2 from __future__ import unicode_literals
17 from ..compat import (
20 compat_etree_fromstring,
27 compat_urllib_parse_unquote,
28 compat_urllib_parse_urlencode,
29 compat_urllib_request,
31 compat_xml_parse_error,
33 from ..downloader.f4m import (
35 remove_encrypted_media,
61 parse_m3u8_attributes,
78 class InfoExtractor(object):
79 """Information Extractor class.
81 Information extractors are the classes that, given a URL, extract
82 information about the video (or videos) the URL refers to. This
83 information includes the real video URL, the video title, author and
84 others. The information is stored in a dictionary which is then
85 passed to the YoutubeDL. The YoutubeDL processes this
86 information possibly downloading the video to the file system, among
87 other possible outcomes.
89 The type field determines the type of the result.
90 By far the most common value (and the default if _type is missing) is
91 "video", which indicates a single video.
93 For a video, the dictionaries must include the following fields:
96 title: Video title, unescaped.
98 Additionally, it must contain either a formats entry or a url one:
100 formats: A list of dictionaries for each format available, ordered
101 from worst to best quality.
104 * url Mandatory. The URL of the video file
106 The URL of the manifest file in case of
107 fragmented media (DASH, hls, hds)
108 * ext Will be calculated from URL if missing
109 * format A human-readable description of the format
110 ("mp4 container with h264/opus").
111 Calculated from the format_id, width, height.
112 and format_note fields if missing.
113 * format_id A short description of the format
114 ("mp4_h264_opus" or "19").
115 Technically optional, but strongly recommended.
116 * format_note Additional info about the format
117 ("3D" or "DASH video")
118 * width Width of the video, if known
119 * height Height of the video, if known
120 * resolution Textual description of width and height
121 * tbr Average bitrate of audio and video in KBit/s
122 * abr Average audio bitrate in KBit/s
123 * acodec Name of the audio codec in use
124 * asr Audio sampling rate in Hertz
125 * vbr Average video bitrate in KBit/s
127 * vcodec Name of the video codec in use
128 * container Name of the container format
129 * filesize The number of bytes, if known in advance
130 * filesize_approx An estimate for the number of bytes
131 * player_url SWF Player URL (used for rtmpdump).
132 * protocol The protocol that will be used for the actual
133 download, lower-case.
134 "http", "https", "rtsp", "rtmp", "rtmpe",
135 "m3u8", "m3u8_native" or "http_dash_segments".
137 Base URL for fragments. Each fragment's path
138 value (if present) will be relative to
140 * fragments A list of fragments of a fragmented media.
141 Each fragment entry must contain either an url
142 or a path. If an url is present it should be
143 considered by a client. Otherwise both path and
144 fragment_base_url must be present. Here is
145 the list of all potential fields:
146 * "url" - fragment's URL
147 * "path" - fragment's path relative to
149 * "duration" (optional, int or float)
150 * "filesize" (optional, int)
151 * preference Order number of this format. If this field is
152 present and not None, the formats get sorted
153 by this field, regardless of all other values.
154 -1 for default (order by other properties),
155 -2 or smaller for less than default.
156 < -1000 to hide the format (if there is
157 another one which is strictly better)
158 * language Language code, e.g. "de" or "en-US".
159 * language_preference Is this in the language mentioned in
161 10 if it's what the URL is about,
162 -1 for default (don't know),
163 -10 otherwise, other values reserved for now.
164 * quality Order number of the video quality of this
165 format, irrespective of the file format.
166 -1 for default (order by other properties),
167 -2 or smaller for less than default.
168 * source_preference Order number for this video source
169 (quality takes higher priority)
170 -1 for default (order by other properties),
171 -2 or smaller for less than default.
172 * http_headers A dictionary of additional HTTP headers
173 to add to the request.
174 * stretched_ratio If given and not 1, indicates that the
175 video's pixels are not square.
176 width : height ratio as float.
177 * no_resume The server does not support resuming the
178 (HTTP or RTMP) download. Boolean.
179 * downloader_options A dictionary of downloader options as
180 described in FileDownloader
182 url: Final video URL.
183 ext: Video filename extension.
184 format: The video format, defaults to ext (used for --get-format)
185 player_url: SWF Player URL (used for rtmpdump).
187 The following fields are optional:
189 alt_title: A secondary title of the video.
190 display_id An alternative identifier for the video, not necessarily
191 unique, but available before title. Typically, id is
192 something like "4234987", title "Dancing naked mole rats",
193 and display_id "dancing-naked-mole-rats"
194 thumbnails: A list of dictionaries, with the following entries:
195 * "id" (optional, string) - Thumbnail format ID
197 * "preference" (optional, int) - quality of the image
198 * "width" (optional, int)
199 * "height" (optional, int)
200 * "resolution" (optional, string "{width}x{height"},
202 * "filesize" (optional, int)
203 thumbnail: Full URL to a video thumbnail image.
204 description: Full video description.
205 uploader: Full name of the video uploader.
206 license: License name the video is licensed under.
207 creator: The creator of the video.
208 release_date: The date (YYYYMMDD) when the video was released.
209 timestamp: UNIX timestamp of the moment the video became available.
210 upload_date: Video upload date (YYYYMMDD).
211 If not explicitly set, calculated from timestamp.
212 uploader_id: Nickname or id of the video uploader.
213 uploader_url: Full URL to a personal webpage of the video uploader.
214 channel: Full name of the channel the video is uploaded on.
215 Note that channel fields may or may not repeat uploader
216 fields. This depends on a particular extractor.
217 channel_id: Id of the channel.
218 channel_url: Full URL to a channel webpage.
219 location: Physical location where the video was filmed.
220 subtitles: The available subtitles as a dictionary in the format
221 {tag: subformats}. "tag" is usually a language code, and
222 "subformats" is a list sorted from lower to higher
223 preference, each element is a dictionary with the "ext"
225 * "data": The subtitles file contents
226 * "url": A URL pointing to the subtitles file
227 "ext" will be calculated from URL if missing
228 automatic_captions: Like 'subtitles', used by the YoutubeIE for
229 automatically generated captions
230 duration: Length of the video in seconds, as an integer or float.
231 view_count: How many users have watched the video on the platform.
232 like_count: Number of positive ratings of the video
233 dislike_count: Number of negative ratings of the video
234 repost_count: Number of reposts of the video
235 average_rating: Average rating give by users, the scale used depends on the webpage
236 comment_count: Number of comments on the video
237 comments: A list of comments, each with one or more of the following
238 properties (all but one of text or html optional):
239 * "author" - human-readable name of the comment author
240 * "author_id" - user ID of the comment author
242 * "html" - Comment as HTML
243 * "text" - Plain text of the comment
244 * "timestamp" - UNIX timestamp of comment
245 * "parent" - ID of the comment this one is replying to.
246 Set to "root" to indicate that this is a
247 comment to the original video.
248 age_limit: Age restriction for the video, as an integer (years)
249 webpage_url: The URL to the video webpage, if given to youtube-dl it
250 should allow to get the same result again. (It will be set
251 by YoutubeDL if it's missing)
252 categories: A list of categories that the video falls in, for example
254 tags: A list of tags assigned to the video, e.g. ["sweden", "pop music"]
255 is_live: True, False, or None (=unknown). Whether this video is a
256 live stream that goes on instead of a fixed-length video.
257 start_time: Time in seconds where the reproduction should start, as
258 specified in the URL.
259 end_time: Time in seconds where the reproduction should end, as
260 specified in the URL.
261 chapters: A list of dictionaries, with the following entries:
262 * "start_time" - The start time of the chapter in seconds
263 * "end_time" - The end time of the chapter in seconds
264 * "title" (optional, string)
266 The following fields should only be used when the video belongs to some logical
269 chapter: Name or title of the chapter the video belongs to.
270 chapter_number: Number of the chapter the video belongs to, as an integer.
271 chapter_id: Id of the chapter the video belongs to, as a unicode string.
273 The following fields should only be used when the video is an episode of some
274 series, programme or podcast:
276 series: Title of the series or programme the video episode belongs to.
277 season: Title of the season the video episode belongs to.
278 season_number: Number of the season the video episode belongs to, as an integer.
279 season_id: Id of the season the video episode belongs to, as a unicode string.
280 episode: Title of the video episode. Unlike mandatory video title field,
281 this field should denote the exact title of the video episode
282 without any kind of decoration.
283 episode_number: Number of the video episode within a season, as an integer.
284 episode_id: Id of the video episode, as a unicode string.
286 The following fields should only be used when the media is a track or a part of
289 track: Title of the track.
290 track_number: Number of the track within an album or a disc, as an integer.
291 track_id: Id of the track (useful in case of custom indexing, e.g. 6.iii),
293 artist: Artist(s) of the track.
294 genre: Genre(s) of the track.
295 album: Title of the album the track belongs to.
296 album_type: Type of the album (e.g. "Demo", "Full-length", "Split", "Compilation", etc).
297 album_artist: List of all artists appeared on the album (e.g.
298 "Ash Borer / Fell Voices" or "Various Artists", useful for splits
300 disc_number: Number of the disc or other physical medium the track belongs to,
302 release_year: Year (YYYY) when the album was released.
304 Unless mentioned otherwise, the fields should be Unicode strings.
306 Unless mentioned otherwise, None is equivalent to absence of information.
309 _type "playlist" indicates multiple videos.
310 There must be a key "entries", which is a list, an iterable, or a PagedList
311 object, each element of which is a valid dictionary by this specification.
313 Additionally, playlists can have "id", "title", "description", "uploader",
314 "uploader_id", "uploader_url" attributes with the same semantics as videos
318 _type "multi_video" indicates that there are multiple videos that
319 form a single show, for examples multiple acts of an opera or TV episode.
320 It must have an entries key like a playlist and contain all the keys
321 required for a video at the same time.
324 _type "url" indicates that the video must be extracted from another
325 location, possibly by a different extractor. Its only required key is:
326 "url" - the next URL to extract.
327 The key "ie_key" can be set to the class name (minus the trailing "IE",
328 e.g. "Youtube") if the extractor class is known in advance.
329 Additionally, the dictionary may have any properties of the resolved entity
330 known in advance, for example "title" if the title of the referred video is
334 _type "url_transparent" entities have the same specification as "url", but
335 indicate that the given additional information is more precise than the one
336 associated with the resolved URL.
337 This is useful when a site employs a video service that hosts the video and
338 its technical metadata, but that video service does not embed a useful
339 title, description etc.
342 Subclasses of this one should re-define the _real_initialize() and
343 _real_extract() methods and define a _VALID_URL regexp.
344 Probably, they should also be added to the list of extractors.
346 _GEO_BYPASS attribute may be set to False in order to disable
347 geo restriction bypass mechanisms for a particular extractor.
348 Though it won't disable explicit geo restriction bypass based on
349 country code provided with geo_bypass_country.
351 _GEO_COUNTRIES attribute may contain a list of presumably geo unrestricted
352 countries for this extractor. One of these countries will be used by
353 geo restriction bypass mechanism right away in order to bypass
354 geo restriction, of course, if the mechanism is not disabled.
356 _GEO_IP_BLOCKS attribute may contain a list of presumably geo unrestricted
357 IP blocks in CIDR notation for this extractor. One of these IP blocks
358 will be used by geo restriction bypass mechanism similarly
361 Finally, the _WORKING attribute should be set to False for broken IEs
362 in order to warn the users and skip the tests.
367 _x_forwarded_for_ip = None
369 _GEO_COUNTRIES = None
370 _GEO_IP_BLOCKS = None
373 def __init__(self, downloader=None):
374 """Constructor. Receives an optional downloader."""
376 self._x_forwarded_for_ip = None
377 self.set_downloader(downloader)
380 def suitable(cls, url):
381 """Receives a URL and returns True if suitable for this IE."""
383 # This does not use has/getattr intentionally - we want to know whether
384 # we have cached the regexp for *this* class, whereas getattr would also
385 # match the superclass
386 if '_VALID_URL_RE' not in cls.__dict__:
387 cls._VALID_URL_RE = re.compile(cls._VALID_URL)
388 return cls._VALID_URL_RE.match(url) is not None
391 def _match_id(cls, url):
392 if '_VALID_URL_RE' not in cls.__dict__:
393 cls._VALID_URL_RE = re.compile(cls._VALID_URL)
394 m = cls._VALID_URL_RE.match(url)
396 return compat_str(m.group('id'))
400 """Getter method for _WORKING."""
403 def initialize(self):
404 """Initializes an instance (authentication, etc)."""
405 self._initialize_geo_bypass({
406 'countries': self._GEO_COUNTRIES,
407 'ip_blocks': self._GEO_IP_BLOCKS,
410 self._real_initialize()
413 def _initialize_geo_bypass(self, geo_bypass_context):
415 Initialize geo restriction bypass mechanism.
417 This method is used to initialize geo bypass mechanism based on faking
418 X-Forwarded-For HTTP header. A random country from provided country list
419 is selected and a random IP belonging to this country is generated. This
420 IP will be passed as X-Forwarded-For HTTP header in all subsequent
423 This method will be used for initial geo bypass mechanism initialization
424 during the instance initialization with _GEO_COUNTRIES and
427 You may also manually call it from extractor's code if geo bypass
428 information is not available beforehand (e.g. obtained during
429 extraction) or due to some other reason. In this case you should pass
430 this information in geo bypass context passed as first argument. It may
431 contain following fields:
433 countries: List of geo unrestricted countries (similar
435 ip_blocks: List of geo unrestricted IP blocks in CIDR notation
436 (similar to _GEO_IP_BLOCKS)
439 if not self._x_forwarded_for_ip:
441 # Geo bypass mechanism is explicitly disabled by user
442 if not self._downloader.params.get('geo_bypass', True):
445 if not geo_bypass_context:
446 geo_bypass_context = {}
448 # Backward compatibility: previously _initialize_geo_bypass
449 # expected a list of countries, some 3rd party code may still use
451 if isinstance(geo_bypass_context, (list, tuple)):
452 geo_bypass_context = {
453 'countries': geo_bypass_context,
456 # The whole point of geo bypass mechanism is to fake IP
457 # as X-Forwarded-For HTTP header based on some IP block or
460 # Path 1: bypassing based on IP block in CIDR notation
462 # Explicit IP block specified by user, use it right away
463 # regardless of whether extractor is geo bypassable or not
464 ip_block = self._downloader.params.get('geo_bypass_ip_block', None)
466 # Otherwise use random IP block from geo bypass context but only
467 # if extractor is known as geo bypassable
469 ip_blocks = geo_bypass_context.get('ip_blocks')
470 if self._GEO_BYPASS and ip_blocks:
471 ip_block = random.choice(ip_blocks)
474 self._x_forwarded_for_ip = GeoUtils.random_ipv4(ip_block)
475 if self._downloader.params.get('verbose', False):
476 self._downloader.to_screen(
477 '[debug] Using fake IP %s as X-Forwarded-For.'
478 % self._x_forwarded_for_ip)
481 # Path 2: bypassing based on country code
483 # Explicit country code specified by user, use it right away
484 # regardless of whether extractor is geo bypassable or not
485 country = self._downloader.params.get('geo_bypass_country', None)
487 # Otherwise use random country code from geo bypass context but
488 # only if extractor is known as geo bypassable
490 countries = geo_bypass_context.get('countries')
491 if self._GEO_BYPASS and countries:
492 country = random.choice(countries)
495 self._x_forwarded_for_ip = GeoUtils.random_ipv4(country)
496 if self._downloader.params.get('verbose', False):
497 self._downloader.to_screen(
498 '[debug] Using fake IP %s (%s) as X-Forwarded-For.'
499 % (self._x_forwarded_for_ip, country.upper()))
501 def extract(self, url):
502 """Extracts URL information and returns it in list of dicts."""
507 ie_result = self._real_extract(url)
508 if self._x_forwarded_for_ip:
509 ie_result['__x_forwarded_for_ip'] = self._x_forwarded_for_ip
511 except GeoRestrictedError as e:
512 if self.__maybe_fake_ip_and_retry(e.countries):
515 except ExtractorError:
517 except compat_http_client.IncompleteRead as e:
518 raise ExtractorError('A network error has occurred.', cause=e, expected=True)
519 except (KeyError, StopIteration) as e:
520 raise ExtractorError('An extractor error has occurred.', cause=e)
522 def __maybe_fake_ip_and_retry(self, countries):
523 if (not self._downloader.params.get('geo_bypass_country', None) and
525 self._downloader.params.get('geo_bypass', True) and
526 not self._x_forwarded_for_ip and
528 country_code = random.choice(countries)
529 self._x_forwarded_for_ip = GeoUtils.random_ipv4(country_code)
530 if self._x_forwarded_for_ip:
532 'Video is geo restricted. Retrying extraction with fake IP %s (%s) as X-Forwarded-For.'
533 % (self._x_forwarded_for_ip, country_code.upper()))
537 def set_downloader(self, downloader):
538 """Sets the downloader for this IE."""
539 self._downloader = downloader
541 def _real_initialize(self):
542 """Real initialization process. Redefine in subclasses."""
545 def _real_extract(self, url):
546 """Real extraction process. Redefine in subclasses."""
551 """A string for getting the InfoExtractor with get_info_extractor"""
552 return compat_str(cls.__name__[:-2])
556 return compat_str(type(self).__name__[:-2])
559 def __can_accept_status_code(err, expected_status):
560 assert isinstance(err, compat_urllib_error.HTTPError)
561 if expected_status is None:
563 if isinstance(expected_status, compat_integer_types):
564 return err.code == expected_status
565 elif isinstance(expected_status, (list, tuple)):
566 return err.code in expected_status
567 elif callable(expected_status):
568 return expected_status(err.code) is True
572 def _request_webpage(self, url_or_request, video_id, note=None, errnote=None, fatal=True, data=None, headers={}, query={}, expected_status=None):
574 Return the response handle.
576 See _download_webpage docstring for arguments specification.
579 self.report_download_webpage(video_id)
580 elif note is not False:
582 self.to_screen('%s' % (note,))
584 self.to_screen('%s: %s' % (video_id, note))
586 # Some sites check X-Forwarded-For HTTP header in order to figure out
587 # the origin of the client behind proxy. This allows bypassing geo
588 # restriction by faking this header's value to IP that belongs to some
589 # geo unrestricted country. We will do so once we encounter any
590 # geo restriction error.
591 if self._x_forwarded_for_ip:
592 if 'X-Forwarded-For' not in headers:
593 headers['X-Forwarded-For'] = self._x_forwarded_for_ip
595 if isinstance(url_or_request, compat_urllib_request.Request):
596 url_or_request = update_Request(
597 url_or_request, data=data, headers=headers, query=query)
600 url_or_request = update_url_query(url_or_request, query)
601 if data is not None or headers:
602 url_or_request = sanitized_Request(url_or_request, data, headers)
604 return self._downloader.urlopen(url_or_request)
605 except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err:
606 if isinstance(err, compat_urllib_error.HTTPError):
607 if self.__can_accept_status_code(err, expected_status):
613 errnote = 'Unable to download webpage'
615 errmsg = '%s: %s' % (errnote, error_to_compat_str(err))
617 raise ExtractorError(errmsg, sys.exc_info()[2], cause=err)
619 self._downloader.report_warning(errmsg)
622 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):
624 Return a tuple (page content as string, URL handle).
626 See _download_webpage docstring for arguments specification.
628 # Strip hashes from the URL (#1038)
629 if isinstance(url_or_request, (compat_str, str)):
630 url_or_request = url_or_request.partition('#')[0]
632 urlh = self._request_webpage(url_or_request, video_id, note, errnote, fatal, data=data, headers=headers, query=query, expected_status=expected_status)
636 content = self._webpage_read_content(urlh, url_or_request, video_id, note, errnote, fatal, encoding=encoding)
637 return (content, urlh)
640 def _guess_encoding_from_content(content_type, webpage_bytes):
641 m = re.match(r'[a-zA-Z0-9_.-]+/[a-zA-Z0-9_.-]+\s*;\s*charset=(.+)', content_type)
643 encoding = m.group(1)
645 m = re.search(br'<meta[^>]+charset=[\'"]?([^\'")]+)[ /\'">]',
646 webpage_bytes[:1024])
648 encoding = m.group(1).decode('ascii')
649 elif webpage_bytes.startswith(b'\xff\xfe'):
656 def __check_blocked(self, content):
657 first_block = content[:512]
658 if ('<title>Access to this site is blocked</title>' in content and
659 'Websense' in first_block):
660 msg = 'Access to this webpage has been blocked by Websense filtering software in your network.'
661 blocked_iframe = self._html_search_regex(
662 r'<iframe src="([^"]+)"', content,
663 'Websense information URL', default=None)
665 msg += ' Visit %s for more details' % blocked_iframe
666 raise ExtractorError(msg, expected=True)
667 if '<title>The URL you requested has been blocked</title>' in first_block:
669 'Access to this webpage has been blocked by Indian censorship. '
670 'Use a VPN or proxy server (with --proxy) to route around it.')
671 block_msg = self._html_search_regex(
672 r'</h1><p>(.*?)</p>',
673 content, 'block message', default=None)
675 msg += ' (Message: "%s")' % block_msg.replace('\n', ' ')
676 raise ExtractorError(msg, expected=True)
677 if ('<title>TTK :: Доступ к ресурсу ограничен</title>' in content and
678 'blocklist.rkn.gov.ru' in content):
679 raise ExtractorError(
680 'Access to this webpage has been blocked by decision of the Russian government. '
681 'Visit http://blocklist.rkn.gov.ru/ for a block reason.',
684 def _webpage_read_content(self, urlh, url_or_request, video_id, note=None, errnote=None, fatal=True, prefix=None, encoding=None):
685 content_type = urlh.headers.get('Content-Type', '')
686 webpage_bytes = urlh.read()
687 if prefix is not None:
688 webpage_bytes = prefix + webpage_bytes
690 encoding = self._guess_encoding_from_content(content_type, webpage_bytes)
691 if self._downloader.params.get('dump_intermediate_pages', False):
692 self.to_screen('Dumping request to ' + urlh.geturl())
693 dump = base64.b64encode(webpage_bytes).decode('ascii')
694 self._downloader.to_screen(dump)
695 if self._downloader.params.get('write_pages', False):
696 basen = '%s_%s' % (video_id, urlh.geturl())
698 h = '___' + hashlib.md5(basen.encode('utf-8')).hexdigest()
699 basen = basen[:240 - len(h)] + h
700 raw_filename = basen + '.dump'
701 filename = sanitize_filename(raw_filename, restricted=True)
702 self.to_screen('Saving request to ' + filename)
703 # Working around MAX_PATH limitation on Windows (see
704 # http://msdn.microsoft.com/en-us/library/windows/desktop/aa365247(v=vs.85).aspx)
705 if compat_os_name == 'nt':
706 absfilepath = os.path.abspath(filename)
707 if len(absfilepath) > 259:
708 filename = '\\\\?\\' + absfilepath
709 with open(filename, 'wb') as outf:
710 outf.write(webpage_bytes)
713 content = webpage_bytes.decode(encoding, 'replace')
715 content = webpage_bytes.decode('utf-8', 'replace')
717 self.__check_blocked(content)
721 def _download_webpage(
722 self, url_or_request, video_id, note=None, errnote=None,
723 fatal=True, tries=1, timeout=5, encoding=None, data=None,
724 headers={}, query={}, expected_status=None):
726 Return the data of the page as a string.
729 url_or_request -- plain text URL as a string or
730 a compat_urllib_request.Requestobject
731 video_id -- Video/playlist/item identifier (string)
734 note -- note printed before downloading (string)
735 errnote -- note printed in case of an error (string)
736 fatal -- flag denoting whether error should be considered fatal,
737 i.e. whether it should cause ExtractionError to be raised,
738 otherwise a warning will be reported and extraction continued
739 tries -- number of tries
740 timeout -- sleep interval between tries
741 encoding -- encoding for a page content decoding, guessed automatically
742 when not explicitly specified
743 data -- POST data (bytes)
744 headers -- HTTP headers (dict)
745 query -- URL query (dict)
746 expected_status -- allows to accept failed HTTP requests (non 2xx
747 status code) by explicitly specifying a set of accepted status
748 codes. Can be any of the following entities:
749 - an integer type specifying an exact failed status code to
751 - a list or a tuple of integer types specifying a list of
752 failed status codes to accept
753 - a callable accepting an actual failed status code and
754 returning True if it should be accepted
755 Note that this argument does not affect success status codes (2xx)
756 which are always accepted.
761 while success is False:
763 res = self._download_webpage_handle(
764 url_or_request, video_id, note, errnote, fatal,
765 encoding=encoding, data=data, headers=headers, query=query,
766 expected_status=expected_status)
768 except compat_http_client.IncompleteRead as e:
770 if try_count >= tries:
772 self._sleep(timeout, video_id)
779 def _download_xml_handle(
780 self, url_or_request, video_id, note='Downloading XML',
781 errnote='Unable to download XML', transform_source=None,
782 fatal=True, encoding=None, data=None, headers={}, query={},
783 expected_status=None):
785 Return a tuple (xml as an xml.etree.ElementTree.Element, URL handle).
787 See _download_webpage docstring for arguments specification.
789 res = self._download_webpage_handle(
790 url_or_request, video_id, note, errnote, fatal=fatal,
791 encoding=encoding, data=data, headers=headers, query=query,
792 expected_status=expected_status)
795 xml_string, urlh = res
796 return self._parse_xml(
797 xml_string, video_id, transform_source=transform_source,
801 self, url_or_request, video_id,
802 note='Downloading XML', errnote='Unable to download XML',
803 transform_source=None, fatal=True, encoding=None,
804 data=None, headers={}, query={}, expected_status=None):
806 Return the xml as an xml.etree.ElementTree.Element.
808 See _download_webpage docstring for arguments specification.
810 res = self._download_xml_handle(
811 url_or_request, video_id, note=note, errnote=errnote,
812 transform_source=transform_source, fatal=fatal, encoding=encoding,
813 data=data, headers=headers, query=query,
814 expected_status=expected_status)
815 return res if res is False else res[0]
817 def _parse_xml(self, xml_string, video_id, transform_source=None, fatal=True):
819 xml_string = transform_source(xml_string)
821 return compat_etree_fromstring(xml_string.encode('utf-8'))
822 except compat_xml_parse_error as ve:
823 errmsg = '%s: Failed to parse XML ' % video_id
825 raise ExtractorError(errmsg, cause=ve)
827 self.report_warning(errmsg + str(ve))
829 def _download_json_handle(
830 self, url_or_request, video_id, note='Downloading JSON metadata',
831 errnote='Unable to download JSON metadata', transform_source=None,
832 fatal=True, encoding=None, data=None, headers={}, query={},
833 expected_status=None):
835 Return a tuple (JSON object, URL handle).
837 See _download_webpage docstring for arguments specification.
839 res = self._download_webpage_handle(
840 url_or_request, video_id, note, errnote, fatal=fatal,
841 encoding=encoding, data=data, headers=headers, query=query,
842 expected_status=expected_status)
845 json_string, urlh = res
846 return self._parse_json(
847 json_string, video_id, transform_source=transform_source,
851 self, url_or_request, video_id, note='Downloading JSON metadata',
852 errnote='Unable to download JSON metadata', transform_source=None,
853 fatal=True, encoding=None, data=None, headers={}, query={},
854 expected_status=None):
856 Return the JSON object as a dict.
858 See _download_webpage docstring for arguments specification.
860 res = self._download_json_handle(
861 url_or_request, video_id, note=note, errnote=errnote,
862 transform_source=transform_source, fatal=fatal, encoding=encoding,
863 data=data, headers=headers, query=query,
864 expected_status=expected_status)
865 return res if res is False else res[0]
867 def _parse_json(self, json_string, video_id, transform_source=None, fatal=True):
869 json_string = transform_source(json_string)
871 return json.loads(json_string)
872 except ValueError as ve:
873 errmsg = '%s: Failed to parse JSON ' % video_id
875 raise ExtractorError(errmsg, cause=ve)
877 self.report_warning(errmsg + str(ve))
879 def report_warning(self, msg, video_id=None):
880 idstr = '' if video_id is None else '%s: ' % video_id
881 self._downloader.report_warning(
882 '[%s] %s%s' % (self.IE_NAME, idstr, msg))
884 def to_screen(self, msg):
885 """Print msg to screen, prefixing it with '[ie_name]'"""
886 self._downloader.to_screen('[%s] %s' % (self.IE_NAME, msg))
888 def report_extraction(self, id_or_name):
889 """Report information extraction."""
890 self.to_screen('%s: Extracting information' % id_or_name)
892 def report_download_webpage(self, video_id):
893 """Report webpage download."""
894 self.to_screen('%s: Downloading webpage' % video_id)
896 def report_age_confirmation(self):
897 """Report attempt to confirm age."""
898 self.to_screen('Confirming age')
900 def report_login(self):
901 """Report attempt to log in."""
902 self.to_screen('Logging in')
905 def raise_login_required(msg='This video is only available for registered users'):
906 raise ExtractorError(
907 '%s. Use --username and --password or --netrc to provide account credentials.' % msg,
911 def raise_geo_restricted(msg='This video is not available from your location due to geo restriction', countries=None):
912 raise GeoRestrictedError(msg, countries=countries)
914 # Methods for following #608
916 def url_result(url, ie=None, video_id=None, video_title=None):
917 """Returns a URL that points to a page that should be processed"""
918 # TODO: ie should be the class used for getting the info
919 video_info = {'_type': 'url',
922 if video_id is not None:
923 video_info['id'] = video_id
924 if video_title is not None:
925 video_info['title'] = video_title
928 def playlist_from_matches(self, matches, playlist_id=None, playlist_title=None, getter=None, ie=None):
930 self.url_result(self._proto_relative_url(getter(m) if getter else m), ie)
932 return self.playlist_result(
933 urls, playlist_id=playlist_id, playlist_title=playlist_title)
936 def playlist_result(entries, playlist_id=None, playlist_title=None, playlist_description=None):
937 """Returns a playlist"""
938 video_info = {'_type': 'playlist',
941 video_info['id'] = playlist_id
943 video_info['title'] = playlist_title
944 if playlist_description:
945 video_info['description'] = playlist_description
948 def _search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
950 Perform a regex search on the given string, using a single or a list of
951 patterns returning the first matching group.
952 In case of failure return a default value or raise a WARNING or a
953 RegexNotFoundError, depending on fatal, specifying the field name.
955 if isinstance(pattern, (str, compat_str, compiled_regex_type)):
956 mobj = re.search(pattern, string, flags)
959 mobj = re.search(p, string, flags)
963 if not self._downloader.params.get('no_color') and compat_os_name != 'nt' and sys.stderr.isatty():
964 _name = '\033[0;34m%s\033[0m' % name
970 # return the first matching group
971 return next(g for g in mobj.groups() if g is not None)
973 return mobj.group(group)
974 elif default is not NO_DEFAULT:
977 raise RegexNotFoundError('Unable to extract %s' % _name)
979 self._downloader.report_warning('unable to extract %s' % _name + bug_reports_message())
982 def _html_search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
984 Like _search_regex, but strips HTML tags and unescapes entities.
986 res = self._search_regex(pattern, string, name, default, fatal, flags, group)
988 return clean_html(res).strip()
992 def _get_netrc_login_info(self, netrc_machine=None):
995 netrc_machine = netrc_machine or self._NETRC_MACHINE
997 if self._downloader.params.get('usenetrc', False):
999 info = netrc.netrc().authenticators(netrc_machine)
1000 if info is not None:
1004 raise netrc.NetrcParseError(
1005 'No authenticators for %s' % netrc_machine)
1006 except (IOError, netrc.NetrcParseError) as err:
1007 self._downloader.report_warning(
1008 'parsing .netrc: %s' % error_to_compat_str(err))
1010 return username, password
1012 def _get_login_info(self, username_option='username', password_option='password', netrc_machine=None):
1014 Get the login info as (username, password)
1015 First look for the manually specified credentials using username_option
1016 and password_option as keys in params dictionary. If no such credentials
1017 available look in the netrc file using the netrc_machine or _NETRC_MACHINE
1019 If there's no info available, return (None, None)
1021 if self._downloader is None:
1024 downloader_params = self._downloader.params
1026 # Attempt to use provided username and password or .netrc data
1027 if downloader_params.get(username_option) is not None:
1028 username = downloader_params[username_option]
1029 password = downloader_params[password_option]
1031 username, password = self._get_netrc_login_info(netrc_machine)
1033 return username, password
1035 def _get_tfa_info(self, note='two-factor verification code'):
1037 Get the two-factor authentication info
1038 TODO - asking the user will be required for sms/phone verify
1039 currently just uses the command line option
1040 If there's no info available, return None
1042 if self._downloader is None:
1044 downloader_params = self._downloader.params
1046 if downloader_params.get('twofactor') is not None:
1047 return downloader_params['twofactor']
1049 return compat_getpass('Type %s and press [Return]: ' % note)
1051 # Helper functions for extracting OpenGraph info
1053 def _og_regexes(prop):
1054 content_re = r'content=(?:"([^"]+?)"|\'([^\']+?)\'|\s*([^\s"\'=<>`]+?))'
1055 property_re = (r'(?:name|property)=(?:\'og:%(prop)s\'|"og:%(prop)s"|\s*og:%(prop)s\b)'
1056 % {'prop': re.escape(prop)})
1057 template = r'<meta[^>]+?%s[^>]+?%s'
1059 template % (property_re, content_re),
1060 template % (content_re, property_re),
1064 def _meta_regex(prop):
1065 return r'''(?isx)<meta
1066 (?=[^>]+(?:itemprop|name|property|id|http-equiv)=(["\']?)%s\1)
1067 [^>]+?content=(["\'])(?P<content>.*?)\2''' % re.escape(prop)
1069 def _og_search_property(self, prop, html, name=None, **kargs):
1070 if not isinstance(prop, (list, tuple)):
1073 name = 'OpenGraph %s' % prop[0]
1076 og_regexes.extend(self._og_regexes(p))
1077 escaped = self._search_regex(og_regexes, html, name, flags=re.DOTALL, **kargs)
1080 return unescapeHTML(escaped)
1082 def _og_search_thumbnail(self, html, **kargs):
1083 return self._og_search_property('image', html, 'thumbnail URL', fatal=False, **kargs)
1085 def _og_search_description(self, html, **kargs):
1086 return self._og_search_property('description', html, fatal=False, **kargs)
1088 def _og_search_title(self, html, **kargs):
1089 return self._og_search_property('title', html, **kargs)
1091 def _og_search_video_url(self, html, name='video url', secure=True, **kargs):
1092 regexes = self._og_regexes('video') + self._og_regexes('video:url')
1094 regexes = self._og_regexes('video:secure_url') + regexes
1095 return self._html_search_regex(regexes, html, name, **kargs)
1097 def _og_search_url(self, html, **kargs):
1098 return self._og_search_property('url', html, **kargs)
1100 def _html_search_meta(self, name, html, display_name=None, fatal=False, **kwargs):
1101 if not isinstance(name, (list, tuple)):
1103 if display_name is None:
1104 display_name = name[0]
1105 return self._html_search_regex(
1106 [self._meta_regex(n) for n in name],
1107 html, display_name, fatal=fatal, group='content', **kwargs)
1109 def _dc_search_uploader(self, html):
1110 return self._html_search_meta('dc.creator', html, 'uploader')
1112 def _rta_search(self, html):
1113 # See http://www.rtalabel.org/index.php?content=howtofaq#single
1114 if re.search(r'(?ix)<meta\s+name="rating"\s+'
1115 r' content="RTA-5042-1996-1400-1577-RTA"',
1120 def _media_rating_search(self, html):
1121 # See http://www.tjg-designs.com/WP/metadata-code-examples-adding-metadata-to-your-web-pages/
1122 rating = self._html_search_meta('rating', html)
1134 return RATING_TABLE.get(rating.lower())
1136 def _family_friendly_search(self, html):
1137 # See http://schema.org/VideoObject
1138 family_friendly = self._html_search_meta(
1139 'isFamilyFriendly', html, default=None)
1141 if not family_friendly:
1150 return RATING_TABLE.get(family_friendly.lower())
1152 def _twitter_search_player(self, html):
1153 return self._html_search_meta('twitter:player', html,
1154 'twitter card player')
1156 def _search_json_ld(self, html, video_id, expected_type=None, **kwargs):
1157 json_ld = self._search_regex(
1158 JSON_LD_RE, html, 'JSON-LD', group='json_ld', **kwargs)
1159 default = kwargs.get('default', NO_DEFAULT)
1161 return default if default is not NO_DEFAULT else {}
1162 # JSON-LD may be malformed and thus `fatal` should be respected.
1163 # At the same time `default` may be passed that assumes `fatal=False`
1164 # for _search_regex. Let's simulate the same behavior here as well.
1165 fatal = kwargs.get('fatal', True) if default == NO_DEFAULT else False
1166 return self._json_ld(json_ld, video_id, fatal=fatal, expected_type=expected_type)
1168 def _json_ld(self, json_ld, video_id, fatal=True, expected_type=None):
1169 if isinstance(json_ld, compat_str):
1170 json_ld = self._parse_json(json_ld, video_id, fatal=fatal)
1174 if not isinstance(json_ld, (list, tuple, dict)):
1176 if isinstance(json_ld, dict):
1179 INTERACTION_TYPE_MAP = {
1180 'CommentAction': 'comment',
1181 'AgreeAction': 'like',
1182 'DisagreeAction': 'dislike',
1183 'LikeAction': 'like',
1184 'DislikeAction': 'dislike',
1185 'ListenAction': 'view',
1186 'WatchAction': 'view',
1187 'ViewAction': 'view',
1190 def extract_interaction_statistic(e):
1191 interaction_statistic = e.get('interactionStatistic')
1192 if not isinstance(interaction_statistic, list):
1194 for is_e in interaction_statistic:
1195 if not isinstance(is_e, dict):
1197 if is_e.get('@type') != 'InteractionCounter':
1199 interaction_type = is_e.get('interactionType')
1200 if not isinstance(interaction_type, compat_str):
1202 interaction_count = int_or_none(is_e.get('userInteractionCount'))
1203 if interaction_count is None:
1205 count_kind = INTERACTION_TYPE_MAP.get(interaction_type.split('/')[-1])
1208 count_key = '%s_count' % count_kind
1209 if info.get(count_key) is not None:
1211 info[count_key] = interaction_count
1213 def extract_video_object(e):
1214 assert e['@type'] == 'VideoObject'
1216 'url': e.get('contentUrl'),
1217 'title': unescapeHTML(e.get('name')),
1218 'description': unescapeHTML(e.get('description')),
1219 'thumbnail': e.get('thumbnailUrl') or e.get('thumbnailURL'),
1220 'duration': parse_duration(e.get('duration')),
1221 'timestamp': unified_timestamp(e.get('uploadDate')),
1222 'filesize': float_or_none(e.get('contentSize')),
1223 'tbr': int_or_none(e.get('bitrate')),
1224 'width': int_or_none(e.get('width')),
1225 'height': int_or_none(e.get('height')),
1226 'view_count': int_or_none(e.get('interactionCount')),
1228 extract_interaction_statistic(e)
1231 if isinstance(e.get('@context'), compat_str) and re.match(r'^https?://schema.org/?$', e.get('@context')):
1232 item_type = e.get('@type')
1233 if expected_type is not None and expected_type != item_type:
1235 if item_type in ('TVEpisode', 'Episode'):
1237 'episode': unescapeHTML(e.get('name')),
1238 'episode_number': int_or_none(e.get('episodeNumber')),
1239 'description': unescapeHTML(e.get('description')),
1241 part_of_season = e.get('partOfSeason')
1242 if isinstance(part_of_season, dict) and part_of_season.get('@type') in ('TVSeason', 'Season', 'CreativeWorkSeason'):
1243 info['season_number'] = int_or_none(part_of_season.get('seasonNumber'))
1244 part_of_series = e.get('partOfSeries') or e.get('partOfTVSeries')
1245 if isinstance(part_of_series, dict) and part_of_series.get('@type') in ('TVSeries', 'Series', 'CreativeWorkSeries'):
1246 info['series'] = unescapeHTML(part_of_series.get('name'))
1247 elif item_type in ('Article', 'NewsArticle'):
1249 'timestamp': parse_iso8601(e.get('datePublished')),
1250 'title': unescapeHTML(e.get('headline')),
1251 'description': unescapeHTML(e.get('articleBody')),
1253 elif item_type == 'VideoObject':
1254 extract_video_object(e)
1256 video = e.get('video')
1257 if isinstance(video, dict) and video.get('@type') == 'VideoObject':
1258 extract_video_object(video)
1260 return dict((k, v) for k, v in info.items() if v is not None)
1263 def _hidden_inputs(html):
1264 html = re.sub(r'<!--(?:(?!<!--).)*-->', '', html)
1266 for input in re.findall(r'(?i)(<input[^>]+>)', html):
1267 attrs = extract_attributes(input)
1270 if attrs.get('type') not in ('hidden', 'submit'):
1272 name = attrs.get('name') or attrs.get('id')
1273 value = attrs.get('value')
1274 if name and value is not None:
1275 hidden_inputs[name] = value
1276 return hidden_inputs
1278 def _form_hidden_inputs(self, form_id, html):
1279 form = self._search_regex(
1280 r'(?is)<form[^>]+?id=(["\'])%s\1[^>]*>(?P<form>.+?)</form>' % form_id,
1281 html, '%s form' % form_id, group='form')
1282 return self._hidden_inputs(form)
1284 def _sort_formats(self, formats, field_preference=None):
1286 raise ExtractorError('No video formats found')
1289 # Automatically determine tbr when missing based on abr and vbr (improves
1290 # formats sorting in some cases)
1291 if 'tbr' not in f and f.get('abr') is not None and f.get('vbr') is not None:
1292 f['tbr'] = f['abr'] + f['vbr']
1294 def _formats_key(f):
1295 # TODO remove the following workaround
1296 from ..utils import determine_ext
1297 if not f.get('ext') and 'url' in f:
1298 f['ext'] = determine_ext(f['url'])
1300 if isinstance(field_preference, (list, tuple)):
1303 if f.get(field) is not None
1304 else ('' if field == 'format_id' else -1)
1305 for field in field_preference)
1307 preference = f.get('preference')
1308 if preference is None:
1310 if f.get('ext') in ['f4f', 'f4m']: # Not yet supported
1313 protocol = f.get('protocol') or determine_protocol(f)
1314 proto_preference = 0 if protocol in ['http', 'https'] else (-0.5 if protocol == 'rtsp' else -0.1)
1316 if f.get('vcodec') == 'none': # audio only
1318 if self._downloader.params.get('prefer_free_formats'):
1319 ORDER = ['aac', 'mp3', 'm4a', 'webm', 'ogg', 'opus']
1321 ORDER = ['webm', 'opus', 'ogg', 'mp3', 'aac', 'm4a']
1324 audio_ext_preference = ORDER.index(f['ext'])
1326 audio_ext_preference = -1
1328 if f.get('acodec') == 'none': # video only
1330 if self._downloader.params.get('prefer_free_formats'):
1331 ORDER = ['flv', 'mp4', 'webm']
1333 ORDER = ['webm', 'flv', 'mp4']
1335 ext_preference = ORDER.index(f['ext'])
1338 audio_ext_preference = 0
1342 f.get('language_preference') if f.get('language_preference') is not None else -1,
1343 f.get('quality') if f.get('quality') is not None else -1,
1344 f.get('tbr') if f.get('tbr') is not None else -1,
1345 f.get('filesize') if f.get('filesize') is not None else -1,
1346 f.get('vbr') if f.get('vbr') is not None else -1,
1347 f.get('height') if f.get('height') is not None else -1,
1348 f.get('width') if f.get('width') is not None else -1,
1351 f.get('abr') if f.get('abr') is not None else -1,
1352 audio_ext_preference,
1353 f.get('fps') if f.get('fps') is not None else -1,
1354 f.get('filesize_approx') if f.get('filesize_approx') is not None else -1,
1355 f.get('source_preference') if f.get('source_preference') is not None else -1,
1356 f.get('format_id') if f.get('format_id') is not None else '',
1358 formats.sort(key=_formats_key)
1360 def _check_formats(self, formats, video_id):
1362 formats[:] = filter(
1363 lambda f: self._is_valid_url(
1365 item='%s video format' % f.get('format_id') if f.get('format_id') else 'video'),
1369 def _remove_duplicate_formats(formats):
1373 if f['url'] not in format_urls:
1374 format_urls.add(f['url'])
1375 unique_formats.append(f)
1376 formats[:] = unique_formats
1378 def _is_valid_url(self, url, video_id, item='video', headers={}):
1379 url = self._proto_relative_url(url, scheme='http:')
1380 # For now assume non HTTP(S) URLs always valid
1381 if not (url.startswith('http://') or url.startswith('https://')):
1384 self._request_webpage(url, video_id, 'Checking %s URL' % item, headers=headers)
1386 except ExtractorError as e:
1387 if isinstance(e.cause, compat_urllib_error.URLError):
1389 '%s: %s URL is invalid, skipping' % (video_id, item))
1393 def http_scheme(self):
1394 """ Either "http:" or "https:", depending on the user's preferences """
1397 if self._downloader.params.get('prefer_insecure', False)
1400 def _proto_relative_url(self, url, scheme=None):
1403 if url.startswith('//'):
1405 scheme = self.http_scheme()
1410 def _sleep(self, timeout, video_id, msg_template=None):
1411 if msg_template is None:
1412 msg_template = '%(video_id)s: Waiting for %(timeout)s seconds'
1413 msg = msg_template % {'video_id': video_id, 'timeout': timeout}
1417 def _extract_f4m_formats(self, manifest_url, video_id, preference=None, f4m_id=None,
1418 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1419 fatal=True, m3u8_id=None):
1420 manifest = self._download_xml(
1421 manifest_url, video_id, 'Downloading f4m manifest',
1422 'Unable to download f4m manifest',
1423 # Some manifests may be malformed, e.g. prosiebensat1 generated manifests
1424 # (see https://github.com/rg3/youtube-dl/issues/6215#issuecomment-121704244)
1425 transform_source=transform_source,
1428 if manifest is False:
1431 return self._parse_f4m_formats(
1432 manifest, manifest_url, video_id, preference=preference, f4m_id=f4m_id,
1433 transform_source=transform_source, fatal=fatal, m3u8_id=m3u8_id)
1435 def _parse_f4m_formats(self, manifest, manifest_url, video_id, preference=None, f4m_id=None,
1436 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1437 fatal=True, m3u8_id=None):
1438 # currently youtube-dl cannot decode the playerVerificationChallenge as Akamai uses Adobe Alchemy
1439 akamai_pv = manifest.find('{http://ns.adobe.com/f4m/1.0}pv-2.0')
1440 if akamai_pv is not None and ';' in akamai_pv.text:
1441 playerVerificationChallenge = akamai_pv.text.split(';')[0]
1442 if playerVerificationChallenge.strip() != '':
1446 manifest_version = '1.0'
1447 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/1.0}media')
1449 manifest_version = '2.0'
1450 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/2.0}media')
1451 # Remove unsupported DRM protected media from final formats
1452 # rendition (see https://github.com/rg3/youtube-dl/issues/8573).
1453 media_nodes = remove_encrypted_media(media_nodes)
1457 manifest_base_url = get_base_url(manifest)
1459 bootstrap_info = xpath_element(
1460 manifest, ['{http://ns.adobe.com/f4m/1.0}bootstrapInfo', '{http://ns.adobe.com/f4m/2.0}bootstrapInfo'],
1461 'bootstrap info', default=None)
1464 mime_type = xpath_text(
1465 manifest, ['{http://ns.adobe.com/f4m/1.0}mimeType', '{http://ns.adobe.com/f4m/2.0}mimeType'],
1466 'base URL', default=None)
1467 if mime_type and mime_type.startswith('audio/'):
1470 for i, media_el in enumerate(media_nodes):
1471 tbr = int_or_none(media_el.attrib.get('bitrate'))
1472 width = int_or_none(media_el.attrib.get('width'))
1473 height = int_or_none(media_el.attrib.get('height'))
1474 format_id = '-'.join(filter(None, [f4m_id, compat_str(i if tbr is None else tbr)]))
1475 # If <bootstrapInfo> is present, the specified f4m is a
1476 # stream-level manifest, and only set-level manifests may refer to
1477 # external resources. See section 11.4 and section 4 of F4M spec
1478 if bootstrap_info is None:
1480 # @href is introduced in 2.0, see section 11.6 of F4M spec
1481 if manifest_version == '2.0':
1482 media_url = media_el.attrib.get('href')
1483 if media_url is None:
1484 media_url = media_el.attrib.get('url')
1488 media_url if media_url.startswith('http://') or media_url.startswith('https://')
1489 else ((manifest_base_url or '/'.join(manifest_url.split('/')[:-1])) + '/' + media_url))
1490 # If media_url is itself a f4m manifest do the recursive extraction
1491 # since bitrates in parent manifest (this one) and media_url manifest
1492 # may differ leading to inability to resolve the format by requested
1493 # bitrate in f4m downloader
1494 ext = determine_ext(manifest_url)
1496 f4m_formats = self._extract_f4m_formats(
1497 manifest_url, video_id, preference=preference, f4m_id=f4m_id,
1498 transform_source=transform_source, fatal=fatal)
1499 # Sometimes stream-level manifest contains single media entry that
1500 # does not contain any quality metadata (e.g. http://matchtv.ru/#live-player).
1501 # At the same time parent's media entry in set-level manifest may
1502 # contain it. We will copy it from parent in such cases.
1503 if len(f4m_formats) == 1:
1506 'tbr': f.get('tbr') or tbr,
1507 'width': f.get('width') or width,
1508 'height': f.get('height') or height,
1509 'format_id': f.get('format_id') if not tbr else format_id,
1512 formats.extend(f4m_formats)
1515 formats.extend(self._extract_m3u8_formats(
1516 manifest_url, video_id, 'mp4', preference=preference,
1517 m3u8_id=m3u8_id, fatal=fatal))
1520 'format_id': format_id,
1521 'url': manifest_url,
1522 'manifest_url': manifest_url,
1523 'ext': 'flv' if bootstrap_info is not None else None,
1529 'preference': preference,
1533 def _m3u8_meta_format(self, m3u8_url, ext=None, preference=None, m3u8_id=None):
1535 'format_id': '-'.join(filter(None, [m3u8_id, 'meta'])),
1539 'preference': preference - 100 if preference else -100,
1540 'resolution': 'multiple',
1541 'format_note': 'Quality selection URL',
1544 def _extract_m3u8_formats(self, m3u8_url, video_id, ext=None,
1545 entry_protocol='m3u8', preference=None,
1546 m3u8_id=None, note=None, errnote=None,
1547 fatal=True, live=False):
1548 res = self._download_webpage_handle(
1550 note=note or 'Downloading m3u8 information',
1551 errnote=errnote or 'Failed to download m3u8 information',
1557 m3u8_doc, urlh = res
1558 m3u8_url = urlh.geturl()
1560 return self._parse_m3u8_formats(
1561 m3u8_doc, m3u8_url, ext=ext, entry_protocol=entry_protocol,
1562 preference=preference, m3u8_id=m3u8_id, live=live)
1564 def _parse_m3u8_formats(self, m3u8_doc, m3u8_url, ext=None,
1565 entry_protocol='m3u8', preference=None,
1566 m3u8_id=None, live=False):
1567 if '#EXT-X-FAXS-CM:' in m3u8_doc: # Adobe Flash Access
1570 if re.search(r'#EXT-X-SESSION-KEY:.*?URI="skd://', m3u8_doc): # Apple FairPlay
1575 format_url = lambda u: (
1577 if re.match(r'^https?://', u)
1578 else compat_urlparse.urljoin(m3u8_url, u))
1581 # 1. https://tools.ietf.org/html/draft-pantos-http-live-streaming-21
1582 # 2. https://github.com/rg3/youtube-dl/issues/12211
1584 # We should try extracting formats only from master playlists [1, 4.3.4],
1585 # i.e. playlists that describe available qualities. On the other hand
1586 # media playlists [1, 4.3.3] should be returned as is since they contain
1587 # just the media without qualities renditions.
1588 # Fortunately, master playlist can be easily distinguished from media
1589 # playlist based on particular tags availability. As of [1, 4.3.3, 4.3.4]
1590 # master playlist tags MUST NOT appear in a media playist and vice versa.
1591 # As of [1, 4.3.3.1] #EXT-X-TARGETDURATION tag is REQUIRED for every
1592 # media playlist and MUST NOT appear in master playlist thus we can
1593 # clearly detect media playlist with this criterion.
1595 if '#EXT-X-TARGETDURATION' in m3u8_doc: # media playlist, return as is
1598 'format_id': m3u8_id,
1600 'protocol': entry_protocol,
1601 'preference': preference,
1605 last_stream_inf = {}
1607 def extract_media(x_media_line):
1608 media = parse_m3u8_attributes(x_media_line)
1609 # As per [1, 4.3.4.1] TYPE, GROUP-ID and NAME are REQUIRED
1610 media_type, group_id, name = media.get('TYPE'), media.get('GROUP-ID'), media.get('NAME')
1611 if not (media_type and group_id and name):
1613 groups.setdefault(group_id, []).append(media)
1614 if media_type not in ('VIDEO', 'AUDIO'):
1616 media_url = media.get('URI')
1619 for v in (m3u8_id, group_id, name):
1623 'format_id': '-'.join(format_id),
1624 'url': format_url(media_url),
1625 'manifest_url': m3u8_url,
1626 'language': media.get('LANGUAGE'),
1628 'protocol': entry_protocol,
1629 'preference': preference,
1631 if media_type == 'AUDIO':
1632 f['vcodec'] = 'none'
1635 def build_stream_name():
1636 # Despite specification does not mention NAME attribute for
1637 # EXT-X-STREAM-INF tag it still sometimes may be present (see [1]
1638 # or vidio test in TestInfoExtractor.test_parse_m3u8_formats)
1639 # 1. http://www.vidio.com/watch/165683-dj_ambred-booyah-live-2015
1640 stream_name = last_stream_inf.get('NAME')
1643 # If there is no NAME in EXT-X-STREAM-INF it will be obtained
1644 # from corresponding rendition group
1645 stream_group_id = last_stream_inf.get('VIDEO')
1646 if not stream_group_id:
1648 stream_group = groups.get(stream_group_id)
1649 if not stream_group:
1650 return stream_group_id
1651 rendition = stream_group[0]
1652 return rendition.get('NAME') or stream_group_id
1654 for line in m3u8_doc.splitlines():
1655 if line.startswith('#EXT-X-STREAM-INF:'):
1656 last_stream_inf = parse_m3u8_attributes(line)
1657 elif line.startswith('#EXT-X-MEDIA:'):
1659 elif line.startswith('#') or not line.strip():
1662 tbr = float_or_none(
1663 last_stream_inf.get('AVERAGE-BANDWIDTH') or
1664 last_stream_inf.get('BANDWIDTH'), scale=1000)
1667 format_id.append(m3u8_id)
1668 stream_name = build_stream_name()
1669 # Bandwidth of live streams may differ over time thus making
1670 # format_id unpredictable. So it's better to keep provided
1673 format_id.append(stream_name if stream_name else '%d' % (tbr if tbr else len(formats)))
1674 manifest_url = format_url(line.strip())
1676 'format_id': '-'.join(format_id),
1677 'url': manifest_url,
1678 'manifest_url': m3u8_url,
1681 'fps': float_or_none(last_stream_inf.get('FRAME-RATE')),
1682 'protocol': entry_protocol,
1683 'preference': preference,
1685 resolution = last_stream_inf.get('RESOLUTION')
1687 mobj = re.search(r'(?P<width>\d+)[xX](?P<height>\d+)', resolution)
1689 f['width'] = int(mobj.group('width'))
1690 f['height'] = int(mobj.group('height'))
1691 # Unified Streaming Platform
1693 r'audio.*?(?:%3D|=)(\d+)(?:-video.*?(?:%3D|=)(\d+))?', f['url'])
1695 abr, vbr = mobj.groups()
1696 abr, vbr = float_or_none(abr, 1000), float_or_none(vbr, 1000)
1701 codecs = parse_codecs(last_stream_inf.get('CODECS'))
1703 audio_group_id = last_stream_inf.get('AUDIO')
1704 # As per [1, 4.3.4.1.1] any EXT-X-STREAM-INF tag which
1705 # references a rendition group MUST have a CODECS attribute.
1706 # However, this is not always respected, for example, [2]
1707 # contains EXT-X-STREAM-INF tag which references AUDIO
1708 # rendition group but does not have CODECS and despite
1709 # referencing an audio group it represents a complete
1710 # (with audio and video) format. So, for such cases we will
1711 # ignore references to rendition groups and treat them
1712 # as complete formats.
1713 if audio_group_id and codecs and f.get('vcodec') != 'none':
1714 audio_group = groups.get(audio_group_id)
1715 if audio_group and audio_group[0].get('URI'):
1716 # TODO: update acodec for audio only formats with
1718 f['acodec'] = 'none'
1720 last_stream_inf = {}
1724 def _xpath_ns(path, namespace=None):
1728 for c in path.split('/'):
1729 if not c or c == '.':
1732 out.append('{%s}%s' % (namespace, c))
1733 return '/'.join(out)
1735 def _extract_smil_formats(self, smil_url, video_id, fatal=True, f4m_params=None, transform_source=None):
1736 smil = self._download_smil(smil_url, video_id, fatal=fatal, transform_source=transform_source)
1742 namespace = self._parse_smil_namespace(smil)
1744 return self._parse_smil_formats(
1745 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
1747 def _extract_smil_info(self, smil_url, video_id, fatal=True, f4m_params=None):
1748 smil = self._download_smil(smil_url, video_id, fatal=fatal)
1751 return self._parse_smil(smil, smil_url, video_id, f4m_params=f4m_params)
1753 def _download_smil(self, smil_url, video_id, fatal=True, transform_source=None):
1754 return self._download_xml(
1755 smil_url, video_id, 'Downloading SMIL file',
1756 'Unable to download SMIL file', fatal=fatal, transform_source=transform_source)
1758 def _parse_smil(self, smil, smil_url, video_id, f4m_params=None):
1759 namespace = self._parse_smil_namespace(smil)
1761 formats = self._parse_smil_formats(
1762 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
1763 subtitles = self._parse_smil_subtitles(smil, namespace=namespace)
1765 video_id = os.path.splitext(url_basename(smil_url))[0]
1769 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
1770 name = meta.attrib.get('name')
1771 content = meta.attrib.get('content')
1772 if not name or not content:
1774 if not title and name == 'title':
1776 elif not description and name in ('description', 'abstract'):
1777 description = content
1778 elif not upload_date and name == 'date':
1779 upload_date = unified_strdate(content)
1782 'id': image.get('type'),
1783 'url': image.get('src'),
1784 'width': int_or_none(image.get('width')),
1785 'height': int_or_none(image.get('height')),
1786 } for image in smil.findall(self._xpath_ns('.//image', namespace)) if image.get('src')]
1790 'title': title or video_id,
1791 'description': description,
1792 'upload_date': upload_date,
1793 'thumbnails': thumbnails,
1795 'subtitles': subtitles,
1798 def _parse_smil_namespace(self, smil):
1799 return self._search_regex(
1800 r'(?i)^{([^}]+)?}smil$', smil.tag, 'namespace', default=None)
1802 def _parse_smil_formats(self, smil, smil_url, video_id, namespace=None, f4m_params=None, transform_rtmp_url=None):
1804 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
1805 b = meta.get('base') or meta.get('httpBase')
1816 media = smil.findall(self._xpath_ns('.//video', namespace)) + smil.findall(self._xpath_ns('.//audio', namespace))
1817 for medium in media:
1818 src = medium.get('src')
1819 if not src or src in srcs:
1823 bitrate = float_or_none(medium.get('system-bitrate') or medium.get('systemBitrate'), 1000)
1824 filesize = int_or_none(medium.get('size') or medium.get('fileSize'))
1825 width = int_or_none(medium.get('width'))
1826 height = int_or_none(medium.get('height'))
1827 proto = medium.get('proto')
1828 ext = medium.get('ext')
1829 src_ext = determine_ext(src)
1830 streamer = medium.get('streamer') or base
1832 if proto == 'rtmp' or streamer.startswith('rtmp'):
1838 'format_id': 'rtmp-%d' % (rtmp_count if bitrate is None else bitrate),
1840 'filesize': filesize,
1844 if transform_rtmp_url:
1845 streamer, src = transform_rtmp_url(streamer, src)
1846 formats[-1].update({
1852 src_url = src if src.startswith('http') else compat_urlparse.urljoin(base, src)
1853 src_url = src_url.strip()
1855 if proto == 'm3u8' or src_ext == 'm3u8':
1856 m3u8_formats = self._extract_m3u8_formats(
1857 src_url, video_id, ext or 'mp4', m3u8_id='hls', fatal=False)
1858 if len(m3u8_formats) == 1:
1860 m3u8_formats[0].update({
1861 'format_id': 'hls-%d' % (m3u8_count if bitrate is None else bitrate),
1866 formats.extend(m3u8_formats)
1867 elif src_ext == 'f4m':
1872 'plugin': 'flowplayer-3.2.0.1',
1874 f4m_url += '&' if '?' in f4m_url else '?'
1875 f4m_url += compat_urllib_parse_urlencode(f4m_params)
1876 formats.extend(self._extract_f4m_formats(f4m_url, video_id, f4m_id='hds', fatal=False))
1877 elif src_ext == 'mpd':
1878 formats.extend(self._extract_mpd_formats(
1879 src_url, video_id, mpd_id='dash', fatal=False))
1880 elif re.search(r'\.ism/[Mm]anifest', src_url):
1881 formats.extend(self._extract_ism_formats(
1882 src_url, video_id, ism_id='mss', fatal=False))
1883 elif src_url.startswith('http') and self._is_valid_url(src, video_id):
1887 'ext': ext or src_ext or 'flv',
1888 'format_id': 'http-%d' % (bitrate or http_count),
1890 'filesize': filesize,
1897 def _parse_smil_subtitles(self, smil, namespace=None, subtitles_lang='en'):
1900 for num, textstream in enumerate(smil.findall(self._xpath_ns('.//textstream', namespace))):
1901 src = textstream.get('src')
1902 if not src or src in urls:
1905 ext = textstream.get('ext') or mimetype2ext(textstream.get('type')) or determine_ext(src)
1906 lang = textstream.get('systemLanguage') or textstream.get('systemLanguageName') or textstream.get('lang') or subtitles_lang
1907 subtitles.setdefault(lang, []).append({
1913 def _extract_xspf_playlist(self, xspf_url, playlist_id, fatal=True):
1914 xspf = self._download_xml(
1915 xspf_url, playlist_id, 'Downloading xpsf playlist',
1916 'Unable to download xspf manifest', fatal=fatal)
1919 return self._parse_xspf(
1920 xspf, playlist_id, xspf_url=xspf_url,
1921 xspf_base_url=base_url(xspf_url))
1923 def _parse_xspf(self, xspf_doc, playlist_id, xspf_url=None, xspf_base_url=None):
1925 'xspf': 'http://xspf.org/ns/0/',
1926 's1': 'http://static.streamone.nl/player/ns/0',
1930 for track in xspf_doc.findall(xpath_with_ns('./xspf:trackList/xspf:track', NS_MAP)):
1932 track, xpath_with_ns('./xspf:title', NS_MAP), 'title', default=playlist_id)
1933 description = xpath_text(
1934 track, xpath_with_ns('./xspf:annotation', NS_MAP), 'description')
1935 thumbnail = xpath_text(
1936 track, xpath_with_ns('./xspf:image', NS_MAP), 'thumbnail')
1937 duration = float_or_none(
1938 xpath_text(track, xpath_with_ns('./xspf:duration', NS_MAP), 'duration'), 1000)
1941 for location in track.findall(xpath_with_ns('./xspf:location', NS_MAP)):
1942 format_url = urljoin(xspf_base_url, location.text)
1947 'manifest_url': xspf_url,
1948 'format_id': location.get(xpath_with_ns('s1:label', NS_MAP)),
1949 'width': int_or_none(location.get(xpath_with_ns('s1:width', NS_MAP))),
1950 'height': int_or_none(location.get(xpath_with_ns('s1:height', NS_MAP))),
1952 self._sort_formats(formats)
1957 'description': description,
1958 'thumbnail': thumbnail,
1959 'duration': duration,
1964 def _extract_mpd_formats(self, mpd_url, video_id, mpd_id=None, note=None, errnote=None, fatal=True, formats_dict={}):
1965 res = self._download_xml_handle(
1967 note=note or 'Downloading MPD manifest',
1968 errnote=errnote or 'Failed to download MPD manifest',
1973 mpd_base_url = base_url(urlh.geturl())
1975 return self._parse_mpd_formats(
1976 mpd_doc, mpd_id=mpd_id, mpd_base_url=mpd_base_url,
1977 formats_dict=formats_dict, mpd_url=mpd_url)
1979 def _parse_mpd_formats(self, mpd_doc, mpd_id=None, mpd_base_url='', formats_dict={}, mpd_url=None):
1981 Parse formats from MPD manifest.
1983 1. MPEG-DASH Standard, ISO/IEC 23009-1:2014(E),
1984 http://standards.iso.org/ittf/PubliclyAvailableStandards/c065274_ISO_IEC_23009-1_2014.zip
1985 2. https://en.wikipedia.org/wiki/Dynamic_Adaptive_Streaming_over_HTTP
1987 if mpd_doc.get('type') == 'dynamic':
1990 namespace = self._search_regex(r'(?i)^{([^}]+)?}MPD$', mpd_doc.tag, 'namespace', default=None)
1993 return self._xpath_ns(path, namespace)
1995 def is_drm_protected(element):
1996 return element.find(_add_ns('ContentProtection')) is not None
1998 def extract_multisegment_info(element, ms_parent_info):
1999 ms_info = ms_parent_info.copy()
2001 # As per [1, 5.3.9.2.2] SegmentList and SegmentTemplate share some
2002 # common attributes and elements. We will only extract relevant
2004 def extract_common(source):
2005 segment_timeline = source.find(_add_ns('SegmentTimeline'))
2006 if segment_timeline is not None:
2007 s_e = segment_timeline.findall(_add_ns('S'))
2009 ms_info['total_number'] = 0
2012 r = int(s.get('r', 0))
2013 ms_info['total_number'] += 1 + r
2014 ms_info['s'].append({
2015 't': int(s.get('t', 0)),
2016 # @d is mandatory (see [1, 5.3.9.6.2, Table 17, page 60])
2017 'd': int(s.attrib['d']),
2020 start_number = source.get('startNumber')
2022 ms_info['start_number'] = int(start_number)
2023 timescale = source.get('timescale')
2025 ms_info['timescale'] = int(timescale)
2026 segment_duration = source.get('duration')
2027 if segment_duration:
2028 ms_info['segment_duration'] = float(segment_duration)
2030 def extract_Initialization(source):
2031 initialization = source.find(_add_ns('Initialization'))
2032 if initialization is not None:
2033 ms_info['initialization_url'] = initialization.attrib['sourceURL']
2035 segment_list = element.find(_add_ns('SegmentList'))
2036 if segment_list is not None:
2037 extract_common(segment_list)
2038 extract_Initialization(segment_list)
2039 segment_urls_e = segment_list.findall(_add_ns('SegmentURL'))
2041 ms_info['segment_urls'] = [segment.attrib['media'] for segment in segment_urls_e]
2043 segment_template = element.find(_add_ns('SegmentTemplate'))
2044 if segment_template is not None:
2045 extract_common(segment_template)
2046 media = segment_template.get('media')
2048 ms_info['media'] = media
2049 initialization = segment_template.get('initialization')
2051 ms_info['initialization'] = initialization
2053 extract_Initialization(segment_template)
2056 mpd_duration = parse_duration(mpd_doc.get('mediaPresentationDuration'))
2058 for period in mpd_doc.findall(_add_ns('Period')):
2059 period_duration = parse_duration(period.get('duration')) or mpd_duration
2060 period_ms_info = extract_multisegment_info(period, {
2064 for adaptation_set in period.findall(_add_ns('AdaptationSet')):
2065 if is_drm_protected(adaptation_set):
2067 adaption_set_ms_info = extract_multisegment_info(adaptation_set, period_ms_info)
2068 for representation in adaptation_set.findall(_add_ns('Representation')):
2069 if is_drm_protected(representation):
2071 representation_attrib = adaptation_set.attrib.copy()
2072 representation_attrib.update(representation.attrib)
2073 # According to [1, 5.3.7.2, Table 9, page 41], @mimeType is mandatory
2074 mime_type = representation_attrib['mimeType']
2075 content_type = mime_type.split('/')[0]
2076 if content_type == 'text':
2077 # TODO implement WebVTT downloading
2079 elif content_type in ('video', 'audio'):
2081 for element in (representation, adaptation_set, period, mpd_doc):
2082 base_url_e = element.find(_add_ns('BaseURL'))
2083 if base_url_e is not None:
2084 base_url = base_url_e.text + base_url
2085 if re.match(r'^https?://', base_url):
2087 if mpd_base_url and not re.match(r'^https?://', base_url):
2088 if not mpd_base_url.endswith('/') and not base_url.startswith('/'):
2090 base_url = mpd_base_url + base_url
2091 representation_id = representation_attrib.get('id')
2092 lang = representation_attrib.get('lang')
2093 url_el = representation.find(_add_ns('BaseURL'))
2094 filesize = int_or_none(url_el.attrib.get('{http://youtube.com/yt/2012/10/10}contentLength') if url_el is not None else None)
2095 bandwidth = int_or_none(representation_attrib.get('bandwidth'))
2097 'format_id': '%s-%s' % (mpd_id, representation_id) if mpd_id else representation_id,
2099 'manifest_url': mpd_url,
2100 'ext': mimetype2ext(mime_type),
2101 'width': int_or_none(representation_attrib.get('width')),
2102 'height': int_or_none(representation_attrib.get('height')),
2103 'tbr': float_or_none(bandwidth, 1000),
2104 'asr': int_or_none(representation_attrib.get('audioSamplingRate')),
2105 'fps': int_or_none(representation_attrib.get('frameRate')),
2106 'language': lang if lang not in ('mul', 'und', 'zxx', 'mis') else None,
2107 'format_note': 'DASH %s' % content_type,
2108 'filesize': filesize,
2109 'container': mimetype2ext(mime_type) + '_dash',
2111 f.update(parse_codecs(representation_attrib.get('codecs')))
2112 representation_ms_info = extract_multisegment_info(representation, adaption_set_ms_info)
2114 def prepare_template(template_name, identifiers):
2115 tmpl = representation_ms_info[template_name]
2116 # First of, % characters outside $...$ templates
2117 # must be escaped by doubling for proper processing
2118 # by % operator string formatting used further (see
2119 # https://github.com/rg3/youtube-dl/issues/16867).
2125 in_template = not in_template
2126 elif c == '%' and not in_template:
2128 # Next, $...$ templates are translated to their
2129 # %(...) counterparts to be used with % operator
2130 t = t.replace('$RepresentationID$', representation_id)
2131 t = re.sub(r'\$(%s)\$' % '|'.join(identifiers), r'%(\1)d', t)
2132 t = re.sub(r'\$(%s)%%([^$]+)\$' % '|'.join(identifiers), r'%(\1)\2', t)
2133 t.replace('$$', '$')
2136 # @initialization is a regular template like @media one
2137 # so it should be handled just the same way (see
2138 # https://github.com/rg3/youtube-dl/issues/11605)
2139 if 'initialization' in representation_ms_info:
2140 initialization_template = prepare_template(
2142 # As per [1, 5.3.9.4.2, Table 15, page 54] $Number$ and
2143 # $Time$ shall not be included for @initialization thus
2144 # only $Bandwidth$ remains
2146 representation_ms_info['initialization_url'] = initialization_template % {
2147 'Bandwidth': bandwidth,
2150 def location_key(location):
2151 return 'url' if re.match(r'^https?://', location) else 'path'
2153 if 'segment_urls' not in representation_ms_info and 'media' in representation_ms_info:
2155 media_template = prepare_template('media', ('Number', 'Bandwidth', 'Time'))
2156 media_location_key = location_key(media_template)
2158 # As per [1, 5.3.9.4.4, Table 16, page 55] $Number$ and $Time$
2159 # can't be used at the same time
2160 if '%(Number' in media_template and 's' not in representation_ms_info:
2161 segment_duration = None
2162 if 'total_number' not in representation_ms_info and 'segment_duration' in representation_ms_info:
2163 segment_duration = float_or_none(representation_ms_info['segment_duration'], representation_ms_info['timescale'])
2164 representation_ms_info['total_number'] = int(math.ceil(float(period_duration) / segment_duration))
2165 representation_ms_info['fragments'] = [{
2166 media_location_key: media_template % {
2167 'Number': segment_number,
2168 'Bandwidth': bandwidth,
2170 'duration': segment_duration,
2171 } for segment_number in range(
2172 representation_ms_info['start_number'],
2173 representation_ms_info['total_number'] + representation_ms_info['start_number'])]
2175 # $Number*$ or $Time$ in media template with S list available
2176 # Example $Number*$: http://www.svtplay.se/klipp/9023742/stopptid-om-bjorn-borg
2177 # Example $Time$: https://play.arkena.com/embed/avp/v2/player/media/b41dda37-d8e7-4d3f-b1b5-9a9db578bdfe/1/129411
2178 representation_ms_info['fragments'] = []
2181 segment_number = representation_ms_info['start_number']
2183 def add_segment_url():
2184 segment_url = media_template % {
2185 'Time': segment_time,
2186 'Bandwidth': bandwidth,
2187 'Number': segment_number,
2189 representation_ms_info['fragments'].append({
2190 media_location_key: segment_url,
2191 'duration': float_or_none(segment_d, representation_ms_info['timescale']),
2194 for num, s in enumerate(representation_ms_info['s']):
2195 segment_time = s.get('t') or segment_time
2199 for r in range(s.get('r', 0)):
2200 segment_time += segment_d
2203 segment_time += segment_d
2204 elif 'segment_urls' in representation_ms_info and 's' in representation_ms_info:
2206 # Example: https://www.youtube.com/watch?v=iXZV5uAYMJI
2207 # or any YouTube dashsegments video
2210 timescale = representation_ms_info['timescale']
2211 for s in representation_ms_info['s']:
2212 duration = float_or_none(s['d'], timescale)
2213 for r in range(s.get('r', 0) + 1):
2214 segment_uri = representation_ms_info['segment_urls'][segment_index]
2216 location_key(segment_uri): segment_uri,
2217 'duration': duration,
2220 representation_ms_info['fragments'] = fragments
2221 elif 'segment_urls' in representation_ms_info:
2222 # Segment URLs with no SegmentTimeline
2223 # Example: https://www.seznam.cz/zpravy/clanek/cesko-zasahne-vitr-o-sile-vichrice-muze-byt-i-zivotu-nebezpecny-39091
2224 # https://github.com/rg3/youtube-dl/pull/14844
2226 segment_duration = float_or_none(
2227 representation_ms_info['segment_duration'],
2228 representation_ms_info['timescale']) if 'segment_duration' in representation_ms_info else None
2229 for segment_url in representation_ms_info['segment_urls']:
2231 location_key(segment_url): segment_url,
2233 if segment_duration:
2234 fragment['duration'] = segment_duration
2235 fragments.append(fragment)
2236 representation_ms_info['fragments'] = fragments
2237 # NB: MPD manifest may contain direct URLs to unfragmented media.
2238 # No fragments key is present in this case.
2239 if 'fragments' in representation_ms_info:
2241 'fragment_base_url': base_url,
2243 'protocol': 'http_dash_segments',
2245 if 'initialization_url' in representation_ms_info:
2246 initialization_url = representation_ms_info['initialization_url']
2247 if not f.get('url'):
2248 f['url'] = initialization_url
2249 f['fragments'].append({location_key(initialization_url): initialization_url})
2250 f['fragments'].extend(representation_ms_info['fragments'])
2251 # According to [1, 5.3.5.2, Table 7, page 35] @id of Representation
2252 # is not necessarily unique within a Period thus formats with
2253 # the same `format_id` are quite possible. There are numerous examples
2254 # of such manifests (see https://github.com/rg3/youtube-dl/issues/15111,
2255 # https://github.com/rg3/youtube-dl/issues/13919)
2256 full_info = formats_dict.get(representation_id, {}).copy()
2258 formats.append(full_info)
2260 self.report_warning('Unknown MIME type %s in DASH manifest' % mime_type)
2263 def _extract_ism_formats(self, ism_url, video_id, ism_id=None, note=None, errnote=None, fatal=True):
2264 res = self._download_xml_handle(
2266 note=note or 'Downloading ISM manifest',
2267 errnote=errnote or 'Failed to download ISM manifest',
2273 return self._parse_ism_formats(ism_doc, urlh.geturl(), ism_id)
2275 def _parse_ism_formats(self, ism_doc, ism_url, ism_id=None):
2277 Parse formats from ISM manifest.
2279 1. [MS-SSTR]: Smooth Streaming Protocol,
2280 https://msdn.microsoft.com/en-us/library/ff469518.aspx
2282 if ism_doc.get('IsLive') == 'TRUE' or ism_doc.find('Protection') is not None:
2285 duration = int(ism_doc.attrib['Duration'])
2286 timescale = int_or_none(ism_doc.get('TimeScale')) or 10000000
2289 for stream in ism_doc.findall('StreamIndex'):
2290 stream_type = stream.get('Type')
2291 if stream_type not in ('video', 'audio'):
2293 url_pattern = stream.attrib['Url']
2294 stream_timescale = int_or_none(stream.get('TimeScale')) or timescale
2295 stream_name = stream.get('Name')
2296 for track in stream.findall('QualityLevel'):
2297 fourcc = track.get('FourCC', 'AACL' if track.get('AudioTag') == '255' else None)
2298 # TODO: add support for WVC1 and WMAP
2299 if fourcc not in ('H264', 'AVC1', 'AACL'):
2300 self.report_warning('%s is not a supported codec' % fourcc)
2302 tbr = int(track.attrib['Bitrate']) // 1000
2303 # [1] does not mention Width and Height attributes. However,
2304 # they're often present while MaxWidth and MaxHeight are
2305 # missing, so should be used as fallbacks
2306 width = int_or_none(track.get('MaxWidth') or track.get('Width'))
2307 height = int_or_none(track.get('MaxHeight') or track.get('Height'))
2308 sampling_rate = int_or_none(track.get('SamplingRate'))
2310 track_url_pattern = re.sub(r'{[Bb]itrate}', track.attrib['Bitrate'], url_pattern)
2311 track_url_pattern = compat_urlparse.urljoin(ism_url, track_url_pattern)
2317 stream_fragments = stream.findall('c')
2318 for stream_fragment_index, stream_fragment in enumerate(stream_fragments):
2319 fragment_ctx['time'] = int_or_none(stream_fragment.get('t')) or fragment_ctx['time']
2320 fragment_repeat = int_or_none(stream_fragment.get('r')) or 1
2321 fragment_ctx['duration'] = int_or_none(stream_fragment.get('d'))
2322 if not fragment_ctx['duration']:
2324 next_fragment_time = int(stream_fragment[stream_fragment_index + 1].attrib['t'])
2326 next_fragment_time = duration
2327 fragment_ctx['duration'] = (next_fragment_time - fragment_ctx['time']) / fragment_repeat
2328 for _ in range(fragment_repeat):
2330 'url': re.sub(r'{start[ _]time}', compat_str(fragment_ctx['time']), track_url_pattern),
2331 'duration': fragment_ctx['duration'] / stream_timescale,
2333 fragment_ctx['time'] += fragment_ctx['duration']
2337 format_id.append(ism_id)
2339 format_id.append(stream_name)
2340 format_id.append(compat_str(tbr))
2343 'format_id': '-'.join(format_id),
2345 'manifest_url': ism_url,
2346 'ext': 'ismv' if stream_type == 'video' else 'isma',
2350 'asr': sampling_rate,
2351 'vcodec': 'none' if stream_type == 'audio' else fourcc,
2352 'acodec': 'none' if stream_type == 'video' else fourcc,
2354 'fragments': fragments,
2355 '_download_params': {
2356 'duration': duration,
2357 'timescale': stream_timescale,
2358 'width': width or 0,
2359 'height': height or 0,
2361 'codec_private_data': track.get('CodecPrivateData'),
2362 'sampling_rate': sampling_rate,
2363 'channels': int_or_none(track.get('Channels', 2)),
2364 'bits_per_sample': int_or_none(track.get('BitsPerSample', 16)),
2365 'nal_unit_length_field': int_or_none(track.get('NALUnitLengthField', 4)),
2370 def _parse_html5_media_entries(self, base_url, webpage, video_id, m3u8_id=None, m3u8_entry_protocol='m3u8', mpd_id=None, preference=None):
2371 def absolute_url(item_url):
2372 return urljoin(base_url, item_url)
2374 def parse_content_type(content_type):
2375 if not content_type:
2377 ctr = re.search(r'(?P<mimetype>[^/]+/[^;]+)(?:;\s*codecs="?(?P<codecs>[^"]+))?', content_type)
2379 mimetype, codecs = ctr.groups()
2380 f = parse_codecs(codecs)
2381 f['ext'] = mimetype2ext(mimetype)
2385 def _media_formats(src, cur_media_type, type_info={}):
2386 full_url = absolute_url(src)
2387 ext = type_info.get('ext') or determine_ext(full_url)
2389 is_plain_url = False
2390 formats = self._extract_m3u8_formats(
2391 full_url, video_id, ext='mp4',
2392 entry_protocol=m3u8_entry_protocol, m3u8_id=m3u8_id,
2393 preference=preference, fatal=False)
2395 is_plain_url = False
2396 formats = self._extract_mpd_formats(
2397 full_url, video_id, mpd_id=mpd_id, fatal=False)
2402 'vcodec': 'none' if cur_media_type == 'audio' else None,
2404 return is_plain_url, formats
2407 # amp-video and amp-audio are very similar to their HTML5 counterparts
2408 # so we wll include them right here (see
2409 # https://www.ampproject.org/docs/reference/components/amp-video)
2410 media_tags = [(media_tag, media_type, '')
2411 for media_tag, media_type
2412 in re.findall(r'(?s)(<(?:amp-)?(video|audio)[^>]*/>)', webpage)]
2413 media_tags.extend(re.findall(
2414 # We only allow video|audio followed by a whitespace or '>'.
2415 # Allowing more characters may end up in significant slow down (see
2416 # https://github.com/rg3/youtube-dl/issues/11979, example URL:
2417 # http://www.porntrex.com/maps/videositemap.xml).
2418 r'(?s)(<(?P<tag>(?:amp-)?(?:video|audio))(?:\s+[^>]*)?>)(.*?)</(?P=tag)>', webpage))
2419 for media_tag, media_type, media_content in media_tags:
2424 media_attributes = extract_attributes(media_tag)
2425 src = media_attributes.get('src')
2427 _, formats = _media_formats(src, media_type)
2428 media_info['formats'].extend(formats)
2429 media_info['thumbnail'] = absolute_url(media_attributes.get('poster'))
2431 for source_tag in re.findall(r'<source[^>]+>', media_content):
2432 source_attributes = extract_attributes(source_tag)
2433 src = source_attributes.get('src')
2436 f = parse_content_type(source_attributes.get('type'))
2437 is_plain_url, formats = _media_formats(src, media_type, f)
2439 # res attribute is not standard but seen several times
2442 'height': int_or_none(source_attributes.get('res')),
2443 'format_id': source_attributes.get('label'),
2445 f.update(formats[0])
2446 media_info['formats'].append(f)
2448 media_info['formats'].extend(formats)
2449 for track_tag in re.findall(r'<track[^>]+>', media_content):
2450 track_attributes = extract_attributes(track_tag)
2451 kind = track_attributes.get('kind')
2452 if not kind or kind in ('subtitles', 'captions'):
2453 src = track_attributes.get('src')
2456 lang = track_attributes.get('srclang') or track_attributes.get('lang') or track_attributes.get('label')
2457 media_info['subtitles'].setdefault(lang, []).append({
2458 'url': absolute_url(src),
2460 for f in media_info['formats']:
2461 f.setdefault('http_headers', {})['Referer'] = base_url
2462 if media_info['formats'] or media_info['subtitles']:
2463 entries.append(media_info)
2466 def _extract_akamai_formats(self, manifest_url, video_id, hosts={}):
2468 hdcore_sign = 'hdcore=3.7.0'
2469 f4m_url = re.sub(r'(https?://[^/]+)/i/', r'\1/z/', manifest_url).replace('/master.m3u8', '/manifest.f4m')
2470 hds_host = hosts.get('hds')
2472 f4m_url = re.sub(r'(https?://)[^/]+', r'\1' + hds_host, f4m_url)
2473 if 'hdcore=' not in f4m_url:
2474 f4m_url += ('&' if '?' in f4m_url else '?') + hdcore_sign
2475 f4m_formats = self._extract_f4m_formats(
2476 f4m_url, video_id, f4m_id='hds', fatal=False)
2477 for entry in f4m_formats:
2478 entry.update({'extra_param_to_segment_url': hdcore_sign})
2479 formats.extend(f4m_formats)
2480 m3u8_url = re.sub(r'(https?://[^/]+)/z/', r'\1/i/', manifest_url).replace('/manifest.f4m', '/master.m3u8')
2481 hls_host = hosts.get('hls')
2483 m3u8_url = re.sub(r'(https?://)[^/]+', r'\1' + hls_host, m3u8_url)
2484 formats.extend(self._extract_m3u8_formats(
2485 m3u8_url, video_id, 'mp4', 'm3u8_native',
2486 m3u8_id='hls', fatal=False))
2489 def _extract_wowza_formats(self, url, video_id, m3u8_entry_protocol='m3u8_native', skip_protocols=[]):
2490 query = compat_urlparse.urlparse(url).query
2491 url = re.sub(r'/(?:manifest|playlist|jwplayer)\.(?:m3u8|f4m|mpd|smil)', '', url)
2493 r'(?:(?:http|rtmp|rtsp)(?P<s>s)?:)?(?P<url>//[^?]+)', url)
2494 url_base = mobj.group('url')
2495 http_base_url = '%s%s:%s' % ('http', mobj.group('s') or '', url_base)
2498 def manifest_url(manifest):
2499 m_url = '%s/%s' % (http_base_url, manifest)
2501 m_url += '?%s' % query
2504 if 'm3u8' not in skip_protocols:
2505 formats.extend(self._extract_m3u8_formats(
2506 manifest_url('playlist.m3u8'), video_id, 'mp4',
2507 m3u8_entry_protocol, m3u8_id='hls', fatal=False))
2508 if 'f4m' not in skip_protocols:
2509 formats.extend(self._extract_f4m_formats(
2510 manifest_url('manifest.f4m'),
2511 video_id, f4m_id='hds', fatal=False))
2512 if 'dash' not in skip_protocols:
2513 formats.extend(self._extract_mpd_formats(
2514 manifest_url('manifest.mpd'),
2515 video_id, mpd_id='dash', fatal=False))
2516 if re.search(r'(?:/smil:|\.smil)', url_base):
2517 if 'smil' not in skip_protocols:
2518 rtmp_formats = self._extract_smil_formats(
2519 manifest_url('jwplayer.smil'),
2520 video_id, fatal=False)
2521 for rtmp_format in rtmp_formats:
2522 rtsp_format = rtmp_format.copy()
2523 rtsp_format['url'] = '%s/%s' % (rtmp_format['url'], rtmp_format['play_path'])
2524 del rtsp_format['play_path']
2525 del rtsp_format['ext']
2526 rtsp_format.update({
2527 'url': rtsp_format['url'].replace('rtmp://', 'rtsp://'),
2528 'format_id': rtmp_format['format_id'].replace('rtmp', 'rtsp'),
2531 formats.extend([rtmp_format, rtsp_format])
2533 for protocol in ('rtmp', 'rtsp'):
2534 if protocol not in skip_protocols:
2536 'url': '%s:%s' % (protocol, url_base),
2537 'format_id': protocol,
2538 'protocol': protocol,
2542 def _find_jwplayer_data(self, webpage, video_id=None, transform_source=js_to_json):
2544 r'(?s)jwplayer\((?P<quote>[\'"])[^\'" ]+(?P=quote)\)(?!</script>).*?\.setup\s*\((?P<options>[^)]+)\)',
2548 jwplayer_data = self._parse_json(mobj.group('options'),
2550 transform_source=transform_source)
2551 except ExtractorError:
2554 if isinstance(jwplayer_data, dict):
2555 return jwplayer_data
2557 def _extract_jwplayer_data(self, webpage, video_id, *args, **kwargs):
2558 jwplayer_data = self._find_jwplayer_data(
2559 webpage, video_id, transform_source=js_to_json)
2560 return self._parse_jwplayer_data(
2561 jwplayer_data, video_id, *args, **kwargs)
2563 def _parse_jwplayer_data(self, jwplayer_data, video_id=None, require_title=True,
2564 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
2565 # JWPlayer backward compatibility: flattened playlists
2566 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/api/config.js#L81-L96
2567 if 'playlist' not in jwplayer_data:
2568 jwplayer_data = {'playlist': [jwplayer_data]}
2572 # JWPlayer backward compatibility: single playlist item
2573 # https://github.com/jwplayer/jwplayer/blob/v7.7.0/src/js/playlist/playlist.js#L10
2574 if not isinstance(jwplayer_data['playlist'], list):
2575 jwplayer_data['playlist'] = [jwplayer_data['playlist']]
2577 for video_data in jwplayer_data['playlist']:
2578 # JWPlayer backward compatibility: flattened sources
2579 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/playlist/item.js#L29-L35
2580 if 'sources' not in video_data:
2581 video_data['sources'] = [video_data]
2583 this_video_id = video_id or video_data['mediaid']
2585 formats = self._parse_jwplayer_formats(
2586 video_data['sources'], video_id=this_video_id, m3u8_id=m3u8_id,
2587 mpd_id=mpd_id, rtmp_params=rtmp_params, base_url=base_url)
2590 tracks = video_data.get('tracks')
2591 if tracks and isinstance(tracks, list):
2592 for track in tracks:
2593 if not isinstance(track, dict):
2595 track_kind = track.get('kind')
2596 if not track_kind or not isinstance(track_kind, compat_str):
2598 if track_kind.lower() not in ('captions', 'subtitles'):
2600 track_url = urljoin(base_url, track.get('file'))
2603 subtitles.setdefault(track.get('label') or 'en', []).append({
2604 'url': self._proto_relative_url(track_url)
2608 'id': this_video_id,
2609 'title': unescapeHTML(video_data['title'] if require_title else video_data.get('title')),
2610 'description': video_data.get('description'),
2611 'thumbnail': self._proto_relative_url(video_data.get('image')),
2612 'timestamp': int_or_none(video_data.get('pubdate')),
2613 'duration': float_or_none(jwplayer_data.get('duration') or video_data.get('duration')),
2614 'subtitles': subtitles,
2616 # https://github.com/jwplayer/jwplayer/blob/master/src/js/utils/validator.js#L32
2617 if len(formats) == 1 and re.search(r'^(?:http|//).*(?:youtube\.com|youtu\.be)/.+', formats[0]['url']):
2619 '_type': 'url_transparent',
2620 'url': formats[0]['url'],
2623 self._sort_formats(formats)
2624 entry['formats'] = formats
2625 entries.append(entry)
2626 if len(entries) == 1:
2629 return self.playlist_result(entries)
2631 def _parse_jwplayer_formats(self, jwplayer_sources_data, video_id=None,
2632 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
2635 for source in jwplayer_sources_data:
2636 if not isinstance(source, dict):
2638 source_url = self._proto_relative_url(source.get('file'))
2642 source_url = compat_urlparse.urljoin(base_url, source_url)
2643 if source_url in urls:
2645 urls.append(source_url)
2646 source_type = source.get('type') or ''
2647 ext = mimetype2ext(source_type) or determine_ext(source_url)
2648 if source_type == 'hls' or ext == 'm3u8':
2649 formats.extend(self._extract_m3u8_formats(
2650 source_url, video_id, 'mp4', entry_protocol='m3u8_native',
2651 m3u8_id=m3u8_id, fatal=False))
2652 elif source_type == 'dash' or ext == 'mpd':
2653 formats.extend(self._extract_mpd_formats(
2654 source_url, video_id, mpd_id=mpd_id, fatal=False))
2656 formats.extend(self._extract_smil_formats(
2657 source_url, video_id, fatal=False))
2658 # https://github.com/jwplayer/jwplayer/blob/master/src/js/providers/default.js#L67
2659 elif source_type.startswith('audio') or ext in (
2660 'oga', 'aac', 'mp3', 'mpeg', 'vorbis'):
2667 height = int_or_none(source.get('height'))
2669 # Often no height is provided but there is a label in
2670 # format like "1080p", "720p SD", or 1080.
2671 height = int_or_none(self._search_regex(
2672 r'^(\d{3,4})[pP]?(?:\b|$)', compat_str(source.get('label') or ''),
2673 'height', default=None))
2676 'width': int_or_none(source.get('width')),
2678 'tbr': int_or_none(source.get('bitrate')),
2681 if source_url.startswith('rtmp'):
2682 a_format['ext'] = 'flv'
2683 # See com/longtailvideo/jwplayer/media/RTMPMediaProvider.as
2684 # of jwplayer.flash.swf
2685 rtmp_url_parts = re.split(
2686 r'((?:mp4|mp3|flv):)', source_url, 1)
2687 if len(rtmp_url_parts) == 3:
2688 rtmp_url, prefix, play_path = rtmp_url_parts
2691 'play_path': prefix + play_path,
2694 a_format.update(rtmp_params)
2695 formats.append(a_format)
2698 def _live_title(self, name):
2699 """ Generate the title for a live video """
2700 now = datetime.datetime.now()
2701 now_str = now.strftime('%Y-%m-%d %H:%M')
2702 return name + ' ' + now_str
2704 def _int(self, v, name, fatal=False, **kwargs):
2705 res = int_or_none(v, **kwargs)
2706 if 'get_attr' in kwargs:
2707 print(getattr(v, kwargs['get_attr']))
2709 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
2711 raise ExtractorError(msg)
2713 self._downloader.report_warning(msg)
2716 def _float(self, v, name, fatal=False, **kwargs):
2717 res = float_or_none(v, **kwargs)
2719 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
2721 raise ExtractorError(msg)
2723 self._downloader.report_warning(msg)
2726 def _set_cookie(self, domain, name, value, expire_time=None, port=None,
2727 path='/', secure=False, discard=False, rest={}, **kwargs):
2728 cookie = compat_cookiejar.Cookie(
2729 0, name, value, port, port is not None, domain, True,
2730 domain.startswith('.'), path, True, secure, expire_time,
2731 discard, None, None, rest)
2732 self._downloader.cookiejar.set_cookie(cookie)
2734 def _get_cookies(self, url):
2735 """ Return a compat_cookies.SimpleCookie with the cookies for the url """
2736 req = sanitized_Request(url)
2737 self._downloader.cookiejar.add_cookie_header(req)
2738 return compat_cookies.SimpleCookie(req.get_header('Cookie'))
2740 def get_testcases(self, include_onlymatching=False):
2741 t = getattr(self, '_TEST', None)
2743 assert not hasattr(self, '_TESTS'), \
2744 '%s has _TEST and _TESTS' % type(self).__name__
2747 tests = getattr(self, '_TESTS', [])
2749 if not include_onlymatching and t.get('only_matching', False):
2751 t['name'] = type(self).__name__[:-len('IE')]
2754 def is_suitable(self, age_limit):
2755 """ Test whether the extractor is generally suitable for the given
2756 age limit (i.e. pornographic sites are not, all others usually are) """
2758 any_restricted = False
2759 for tc in self.get_testcases(include_onlymatching=False):
2760 if tc.get('playlist', []):
2761 tc = tc['playlist'][0]
2762 is_restricted = age_restricted(
2763 tc.get('info_dict', {}).get('age_limit'), age_limit)
2764 if not is_restricted:
2766 any_restricted = any_restricted or is_restricted
2767 return not any_restricted
2769 def extract_subtitles(self, *args, **kwargs):
2770 if (self._downloader.params.get('writesubtitles', False) or
2771 self._downloader.params.get('listsubtitles')):
2772 return self._get_subtitles(*args, **kwargs)
2775 def _get_subtitles(self, *args, **kwargs):
2776 raise NotImplementedError('This method must be implemented by subclasses')
2779 def _merge_subtitle_items(subtitle_list1, subtitle_list2):
2780 """ Merge subtitle items for one language. Items with duplicated URLs
2781 will be dropped. """
2782 list1_urls = set([item['url'] for item in subtitle_list1])
2783 ret = list(subtitle_list1)
2784 ret.extend([item for item in subtitle_list2 if item['url'] not in list1_urls])
2788 def _merge_subtitles(cls, subtitle_dict1, subtitle_dict2):
2789 """ Merge two subtitle dictionaries, language by language. """
2790 ret = dict(subtitle_dict1)
2791 for lang in subtitle_dict2:
2792 ret[lang] = cls._merge_subtitle_items(subtitle_dict1.get(lang, []), subtitle_dict2[lang])
2795 def extract_automatic_captions(self, *args, **kwargs):
2796 if (self._downloader.params.get('writeautomaticsub', False) or
2797 self._downloader.params.get('listsubtitles')):
2798 return self._get_automatic_captions(*args, **kwargs)
2801 def _get_automatic_captions(self, *args, **kwargs):
2802 raise NotImplementedError('This method must be implemented by subclasses')
2804 def mark_watched(self, *args, **kwargs):
2805 if (self._downloader.params.get('mark_watched', False) and
2806 (self._get_login_info()[0] is not None or
2807 self._downloader.params.get('cookiefile') is not None)):
2808 self._mark_watched(*args, **kwargs)
2810 def _mark_watched(self, *args, **kwargs):
2811 raise NotImplementedError('This method must be implemented by subclasses')
2813 def geo_verification_headers(self):
2815 geo_verification_proxy = self._downloader.params.get('geo_verification_proxy')
2816 if geo_verification_proxy:
2817 headers['Ytdl-request-proxy'] = geo_verification_proxy
2820 def _generic_id(self, url):
2821 return compat_urllib_parse_unquote(os.path.splitext(url.rstrip('/').split('/')[-1])[0])
2823 def _generic_title(self, url):
2824 return compat_urllib_parse_unquote(os.path.splitext(url_basename(url))[0])
2827 class SearchInfoExtractor(InfoExtractor):
2829 Base class for paged search queries extractors.
2830 They accept URLs in the format _SEARCH_KEY(|all|[0-9]):{query}
2831 Instances should define _SEARCH_KEY and _MAX_RESULTS.
2835 def _make_valid_url(cls):
2836 return r'%s(?P<prefix>|[1-9][0-9]*|all):(?P<query>[\s\S]+)' % cls._SEARCH_KEY
2839 def suitable(cls, url):
2840 return re.match(cls._make_valid_url(), url) is not None
2842 def _real_extract(self, query):
2843 mobj = re.match(self._make_valid_url(), query)
2845 raise ExtractorError('Invalid search query "%s"' % query)
2847 prefix = mobj.group('prefix')
2848 query = mobj.group('query')
2850 return self._get_n_results(query, 1)
2851 elif prefix == 'all':
2852 return self._get_n_results(query, self._MAX_RESULTS)
2856 raise ExtractorError('invalid download number %s for query "%s"' % (n, query))
2857 elif n > self._MAX_RESULTS:
2858 self._downloader.report_warning('%s returns max %i results (you requested %i)' % (self._SEARCH_KEY, self._MAX_RESULTS, n))
2859 n = self._MAX_RESULTS
2860 return self._get_n_results(query, n)
2862 def _get_n_results(self, query, n):
2863 """Get a specified number of results for a query"""
2864 raise NotImplementedError('This method must be implemented by subclasses')
2867 def SEARCH_KEY(self):
2868 return self._SEARCH_KEY