'VideoObject':
1299 extract_video_object(video)
1301 return dict((k, v) for k, v in info.items() if v is not None)
1304 def _hidden_inputs(html):
1305 html = re.sub(r'<!--(?:(?!<!--).)*-->', '', html)
1307 for input in re.findall(r'(?i)(<input[^>]+>)', html):
1308 attrs = extract_attributes(input)
1311 if attrs.get('type') not in ('hidden', 'submit'):
1313 name = attrs.get('name') or attrs.get('id')
1314 value = attrs.get('value')
1315 if name and value is not None:
1316 hidden_inputs[name] = value
1317 return hidden_inputs
1319 def _form_hidden_inputs(self, form_id, html):
1320 form = self._search_regex(
1321 r'(?is)<form[^>]+?id=(["\'])%s\1[^>]*>(?P<form>.+?)</form>' % form_id,
1322 html, '%s form' % form_id, group='form')
1323 return self._hidden_inputs(form)
1325 def _sort_formats(self, formats, field_preference=None):
1327 raise ExtractorError('No video formats found')
1330 # Automatically determine tbr when missing based on abr and vbr (improves
1331 # formats sorting in some cases)
1332 if 'tbr' not in f and f.get('abr') is not None and f.get('vbr') is not None:
1333 f['tbr'] = f['abr'] + f['vbr']
1335 def _formats_key(f):
1336 # TODO remove the following workaround
1337 from ..utils import determine_ext
1338 if not f.get('ext') and 'url' in f:
1339 f['ext'] = determine_ext(f['url'])
1341 if isinstance(field_preference, (list, tuple)):
1344 if f.get(field) is not None
1345 else ('' if field == 'format_id' else -1)
1346 for field in field_preference)
1348 preference = f.get('preference')
1349 if preference is None:
1351 if f.get('ext') in ['f4f', 'f4m']: # Not yet supported
1354 protocol = f.get('protocol') or determine_protocol(f)
1355 proto_preference = 0 if protocol in ['http', 'https'] else (-0.5 if protocol == 'rtsp' else -0.1)
1357 if f.get('vcodec') == 'none': # audio only
1359 if self._downloader.params.get('prefer_free_formats'):
1360 ORDER = ['aac', 'mp3', 'm4a', 'webm', 'ogg', 'opus']
1362 ORDER = ['webm', 'opus', 'ogg', 'mp3', 'aac', 'm4a']
1365 audio_ext_preference = ORDER.index(f['ext'])
1367 audio_ext_preference = -1
1369 if f.get('acodec') == 'none': # video only
1371 if self._downloader.params.get('prefer_free_formats'):
1372 ORDER = ['flv', 'mp4', 'webm']
1374 ORDER = ['webm', 'flv', 'mp4']
1376 ext_preference = ORDER.index(f['ext'])
1379 audio_ext_preference = 0
1383 f.get('language_preference') if f.get('language_preference') is not None else -1,
1384 f.get('quality') if f.get('quality') is not None else -1,
1385 f.get('tbr') if f.get('tbr') is not None else -1,
1386 f.get('filesize') if f.get('filesize') is not None else -1,
1387 f.get('vbr') if f.get('vbr') is not None else -1,
1388 f.get('height') if f.get('height') is not None else -1,
1389 f.get('width') if f.get('width') is not None else -1,
1392 f.get('abr') if f.get('abr') is not None else -1,
1393 audio_ext_preference,
1394 f.get('fps') if f.get('fps') is not None else -1,
1395 f.get('filesize_approx') if f.get('filesize_approx') is not None else -1,
1396 f.get('source_preference') if f.get('source_preference') is not None else -1,
1397 f.get('format_id') if f.get('format_id') is not None else '',
1399 formats.sort(key=_formats_key)
1401 def _check_formats(self, formats, video_id):
1403 formats[:] = filter(
1404 lambda f: self._is_valid_url(
1406 item='%s video format' % f.get('format_id') if f.get('format_id') else 'video'),
1410 def _remove_duplicate_formats(formats):
1414 if f['url'] not in format_urls:
1415 format_urls.add(f['url'])
1416 unique_formats.append(f)
1417 formats[:] = unique_formats
1419 def _is_valid_url(self, url, video_id, item='video', headers={}):
1420 url = self._proto_relative_url(url, scheme='http:')
1421 # For now assume non HTTP(S) URLs always valid
1422 if not (url.startswith('http://') or url.startswith('https://')):
1425 self._request_webpage(url, video_id, 'Checking %s URL' % item, headers=headers)
1427 except ExtractorError:
1429 '%s: %s URL is invalid, skipping' % (video_id, item))
1432 def http_scheme(self):
1433 """ Either "http:" or "https:", depending on the user's preferences """
1436 if self._downloader.params.get('prefer_insecure', False)
1439 def _proto_relative_url(self, url, scheme=None):
1442 if url.startswith('//'):
1444 scheme = self.http_scheme()
1449 def _sleep(self, timeout, video_id, msg_template=None):
1450 if msg_template is None:
1451 msg_template = '%(video_id)s: Waiting for %(timeout)s seconds'
1452 msg = msg_template % {'video_id': video_id, 'timeout': timeout}
1456 def _extract_f4m_formats(self, manifest_url, video_id, preference=None, f4m_id=None,
1457 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1458 fatal=True, m3u8_id=None, data=None, headers={}, query={}):
1459 manifest = self._download_xml(
1460 manifest_url, video_id, 'Downloading f4m manifest',
1461 'Unable to download f4m manifest',
1462 # Some manifests may be malformed, e.g. prosiebensat1 generated manifests
1463 # (see https://github.com/ytdl-org/youtube-dl/issues/6215#issuecomment-121704244)
1464 transform_source=transform_source,
1465 fatal=fatal, data=data, headers=headers, query=query)
1467 if manifest is False:
1470 return self._parse_f4m_formats(
1471 manifest, manifest_url, video_id, preference=preference, f4m_id=f4m_id,
1472 transform_source=transform_source, fatal=fatal, m3u8_id=m3u8_id)
1474 def _parse_f4m_formats(self, manifest, manifest_url, video_id, preference=None, f4m_id=None,
1475 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1476 fatal=True, m3u8_id=None):
1477 if not isinstance(manifest, compat_etree_Element) and not fatal:
1480 # currently youtube-dl cannot decode the playerVerificationChallenge as Akamai uses Adobe Alchemy
1481 akamai_pv = manifest.find('{http://ns.adobe.com/f4m/1.0}pv-2.0')
1482 if akamai_pv is not None and ';' in akamai_pv.text:
1483 playerVerificationChallenge = akamai_pv.text.split(';')[0]
1484 if playerVerificationChallenge.strip() != '':
1488 manifest_version = '1.0'
1489 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/1.0}media')
1491 manifest_version = '2.0'
1492 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/2.0}media')
1493 # Remove unsupported DRM protected media from final formats
1494 # rendition (see https://github.com/ytdl-org/youtube-dl/issues/8573).
1495 media_nodes = remove_encrypted_media(media_nodes)
1499 manifest_base_url = get_base_url(manifest)
1501 bootstrap_info = xpath_element(
1502 manifest, ['{http://ns.adobe.com/f4m/1.0}bootstrapInfo', '{http://ns.adobe.com/f4m/2.0}bootstrapInfo'],
1503 'bootstrap info', default=None)
1506 mime_type = xpath_text(
1507 manifest, ['{http://ns.adobe.com/f4m/1.0}mimeType', '{http://ns.adobe.com/f4m/2.0}mimeType'],
1508 'base URL', default=None)
1509 if mime_type and mime_type.startswith('audio/'):
1512 for i, media_el in enumerate(media_nodes):
1513 tbr = int_or_none(media_el.attrib.get('bitrate'))
1514 width = int_or_none(media_el.attrib.get('width'))
1515 height = int_or_none(media_el.attrib.get('height'))
1516 format_id = '-'.join(filter(None, [f4m_id, compat_str(i if tbr is None else tbr)]))
1517 # If <bootstrapInfo> is present, the specified f4m is a
1518 # stream-level manifest, and only set-level manifests may refer to
1519 # external resources. See section 11.4 and section 4 of F4M spec
1520 if bootstrap_info is None:
1522 # @href is introduced in 2.0, see section 11.6 of F4M spec
1523 if manifest_version == '2.0':
1524 media_url = media_el.attrib.get('href')
1525 if media_url is None:
1526 media_url = media_el.attrib.get('url')
1530 media_url if media_url.startswith('http://') or media_url.startswith('https://')
1531 else ((manifest_base_url or '/'.join(manifest_url.split('/')[:-1])) + '/' + media_url))
1532 # If media_url is itself a f4m manifest do the recursive extraction
1533 # since bitrates in parent manifest (this one) and media_url manifest
1534 # may differ leading to inability to resolve the format by requested
1535 # bitrate in f4m downloader
1536 ext = determine_ext(manifest_url)
1538 f4m_formats = self._extract_f4m_formats(
1539 manifest_url, video_id, preference=preference, f4m_id=f4m_id,
1540 transform_source=transform_source, fatal=fatal)
1541 # Sometimes stream-level manifest contains single media entry that
1542 # does not contain any quality metadata (e.g. http://matchtv.ru/#live-player).
1543 # At the same time parent's media entry in set-level manifest may
1544 # contain it. We will copy it from parent in such cases.
1545 if len(f4m_formats) == 1:
1548 'tbr': f.get('tbr') or tbr,
1549 'width': f.get('width') or width,
1550 'height': f.get('height') or height,
1551 'format_id': f.get('format_id') if not tbr else format_id,
1554 formats.extend(f4m_formats)
1557 formats.extend(self._extract_m3u8_formats(
1558 manifest_url, video_id, 'mp4', preference=preference,
1559 m3u8_id=m3u8_id, fatal=fatal))
1562 'format_id': format_id,
1563 'url': manifest_url,
1564 'manifest_url': manifest_url,
1565 'ext': 'flv' if bootstrap_info is not None else None,
1571 'preference': preference,
1575 def _m3u8_meta_format(self, m3u8_url, ext=None, preference=None, m3u8_id=None):
1577 'format_id': '-'.join(filter(None, [m3u8_id, 'meta'])),
1581 'preference': preference - 100 if preference else -100,
1582 'resolution': 'multiple',
1583 'format_note': 'Quality selection URL',
1586 def _extract_m3u8_formats(self, m3u8_url, video_id, ext=None,
1587 entry_protocol='m3u8', preference=None,
1588 m3u8_id=None, note=None, errnote=None,
1589 fatal=True, live=False, data=None, headers={},
1591 res = self._download_webpage_handle(
1593 note=note or 'Downloading m3u8 information',
1594 errnote=errnote or 'Failed to download m3u8 information',
1595 fatal=fatal, data=data, headers=headers, query=query)
1600 m3u8_doc, urlh = res
1601 m3u8_url = urlh.geturl()
1603 return self._parse_m3u8_formats(
1604 m3u8_doc, m3u8_url, ext=ext, entry_protocol=entry_protocol,
1605 preference=preference, m3u8_id=m3u8_id, live=live)
1607 def _parse_m3u8_formats(self, m3u8_doc, m3u8_url, ext=None,
1608 entry_protocol='m3u8', preference=None,
1609 m3u8_id=None, live=False):
1610 if '#EXT-X-FAXS-CM:' in m3u8_doc: # Adobe Flash Access
1613 if re.search(r'#EXT-X-SESSION-KEY:.*?URI="skd://', m3u8_doc): # Apple FairPlay
1618 format_url = lambda u: (
1620 if re.match(r'^https?://', u)
1621 else compat_urlparse.urljoin(m3u8_url, u))
1624 # 1. https://tools.ietf.org/html/draft-pantos-http-live-streaming-21
1625 # 2. https://github.com/ytdl-org/youtube-dl/issues/12211
1626 # 3. https://github.com/ytdl-org/youtube-dl/issues/18923
1628 # We should try extracting formats only from master playlists [1, 4.3.4],
1629 # i.e. playlists that describe available qualities. On the other hand
1630 # media playlists [1, 4.3.3] should be returned as is since they contain
1631 # just the media without qualities renditions.
1632 # Fortunately, master playlist can be easily distinguished from media
1633 # playlist based on particular tags availability. As of [1, 4.3.3, 4.3.4]
1634 # master playlist tags MUST NOT appear in a media playist and vice versa.
1635 # As of [1, 4.3.3.1] #EXT-X-TARGETDURATION tag is REQUIRED for every
1636 # media playlist and MUST NOT appear in master playlist thus we can
1637 # clearly detect media playlist with this criterion.
1639 if '#EXT-X-TARGETDURATION' in m3u8_doc: # media playlist, return as is
1642 'format_id': m3u8_id,
1644 'protocol': entry_protocol,
1645 'preference': preference,
1649 last_stream_inf = {}
1651 def extract_media(x_media_line):
1652 media = parse_m3u8_attributes(x_media_line)
1653 # As per [1, 4.3.4.1] TYPE, GROUP-ID and NAME are REQUIRED
1654 media_type, group_id, name = media.get('TYPE'), media.get('GROUP-ID'), media.get('NAME')
1655 if not (media_type and group_id and name):
1657 groups.setdefault(group_id, []).append(media)
1658 if media_type not in ('VIDEO', 'AUDIO'):
1660 media_url = media.get('URI')
1663 for v in (m3u8_id, group_id, name):
1667 'format_id': '-'.join(format_id),
1668 'url': format_url(media_url),
1669 'manifest_url': m3u8_url,
1670 'language': media.get('LANGUAGE'),
1672 'protocol': entry_protocol,
1673 'preference': preference,
1675 if media_type == 'AUDIO':
1676 f['vcodec'] = 'none'
1679 def build_stream_name():
1680 # Despite specification does not mention NAME attribute for
1681 # EXT-X-STREAM-INF tag it still sometimes may be present (see [1]
1682 # or vidio test in TestInfoExtractor.test_parse_m3u8_formats)
1683 # 1. http://www.vidio.com/watch/165683-dj_ambred-booyah-live-2015
1684 stream_name = last_stream_inf.get('NAME')
1687 # If there is no NAME in EXT-X-STREAM-INF it will be obtained
1688 # from corresponding rendition group
1689 stream_group_id = last_stream_inf.get('VIDEO')
1690 if not stream_group_id:
1692 stream_group = groups.get(stream_group_id)
1693 if not stream_group:
1694 return stream_group_id
1695 rendition = stream_group[0]
1696 return rendition.get('NAME') or stream_group_id
1698 # parse EXT-X-MEDIA tags before EXT-X-STREAM-INF in order to have the
1699 # chance to detect video only formats when EXT-X-STREAM-INF tags
1700 # precede EXT-X-MEDIA tags in HLS manifest such as [3].
1701 for line in m3u8_doc.splitlines():
1702 if line.startswith('#EXT-X-MEDIA:'):
1705 for line in m3u8_doc.splitlines():
1706 if line.startswith('#EXT-X-STREAM-INF:'):
1707 last_stream_inf = parse_m3u8_attributes(line)
1708 elif line.startswith('#') or not line.strip():
1711 tbr = float_or_none(
1712 last_stream_inf.get('AVERAGE-BANDWIDTH')
1713 or last_stream_inf.get('BANDWIDTH'), scale=1000)
1716 format_id.append(m3u8_id)
1717 stream_name = build_stream_name()
1718 # Bandwidth of live streams may differ over time thus making
1719 # format_id unpredictable. So it's better to keep provided
1722 format_id.append(stream_name if stream_name else '%d' % (tbr if tbr else len(formats)))
1723 manifest_url = format_url(line.strip())
1725 'format_id': '-'.join(format_id),
1726 'url': manifest_url,
1727 'manifest_url': m3u8_url,
1730 'fps': float_or_none(last_stream_inf.get('FRAME-RATE')),
1731 'protocol': entry_protocol,
1732 'preference': preference,
1734 resolution = last_stream_inf.get('RESOLUTION')
1736 mobj = re.search(r'(?P<width>\d+)[xX](?P<height>\d+)', resolution)
1738 f['width'] = int(mobj.group('width'))
1739 f['height'] = int(mobj.group('height'))
1740 # Unified Streaming Platform
1742 r'audio.*?(?:%3D|=)(\d+)(?:-video.*?(?:%3D|=)(\d+))?', f['url'])
1744 abr, vbr = mobj.groups()
1745 abr, vbr = float_or_none(abr, 1000), float_or_none(vbr, 1000)
1750 codecs = parse_codecs(last_stream_inf.get('CODECS'))
1752 audio_group_id = last_stream_inf.get('AUDIO')
1753 # As per [1, 4.3.4.1.1] any EXT-X-STREAM-INF tag which
1754 # references a rendition group MUST have a CODECS attribute.
1755 # However, this is not always respected, for example, [2]
1756 # contains EXT-X-STREAM-INF tag which references AUDIO
1757 # rendition group but does not have CODECS and despite
1758 # referencing an audio group it represents a complete
1759 # (with audio and video) format. So, for such cases we will
1760 # ignore references to rendition groups and treat them
1761 # as complete formats.
1762 if audio_group_id and codecs and f.get('vcodec') != 'none':
1763 audio_group = groups.get(audio_group_id)
1764 if audio_group and audio_group[0].get('URI'):
1765 # TODO: update acodec for audio only formats with
1767 f['acodec'] = 'none'
1771 progressive_uri = last_stream_inf.get('PROGRESSIVE-URI')
1774 del http_f['manifest_url']
1776 'format_id': f['format_id'].replace('hls-', 'http-'),
1778 'url': progressive_uri,
1780 formats.append(http_f)
1782 last_stream_inf = {}
1786 def _xpath_ns(path, namespace=None):
1790 for c in path.split('/'):
1791 if not c or c == '.':
1794 out.append('{%s}%s' % (namespace, c))
1795 return '/'.join(out)
1797 def _extract_smil_formats(self, smil_url, video_id, fatal=True, f4m_params=None, transform_source=None):
1798 smil = self._download_smil(smil_url, video_id, fatal=fatal, transform_source=transform_source)
1804 namespace = self._parse_smil_namespace(smil)
1806 return self._parse_smil_formats(
1807 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
1809 def _extract_smil_info(self, smil_url, video_id, fatal=True, f4m_params=None):
1810 smil = self._download_smil(smil_url, video_id, fatal=fatal)
1813 return self._parse_smil(smil, smil_url, video_id, f4m_params=f4m_params)
1815 def _download_smil(self, smil_url, video_id, fatal=True, transform_source=None):
1816 return self._download_xml(
1817 smil_url, video_id, 'Downloading SMIL file',
1818 'Unable to download SMIL file', fatal=fatal, transform_source=transform_source)
1820 def _parse_smil(self, smil, smil_url, video_id, f4m_params=None):
1821 namespace = self._parse_smil_namespace(smil)
1823 formats = self._parse_smil_formats(
1824 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
1825 subtitles = self._parse_smil_subtitles(smil, namespace=namespace)
1827 video_id = os.path.splitext(url_basename(smil_url))[0]
1831 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
1832 name = meta.attrib.get('name')
1833 content = meta.attrib.get('content')
1834 if not name or not content:
1836 if not title and name == 'title':
1838 elif not description and name in ('description', 'abstract'):
1839 description = content
1840 elif not upload_date and name == 'date':
1841 upload_date = unified_strdate(content)
1844 'id': image.get('type'),
1845 'url': image.get('src'),
1846 'width': int_or_none(image.get('width')),
1847 'height': int_or_none(image.get('height')),
1848 } for image in smil.findall(self._xpath_ns('.//image', namespace)) if image.get('src')]
1852 'title': title or video_id,
1853 'description': description,
1854 'upload_date': upload_date,
1855 'thumbnails': thumbnails,
1857 'subtitles': subtitles,
1860 def _parse_smil_namespace(self, smil):
1861 return self._search_regex(
1862 r'(?i)^{([^}]+)?}smil$', smil.tag, 'namespace', default=None)
1864 def _parse_smil_formats(self, smil, smil_url, video_id, namespace=None, f4m_params=None, transform_rtmp_url=None):
1866 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
1867 b = meta.get('base') or meta.get('httpBase')
1878 media = smil.findall(self._xpath_ns('.//video', namespace)) + smil.findall(self._xpath_ns('.//audio', namespace))
1879 for medium in media:
1880 src = medium.get('src')
1881 if not src or src in srcs:
1885 bitrate = float_or_none(medium.get('system-bitrate') or medium.get('systemBitrate'), 1000)
1886 filesize = int_or_none(medium.get('size') or medium.get('fileSize'))
1887 width = int_or_none(medium.get('width'))
1888 height = int_or_none(medium.get('height'))
1889 proto = medium.get('proto')
1890 ext = medium.get('ext')
1891 src_ext = determine_ext(src)
1892 streamer = medium.get('streamer') or base
1894 if proto == 'rtmp' or streamer.startswith('rtmp'):
1900 'format_id': 'rtmp-%d' % (rtmp_count if bitrate is None else bitrate),
1902 'filesize': filesize,
1906 if transform_rtmp_url:
1907 streamer, src = transform_rtmp_url(streamer, src)
1908 formats[-1].update({
1914 src_url = src if src.startswith('http') else compat_urlparse.urljoin(base, src)
1915 src_url = src_url.strip()
1917 if proto == 'm3u8' or src_ext == 'm3u8':
1918 m3u8_formats = self._extract_m3u8_formats(
1919 src_url, video_id, ext or 'mp4', m3u8_id='hls', fatal=False)
1920 if len(m3u8_formats) == 1:
1922 m3u8_formats[0].update({
1923 'format_id': 'hls-%d' % (m3u8_count if bitrate is None else bitrate),
1928 formats.extend(m3u8_formats)
1929 elif src_ext == 'f4m':
1934 'plugin': 'flowplayer-3.2.0.1',
1936 f4m_url += '&' if '?' in f4m_url else '?'
1937 f4m_url += compat_urllib_parse_urlencode(f4m_params)
1938 formats.extend(self._extract_f4m_formats(f4m_url, video_id, f4m_id='hds', fatal=False))
1939 elif src_ext == 'mpd':
1940 formats.extend(self._extract_mpd_formats(
1941 src_url, video_id, mpd_id='dash', fatal=False))
1942 elif re.search(r'\.ism/[Mm]anifest', src_url):
1943 formats.extend(self._extract_ism_formats(
1944 src_url, video_id, ism_id='mss', fatal=False))
1945 elif src_url.startswith('http') and self._is_valid_url(src, video_id):
1949 'ext': ext or src_ext or 'flv',
1950 'format_id': 'http-%d' % (bitrate or http_count),
1952 'filesize': filesize,
1959 def _parse_smil_subtitles(self, smil, namespace=None, subtitles_lang='en'):
1962 for num, textstream in enumerate(smil.findall(self._xpath_ns('.//textstream', namespace))):
1963 src = textstream.get('src')
1964 if not src or src in urls:
1967 ext = textstream.get('ext') or mimetype2ext(textstream.get('type')) or determine_ext(src)
1968 lang = textstream.get('systemLanguage') or textstream.get('systemLanguageName') or textstream.get('lang') or subtitles_lang
1969 subtitles.setdefault(lang, []).append({
1975 def _extract_xspf_playlist(self, xspf_url, playlist_id, fatal=True):
1976 xspf = self._download_xml(
1977 xspf_url, playlist_id, 'Downloading xpsf playlist',
1978 'Unable to download xspf manifest', fatal=fatal)
1981 return self._parse_xspf(
1982 xspf, playlist_id, xspf_url=xspf_url,
1983 xspf_base_url=base_url(xspf_url))
1985 def _parse_xspf(self, xspf_doc, playlist_id, xspf_url=None, xspf_base_url=None):
1987 'xspf': 'http://xspf.org/ns/0/',
1988 's1': 'http://static.streamone.nl/player/ns/0',
1992 for track in xspf_doc.findall(xpath_with_ns('./xspf:trackList/xspf:track', NS_MAP)):
1994 track, xpath_with_ns('./xspf:title', NS_MAP), 'title', default=playlist_id)
1995 description = xpath_text(
1996 track, xpath_with_ns('./xspf:annotation', NS_MAP), 'description')
1997 thumbnail = xpath_text(
1998 track, xpath_with_ns('./xspf:image', NS_MAP), 'thumbnail')
1999 duration = float_or_none(
2000 xpath_text(track, xpath_with_ns('./xspf:duration', NS_MAP), 'duration'), 1000)
2003 for location in track.findall(xpath_with_ns('./xspf:location', NS_MAP)):
2004 format_url = urljoin(xspf_base_url, location.text)
2009 'manifest_url': xspf_url,
2010 'format_id': location.get(xpath_with_ns('s1:label', NS_MAP)),
2011 'width': int_or_none(location.get(xpath_with_ns('s1:width', NS_MAP))),
2012 'height': int_or_none(location.get(xpath_with_ns('s1:height', NS_MAP))),
2014 self._sort_formats(formats)
2019 'description': description,
2020 'thumbnail': thumbnail,
2021 'duration': duration,
2026 def _extract_mpd_formats(self, mpd_url, video_id, mpd_id=None, note=None, errnote=None, fatal=True, formats_dict={}, data=None, headers={}, query={}):
2027 res = self._download_xml_handle(
2029 note=note or 'Downloading MPD manifest',
2030 errnote=errnote or 'Failed to download MPD manifest',
2031 fatal=fatal, data=data, headers=headers, query=query)
2037 mpd_base_url = base_url(urlh.geturl())
2039 return self._parse_mpd_formats(
2040 mpd_doc, mpd_id=mpd_id, mpd_base_url=mpd_base_url,
2041 formats_dict=formats_dict, mpd_url=mpd_url)
2043 def _parse_mpd_formats(self, mpd_doc, mpd_id=None, mpd_base_url='', formats_dict={}, mpd_url=None):
2045 Parse formats from MPD manifest.
2047 1. MPEG-DASH Standard, ISO/IEC 23009-1:2014(E),
2048 http://standards.iso.org/ittf/PubliclyAvailableStandards/c065274_ISO_IEC_23009-1_2014.zip
2049 2. https://en.wikipedia.org/wiki/Dynamic_Adaptive_Streaming_over_HTTP
2051 if mpd_doc.get('type') == 'dynamic':
2054 namespace = self._search_regex(r'(?i)^{([^}]+)?}MPD$', mpd_doc.tag, 'namespace', default=None)
2057 return self._xpath_ns(path, namespace)
2059 def is_drm_protected(element):
2060 return element.find(_add_ns('ContentProtection')) is not None
2062 def extract_multisegment_info(element, ms_parent_info):
2063 ms_info = ms_parent_info.copy()
2065 # As per [1, 5.3.9.2.2] SegmentList and SegmentTemplate share some
2066 # common attributes and elements. We will only extract relevant
2068 def extract_common(source):
2069 segment_timeline = source.find(_add_ns('SegmentTimeline'))
2070 if segment_timeline is not None:
2071 s_e = segment_timeline.findall(_add_ns('S'))
2073 ms_info['total_number'] = 0
2076 r = int(s.get('r', 0))
2077 ms_info['total_number'] += 1 + r
2078 ms_info['s'].append({
2079 't': int(s.get('t', 0)),
2080 # @d is mandatory (see [1, 5.3.9.6.2, Table 17, page 60])
2081 'd': int(s.attrib['d']),
2084 start_number = source.get('startNumber')
2086 ms_info['start_number'] = int(start_number)
2087 timescale = source.get('timescale')
2089 ms_info['timescale'] = int(timescale)
2090 segment_duration = source.get('duration')
2091 if segment_duration:
2092 ms_info['segment_duration'] = float(segment_duration)
2094 def extract_Initialization(source):
2095 initialization = source.find(_add_ns('Initialization'))
2096 if initialization is not None:
2097 ms_info['initialization_url'] = initialization.attrib['sourceURL']
2099 segment_list = element.find(_add_ns('SegmentList'))
2100 if segment_list is not None:
2101 extract_common(segment_list)
2102 extract_Initialization(segment_list)
2103 segment_urls_e = segment_list.findall(_add_ns('SegmentURL'))
2105 ms_info['segment_urls'] = [segment.attrib['media'] for segment in segment_urls_e]
2107 segment_template = element.find(_add_ns('SegmentTemplate'))
2108 if segment_template is not None:
2109 extract_common(segment_template)
2110 media = segment_template.get('media')
2112 ms_info['media'] = media
2113 initialization = segment_template.get('initialization')
2115 ms_info['initialization'] = initialization
2117 extract_Initialization(segment_template)
2120 mpd_duration = parse_duration(mpd_doc.get('mediaPresentationDuration'))
2122 for period in mpd_doc.findall(_add_ns('Period')):
2123 period_duration = parse_duration(period.get('duration')) or mpd_duration
2124 period_ms_info = extract_multisegment_info(period, {
2128 for adaptation_set in period.findall(_add_ns('AdaptationSet')):
2129 if is_drm_protected(adaptation_set):
2131 adaption_set_ms_info = extract_multisegment_info(adaptation_set, period_ms_info)
2132 for representation in adaptation_set.findall(_add_ns('Representation')):
2133 if is_drm_protected(representation):
2135 representation_attrib = adaptation_set.attrib.copy()
2136 representation_attrib.update(representation.attrib)
2137 # According to [1, 5.3.7.2, Table 9, page 41], @mimeType is mandatory
2138 mime_type = representation_attrib['mimeType']
2139 content_type = mime_type.split('/')[0]
2140 if content_type == 'text':
2141 # TODO implement WebVTT downloading
2143 elif content_type in ('video', 'audio'):
2145 for element in (representation, adaptation_set, period, mpd_doc):
2146 base_url_e = element.find(_add_ns('BaseURL'))
2147 if base_url_e is not None:
2148 base_url = base_url_e.text + base_url
2149 if re.match(r'^https?://', base_url):
2151 if mpd_base_url and not re.match(r'^https?://', base_url):
2152 if not mpd_base_url.endswith('/') and not base_url.startswith('/'):
2154 base_url = mpd_base_url + base_url
2155 representation_id = representation_attrib.get('id')
2156 lang = representation_attrib.get('lang')
2157 url_el = representation.find(_add_ns('BaseURL'))
2158 filesize = int_or_none(url_el.attrib.get('{http://youtube.com/yt/2012/10/10}contentLength') if url_el is not None else None)
2159 bandwidth = int_or_none(representation_attrib.get('bandwidth'))
2161 'format_id': '%s-%s' % (mpd_id, representation_id) if mpd_id else representation_id,
2162 'manifest_url': mpd_url,
2163 'ext': mimetype2ext(mime_type),
2164 'width': int_or_none(representation_attrib.get('width')),
2165 'height': int_or_none(representation_attrib.get('height')),
2166 'tbr': float_or_none(bandwidth, 1000),
2167 'asr': int_or_none(representation_attrib.get('audioSamplingRate')),
2168 'fps': int_or_none(representation_attrib.get('frameRate')),
2169 'language': lang if lang not in ('mul', 'und', 'zxx', 'mis') else None,
2170 'format_note': 'DASH %s' % content_type,
2171 'filesize': filesize,
2172 'container': mimetype2ext(mime_type) + '_dash',
2174 f.update(parse_codecs(representation_attrib.get('codecs')))
2175 representation_ms_info = extract_multisegment_info(representation, adaption_set_ms_info)
2177 def prepare_template(template_name, identifiers):
2178 tmpl = representation_ms_info[template_name]
2179 # First of, % characters outside $...$ templates
2180 # must be escaped by doubling for proper processing
2181 # by % operator string formatting used further (see
2182 # https://github.com/ytdl-org/youtube-dl/issues/16867).
2188 in_template = not in_template
2189 elif c == '%' and not in_template:
2191 # Next, $...$ templates are translated to their
2192 # %(...) counterparts to be used with % operator
2193 t = t.replace('$RepresentationID$', representation_id)
2194 t = re.sub(r'\$(%s)\$' % '|'.join(identifiers), r'%(\1)d', t)
2195 t = re.sub(r'\$(%s)%%([^$]+)\$' % '|'.join(identifiers), r'%(\1)\2', t)
2196 t.replace('$$', '$')
2199 # @initialization is a regular template like @media one
2200 # so it should be handled just the same way (see
2201 # https://github.com/ytdl-org/youtube-dl/issues/11605)
2202 if 'initialization' in representation_ms_info:
2203 initialization_template = prepare_template(
2205 # As per [1, 5.3.9.4.2, Table 15, page 54] $Number$ and
2206 # $Time$ shall not be included for @initialization thus
2207 # only $Bandwidth$ remains
2209 representation_ms_info['initialization_url'] = initialization_template % {
2210 'Bandwidth': bandwidth,
2213 def location_key(location):
2214 return 'url' if re.match(r'^https?://', location) else 'path'
2216 if 'segment_urls' not in representation_ms_info and 'media' in representation_ms_info:
2218 media_template = prepare_template('media', ('Number', 'Bandwidth', 'Time'))
2219 media_location_key = location_key(media_template)
2221 # As per [1, 5.3.9.4.4, Table 16, page 55] $Number$ and $Time$
2222 # can't be used at the same time
2223 if '%(Number' in media_template and 's' not in representation_ms_info:
2224 segment_duration = None
2225 if 'total_number' not in representation_ms_info and 'segment_duration' in representation_ms_info:
2226 segment_duration = float_or_none(representation_ms_info['segment_duration'], representation_ms_info['timescale'])
2227 representation_ms_info['total_number'] = int(math.ceil(float(period_duration) / segment_duration))
2228 representation_ms_info['fragments'] = [{
2229 media_location_key: media_template % {
2230 'Number': segment_number,
2231 'Bandwidth': bandwidth,
2233 'duration': segment_duration,
2234 } for segment_number in range(
2235 representation_ms_info['start_number'],
2236 representation_ms_info['total_number'] + representation_ms_info['start_number'])]
2238 # $Number*$ or $Time$ in media template with S list available
2239 # Example $Number*$: http://www.svtplay.se/klipp/9023742/stopptid-om-bjorn-borg
2240 # Example $Time$: https://play.arkena.com/embed/avp/v2/player/media/b41dda37-d8e7-4d3f-b1b5-9a9db578bdfe/1/129411
2241 representation_ms_info['fragments'] = []
2244 segment_number = representation_ms_info['start_number']
2246 def add_segment_url():
2247 segment_url = media_template % {
2248 'Time': segment_time,
2249 'Bandwidth': bandwidth,
2250 'Number': segment_number,
2252 representation_ms_info['fragments'].append({
2253 media_location_key: segment_url,
2254 'duration': float_or_none(segment_d, representation_ms_info['timescale']),
2257 for num, s in enumerate(representation_ms_info['s']):
2258 segment_time = s.get('t') or segment_time
2262 for r in range(s.get('r', 0)):
2263 segment_time += segment_d
2266 segment_time += segment_d
2267 elif 'segment_urls' in representation_ms_info and 's' in representation_ms_info:
2269 # Example: https://www.youtube.com/watch?v=iXZV5uAYMJI
2270 # or any YouTube dashsegments video
2273 timescale = representation_ms_info['timescale']
2274 for s in representation_ms_info['s']:
2275 duration = float_or_none(s['d'], timescale)
2276 for r in range(s.get('r', 0) + 1):
2277 segment_uri = representation_ms_info['segment_urls'][segment_index]
2279 location_key(segment_uri): segment_uri,
2280 'duration': duration,
2283 representation_ms_info['fragments'] = fragments
2284 elif 'segment_urls' in representation_ms_info:
2285 # Segment URLs with no SegmentTimeline
2286 # Example: https://www.seznam.cz/zpravy/clanek/cesko-zasahne-vitr-o-sile-vichrice-muze-byt-i-zivotu-nebezpecny-39091
2287 # https://github.com/ytdl-org/youtube-dl/pull/14844
2289 segment_duration = float_or_none(
2290 representation_ms_info['segment_duration'],
2291 representation_ms_info['timescale']) if 'segment_duration' in representation_ms_info else None
2292 for segment_url in representation_ms_info['segment_urls']:
2294 location_key(segment_url): segment_url,
2296 if segment_duration:
2297 fragment['duration'] = segment_duration
2298 fragments.append(fragment)
2299 representation_ms_info['fragments'] = fragments
2300 # If there is a fragments key available then we correctly recognized fragmented media.
2301 # Otherwise we will assume unfragmented media with direct access. Technically, such
2302 # assumption is not necessarily correct since we may simply have no support for
2303 # some forms of fragmented media renditions yet, but for now we'll use this fallback.
2304 if 'fragments' in representation_ms_info:
2306 # NB: mpd_url may be empty when MPD manifest is parsed from a string
2307 'url': mpd_url or base_url,
2308 'fragment_base_url': base_url,
2310 'protocol': 'http_dash_segments',
2312 if 'initialization_url' in representation_ms_info:
2313 initialization_url = representation_ms_info['initialization_url']
2314 if not f.get('url'):
2315 f['url'] = initialization_url
2316 f['fragments'].append({location_key(initialization_url): initialization_url})
2317 f['fragments'].extend(representation_ms_info['fragments'])
2319 # Assuming direct URL to unfragmented media.
2322 # According to [1, 5.3.5.2, Table 7, page 35] @id of Representation
2323 # is not necessarily unique within a Period thus formats with
2324 # the same `format_id` are quite possible. There are numerous examples
2325 # of such manifests (see https://github.com/ytdl-org/youtube-dl/issues/15111,
2326 # https://github.com/ytdl-org/youtube-dl/issues/13919)
2327 full_info = formats_dict.get(representation_id, {}).copy()
2329 formats.append(full_info)
2331 self.report_warning('Unknown MIME type %s in DASH manifest' % mime_type)
2334 def _extract_ism_formats(self, ism_url, video_id, ism_id=None, note=None, errnote=None, fatal=True, data=None, headers={}, query={}):
2335 res = self._download_xml_handle(
2337 note=note or 'Downloading ISM manifest',
2338 errnote=errnote or 'Failed to download ISM manifest',
2339 fatal=fatal, data=data, headers=headers, query=query)
2346 return self._parse_ism_formats(ism_doc, urlh.geturl(), ism_id)
2348 def _parse_ism_formats(self, ism_doc, ism_url, ism_id=None):
2350 Parse formats from ISM manifest.
2352 1. [MS-SSTR]: Smooth Streaming Protocol,
2353 https://msdn.microsoft.com/en-us/library/ff469518.aspx
2355 if ism_doc.get('IsLive') == 'TRUE' or ism_doc.find('Protection') is not None:
2358 duration = int(ism_doc.attrib['Duration'])
2359 timescale = int_or_none(ism_doc.get('TimeScale')) or 10000000
2362 for stream in ism_doc.findall('StreamIndex'):
2363 stream_type = stream.get('Type')
2364 if stream_type not in ('video', 'audio'):
2366 url_pattern = stream.attrib['Url']
2367 stream_timescale = int_or_none(stream.get('TimeScale')) or timescale
2368 stream_name = stream.get('Name')
2369 for track in stream.findall('QualityLevel'):
2370 fourcc = track.get('FourCC', 'AACL' if track.get('AudioTag') == '255' else None)
2371 # TODO: add support for WVC1 and WMAP
2372 if fourcc not in ('H264', 'AVC1', 'AACL'):
2373 self.report_warning('%s is not a supported codec' % fourcc)
2375 tbr = int(track.attrib['Bitrate']) // 1000
2376 # [1] does not mention Width and Height attributes. However,
2377 # they're often present while MaxWidth and MaxHeight are
2378 # missing, so should be used as fallbacks
2379 width = int_or_none(track.get('MaxWidth') or track.get('Width'))
2380 height = int_or_none(track.get('MaxHeight') or track.get('Height'))
2381 sampling_rate = int_or_none(track.get('SamplingRate'))
2383 track_url_pattern = re.sub(r'{[Bb]itrate}', track.attrib['Bitrate'], url_pattern)
2384 track_url_pattern = compat_urlparse.urljoin(ism_url, track_url_pattern)
2390 stream_fragments = stream.findall('c')
2391 for stream_fragment_index, stream_fragment in enumerate(stream_fragments):
2392 fragment_ctx['time'] = int_or_none(stream_fragment.get('t')) or fragment_ctx['time']
2393 fragment_repeat = int_or_none(stream_fragment.get('r')) or 1
2394 fragment_ctx['duration'] = int_or_none(stream_fragment.get('d'))
2395 if not fragment_ctx['duration']:
2397 next_fragment_time = int(stream_fragment[stream_fragment_index + 1].attrib['t'])
2399 next_fragment_time = duration
2400 fragment_ctx['duration'] = (next_fragment_time - fragment_ctx['time']) / fragment_repeat
2401 for _ in range(fragment_repeat):
2403 'url': re.sub(r'{start[ _]time}', compat_str(fragment_ctx['time']), track_url_pattern),
2404 'duration': fragment_ctx['duration'] / stream_timescale,
2406 fragment_ctx['time'] += fragment_ctx['duration']
2410 format_id.append(ism_id)
2412 format_id.append(stream_name)
2413 format_id.append(compat_str(tbr))
2416 'format_id': '-'.join(format_id),
2418 'manifest_url': ism_url,
2419 'ext': 'ismv' if stream_type == 'video' else 'isma',
2423 'asr': sampling_rate,
2424 'vcodec': 'none' if stream_type == 'audio' else fourcc,
2425 'acodec': 'none' if stream_type == 'video' else fourcc,
2427 'fragments': fragments,
2428 '_download_params': {
2429 'duration': duration,
2430 'timescale': stream_timescale,
2431 'width': width or 0,
2432 'height': height or 0,
2434 'codec_private_data': track.get('CodecPrivateData'),
2435 'sampling_rate': sampling_rate,
2436 'channels': int_or_none(track.get('Channels', 2)),
2437 'bits_per_sample': int_or_none(track.get('BitsPerSample', 16)),
2438 'nal_unit_length_field': int_or_none(track.get('NALUnitLengthField', 4)),
2443 def _parse_html5_media_entries(self, base_url, webpage, video_id, m3u8_id=None, m3u8_entry_protocol='m3u8', mpd_id=None, preference=None):
2444 def absolute_url(item_url):
2445 return urljoin(base_url, item_url)
2447 def parse_content_type(content_type):
2448 if not content_type:
2450 ctr = re.search(r'(?P<mimetype>[^/]+/[^;]+)(?:;\s*codecs="?(?P<codecs>[^"]+))?', content_type)
2452 mimetype, codecs = ctr.groups()
2453 f = parse_codecs(codecs)
2454 f['ext'] = mimetype2ext(mimetype)
2458 def _media_formats(src, cur_media_type, type_info={}):
2459 full_url = absolute_url(src)
2460 ext = type_info.get('ext') or determine_ext(full_url)
2462 is_plain_url = False
2463 formats = self._extract_m3u8_formats(
2464 full_url, video_id, ext='mp4',
2465 entry_protocol=m3u8_entry_protocol, m3u8_id=m3u8_id,
2466 preference=preference, fatal=False)
2468 is_plain_url = False
2469 formats = self._extract_mpd_formats(
2470 full_url, video_id, mpd_id=mpd_id, fatal=False)
2475 'vcodec': 'none' if cur_media_type == 'audio' else None,
2477 return is_plain_url, formats
2480 # amp-video and amp-audio are very similar to their HTML5 counterparts
2481 # so we wll include them right here (see
2482 # https://www.ampproject.org/docs/reference/components/amp-video)
2483 media_tags = [(media_tag, media_type, '')
2484 for media_tag, media_type
2485 in re.findall(r'(?s)(<(?:amp-)?(video|audio)[^>]*/>)', webpage)]
2486 media_tags.extend(re.findall(
2487 # We only allow video|audio followed by a whitespace or '>'.
2488 # Allowing more characters may end up in significant slow down (see
2489 # https://github.com/ytdl-org/youtube-dl/issues/11979, example URL:
2490 # http://www.porntrex.com/maps/videositemap.xml).
2491 r'(?s)(<(?P<tag>(?:amp-)?(?:video|audio))(?:\s+[^>]*)?>)(.*?)</(?P=tag)>', webpage))
2492 for media_tag, media_type, media_content in media_tags:
2497 media_attributes = extract_attributes(media_tag)
2498 src = strip_or_none(media_attributes.get('src'))
2500 _, formats = _media_formats(src, media_type)
2501 media_info['formats'].extend(formats)
2502 media_info['thumbnail'] = absolute_url(media_attributes.get('poster'))
2504 for source_tag in re.findall(r'<source[^>]+>', media_content):
2505 s_attr = extract_attributes(source_tag)
2506 # data-video-src and data-src are non standard but seen
2507 # several times in the wild
2508 src = strip_or_none(dict_get(s_attr, ('src', 'data-video-src', 'data-src')))
2511 f = parse_content_type(s_attr.get('type'))
2512 is_plain_url, formats = _media_formats(src, media_type, f)
2514 # width, height, res, label and title attributes are
2515 # all not standard but seen several times in the wild
2518 for lbl in ('label', 'title')
2519 if str_or_none(s_attr.get(lbl))
2521 width = int_or_none(s_attr.get('width'))
2522 height = (int_or_none(s_attr.get('height'))
2523 or int_or_none(s_attr.get('res')))
2524 if not width or not height:
2526 resolution = parse_resolution(lbl)
2529 width = width or resolution.get('width')
2530 height = height or resolution.get('height')
2532 tbr = parse_bitrate(lbl)
2541 'format_id': s_attr.get('label') or s_attr.get('title'),
2543 f.update(formats[0])
2544 media_info['formats'].append(f)
2546 media_info['formats'].extend(formats)
2547 for track_tag in re.findall(r'<track[^>]+>', media_content):
2548 track_attributes = extract_attributes(track_tag)
2549 kind = track_attributes.get('kind')
2550 if not kind or kind in ('subtitles', 'captions'):
2551 src = strip_or_none(track_attributes.get('src'))
2554 lang = track_attributes.get('srclang') or track_attributes.get('lang') or track_attributes.get('label')
2555 media_info['subtitles'].setdefault(lang, []).append({
2556 'url': absolute_url(src),
2558 for f in media_info['formats']:
2559 f.setdefault('http_headers', {})['Referer'] = base_url
2560 if media_info['formats'] or media_info['subtitles']:
2561 entries.append(media_info)
2564 def _extract_akamai_formats(self, manifest_url, video_id, hosts={}):
2566 hdcore_sign = 'hdcore=3.7.0'
2567 f4m_url = re.sub(r'(https?://[^/]+)/i/', r'\1/z/', manifest_url).replace('/master.m3u8', '/manifest.f4m')
2568 hds_host = hosts.get('hds')
2570 f4m_url = re.sub(r'(https?://)[^/]+', r'\1' + hds_host, f4m_url)
2571 if 'hdcore=' not in f4m_url:
2572 f4m_url += ('&' if '?' in f4m_url else '?') + hdcore_sign
2573 f4m_formats = self._extract_f4m_formats(
2574 f4m_url, video_id, f4m_id='hds', fatal=False)
2575 for entry in f4m_formats:
2576 entry.update({'extra_param_to_segment_url': hdcore_sign})
2577 formats.extend(f4m_formats)
2578 m3u8_url = re.sub(r'(https?://[^/]+)/z/', r'\1/i/', manifest_url).replace('/manifest.f4m', '/master.m3u8')
2579 hls_host = hosts.get('hls')
2581 m3u8_url = re.sub(r'(https?://)[^/]+', r'\1' + hls_host, m3u8_url)
2582 formats.extend(self._extract_m3u8_formats(
2583 m3u8_url, video_id, 'mp4', 'm3u8_native',
2584 m3u8_id='hls', fatal=False))
2587 def _extract_wowza_formats(self, url, video_id, m3u8_entry_protocol='m3u8_native', skip_protocols=[]):
2588 query = compat_urlparse.urlparse(url).query
2589 url = re.sub(r'/(?:manifest|playlist|jwplayer)\.(?:m3u8|f4m|mpd|smil)', '', url)
2591 r'(?:(?:http|rtmp|rtsp)(?P<s>s)?:)?(?P<url>//[^?]+)', url)
2592 url_base = mobj.group('url')
2593 http_base_url = '%s%s:%s' % ('http', mobj.group('s') or '', url_base)
2596 def manifest_url(manifest):
2597 m_url = '%s/%s' % (http_base_url, manifest)
2599 m_url += '?%s' % query
2602 if 'm3u8' not in skip_protocols:
2603 formats.extend(self._extract_m3u8_formats(
2604 manifest_url('playlist.m3u8'), video_id, 'mp4',
2605 m3u8_entry_protocol, m3u8_id='hls', fatal=False))
2606 if 'f4m' not in skip_protocols:
2607 formats.extend(self._extract_f4m_formats(
2608 manifest_url('manifest.f4m'),
2609 video_id, f4m_id='hds', fatal=False))
2610 if 'dash' not in skip_protocols:
2611 formats.extend(self._extract_mpd_formats(
2612 manifest_url('manifest.mpd'),
2613 video_id, mpd_id='dash', fatal=False))
2614 if re.search(r'(?:/smil:|\.smil)', url_base):
2615 if 'smil' not in skip_protocols:
2616 rtmp_formats = self._extract_smil_formats(
2617 manifest_url('jwplayer.smil'),
2618 video_id, fatal=False)
2619 for rtmp_format in rtmp_formats:
2620 rtsp_format = rtmp_format.copy()
2621 rtsp_format['url'] = '%s/%s' % (rtmp_format['url'], rtmp_format['play_path'])
2622 del rtsp_format['play_path']
2623 del rtsp_format['ext']
2624 rtsp_format.update({
2625 'url': rtsp_format['url'].replace('rtmp://', 'rtsp://'),
2626 'format_id': rtmp_format['format_id'].replace('rtmp', 'rtsp'),
2629 formats.extend([rtmp_format, rtsp_format])
2631 for protocol in ('rtmp', 'rtsp'):
2632 if protocol not in skip_protocols:
2634 'url': '%s:%s' % (protocol, url_base),
2635 'format_id': protocol,
2636 'protocol': protocol,
2640 def _find_jwplayer_data(self, webpage, video_id=None, transform_source=js_to_json):
2642 r'(?s)jwplayer\((?P<quote>[\'"])[^\'" ]+(?P=quote)\)(?!</script>).*?\.setup\s*\((?P<options>[^)]+)\)',
2646 jwplayer_data = self._parse_json(mobj.group('options'),
2648 transform_source=transform_source)
2649 except ExtractorError:
2652 if isinstance(jwplayer_data, dict):
2653 return jwplayer_data
2655 def _extract_jwplayer_data(self, webpage, video_id, *args, **kwargs):
2656 jwplayer_data = self._find_jwplayer_data(
2657 webpage, video_id, transform_source=js_to_json)
2658 return self._parse_jwplayer_data(
2659 jwplayer_data, video_id, *args, **kwargs)
2661 def _parse_jwplayer_data(self, jwplayer_data, video_id=None, require_title=True,
2662 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
2663 # JWPlayer backward compatibility: flattened playlists
2664 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/api/config.js#L81-L96
2665 if 'playlist' not in jwplayer_data:
2666 jwplayer_data = {'playlist': [jwplayer_data]}
2670 # JWPlayer backward compatibility: single playlist item
2671 # https://github.com/jwplayer/jwplayer/blob/v7.7.0/src/js/playlist/playlist.js#L10
2672 if not isinstance(jwplayer_data['playlist'], list):
2673 jwplayer_data['playlist'] = [jwplayer_data['playlist']]
2675 for video_data in jwplayer_data['playlist']:
2676 # JWPlayer backward compatibility: flattened sources
2677 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/playlist/item.js#L29-L35
2678 if 'sources' not in video_data:
2679 video_data['sources'] = [video_data]
2681 this_video_id = video_id or video_data['mediaid']
2683 formats = self._parse_jwplayer_formats(
2684 video_data['sources'], video_id=this_video_id, m3u8_id=m3u8_id,
2685 mpd_id=mpd_id, rtmp_params=rtmp_params, base_url=base_url)
2688 tracks = video_data.get('tracks')
2689 if tracks and isinstance(tracks, list):
2690 for track in tracks:
2691 if not isinstance(track, dict):
2693 track_kind = track.get('kind')
2694 if not track_kind or not isinstance(track_kind, compat_str):
2696 if track_kind.lower() not in ('captions', 'subtitles'):
2698 track_url = urljoin(base_url, track.get('file'))
2701 subtitles.setdefault(track.get('label') or 'en', []).append({
2702 'url': self._proto_relative_url(track_url)
2706 'id': this_video_id,
2707 'title': unescapeHTML(video_data['title'] if require_title else video_data.get('title')),
2708 'description': clean_html(video_data.get('description')),
2709 'thumbnail': urljoin(base_url, self._proto_relative_url(video_data.get('image'))),
2710 'timestamp': int_or_none(video_data.get('pubdate')),
2711 'duration': float_or_none(jwplayer_data.get('duration') or video_data.get('duration')),
2712 'subtitles': subtitles,
2714 # https://github.com/jwplayer/jwplayer/blob/master/src/js/utils/validator.js#L32
2715 if len(formats) == 1 and re.search(r'^(?:http|//).*(?:youtube\.com|youtu\.be)/.+', formats[0]['url']):
2717 '_type': 'url_transparent',
2718 'url': formats[0]['url'],
2721 self._sort_formats(formats)
2722 entry['formats'] = formats
2723 entries.append(entry)
2724 if len(entries) == 1:
2727 return self.playlist_result(entries)
2729 def _parse_jwplayer_formats(self, jwplayer_sources_data, video_id=None,
2730 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
2733 for source in jwplayer_sources_data:
2734 if not isinstance(source, dict):
2736 source_url = urljoin(
2737 base_url, self._proto_relative_url(source.get('file')))
2738 if not source_url or source_url in urls:
2740 urls.append(source_url)
2741 source_type = source.get('type') or ''
2742 ext = mimetype2ext(source_type) or determine_ext(source_url)
2743 if source_type == 'hls' or ext == 'm3u8':
2744 formats.extend(self._extract_m3u8_formats(
2745 source_url, video_id, 'mp4', entry_protocol='m3u8_native',
2746 m3u8_id=m3u8_id, fatal=False))
2747 elif source_type == 'dash' or ext == 'mpd':
2748 formats.extend(self._extract_mpd_formats(
2749 source_url, video_id, mpd_id=mpd_id, fatal=False))
2751 formats.extend(self._extract_smil_formats(
2752 source_url, video_id, fatal=False))
2753 # https://github.com/jwplayer/jwplayer/blob/master/src/js/providers/default.js#L67
2754 elif source_type.startswith('audio') or ext in (
2755 'oga', 'aac', 'mp3', 'mpeg', 'vorbis'):
2762 height = int_or_none(source.get('height'))
2764 # Often no height is provided but there is a label in
2765 # format like "1080p", "720p SD", or 1080.
2766 height = int_or_none(self._search_regex(
2767 r'^(\d{3,4})[pP]?(?:\b|$)', compat_str(source.get('label') or ''),
2768 'height', default=None))
2771 'width': int_or_none(source.get('width')),
2773 'tbr': int_or_none(source.get('bitrate')),
2776 if source_url.startswith('rtmp'):
2777 a_format['ext'] = 'flv'
2778 # See com/longtailvideo/jwplayer/media/RTMPMediaProvider.as
2779 # of jwplayer.flash.swf
2780 rtmp_url_parts = re.split(
2781 r'((?:mp4|mp3|flv):)', source_url, 1)
2782 if len(rtmp_url_parts) == 3:
2783 rtmp_url, prefix, play_path = rtmp_url_parts
2786 'play_path': prefix + play_path,
2789 a_format.update(rtmp_params)
2790 formats.append(a_format)
2793 def _live_title(self, name):
2794 """ Generate the title for a live video """
2795 now = datetime.datetime.now()
2796 now_str = now.strftime('%Y-%m-%d %H:%M')
2797 return name + ' ' + now_str
2799 def _int(self, v, name, fatal=False, **kwargs):
2800 res = int_or_none(v, **kwargs)
2801 if 'get_attr' in kwargs:
2802 print(getattr(v, kwargs['get_attr']))
2804 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
2806 raise ExtractorError(msg)
2808 self._downloader.report_warning(msg)
2811 def _float(self, v, name, fatal=False, **kwargs):
2812 res = float_or_none(v, **kwargs)
2814 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
2816 raise ExtractorError(msg)
2818 self._downloader.report_warning(msg)
2821 def _set_cookie(self, domain, name, value, expire_time=None, port=None,
2822 path='/', secure=False, discard=False, rest={}, **kwargs):
2823 cookie = compat_cookiejar.Cookie(
2824 0, name, value, port, port is not None, domain, True,
2825 domain.startswith('.'), path, True, secure, expire_time,
2826 discard, None, None, rest)
2827 self._downloader.cookiejar.set_cookie(cookie)
2829 def _get_cookies(self, url):
2830 """ Return a compat_cookies.SimpleCookie with the cookies for the url """
2831 req = sanitized_Request(url)
2832 self._downloader.cookiejar.add_cookie_header(req)
2833 return compat_cookies.SimpleCookie(req.get_header('Cookie'))
2835 def _apply_first_set_cookie_header(self, url_handle, cookie):
2837 Apply first Set-Cookie header instead of the last. Experimental.
2839 Some sites (e.g. [1-3]) may serve two cookies under the same name
2840 in Set-Cookie header and expect the first (old) one to be set rather
2841 than second (new). However, as of RFC6265 the newer one cookie
2842 should be set into cookie store what actually happens.
2843 We will workaround this issue by resetting the cookie to
2844 the first one manually.
2845 1. https://new.vk.com/
2846 2. https://github.com/ytdl-org/youtube-dl/issues/9841#issuecomment-227871201
2847 3. https://learning.oreilly.com/
2849 for header, cookies in url_handle.headers.items():
2850 if header.lower() != 'set-cookie':
2852 if sys.version_info[0] >= 3:
2853 cookies = cookies.encode('iso-8859-1')
2854 cookies = cookies.decode('utf-8')
2855 cookie_value = re.search(
2856 r'%s=(.+?);.*?\b[Dd]omain=(.+?)(?:[,;]|$)' % cookie, cookies)
2858 value, domain = cookie_value.groups()
2859 self._set_cookie(domain, cookie, value)
2862 def get_testcases(self, include_onlymatching=False):
2863 t = getattr(self, '_TEST', None)
2865 assert not hasattr(self, '_TESTS'), \
2866 '%s has _TEST and _TESTS' % type(self).__name__
2869 tests = getattr(self, '_TESTS', [])
2871 if not include_onlymatching and t.get('only_matching', False):
2873 t['name'] = type(self).__name__[:-len('IE')]
2876 def is_suitable(self, age_limit):
2877 """ Test whether the extractor is generally suitable for the given
2878 age limit (i.e. pornographic sites are not, all others usually are) """
2880 any_restricted = False
2881 for tc in self.get_testcases(include_onlymatching=False):
2882 if tc.get('playlist', []):
2883 tc = tc['playlist'][0]
2884 is_restricted = age_restricted(
2885 tc.get('info_dict', {}).get('age_limit'), age_limit)
2886 if not is_restricted:
2888 any_restricted = any_restricted or is_restricted
2889 return not any_restricted
2891 def extract_subtitles(self, *args, **kwargs):
2892 if (self._downloader.params.get('writesubtitles', False)
2893 or self._downloader.params.get('listsubtitles')):
2894 return self._get_subtitles(*args, **kwargs)
2897 def _get_subtitles(self, *args, **kwargs):
2898 raise NotImplementedError('This method must be implemented by subclasses')
2901 def _merge_subtitle_items(subtitle_list1, subtitle_list2):
2902 """ Merge subtitle items for one language. Items with duplicated URLs
2903 will be dropped. """
2904 list1_urls = set([item['url'] for item in subtitle_list1])
2905 ret = list(subtitle_list1)
2906 ret.extend([item for item in subtitle_list2 if item['url'] not in list1_urls])
2910 def _merge_subtitles(cls, subtitle_dict1, subtitle_dict2):
2911 """ Merge two subtitle dictionaries, language by language. """
2912 ret = dict(subtitle_dict1)
2913 for lang in subtitle_dict2:
2914 ret[lang] = cls._merge_subtitle_items(subtitle_dict1.get(lang, []), subtitle_dict2[lang])
2917 def extract_automatic_captions(self, *args, **kwargs):
2918 if (self._downloader.params.get('writeautomaticsub', False)
2919 or self._downloader.params.get('listsubtitles')):
2920 return self._get_automatic_captions(*args, **kwargs)
2923 def _get_automatic_captions(self, *args, **kwargs):
2924 raise NotImplementedError('This method must be implemented by subclasses')
2926 def mark_watched(self, *args, **kwargs):
2927 if (self._downloader.params.get('mark_watched', False)
2928 and (self._get_login_info()[0] is not None
2929 or self._downloader.params.get('cookiefile') is not None)):
2930 self._mark_watched(*args, **kwargs)
2932 def _mark_watched(self, *args, **kwargs):
2933 raise NotImplementedError('This method must be implemented by subclasses')
2935 def geo_verification_headers(self):
2937 geo_verification_proxy = self._downloader.params.get('geo_verification_proxy')
2938 if geo_verification_proxy:
2939 headers['Ytdl-request-proxy'] = geo_verification_proxy
2942 def _generic_id(self, url):
2943 return compat_urllib_parse_unquote(os.path.splitext(url.rstrip('/').split('/')[-1])[0])
2945 def _generic_title(self, url):
2946 return compat_urllib_parse_unquote(os.path.splitext(url_basename(url))[0])
2949 class SearchInfoExtractor(InfoExtractor):
2951 Base class for paged search queries extractors.
2952 They accept URLs in the format _SEARCH_KEY(|all|[0-9]):{query}
2953 Instances should define _SEARCH_KEY and _MAX_RESULTS.
2957 def _make_valid_url(cls):
2958 return r'%s(?P<prefix>|[1-9][0-9]*|all):(?P<query>[\s\S]+)' % cls._SEARCH_KEY
2961 def suitable(cls, url):
2962 return re.match(cls._make_valid_url(), url) is not None
2964 def _real_extract(self, query):
2965 mobj = re.match(self._make_valid_url(), query)
2967 raise ExtractorError('Invalid search query "%s"' % query)
2969 prefix = mobj.group('prefix')
2970 query = mobj.group('query')
2972 return self._get_n_results(query, 1)
2973 elif prefix == 'all':
2974 return self._get_n_results(query, self._MAX_RESULTS)
2978 raise ExtractorError('invalid download number %s for query "%s"' % (n, query))
2979 elif n > self._MAX_RESULTS:
2980 self._downloader.report_warning('%s returns max %i results (you requested %i)' % (self._SEARCH_KEY, self._MAX_RESULTS, n))
2981 n = self._MAX_RESULTS
2982 return self._get_n_results(query, n)
2984 def _get_n_results(self, query, n):
2985 """Get a specified number of results for a query"""
2986 raise NotImplementedError('This method must be implemented by subclasses')
2989 def SEARCH_KEY(self):
2990 return self._SEARCH_KEY