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,
79 class InfoExtractor(object):
80 """Information Extractor class.
82 Information extractors are the classes that, given a URL, extract
83 information about the video (or videos) the URL refers to. This
84 information includes the real video URL, the video title, author and
85 others. The information is stored in a dictionary which is then
86 passed to the YoutubeDL. The YoutubeDL processes this
87 information possibly downloading the video to the file system, among
88 other possible outcomes.
90 The type field determines the type of the result.
91 By far the most common value (and the default if _type is missing) is
92 "video", which indicates a single video.
94 For a video, the dictionaries must include the following fields:
97 title: Video title, unescaped.
99 Additionally, it must contain either a formats entry or a url one:
101 formats: A list of dictionaries for each format available, ordered
102 from worst to best quality.
105 * url Mandatory. The URL of the video file
107 The URL of the manifest file in case of
108 fragmented media (DASH, hls, hds)
109 * ext Will be calculated from URL if missing
110 * format A human-readable description of the format
111 ("mp4 container with h264/opus").
112 Calculated from the format_id, width, height.
113 and format_note fields if missing.
114 * format_id A short description of the format
115 ("mp4_h264_opus" or "19").
116 Technically optional, but strongly recommended.
117 * format_note Additional info about the format
118 ("3D" or "DASH video")
119 * width Width of the video, if known
120 * height Height of the video, if known
121 * resolution Textual description of width and height
122 * tbr Average bitrate of audio and video in KBit/s
123 * abr Average audio bitrate in KBit/s
124 * acodec Name of the audio codec in use
125 * asr Audio sampling rate in Hertz
126 * vbr Average video bitrate in KBit/s
128 * vcodec Name of the video codec in use
129 * container Name of the container format
130 * filesize The number of bytes, if known in advance
131 * filesize_approx An estimate for the number of bytes
132 * player_url SWF Player URL (used for rtmpdump).
133 * protocol The protocol that will be used for the actual
134 download, lower-case.
135 "http", "https", "rtsp", "rtmp", "rtmpe",
136 "m3u8", "m3u8_native" or "http_dash_segments".
138 Base URL for fragments. Each fragment's path
139 value (if present) will be relative to
141 * fragments A list of fragments of a fragmented media.
142 Each fragment entry must contain either an url
143 or a path. If an url is present it should be
144 considered by a client. Otherwise both path and
145 fragment_base_url must be present. Here is
146 the list of all potential fields:
147 * "url" - fragment's URL
148 * "path" - fragment's path relative to
150 * "duration" (optional, int or float)
151 * "filesize" (optional, int)
152 * preference Order number of this format. If this field is
153 present and not None, the formats get sorted
154 by this field, regardless of all other values.
155 -1 for default (order by other properties),
156 -2 or smaller for less than default.
157 < -1000 to hide the format (if there is
158 another one which is strictly better)
159 * language Language code, e.g. "de" or "en-US".
160 * language_preference Is this in the language mentioned in
162 10 if it's what the URL is about,
163 -1 for default (don't know),
164 -10 otherwise, other values reserved for now.
165 * quality Order number of the video quality of this
166 format, irrespective of the file format.
167 -1 for default (order by other properties),
168 -2 or smaller for less than default.
169 * source_preference Order number for this video source
170 (quality takes higher priority)
171 -1 for default (order by other properties),
172 -2 or smaller for less than default.
173 * http_headers A dictionary of additional HTTP headers
174 to add to the request.
175 * stretched_ratio If given and not 1, indicates that the
176 video's pixels are not square.
177 width : height ratio as float.
178 * no_resume The server does not support resuming the
179 (HTTP or RTMP) download. Boolean.
180 * downloader_options A dictionary of downloader options as
181 described in FileDownloader
183 url: Final video URL.
184 ext: Video filename extension.
185 format: The video format, defaults to ext (used for --get-format)
186 player_url: SWF Player URL (used for rtmpdump).
188 The following fields are optional:
190 alt_title: A secondary title of the video.
191 display_id An alternative identifier for the video, not necessarily
192 unique, but available before title. Typically, id is
193 something like "4234987", title "Dancing naked mole rats",
194 and display_id "dancing-naked-mole-rats"
195 thumbnails: A list of dictionaries, with the following entries:
196 * "id" (optional, string) - Thumbnail format ID
198 * "preference" (optional, int) - quality of the image
199 * "width" (optional, int)
200 * "height" (optional, int)
201 * "resolution" (optional, string "{width}x{height"},
203 * "filesize" (optional, int)
204 thumbnail: Full URL to a video thumbnail image.
205 description: Full video description.
206 uploader: Full name of the video uploader.
207 license: License name the video is licensed under.
208 creator: The creator of the video.
209 release_date: The date (YYYYMMDD) when the video was released.
210 timestamp: UNIX timestamp of the moment the video became available.
211 upload_date: Video upload date (YYYYMMDD).
212 If not explicitly set, calculated from timestamp.
213 uploader_id: Nickname or id of the video uploader.
214 uploader_url: Full URL to a personal webpage of the video uploader.
215 channel: Full name of the channel the video is uploaded on.
216 Note that channel fields may or may not repeat uploader
217 fields. This depends on a particular extractor.
218 channel_id: Id of the channel.
219 channel_url: Full URL to a channel webpage.
220 location: Physical location where the video was filmed.
221 subtitles: The available subtitles as a dictionary in the format
222 {tag: subformats}. "tag" is usually a language code, and
223 "subformats" is a list sorted from lower to higher
224 preference, each element is a dictionary with the "ext"
226 * "data": The subtitles file contents
227 * "url": A URL pointing to the subtitles file
228 "ext" will be calculated from URL if missing
229 automatic_captions: Like 'subtitles', used by the YoutubeIE for
230 automatically generated captions
231 duration: Length of the video in seconds, as an integer or float.
232 view_count: How many users have watched the video on the platform.
233 like_count: Number of positive ratings of the video
234 dislike_count: Number of negative ratings of the video
235 repost_count: Number of reposts of the video
236 average_rating: Average rating give by users, the scale used depends on the webpage
237 comment_count: Number of comments on the video
238 comments: A list of comments, each with one or more of the following
239 properties (all but one of text or html optional):
240 * "author" - human-readable name of the comment author
241 * "author_id" - user ID of the comment author
243 * "html" - Comment as HTML
244 * "text" - Plain text of the comment
245 * "timestamp" - UNIX timestamp of comment
246 * "parent" - ID of the comment this one is replying to.
247 Set to "root" to indicate that this is a
248 comment to the original video.
249 age_limit: Age restriction for the video, as an integer (years)
250 webpage_url: The URL to the video webpage, if given to youtube-dl it
251 should allow to get the same result again. (It will be set
252 by YoutubeDL if it's missing)
253 categories: A list of categories that the video falls in, for example
255 tags: A list of tags assigned to the video, e.g. ["sweden", "pop music"]
256 is_live: True, False, or None (=unknown). Whether this video is a
257 live stream that goes on instead of a fixed-length video.
258 start_time: Time in seconds where the reproduction should start, as
259 specified in the URL.
260 end_time: Time in seconds where the reproduction should end, as
261 specified in the URL.
262 chapters: A list of dictionaries, with the following entries:
263 * "start_time" - The start time of the chapter in seconds
264 * "end_time" - The end time of the chapter in seconds
265 * "title" (optional, string)
267 The following fields should only be used when the video belongs to some logical
270 chapter: Name or title of the chapter the video belongs to.
271 chapter_number: Number of the chapter the video belongs to, as an integer.
272 chapter_id: Id of the chapter the video belongs to, as a unicode string.
274 The following fields should only be used when the video is an episode of some
275 series, programme or podcast:
277 series: Title of the series or programme the video episode belongs to.
278 season: Title of the season the video episode belongs to.
279 season_number: Number of the season the video episode belongs to, as an integer.
280 season_id: Id of the season the video episode belongs to, as a unicode string.
281 episode: Title of the video episode. Unlike mandatory video title field,
282 this field should denote the exact title of the video episode
283 without any kind of decoration.
284 episode_number: Number of the video episode within a season, as an integer.
285 episode_id: Id of the video episode, as a unicode string.
287 The following fields should only be used when the media is a track or a part of
290 track: Title of the track.
291 track_number: Number of the track within an album or a disc, as an integer.
292 track_id: Id of the track (useful in case of custom indexing, e.g. 6.iii),
294 artist: Artist(s) of the track.
295 genre: Genre(s) of the track.
296 album: Title of the album the track belongs to.
297 album_type: Type of the album (e.g. "Demo", "Full-length", "Split", "Compilation", etc).
298 album_artist: List of all artists appeared on the album (e.g.
299 "Ash Borer / Fell Voices" or "Various Artists", useful for splits
301 disc_number: Number of the disc or other physical medium the track belongs to,
303 release_year: Year (YYYY) when the album was released.
305 Unless mentioned otherwise, the fields should be Unicode strings.
307 Unless mentioned otherwise, None is equivalent to absence of information.
310 _type "playlist" indicates multiple videos.
311 There must be a key "entries", which is a list, an iterable, or a PagedList
312 object, each element of which is a valid dictionary by this specification.
314 Additionally, playlists can have "id", "title", "description", "uploader",
315 "uploader_id", "uploader_url" attributes with the same semantics as videos
319 _type "multi_video" indicates that there are multiple videos that
320 form a single show, for examples multiple acts of an opera or TV episode.
321 It must have an entries key like a playlist and contain all the keys
322 required for a video at the same time.
325 _type "url" indicates that the video must be extracted from another
326 location, possibly by a different extractor. Its only required key is:
327 "url" - the next URL to extract.
328 The key "ie_key" can be set to the class name (minus the trailing "IE",
329 e.g. "Youtube") if the extractor class is known in advance.
330 Additionally, the dictionary may have any properties of the resolved entity
331 known in advance, for example "title" if the title of the referred video is
335 _type "url_transparent" entities have the same specification as "url", but
336 indicate that the given additional information is more precise than the one
337 associated with the resolved URL.
338 This is useful when a site employs a video service that hosts the video and
339 its technical metadata, but that video service does not embed a useful
340 title, description etc.
343 Subclasses of this one should re-define the _real_initialize() and
344 _real_extract() methods and define a _VALID_URL regexp.
345 Probably, they should also be added to the list of extractors.
347 _GEO_BYPASS attribute may be set to False in order to disable
348 geo restriction bypass mechanisms for a particular extractor.
349 Though it won't disable explicit geo restriction bypass based on
350 country code provided with geo_bypass_country.
352 _GEO_COUNTRIES attribute may contain a list of presumably geo unrestricted
353 countries for this extractor. One of these countries will be used by
354 geo restriction bypass mechanism right away in order to bypass
355 geo restriction, of course, if the mechanism is not disabled.
357 _GEO_IP_BLOCKS attribute may contain a list of presumably geo unrestricted
358 IP blocks in CIDR notation for this extractor. One of these IP blocks
359 will be used by geo restriction bypass mechanism similarly
362 Finally, the _WORKING attribute should be set to False for broken IEs
363 in order to warn the users and skip the tests.
368 _x_forwarded_for_ip = None
370 _GEO_COUNTRIES = None
371 _GEO_IP_BLOCKS = None
374 def __init__(self, downloader=None):
375 """Constructor. Receives an optional downloader."""
377 self._x_forwarded_for_ip = None
378 self.set_downloader(downloader)
381 def suitable(cls, url):
382 """Receives a URL and returns True if suitable for this IE."""
384 # This does not use has/getattr intentionally - we want to know whether
385 # we have cached the regexp for *this* class, whereas getattr would also
386 # match the superclass
387 if '_VALID_URL_RE' not in cls.__dict__:
388 cls._VALID_URL_RE = re.compile(cls._VALID_URL)
389 return cls._VALID_URL_RE.match(url) is not None
392 def _match_id(cls, url):
393 if '_VALID_URL_RE' not in cls.__dict__:
394 cls._VALID_URL_RE = re.compile(cls._VALID_URL)
395 m = cls._VALID_URL_RE.match(url)
397 return compat_str(m.group('id'))
401 """Getter method for _WORKING."""
404 def initialize(self):
405 """Initializes an instance (authentication, etc)."""
406 self._initialize_geo_bypass({
407 'countries': self._GEO_COUNTRIES,
408 'ip_blocks': self._GEO_IP_BLOCKS,
411 self._real_initialize()
414 def _initialize_geo_bypass(self, geo_bypass_context):
416 Initialize geo restriction bypass mechanism.
418 This method is used to initialize geo bypass mechanism based on faking
419 X-Forwarded-For HTTP header. A random country from provided country list
420 is selected and a random IP belonging to this country is generated. This
421 IP will be passed as X-Forwarded-For HTTP header in all subsequent
424 This method will be used for initial geo bypass mechanism initialization
425 during the instance initialization with _GEO_COUNTRIES and
428 You may also manually call it from extractor's code if geo bypass
429 information is not available beforehand (e.g. obtained during
430 extraction) or due to some other reason. In this case you should pass
431 this information in geo bypass context passed as first argument. It may
432 contain following fields:
434 countries: List of geo unrestricted countries (similar
436 ip_blocks: List of geo unrestricted IP blocks in CIDR notation
437 (similar to _GEO_IP_BLOCKS)
440 if not self._x_forwarded_for_ip:
442 # Geo bypass mechanism is explicitly disabled by user
443 if not self._downloader.params.get('geo_bypass', True):
446 if not geo_bypass_context:
447 geo_bypass_context = {}
449 # Backward compatibility: previously _initialize_geo_bypass
450 # expected a list of countries, some 3rd party code may still use
452 if isinstance(geo_bypass_context, (list, tuple)):
453 geo_bypass_context = {
454 'countries': geo_bypass_context,
457 # The whole point of geo bypass mechanism is to fake IP
458 # as X-Forwarded-For HTTP header based on some IP block or
461 # Path 1: bypassing based on IP block in CIDR notation
463 # Explicit IP block specified by user, use it right away
464 # regardless of whether extractor is geo bypassable or not
465 ip_block = self._downloader.params.get('geo_bypass_ip_block', None)
467 # Otherwise use random IP block from geo bypass context but only
468 # if extractor is known as geo bypassable
470 ip_blocks = geo_bypass_context.get('ip_blocks')
471 if self._GEO_BYPASS and ip_blocks:
472 ip_block = random.choice(ip_blocks)
475 self._x_forwarded_for_ip = GeoUtils.random_ipv4(ip_block)
476 if self._downloader.params.get('verbose', False):
477 self._downloader.to_screen(
478 '[debug] Using fake IP %s as X-Forwarded-For.'
479 % self._x_forwarded_for_ip)
482 # Path 2: bypassing based on country code
484 # Explicit country code specified by user, use it right away
485 # regardless of whether extractor is geo bypassable or not
486 country = self._downloader.params.get('geo_bypass_country', None)
488 # Otherwise use random country code from geo bypass context but
489 # only if extractor is known as geo bypassable
491 countries = geo_bypass_context.get('countries')
492 if self._GEO_BYPASS and countries:
493 country = random.choice(countries)
496 self._x_forwarded_for_ip = GeoUtils.random_ipv4(country)
497 if self._downloader.params.get('verbose', False):
498 self._downloader.to_screen(
499 '[debug] Using fake IP %s (%s) as X-Forwarded-For.'
500 % (self._x_forwarded_for_ip, country.upper()))
502 def extract(self, url):
503 """Extracts URL information and returns it in list of dicts."""
508 ie_result = self._real_extract(url)
509 if self._x_forwarded_for_ip:
510 ie_result['__x_forwarded_for_ip'] = self._x_forwarded_for_ip
512 except GeoRestrictedError as e:
513 if self.__maybe_fake_ip_and_retry(e.countries):
516 except ExtractorError:
518 except compat_http_client.IncompleteRead as e:
519 raise ExtractorError('A network error has occurred.', cause=e, expected=True)
520 except (KeyError, StopIteration) as e:
521 raise ExtractorError('An extractor error has occurred.', cause=e)
523 def __maybe_fake_ip_and_retry(self, countries):
524 if (not self._downloader.params.get('geo_bypass_country', None) and
526 self._downloader.params.get('geo_bypass', True) and
527 not self._x_forwarded_for_ip and
529 country_code = random.choice(countries)
530 self._x_forwarded_for_ip = GeoUtils.random_ipv4(country_code)
531 if self._x_forwarded_for_ip:
533 'Video is geo restricted. Retrying extraction with fake IP %s (%s) as X-Forwarded-For.'
534 % (self._x_forwarded_for_ip, country_code.upper()))
538 def set_downloader(self, downloader):
539 """Sets the downloader for this IE."""
540 self._downloader = downloader
542 def _real_initialize(self):
543 """Real initialization process. Redefine in subclasses."""
546 def _real_extract(self, url):
547 """Real extraction process. Redefine in subclasses."""
552 """A string for getting the InfoExtractor with get_info_extractor"""
553 return compat_str(cls.__name__[:-2])
557 return compat_str(type(self).__name__[:-2])
560 def __can_accept_status_code(err, expected_status):
561 assert isinstance(err, compat_urllib_error.HTTPError)
562 if expected_status is None:
564 if isinstance(expected_status, compat_integer_types):
565 return err.code == expected_status
566 elif isinstance(expected_status, (list, tuple)):
567 return err.code in expected_status
568 elif callable(expected_status):
569 return expected_status(err.code) is True
573 def _request_webpage(self, url_or_request, video_id, note=None, errnote=None, fatal=True, data=None, headers={}, query={}, expected_status=None):
575 Return the response handle.
577 See _download_webpage docstring for arguments specification.
580 self.report_download_webpage(video_id)
581 elif note is not False:
583 self.to_screen('%s' % (note,))
585 self.to_screen('%s: %s' % (video_id, note))
587 # Some sites check X-Forwarded-For HTTP header in order to figure out
588 # the origin of the client behind proxy. This allows bypassing geo
589 # restriction by faking this header's value to IP that belongs to some
590 # geo unrestricted country. We will do so once we encounter any
591 # geo restriction error.
592 if self._x_forwarded_for_ip:
593 if 'X-Forwarded-For' not in headers:
594 headers['X-Forwarded-For'] = self._x_forwarded_for_ip
596 if isinstance(url_or_request, compat_urllib_request.Request):
597 url_or_request = update_Request(
598 url_or_request, data=data, headers=headers, query=query)
601 url_or_request = update_url_query(url_or_request, query)
602 if data is not None or headers:
603 url_or_request = sanitized_Request(url_or_request, data, headers)
605 return self._downloader.urlopen(url_or_request)
606 except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err:
607 if isinstance(err, compat_urllib_error.HTTPError):
608 if self.__can_accept_status_code(err, expected_status):
609 # Retain reference to error to prevent file object from
610 # being closed before it can be read. Works around the
611 # effects of <https://bugs.python.org/issue15002>
612 # introduced in Python 3.4.1.
619 errnote = 'Unable to download webpage'
621 errmsg = '%s: %s' % (errnote, error_to_compat_str(err))
623 raise ExtractorError(errmsg, sys.exc_info()[2], cause=err)
625 self._downloader.report_warning(errmsg)
628 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):
630 Return a tuple (page content as string, URL handle).
632 See _download_webpage docstring for arguments specification.
634 # Strip hashes from the URL (#1038)
635 if isinstance(url_or_request, (compat_str, str)):
636 url_or_request = url_or_request.partition('#')[0]
638 urlh = self._request_webpage(url_or_request, video_id, note, errnote, fatal, data=data, headers=headers, query=query, expected_status=expected_status)
642 content = self._webpage_read_content(urlh, url_or_request, video_id, note, errnote, fatal, encoding=encoding)
643 return (content, urlh)
646 def _guess_encoding_from_content(content_type, webpage_bytes):
647 m = re.match(r'[a-zA-Z0-9_.-]+/[a-zA-Z0-9_.-]+\s*;\s*charset=(.+)', content_type)
649 encoding = m.group(1)
651 m = re.search(br'<meta[^>]+charset=[\'"]?([^\'")]+)[ /\'">]',
652 webpage_bytes[:1024])
654 encoding = m.group(1).decode('ascii')
655 elif webpage_bytes.startswith(b'\xff\xfe'):
662 def __check_blocked(self, content):
663 first_block = content[:512]
664 if ('<title>Access to this site is blocked</title>' in content and
665 'Websense' in first_block):
666 msg = 'Access to this webpage has been blocked by Websense filtering software in your network.'
667 blocked_iframe = self._html_search_regex(
668 r'<iframe src="([^"]+)"', content,
669 'Websense information URL', default=None)
671 msg += ' Visit %s for more details' % blocked_iframe
672 raise ExtractorError(msg, expected=True)
673 if '<title>The URL you requested has been blocked</title>' in first_block:
675 'Access to this webpage has been blocked by Indian censorship. '
676 'Use a VPN or proxy server (with --proxy) to route around it.')
677 block_msg = self._html_search_regex(
678 r'</h1><p>(.*?)</p>',
679 content, 'block message', default=None)
681 msg += ' (Message: "%s")' % block_msg.replace('\n', ' ')
682 raise ExtractorError(msg, expected=True)
683 if ('<title>TTK :: Доступ к ресурсу ограничен</title>' in content and
684 'blocklist.rkn.gov.ru' in content):
685 raise ExtractorError(
686 'Access to this webpage has been blocked by decision of the Russian government. '
687 'Visit http://blocklist.rkn.gov.ru/ for a block reason.',
690 def _webpage_read_content(self, urlh, url_or_request, video_id, note=None, errnote=None, fatal=True, prefix=None, encoding=None):
691 content_type = urlh.headers.get('Content-Type', '')
692 webpage_bytes = urlh.read()
693 if prefix is not None:
694 webpage_bytes = prefix + webpage_bytes
696 encoding = self._guess_encoding_from_content(content_type, webpage_bytes)
697 if self._downloader.params.get('dump_intermediate_pages', False):
698 self.to_screen('Dumping request to ' + urlh.geturl())
699 dump = base64.b64encode(webpage_bytes).decode('ascii')
700 self._downloader.to_screen(dump)
701 if self._downloader.params.get('write_pages', False):
702 basen = '%s_%s' % (video_id, urlh.geturl())
704 h = '___' + hashlib.md5(basen.encode('utf-8')).hexdigest()
705 basen = basen[:240 - len(h)] + h
706 raw_filename = basen + '.dump'
707 filename = sanitize_filename(raw_filename, restricted=True)
708 self.to_screen('Saving request to ' + filename)
709 # Working around MAX_PATH limitation on Windows (see
710 # http://msdn.microsoft.com/en-us/library/windows/desktop/aa365247(v=vs.85).aspx)
711 if compat_os_name == 'nt':
712 absfilepath = os.path.abspath(filename)
713 if len(absfilepath) > 259:
714 filename = '\\\\?\\' + absfilepath
715 with open(filename, 'wb') as outf:
716 outf.write(webpage_bytes)
719 content = webpage_bytes.decode(encoding, 'replace')
721 content = webpage_bytes.decode('utf-8', 'replace')
723 self.__check_blocked(content)
727 def _download_webpage(
728 self, url_or_request, video_id, note=None, errnote=None,
729 fatal=True, tries=1, timeout=5, encoding=None, data=None,
730 headers={}, query={}, expected_status=None):
732 Return the data of the page as a string.
735 url_or_request -- plain text URL as a string or
736 a compat_urllib_request.Requestobject
737 video_id -- Video/playlist/item identifier (string)
740 note -- note printed before downloading (string)
741 errnote -- note printed in case of an error (string)
742 fatal -- flag denoting whether error should be considered fatal,
743 i.e. whether it should cause ExtractionError to be raised,
744 otherwise a warning will be reported and extraction continued
745 tries -- number of tries
746 timeout -- sleep interval between tries
747 encoding -- encoding for a page content decoding, guessed automatically
748 when not explicitly specified
749 data -- POST data (bytes)
750 headers -- HTTP headers (dict)
751 query -- URL query (dict)
752 expected_status -- allows to accept failed HTTP requests (non 2xx
753 status code) by explicitly specifying a set of accepted status
754 codes. Can be any of the following entities:
755 - an integer type specifying an exact failed status code to
757 - a list or a tuple of integer types specifying a list of
758 failed status codes to accept
759 - a callable accepting an actual failed status code and
760 returning True if it should be accepted
761 Note that this argument does not affect success status codes (2xx)
762 which are always accepted.
767 while success is False:
769 res = self._download_webpage_handle(
770 url_or_request, video_id, note, errnote, fatal,
771 encoding=encoding, data=data, headers=headers, query=query,
772 expected_status=expected_status)
774 except compat_http_client.IncompleteRead as e:
776 if try_count >= tries:
778 self._sleep(timeout, video_id)
785 def _download_xml_handle(
786 self, url_or_request, video_id, note='Downloading XML',
787 errnote='Unable to download XML', transform_source=None,
788 fatal=True, encoding=None, data=None, headers={}, query={},
789 expected_status=None):
791 Return a tuple (xml as an xml.etree.ElementTree.Element, URL handle).
793 See _download_webpage docstring for arguments specification.
795 res = self._download_webpage_handle(
796 url_or_request, video_id, note, errnote, fatal=fatal,
797 encoding=encoding, data=data, headers=headers, query=query,
798 expected_status=expected_status)
801 xml_string, urlh = res
802 return self._parse_xml(
803 xml_string, video_id, transform_source=transform_source,
807 self, url_or_request, video_id,
808 note='Downloading XML', errnote='Unable to download XML',
809 transform_source=None, fatal=True, encoding=None,
810 data=None, headers={}, query={}, expected_status=None):
812 Return the xml as an xml.etree.ElementTree.Element.
814 See _download_webpage docstring for arguments specification.
816 res = self._download_xml_handle(
817 url_or_request, video_id, note=note, errnote=errnote,
818 transform_source=transform_source, fatal=fatal, encoding=encoding,
819 data=data, headers=headers, query=query,
820 expected_status=expected_status)
821 return res if res is False else res[0]
823 def _parse_xml(self, xml_string, video_id, transform_source=None, fatal=True):
825 xml_string = transform_source(xml_string)
827 return compat_etree_fromstring(xml_string.encode('utf-8'))
828 except compat_xml_parse_error as ve:
829 errmsg = '%s: Failed to parse XML ' % video_id
831 raise ExtractorError(errmsg, cause=ve)
833 self.report_warning(errmsg + str(ve))
835 def _download_json_handle(
836 self, url_or_request, video_id, note='Downloading JSON metadata',
837 errnote='Unable to download JSON metadata', transform_source=None,
838 fatal=True, encoding=None, data=None, headers={}, query={},
839 expected_status=None):
841 Return a tuple (JSON object, URL handle).
843 See _download_webpage docstring for arguments specification.
845 res = self._download_webpage_handle(
846 url_or_request, video_id, note, errnote, fatal=fatal,
847 encoding=encoding, data=data, headers=headers, query=query,
848 expected_status=expected_status)
851 json_string, urlh = res
852 return self._parse_json(
853 json_string, video_id, transform_source=transform_source,
857 self, url_or_request, video_id, note='Downloading JSON metadata',
858 errnote='Unable to download JSON metadata', transform_source=None,
859 fatal=True, encoding=None, data=None, headers={}, query={},
860 expected_status=None):
862 Return the JSON object as a dict.
864 See _download_webpage docstring for arguments specification.
866 res = self._download_json_handle(
867 url_or_request, video_id, note=note, errnote=errnote,
868 transform_source=transform_source, fatal=fatal, encoding=encoding,
869 data=data, headers=headers, query=query,
870 expected_status=expected_status)
871 return res if res is False else res[0]
873 def _parse_json(self, json_string, video_id, transform_source=None, fatal=True):
875 json_string = transform_source(json_string)
877 return json.loads(json_string)
878 except ValueError as ve:
879 errmsg = '%s: Failed to parse JSON ' % video_id
881 raise ExtractorError(errmsg, cause=ve)
883 self.report_warning(errmsg + str(ve))
885 def report_warning(self, msg, video_id=None):
886 idstr = '' if video_id is None else '%s: ' % video_id
887 self._downloader.report_warning(
888 '[%s] %s%s' % (self.IE_NAME, idstr, msg))
890 def to_screen(self, msg):
891 """Print msg to screen, prefixing it with '[ie_name]'"""
892 self._downloader.to_screen('[%s] %s' % (self.IE_NAME, msg))
894 def report_extraction(self, id_or_name):
895 """Report information extraction."""
896 self.to_screen('%s: Extracting information' % id_or_name)
898 def report_download_webpage(self, video_id):
899 """Report webpage download."""
900 self.to_screen('%s: Downloading webpage' % video_id)
902 def report_age_confirmation(self):
903 """Report attempt to confirm age."""
904 self.to_screen('Confirming age')
906 def report_login(self):
907 """Report attempt to log in."""
908 self.to_screen('Logging in')
911 def raise_login_required(msg='This video is only available for registered users'):
912 raise ExtractorError(
913 '%s. Use --username and --password or --netrc to provide account credentials.' % msg,
917 def raise_geo_restricted(msg='This video is not available from your location due to geo restriction', countries=None):
918 raise GeoRestrictedError(msg, countries=countries)
920 # Methods for following #608
922 def url_result(url, ie=None, video_id=None, video_title=None):
923 """Returns a URL that points to a page that should be processed"""
924 # TODO: ie should be the class used for getting the info
925 video_info = {'_type': 'url',
928 if video_id is not None:
929 video_info['id'] = video_id
930 if video_title is not None:
931 video_info['title'] = video_title
934 def playlist_from_matches(self, matches, playlist_id=None, playlist_title=None, getter=None, ie=None):
936 self.url_result(self._proto_relative_url(getter(m) if getter else m), ie)
938 return self.playlist_result(
939 urls, playlist_id=playlist_id, playlist_title=playlist_title)
942 def playlist_result(entries, playlist_id=None, playlist_title=None, playlist_description=None):
943 """Returns a playlist"""
944 video_info = {'_type': 'playlist',
947 video_info['id'] = playlist_id
949 video_info['title'] = playlist_title
950 if playlist_description:
951 video_info['description'] = playlist_description
954 def _search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
956 Perform a regex search on the given string, using a single or a list of
957 patterns returning the first matching group.
958 In case of failure return a default value or raise a WARNING or a
959 RegexNotFoundError, depending on fatal, specifying the field name.
961 if isinstance(pattern, (str, compat_str, compiled_regex_type)):
962 mobj = re.search(pattern, string, flags)
965 mobj = re.search(p, string, flags)
969 if not self._downloader.params.get('no_color') and compat_os_name != 'nt' and sys.stderr.isatty():
970 _name = '\033[0;34m%s\033[0m' % name
976 # return the first matching group
977 return next(g for g in mobj.groups() if g is not None)
979 return mobj.group(group)
980 elif default is not NO_DEFAULT:
983 raise RegexNotFoundError('Unable to extract %s' % _name)
985 self._downloader.report_warning('unable to extract %s' % _name + bug_reports_message())
988 def _html_search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
990 Like _search_regex, but strips HTML tags and unescapes entities.
992 res = self._search_regex(pattern, string, name, default, fatal, flags, group)
994 return clean_html(res).strip()
998 def _get_netrc_login_info(self, netrc_machine=None):
1001 netrc_machine = netrc_machine or self._NETRC_MACHINE
1003 if self._downloader.params.get('usenetrc', False):
1005 info = netrc.netrc().authenticators(netrc_machine)
1006 if info is not None:
1010 raise netrc.NetrcParseError(
1011 'No authenticators for %s' % netrc_machine)
1012 except (IOError, netrc.NetrcParseError) as err:
1013 self._downloader.report_warning(
1014 'parsing .netrc: %s' % error_to_compat_str(err))
1016 return username, password
1018 def _get_login_info(self, username_option='username', password_option='password', netrc_machine=None):
1020 Get the login info as (username, password)
1021 First look for the manually specified credentials using username_option
1022 and password_option as keys in params dictionary. If no such credentials
1023 available look in the netrc file using the netrc_machine or _NETRC_MACHINE
1025 If there's no info available, return (None, None)
1027 if self._downloader is None:
1030 downloader_params = self._downloader.params
1032 # Attempt to use provided username and password or .netrc data
1033 if downloader_params.get(username_option) is not None:
1034 username = downloader_params[username_option]
1035 password = downloader_params[password_option]
1037 username, password = self._get_netrc_login_info(netrc_machine)
1039 return username, password
1041 def _get_tfa_info(self, note='two-factor verification code'):
1043 Get the two-factor authentication info
1044 TODO - asking the user will be required for sms/phone verify
1045 currently just uses the command line option
1046 If there's no info available, return None
1048 if self._downloader is None:
1050 downloader_params = self._downloader.params
1052 if downloader_params.get('twofactor') is not None:
1053 return downloader_params['twofactor']
1055 return compat_getpass('Type %s and press [Return]: ' % note)
1057 # Helper functions for extracting OpenGraph info
1059 def _og_regexes(prop):
1060 content_re = r'content=(?:"([^"]+?)"|\'([^\']+?)\'|\s*([^\s"\'=<>`]+?))'
1061 property_re = (r'(?:name|property)=(?:\'og:%(prop)s\'|"og:%(prop)s"|\s*og:%(prop)s\b)'
1062 % {'prop': re.escape(prop)})
1063 template = r'<meta[^>]+?%s[^>]+?%s'
1065 template % (property_re, content_re),
1066 template % (content_re, property_re),
1070 def _meta_regex(prop):
1071 return r'''(?isx)<meta
1072 (?=[^>]+(?:itemprop|name|property|id|http-equiv)=(["\']?)%s\1)
1073 [^>]+?content=(["\'])(?P<content>.*?)\2''' % re.escape(prop)
1075 def _og_search_property(self, prop, html, name=None, **kargs):
1076 if not isinstance(prop, (list, tuple)):
1079 name = 'OpenGraph %s' % prop[0]
1082 og_regexes.extend(self._og_regexes(p))
1083 escaped = self._search_regex(og_regexes, html, name, flags=re.DOTALL, **kargs)
1086 return unescapeHTML(escaped)
1088 def _og_search_thumbnail(self, html, **kargs):
1089 return self._og_search_property('image', html, 'thumbnail URL', fatal=False, **kargs)
1091 def _og_search_description(self, html, **kargs):
1092 return self._og_search_property('description', html, fatal=False, **kargs)
1094 def _og_search_title(self, html, **kargs):
1095 return self._og_search_property('title', html, **kargs)
1097 def _og_search_video_url(self, html, name='video url', secure=True, **kargs):
1098 regexes = self._og_regexes('video') + self._og_regexes('video:url')
1100 regexes = self._og_regexes('video:secure_url') + regexes
1101 return self._html_search_regex(regexes, html, name, **kargs)
1103 def _og_search_url(self, html, **kargs):
1104 return self._og_search_property('url', html, **kargs)
1106 def _html_search_meta(self, name, html, display_name=None, fatal=False, **kwargs):
1107 if not isinstance(name, (list, tuple)):
1109 if display_name is None:
1110 display_name = name[0]
1111 return self._html_search_regex(
1112 [self._meta_regex(n) for n in name],
1113 html, display_name, fatal=fatal, group='content', **kwargs)
1115 def _dc_search_uploader(self, html):
1116 return self._html_search_meta('dc.creator', html, 'uploader')
1118 def _rta_search(self, html):
1119 # See http://www.rtalabel.org/index.php?content=howtofaq#single
1120 if re.search(r'(?ix)<meta\s+name="rating"\s+'
1121 r' content="RTA-5042-1996-1400-1577-RTA"',
1126 def _media_rating_search(self, html):
1127 # See http://www.tjg-designs.com/WP/metadata-code-examples-adding-metadata-to-your-web-pages/
1128 rating = self._html_search_meta('rating', html)
1140 return RATING_TABLE.get(rating.lower())
1142 def _family_friendly_search(self, html):
1143 # See http://schema.org/VideoObject
1144 family_friendly = self._html_search_meta(
1145 'isFamilyFriendly', html, default=None)
1147 if not family_friendly:
1156 return RATING_TABLE.get(family_friendly.lower())
1158 def _twitter_search_player(self, html):
1159 return self._html_search_meta('twitter:player', html,
1160 'twitter card player')
1162 def _search_json_ld(self, html, video_id, expected_type=None, **kwargs):
1163 json_ld = self._search_regex(
1164 JSON_LD_RE, html, 'JSON-LD', group='json_ld', **kwargs)
1165 default = kwargs.get('default', NO_DEFAULT)
1167 return default if default is not NO_DEFAULT else {}
1168 # JSON-LD may be malformed and thus `fatal` should be respected.
1169 # At the same time `default` may be passed that assumes `fatal=False`
1170 # for _search_regex. Let's simulate the same behavior here as well.
1171 fatal = kwargs.get('fatal', True) if default == NO_DEFAULT else False
1172 return self._json_ld(json_ld, video_id, fatal=fatal, expected_type=expected_type)
1174 def _json_ld(self, json_ld, video_id, fatal=True, expected_type=None):
1175 if isinstance(json_ld, compat_str):
1176 json_ld = self._parse_json(json_ld, video_id, fatal=fatal)
1180 if not isinstance(json_ld, (list, tuple, dict)):
1182 if isinstance(json_ld, dict):
1185 INTERACTION_TYPE_MAP = {
1186 'CommentAction': 'comment',
1187 'AgreeAction': 'like',
1188 'DisagreeAction': 'dislike',
1189 'LikeAction': 'like',
1190 'DislikeAction': 'dislike',
1191 'ListenAction': 'view',
1192 'WatchAction': 'view',
1193 'ViewAction': 'view',
1196 def extract_interaction_statistic(e):
1197 interaction_statistic = e.get('interactionStatistic')
1198 if not isinstance(interaction_statistic, list):
1200 for is_e in interaction_statistic:
1201 if not isinstance(is_e, dict):
1203 if is_e.get('@type') != 'InteractionCounter':
1205 interaction_type = is_e.get('interactionType')
1206 if not isinstance(interaction_type, compat_str):
1208 interaction_count = int_or_none(is_e.get('userInteractionCount'))
1209 if interaction_count is None:
1211 count_kind = INTERACTION_TYPE_MAP.get(interaction_type.split('/')[-1])
1214 count_key = '%s_count' % count_kind
1215 if info.get(count_key) is not None:
1217 info[count_key] = interaction_count
1219 def extract_video_object(e):
1220 assert e['@type'] == 'VideoObject'
1222 'url': url_or_none(e.get('contentUrl')),
1223 'title': unescapeHTML(e.get('name')),
1224 'description': unescapeHTML(e.get('description')),
1225 'thumbnail': url_or_none(e.get('thumbnailUrl') or e.get('thumbnailURL')),
1226 'duration': parse_duration(e.get('duration')),
1227 'timestamp': unified_timestamp(e.get('uploadDate')),
1228 'filesize': float_or_none(e.get('contentSize')),
1229 'tbr': int_or_none(e.get('bitrate')),
1230 'width': int_or_none(e.get('width')),
1231 'height': int_or_none(e.get('height')),
1232 'view_count': int_or_none(e.get('interactionCount')),
1234 extract_interaction_statistic(e)
1237 if isinstance(e.get('@context'), compat_str) and re.match(r'^https?://schema.org/?$', e.get('@context')):
1238 item_type = e.get('@type')
1239 if expected_type is not None and expected_type != item_type:
1241 if item_type in ('TVEpisode', 'Episode'):
1242 episode_name = unescapeHTML(e.get('name'))
1244 'episode': episode_name,
1245 'episode_number': int_or_none(e.get('episodeNumber')),
1246 'description': unescapeHTML(e.get('description')),
1248 if not info.get('title') and episode_name:
1249 info['title'] = episode_name
1250 part_of_season = e.get('partOfSeason')
1251 if isinstance(part_of_season, dict) and part_of_season.get('@type') in ('TVSeason', 'Season', 'CreativeWorkSeason'):
1253 'season': unescapeHTML(part_of_season.get('name')),
1254 'season_number': int_or_none(part_of_season.get('seasonNumber')),
1256 part_of_series = e.get('partOfSeries') or e.get('partOfTVSeries')
1257 if isinstance(part_of_series, dict) and part_of_series.get('@type') in ('TVSeries', 'Series', 'CreativeWorkSeries'):
1258 info['series'] = unescapeHTML(part_of_series.get('name'))
1259 elif item_type == 'Movie':
1261 'title': unescapeHTML(e.get('name')),
1262 'description': unescapeHTML(e.get('description')),
1263 'duration': parse_duration(e.get('duration')),
1264 'timestamp': unified_timestamp(e.get('dateCreated')),
1266 elif item_type in ('Article', 'NewsArticle'):
1268 'timestamp': parse_iso8601(e.get('datePublished')),
1269 'title': unescapeHTML(e.get('headline')),
1270 'description': unescapeHTML(e.get('articleBody')),
1272 elif item_type == 'VideoObject':
1273 extract_video_object(e)
1275 video = e.get('video')
1276 if isinstance(video, dict) and video.get('@type') == 'VideoObject':
1277 extract_video_object(video)
1279 return dict((k, v) for k, v in info.items() if v is not None)
1282 def _hidden_inputs(html):
1283 html = re.sub(r'<!--(?:(?!<!--).)*-->', '', html)
1285 for input in re.findall(r'(?i)(<input[^>]+>)', html):
1286 attrs = extract_attributes(input)
1289 if attrs.get('type') not in ('hidden', 'submit'):
1291 name = attrs.get('name') or attrs.get('id')
1292 value = attrs.get('value')
1293 if name and value is not None:
1294 hidden_inputs[name] = value
1295 return hidden_inputs
1297 def _form_hidden_inputs(self, form_id, html):
1298 form = self._search_regex(
1299 r'(?is)<form[^>]+?id=(["\'])%s\1[^>]*>(?P<form>.+?)</form>' % form_id,
1300 html, '%s form' % form_id, group='form')
1301 return self._hidden_inputs(form)
1303 def _sort_formats(self, formats, field_preference=None):
1305 raise ExtractorError('No video formats found')
1308 # Automatically determine tbr when missing based on abr and vbr (improves
1309 # formats sorting in some cases)
1310 if 'tbr' not in f and f.get('abr') is not None and f.get('vbr') is not None:
1311 f['tbr'] = f['abr'] + f['vbr']
1313 def _formats_key(f):
1314 # TODO remove the following workaround
1315 from ..utils import determine_ext
1316 if not f.get('ext') and 'url' in f:
1317 f['ext'] = determine_ext(f['url'])
1319 if isinstance(field_preference, (list, tuple)):
1322 if f.get(field) is not None
1323 else ('' if field == 'format_id' else -1)
1324 for field in field_preference)
1326 preference = f.get('preference')
1327 if preference is None:
1329 if f.get('ext') in ['f4f', 'f4m']: # Not yet supported
1332 protocol = f.get('protocol') or determine_protocol(f)
1333 proto_preference = 0 if protocol in ['http', 'https'] else (-0.5 if protocol == 'rtsp' else -0.1)
1335 if f.get('vcodec') == 'none': # audio only
1337 if self._downloader.params.get('prefer_free_formats'):
1338 ORDER = ['aac', 'mp3', 'm4a', 'webm', 'ogg', 'opus']
1340 ORDER = ['webm', 'opus', 'ogg', 'mp3', 'aac', 'm4a']
1343 audio_ext_preference = ORDER.index(f['ext'])
1345 audio_ext_preference = -1
1347 if f.get('acodec') == 'none': # video only
1349 if self._downloader.params.get('prefer_free_formats'):
1350 ORDER = ['flv', 'mp4', 'webm']
1352 ORDER = ['webm', 'flv', 'mp4']
1354 ext_preference = ORDER.index(f['ext'])
1357 audio_ext_preference = 0
1361 f.get('language_preference') if f.get('language_preference') is not None else -1,
1362 f.get('quality') if f.get('quality') is not None else -1,
1363 f.get('tbr') if f.get('tbr') is not None else -1,
1364 f.get('filesize') if f.get('filesize') is not None else -1,
1365 f.get('vbr') if f.get('vbr') is not None else -1,
1366 f.get('height') if f.get('height') is not None else -1,
1367 f.get('width') if f.get('width') is not None else -1,
1370 f.get('abr') if f.get('abr') is not None else -1,
1371 audio_ext_preference,
1372 f.get('fps') if f.get('fps') is not None else -1,
1373 f.get('filesize_approx') if f.get('filesize_approx') is not None else -1,
1374 f.get('source_preference') if f.get('source_preference') is not None else -1,
1375 f.get('format_id') if f.get('format_id') is not None else '',
1377 formats.sort(key=_formats_key)
1379 def _check_formats(self, formats, video_id):
1381 formats[:] = filter(
1382 lambda f: self._is_valid_url(
1384 item='%s video format' % f.get('format_id') if f.get('format_id') else 'video'),
1388 def _remove_duplicate_formats(formats):
1392 if f['url'] not in format_urls:
1393 format_urls.add(f['url'])
1394 unique_formats.append(f)
1395 formats[:] = unique_formats
1397 def _is_valid_url(self, url, video_id, item='video', headers={}):
1398 url = self._proto_relative_url(url, scheme='http:')
1399 # For now assume non HTTP(S) URLs always valid
1400 if not (url.startswith('http://') or url.startswith('https://')):
1403 self._request_webpage(url, video_id, 'Checking %s URL' % item, headers=headers)
1405 except ExtractorError as e:
1406 if isinstance(e.cause, compat_urllib_error.URLError):
1408 '%s: %s URL is invalid, skipping' % (video_id, item))
1412 def http_scheme(self):
1413 """ Either "http:" or "https:", depending on the user's preferences """
1416 if self._downloader.params.get('prefer_insecure', False)
1419 def _proto_relative_url(self, url, scheme=None):
1422 if url.startswith('//'):
1424 scheme = self.http_scheme()
1429 def _sleep(self, timeout, video_id, msg_template=None):
1430 if msg_template is None:
1431 msg_template = '%(video_id)s: Waiting for %(timeout)s seconds'
1432 msg = msg_template % {'video_id': video_id, 'timeout': timeout}
1436 def _extract_f4m_formats(self, manifest_url, video_id, preference=None, f4m_id=None,
1437 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1438 fatal=True, m3u8_id=None):
1439 manifest = self._download_xml(
1440 manifest_url, video_id, 'Downloading f4m manifest',
1441 'Unable to download f4m manifest',
1442 # Some manifests may be malformed, e.g. prosiebensat1 generated manifests
1443 # (see https://github.com/rg3/youtube-dl/issues/6215#issuecomment-121704244)
1444 transform_source=transform_source,
1447 if manifest is False:
1450 return self._parse_f4m_formats(
1451 manifest, manifest_url, video_id, preference=preference, f4m_id=f4m_id,
1452 transform_source=transform_source, fatal=fatal, m3u8_id=m3u8_id)
1454 def _parse_f4m_formats(self, manifest, manifest_url, video_id, preference=None, f4m_id=None,
1455 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1456 fatal=True, m3u8_id=None):
1457 # currently youtube-dl cannot decode the playerVerificationChallenge as Akamai uses Adobe Alchemy
1458 akamai_pv = manifest.find('{http://ns.adobe.com/f4m/1.0}pv-2.0')
1459 if akamai_pv is not None and ';' in akamai_pv.text:
1460 playerVerificationChallenge = akamai_pv.text.split(';')[0]
1461 if playerVerificationChallenge.strip() != '':
1465 manifest_version = '1.0'
1466 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/1.0}media')
1468 manifest_version = '2.0'
1469 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/2.0}media')
1470 # Remove unsupported DRM protected media from final formats
1471 # rendition (see https://github.com/rg3/youtube-dl/issues/8573).
1472 media_nodes = remove_encrypted_media(media_nodes)
1476 manifest_base_url = get_base_url(manifest)
1478 bootstrap_info = xpath_element(
1479 manifest, ['{http://ns.adobe.com/f4m/1.0}bootstrapInfo', '{http://ns.adobe.com/f4m/2.0}bootstrapInfo'],
1480 'bootstrap info', default=None)
1483 mime_type = xpath_text(
1484 manifest, ['{http://ns.adobe.com/f4m/1.0}mimeType', '{http://ns.adobe.com/f4m/2.0}mimeType'],
1485 'base URL', default=None)
1486 if mime_type and mime_type.startswith('audio/'):
1489 for i, media_el in enumerate(media_nodes):
1490 tbr = int_or_none(media_el.attrib.get('bitrate'))
1491 width = int_or_none(media_el.attrib.get('width'))
1492 height = int_or_none(media_el.attrib.get('height'))
1493 format_id = '-'.join(filter(None, [f4m_id, compat_str(i if tbr is None else tbr)]))
1494 # If <bootstrapInfo> is present, the specified f4m is a
1495 # stream-level manifest, and only set-level manifests may refer to
1496 # external resources. See section 11.4 and section 4 of F4M spec
1497 if bootstrap_info is None:
1499 # @href is introduced in 2.0, see section 11.6 of F4M spec
1500 if manifest_version == '2.0':
1501 media_url = media_el.attrib.get('href')
1502 if media_url is None:
1503 media_url = media_el.attrib.get('url')
1507 media_url if media_url.startswith('http://') or media_url.startswith('https://')
1508 else ((manifest_base_url or '/'.join(manifest_url.split('/')[:-1])) + '/' + media_url))
1509 # If media_url is itself a f4m manifest do the recursive extraction
1510 # since bitrates in parent manifest (this one) and media_url manifest
1511 # may differ leading to inability to resolve the format by requested
1512 # bitrate in f4m downloader
1513 ext = determine_ext(manifest_url)
1515 f4m_formats = self._extract_f4m_formats(
1516 manifest_url, video_id, preference=preference, f4m_id=f4m_id,
1517 transform_source=transform_source, fatal=fatal)
1518 # Sometimes stream-level manifest contains single media entry that
1519 # does not contain any quality metadata (e.g. http://matchtv.ru/#live-player).
1520 # At the same time parent's media entry in set-level manifest may
1521 # contain it. We will copy it from parent in such cases.
1522 if len(f4m_formats) == 1:
1525 'tbr': f.get('tbr') or tbr,
1526 'width': f.get('width') or width,
1527 'height': f.get('height') or height,
1528 'format_id': f.get('format_id') if not tbr else format_id,
1531 formats.extend(f4m_formats)
1534 formats.extend(self._extract_m3u8_formats(
1535 manifest_url, video_id, 'mp4', preference=preference,
1536 m3u8_id=m3u8_id, fatal=fatal))
1539 'format_id': format_id,
1540 'url': manifest_url,
1541 'manifest_url': manifest_url,
1542 'ext': 'flv' if bootstrap_info is not None else None,
1548 'preference': preference,
1552 def _m3u8_meta_format(self, m3u8_url, ext=None, preference=None, m3u8_id=None):
1554 'format_id': '-'.join(filter(None, [m3u8_id, 'meta'])),
1558 'preference': preference - 100 if preference else -100,
1559 'resolution': 'multiple',
1560 'format_note': 'Quality selection URL',
1563 def _extract_m3u8_formats(self, m3u8_url, video_id, ext=None,
1564 entry_protocol='m3u8', preference=None,
1565 m3u8_id=None, note=None, errnote=None,
1566 fatal=True, live=False):
1567 res = self._download_webpage_handle(
1569 note=note or 'Downloading m3u8 information',
1570 errnote=errnote or 'Failed to download m3u8 information',
1576 m3u8_doc, urlh = res
1577 m3u8_url = urlh.geturl()
1579 return self._parse_m3u8_formats(
1580 m3u8_doc, m3u8_url, ext=ext, entry_protocol=entry_protocol,
1581 preference=preference, m3u8_id=m3u8_id, live=live)
1583 def _parse_m3u8_formats(self, m3u8_doc, m3u8_url, ext=None,
1584 entry_protocol='m3u8', preference=None,
1585 m3u8_id=None, live=False):
1586 if '#EXT-X-FAXS-CM:' in m3u8_doc: # Adobe Flash Access
1589 if re.search(r'#EXT-X-SESSION-KEY:.*?URI="skd://', m3u8_doc): # Apple FairPlay
1594 format_url = lambda u: (
1596 if re.match(r'^https?://', u)
1597 else compat_urlparse.urljoin(m3u8_url, u))
1600 # 1. https://tools.ietf.org/html/draft-pantos-http-live-streaming-21
1601 # 2. https://github.com/rg3/youtube-dl/issues/12211
1602 # 3. https://github.com/rg3/youtube-dl/issues/18923
1604 # We should try extracting formats only from master playlists [1, 4.3.4],
1605 # i.e. playlists that describe available qualities. On the other hand
1606 # media playlists [1, 4.3.3] should be returned as is since they contain
1607 # just the media without qualities renditions.
1608 # Fortunately, master playlist can be easily distinguished from media
1609 # playlist based on particular tags availability. As of [1, 4.3.3, 4.3.4]
1610 # master playlist tags MUST NOT appear in a media playist and vice versa.
1611 # As of [1, 4.3.3.1] #EXT-X-TARGETDURATION tag is REQUIRED for every
1612 # media playlist and MUST NOT appear in master playlist thus we can
1613 # clearly detect media playlist with this criterion.
1615 if '#EXT-X-TARGETDURATION' in m3u8_doc: # media playlist, return as is
1618 'format_id': m3u8_id,
1620 'protocol': entry_protocol,
1621 'preference': preference,
1625 last_stream_inf = {}
1627 def extract_media(x_media_line):
1628 media = parse_m3u8_attributes(x_media_line)
1629 # As per [1, 4.3.4.1] TYPE, GROUP-ID and NAME are REQUIRED
1630 media_type, group_id, name = media.get('TYPE'), media.get('GROUP-ID'), media.get('NAME')
1631 if not (media_type and group_id and name):
1633 groups.setdefault(group_id, []).append(media)
1634 if media_type not in ('VIDEO', 'AUDIO'):
1636 media_url = media.get('URI')
1639 for v in (m3u8_id, group_id, name):
1643 'format_id': '-'.join(format_id),
1644 'url': format_url(media_url),
1645 'manifest_url': m3u8_url,
1646 'language': media.get('LANGUAGE'),
1648 'protocol': entry_protocol,
1649 'preference': preference,
1651 if media_type == 'AUDIO':
1652 f['vcodec'] = 'none'
1655 def build_stream_name():
1656 # Despite specification does not mention NAME attribute for
1657 # EXT-X-STREAM-INF tag it still sometimes may be present (see [1]
1658 # or vidio test in TestInfoExtractor.test_parse_m3u8_formats)
1659 # 1. http://www.vidio.com/watch/165683-dj_ambred-booyah-live-2015
1660 stream_name = last_stream_inf.get('NAME')
1663 # If there is no NAME in EXT-X-STREAM-INF it will be obtained
1664 # from corresponding rendition group
1665 stream_group_id = last_stream_inf.get('VIDEO')
1666 if not stream_group_id:
1668 stream_group = groups.get(stream_group_id)
1669 if not stream_group:
1670 return stream_group_id
1671 rendition = stream_group[0]
1672 return rendition.get('NAME') or stream_group_id
1674 # parse EXT-X-MEDIA tags before EXT-X-STREAM-INF in order to have the
1675 # chance to detect video only formats when EXT-X-STREAM-INF tags
1676 # precede EXT-X-MEDIA tags in HLS manifest such as [3].
1677 for line in m3u8_doc.splitlines():
1678 if line.startswith('#EXT-X-MEDIA:'):
1681 for line in m3u8_doc.splitlines():
1682 if line.startswith('#EXT-X-STREAM-INF:'):
1683 last_stream_inf = parse_m3u8_attributes(line)
1684 elif line.startswith('#') or not line.strip():
1687 tbr = float_or_none(
1688 last_stream_inf.get('AVERAGE-BANDWIDTH') or
1689 last_stream_inf.get('BANDWIDTH'), scale=1000)
1692 format_id.append(m3u8_id)
1693 stream_name = build_stream_name()
1694 # Bandwidth of live streams may differ over time thus making
1695 # format_id unpredictable. So it's better to keep provided
1698 format_id.append(stream_name if stream_name else '%d' % (tbr if tbr else len(formats)))
1699 manifest_url = format_url(line.strip())
1701 'format_id': '-'.join(format_id),
1702 'url': manifest_url,
1703 'manifest_url': m3u8_url,
1706 'fps': float_or_none(last_stream_inf.get('FRAME-RATE')),
1707 'protocol': entry_protocol,
1708 'preference': preference,
1710 resolution = last_stream_inf.get('RESOLUTION')
1712 mobj = re.search(r'(?P<width>\d+)[xX](?P<height>\d+)', resolution)
1714 f['width'] = int(mobj.group('width'))
1715 f['height'] = int(mobj.group('height'))
1716 # Unified Streaming Platform
1718 r'audio.*?(?:%3D|=)(\d+)(?:-video.*?(?:%3D|=)(\d+))?', f['url'])
1720 abr, vbr = mobj.groups()
1721 abr, vbr = float_or_none(abr, 1000), float_or_none(vbr, 1000)
1726 codecs = parse_codecs(last_stream_inf.get('CODECS'))
1728 audio_group_id = last_stream_inf.get('AUDIO')
1729 # As per [1, 4.3.4.1.1] any EXT-X-STREAM-INF tag which
1730 # references a rendition group MUST have a CODECS attribute.
1731 # However, this is not always respected, for example, [2]
1732 # contains EXT-X-STREAM-INF tag which references AUDIO
1733 # rendition group but does not have CODECS and despite
1734 # referencing an audio group it represents a complete
1735 # (with audio and video) format. So, for such cases we will
1736 # ignore references to rendition groups and treat them
1737 # as complete formats.
1738 if audio_group_id and codecs and f.get('vcodec') != 'none':
1739 audio_group = groups.get(audio_group_id)
1740 if audio_group and audio_group[0].get('URI'):
1741 # TODO: update acodec for audio only formats with
1743 f['acodec'] = 'none'
1745 last_stream_inf = {}
1749 def _xpath_ns(path, namespace=None):
1753 for c in path.split('/'):
1754 if not c or c == '.':
1757 out.append('{%s}%s' % (namespace, c))
1758 return '/'.join(out)
1760 def _extract_smil_formats(self, smil_url, video_id, fatal=True, f4m_params=None, transform_source=None):
1761 smil = self._download_smil(smil_url, video_id, fatal=fatal, transform_source=transform_source)
1767 namespace = self._parse_smil_namespace(smil)
1769 return self._parse_smil_formats(
1770 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
1772 def _extract_smil_info(self, smil_url, video_id, fatal=True, f4m_params=None):
1773 smil = self._download_smil(smil_url, video_id, fatal=fatal)
1776 return self._parse_smil(smil, smil_url, video_id, f4m_params=f4m_params)
1778 def _download_smil(self, smil_url, video_id, fatal=True, transform_source=None):
1779 return self._download_xml(
1780 smil_url, video_id, 'Downloading SMIL file',
1781 'Unable to download SMIL file', fatal=fatal, transform_source=transform_source)
1783 def _parse_smil(self, smil, smil_url, video_id, f4m_params=None):
1784 namespace = self._parse_smil_namespace(smil)
1786 formats = self._parse_smil_formats(
1787 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
1788 subtitles = self._parse_smil_subtitles(smil, namespace=namespace)
1790 video_id = os.path.splitext(url_basename(smil_url))[0]
1794 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
1795 name = meta.attrib.get('name')
1796 content = meta.attrib.get('content')
1797 if not name or not content:
1799 if not title and name == 'title':
1801 elif not description and name in ('description', 'abstract'):
1802 description = content
1803 elif not upload_date and name == 'date':
1804 upload_date = unified_strdate(content)
1807 'id': image.get('type'),
1808 'url': image.get('src'),
1809 'width': int_or_none(image.get('width')),
1810 'height': int_or_none(image.get('height')),
1811 } for image in smil.findall(self._xpath_ns('.//image', namespace)) if image.get('src')]
1815 'title': title or video_id,
1816 'description': description,
1817 'upload_date': upload_date,
1818 'thumbnails': thumbnails,
1820 'subtitles': subtitles,
1823 def _parse_smil_namespace(self, smil):
1824 return self._search_regex(
1825 r'(?i)^{([^}]+)?}smil$', smil.tag, 'namespace', default=None)
1827 def _parse_smil_formats(self, smil, smil_url, video_id, namespace=None, f4m_params=None, transform_rtmp_url=None):
1829 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
1830 b = meta.get('base') or meta.get('httpBase')
1841 media = smil.findall(self._xpath_ns('.//video', namespace)) + smil.findall(self._xpath_ns('.//audio', namespace))
1842 for medium in media:
1843 src = medium.get('src')
1844 if not src or src in srcs:
1848 bitrate = float_or_none(medium.get('system-bitrate') or medium.get('systemBitrate'), 1000)
1849 filesize = int_or_none(medium.get('size') or medium.get('fileSize'))
1850 width = int_or_none(medium.get('width'))
1851 height = int_or_none(medium.get('height'))
1852 proto = medium.get('proto')
1853 ext = medium.get('ext')
1854 src_ext = determine_ext(src)
1855 streamer = medium.get('streamer') or base
1857 if proto == 'rtmp' or streamer.startswith('rtmp'):
1863 'format_id': 'rtmp-%d' % (rtmp_count if bitrate is None else bitrate),
1865 'filesize': filesize,
1869 if transform_rtmp_url:
1870 streamer, src = transform_rtmp_url(streamer, src)
1871 formats[-1].update({
1877 src_url = src if src.startswith('http') else compat_urlparse.urljoin(base, src)
1878 src_url = src_url.strip()
1880 if proto == 'm3u8' or src_ext == 'm3u8':
1881 m3u8_formats = self._extract_m3u8_formats(
1882 src_url, video_id, ext or 'mp4', m3u8_id='hls', fatal=False)
1883 if len(m3u8_formats) == 1:
1885 m3u8_formats[0].update({
1886 'format_id': 'hls-%d' % (m3u8_count if bitrate is None else bitrate),
1891 formats.extend(m3u8_formats)
1892 elif src_ext == 'f4m':
1897 'plugin': 'flowplayer-3.2.0.1',
1899 f4m_url += '&' if '?' in f4m_url else '?'
1900 f4m_url += compat_urllib_parse_urlencode(f4m_params)
1901 formats.extend(self._extract_f4m_formats(f4m_url, video_id, f4m_id='hds', fatal=False))
1902 elif src_ext == 'mpd':
1903 formats.extend(self._extract_mpd_formats(
1904 src_url, video_id, mpd_id='dash', fatal=False))
1905 elif re.search(r'\.ism/[Mm]anifest', src_url):
1906 formats.extend(self._extract_ism_formats(
1907 src_url, video_id, ism_id='mss', fatal=False))
1908 elif src_url.startswith('http') and self._is_valid_url(src, video_id):
1912 'ext': ext or src_ext or 'flv',
1913 'format_id': 'http-%d' % (bitrate or http_count),
1915 'filesize': filesize,
1922 def _parse_smil_subtitles(self, smil, namespace=None, subtitles_lang='en'):
1925 for num, textstream in enumerate(smil.findall(self._xpath_ns('.//textstream', namespace))):
1926 src = textstream.get('src')
1927 if not src or src in urls:
1930 ext = textstream.get('ext') or mimetype2ext(textstream.get('type')) or determine_ext(src)
1931 lang = textstream.get('systemLanguage') or textstream.get('systemLanguageName') or textstream.get('lang') or subtitles_lang
1932 subtitles.setdefault(lang, []).append({
1938 def _extract_xspf_playlist(self, xspf_url, playlist_id, fatal=True):
1939 xspf = self._download_xml(
1940 xspf_url, playlist_id, 'Downloading xpsf playlist',
1941 'Unable to download xspf manifest', fatal=fatal)
1944 return self._parse_xspf(
1945 xspf, playlist_id, xspf_url=xspf_url,
1946 xspf_base_url=base_url(xspf_url))
1948 def _parse_xspf(self, xspf_doc, playlist_id, xspf_url=None, xspf_base_url=None):
1950 'xspf': 'http://xspf.org/ns/0/',
1951 's1': 'http://static.streamone.nl/player/ns/0',
1955 for track in xspf_doc.findall(xpath_with_ns('./xspf:trackList/xspf:track', NS_MAP)):
1957 track, xpath_with_ns('./xspf:title', NS_MAP), 'title', default=playlist_id)
1958 description = xpath_text(
1959 track, xpath_with_ns('./xspf:annotation', NS_MAP), 'description')
1960 thumbnail = xpath_text(
1961 track, xpath_with_ns('./xspf:image', NS_MAP), 'thumbnail')
1962 duration = float_or_none(
1963 xpath_text(track, xpath_with_ns('./xspf:duration', NS_MAP), 'duration'), 1000)
1966 for location in track.findall(xpath_with_ns('./xspf:location', NS_MAP)):
1967 format_url = urljoin(xspf_base_url, location.text)
1972 'manifest_url': xspf_url,
1973 'format_id': location.get(xpath_with_ns('s1:label', NS_MAP)),
1974 'width': int_or_none(location.get(xpath_with_ns('s1:width', NS_MAP))),
1975 'height': int_or_none(location.get(xpath_with_ns('s1:height', NS_MAP))),
1977 self._sort_formats(formats)
1982 'description': description,
1983 'thumbnail': thumbnail,
1984 'duration': duration,
1989 def _extract_mpd_formats(self, mpd_url, video_id, mpd_id=None, note=None, errnote=None, fatal=True, formats_dict={}):
1990 res = self._download_xml_handle(
1992 note=note or 'Downloading MPD manifest',
1993 errnote=errnote or 'Failed to download MPD manifest',
1998 mpd_base_url = base_url(urlh.geturl())
2000 return self._parse_mpd_formats(
2001 mpd_doc, mpd_id=mpd_id, mpd_base_url=mpd_base_url,
2002 formats_dict=formats_dict, mpd_url=mpd_url)
2004 def _parse_mpd_formats(self, mpd_doc, mpd_id=None, mpd_base_url='', formats_dict={}, mpd_url=None):
2006 Parse formats from MPD manifest.
2008 1. MPEG-DASH Standard, ISO/IEC 23009-1:2014(E),
2009 http://standards.iso.org/ittf/PubliclyAvailableStandards/c065274_ISO_IEC_23009-1_2014.zip
2010 2. https://en.wikipedia.org/wiki/Dynamic_Adaptive_Streaming_over_HTTP
2012 if mpd_doc.get('type') == 'dynamic':
2015 namespace = self._search_regex(r'(?i)^{([^}]+)?}MPD$', mpd_doc.tag, 'namespace', default=None)
2018 return self._xpath_ns(path, namespace)
2020 def is_drm_protected(element):
2021 return element.find(_add_ns('ContentProtection')) is not None
2023 def extract_multisegment_info(element, ms_parent_info):
2024 ms_info = ms_parent_info.copy()
2026 # As per [1, 5.3.9.2.2] SegmentList and SegmentTemplate share some
2027 # common attributes and elements. We will only extract relevant
2029 def extract_common(source):
2030 segment_timeline = source.find(_add_ns('SegmentTimeline'))
2031 if segment_timeline is not None:
2032 s_e = segment_timeline.findall(_add_ns('S'))
2034 ms_info['total_number'] = 0
2037 r = int(s.get('r', 0))
2038 ms_info['total_number'] += 1 + r
2039 ms_info['s'].append({
2040 't': int(s.get('t', 0)),
2041 # @d is mandatory (see [1, 5.3.9.6.2, Table 17, page 60])
2042 'd': int(s.attrib['d']),
2045 start_number = source.get('startNumber')
2047 ms_info['start_number'] = int(start_number)
2048 timescale = source.get('timescale')
2050 ms_info['timescale'] = int(timescale)
2051 segment_duration = source.get('duration')
2052 if segment_duration:
2053 ms_info['segment_duration'] = float(segment_duration)
2055 def extract_Initialization(source):
2056 initialization = source.find(_add_ns('Initialization'))
2057 if initialization is not None:
2058 ms_info['initialization_url'] = initialization.attrib['sourceURL']
2060 segment_list = element.find(_add_ns('SegmentList'))
2061 if segment_list is not None:
2062 extract_common(segment_list)
2063 extract_Initialization(segment_list)
2064 segment_urls_e = segment_list.findall(_add_ns('SegmentURL'))
2066 ms_info['segment_urls'] = [segment.attrib['media'] for segment in segment_urls_e]
2068 segment_template = element.find(_add_ns('SegmentTemplate'))
2069 if segment_template is not None:
2070 extract_common(segment_template)
2071 media = segment_template.get('media')
2073 ms_info['media'] = media
2074 initialization = segment_template.get('initialization')
2076 ms_info['initialization'] = initialization
2078 extract_Initialization(segment_template)
2081 mpd_duration = parse_duration(mpd_doc.get('mediaPresentationDuration'))
2083 for period in mpd_doc.findall(_add_ns('Period')):
2084 period_duration = parse_duration(period.get('duration')) or mpd_duration
2085 period_ms_info = extract_multisegment_info(period, {
2089 for adaptation_set in period.findall(_add_ns('AdaptationSet')):
2090 if is_drm_protected(adaptation_set):
2092 adaption_set_ms_info = extract_multisegment_info(adaptation_set, period_ms_info)
2093 for representation in adaptation_set.findall(_add_ns('Representation')):
2094 if is_drm_protected(representation):
2096 representation_attrib = adaptation_set.attrib.copy()
2097 representation_attrib.update(representation.attrib)
2098 # According to [1, 5.3.7.2, Table 9, page 41], @mimeType is mandatory
2099 mime_type = representation_attrib['mimeType']
2100 content_type = mime_type.split('/')[0]
2101 if content_type == 'text':
2102 # TODO implement WebVTT downloading
2104 elif content_type in ('video', 'audio'):
2106 for element in (representation, adaptation_set, period, mpd_doc):
2107 base_url_e = element.find(_add_ns('BaseURL'))
2108 if base_url_e is not None:
2109 base_url = base_url_e.text + base_url
2110 if re.match(r'^https?://', base_url):
2112 if mpd_base_url and not re.match(r'^https?://', base_url):
2113 if not mpd_base_url.endswith('/') and not base_url.startswith('/'):
2115 base_url = mpd_base_url + base_url
2116 representation_id = representation_attrib.get('id')
2117 lang = representation_attrib.get('lang')
2118 url_el = representation.find(_add_ns('BaseURL'))
2119 filesize = int_or_none(url_el.attrib.get('{http://youtube.com/yt/2012/10/10}contentLength') if url_el is not None else None)
2120 bandwidth = int_or_none(representation_attrib.get('bandwidth'))
2122 'format_id': '%s-%s' % (mpd_id, representation_id) if mpd_id else representation_id,
2124 'manifest_url': mpd_url,
2125 'ext': mimetype2ext(mime_type),
2126 'width': int_or_none(representation_attrib.get('width')),
2127 'height': int_or_none(representation_attrib.get('height')),
2128 'tbr': float_or_none(bandwidth, 1000),
2129 'asr': int_or_none(representation_attrib.get('audioSamplingRate')),
2130 'fps': int_or_none(representation_attrib.get('frameRate')),
2131 'language': lang if lang not in ('mul', 'und', 'zxx', 'mis') else None,
2132 'format_note': 'DASH %s' % content_type,
2133 'filesize': filesize,
2134 'container': mimetype2ext(mime_type) + '_dash',
2136 f.update(parse_codecs(representation_attrib.get('codecs')))
2137 representation_ms_info = extract_multisegment_info(representation, adaption_set_ms_info)
2139 def prepare_template(template_name, identifiers):
2140 tmpl = representation_ms_info[template_name]
2141 # First of, % characters outside $...$ templates
2142 # must be escaped by doubling for proper processing
2143 # by % operator string formatting used further (see
2144 # https://github.com/rg3/youtube-dl/issues/16867).
2150 in_template = not in_template
2151 elif c == '%' and not in_template:
2153 # Next, $...$ templates are translated to their
2154 # %(...) counterparts to be used with % operator
2155 t = t.replace('$RepresentationID$', representation_id)
2156 t = re.sub(r'\$(%s)\$' % '|'.join(identifiers), r'%(\1)d', t)
2157 t = re.sub(r'\$(%s)%%([^$]+)\$' % '|'.join(identifiers), r'%(\1)\2', t)
2158 t.replace('$$', '$')
2161 # @initialization is a regular template like @media one
2162 # so it should be handled just the same way (see
2163 # https://github.com/rg3/youtube-dl/issues/11605)
2164 if 'initialization' in representation_ms_info:
2165 initialization_template = prepare_template(
2167 # As per [1, 5.3.9.4.2, Table 15, page 54] $Number$ and
2168 # $Time$ shall not be included for @initialization thus
2169 # only $Bandwidth$ remains
2171 representation_ms_info['initialization_url'] = initialization_template % {
2172 'Bandwidth': bandwidth,
2175 def location_key(location):
2176 return 'url' if re.match(r'^https?://', location) else 'path'
2178 if 'segment_urls' not in representation_ms_info and 'media' in representation_ms_info:
2180 media_template = prepare_template('media', ('Number', 'Bandwidth', 'Time'))
2181 media_location_key = location_key(media_template)
2183 # As per [1, 5.3.9.4.4, Table 16, page 55] $Number$ and $Time$
2184 # can't be used at the same time
2185 if '%(Number' in media_template and 's' not in representation_ms_info:
2186 segment_duration = None
2187 if 'total_number' not in representation_ms_info and 'segment_duration' in representation_ms_info:
2188 segment_duration = float_or_none(representation_ms_info['segment_duration'], representation_ms_info['timescale'])
2189 representation_ms_info['total_number'] = int(math.ceil(float(period_duration) / segment_duration))
2190 representation_ms_info['fragments'] = [{
2191 media_location_key: media_template % {
2192 'Number': segment_number,
2193 'Bandwidth': bandwidth,
2195 'duration': segment_duration,
2196 } for segment_number in range(
2197 representation_ms_info['start_number'],
2198 representation_ms_info['total_number'] + representation_ms_info['start_number'])]
2200 # $Number*$ or $Time$ in media template with S list available
2201 # Example $Number*$: http://www.svtplay.se/klipp/9023742/stopptid-om-bjorn-borg
2202 # Example $Time$: https://play.arkena.com/embed/avp/v2/player/media/b41dda37-d8e7-4d3f-b1b5-9a9db578bdfe/1/129411
2203 representation_ms_info['fragments'] = []
2206 segment_number = representation_ms_info['start_number']
2208 def add_segment_url():
2209 segment_url = media_template % {
2210 'Time': segment_time,
2211 'Bandwidth': bandwidth,
2212 'Number': segment_number,
2214 representation_ms_info['fragments'].append({
2215 media_location_key: segment_url,
2216 'duration': float_or_none(segment_d, representation_ms_info['timescale']),
2219 for num, s in enumerate(representation_ms_info['s']):
2220 segment_time = s.get('t') or segment_time
2224 for r in range(s.get('r', 0)):
2225 segment_time += segment_d
2228 segment_time += segment_d
2229 elif 'segment_urls' in representation_ms_info and 's' in representation_ms_info:
2231 # Example: https://www.youtube.com/watch?v=iXZV5uAYMJI
2232 # or any YouTube dashsegments video
2235 timescale = representation_ms_info['timescale']
2236 for s in representation_ms_info['s']:
2237 duration = float_or_none(s['d'], timescale)
2238 for r in range(s.get('r', 0) + 1):
2239 segment_uri = representation_ms_info['segment_urls'][segment_index]
2241 location_key(segment_uri): segment_uri,
2242 'duration': duration,
2245 representation_ms_info['fragments'] = fragments
2246 elif 'segment_urls' in representation_ms_info:
2247 # Segment URLs with no SegmentTimeline
2248 # Example: https://www.seznam.cz/zpravy/clanek/cesko-zasahne-vitr-o-sile-vichrice-muze-byt-i-zivotu-nebezpecny-39091
2249 # https://github.com/rg3/youtube-dl/pull/14844
2251 segment_duration = float_or_none(
2252 representation_ms_info['segment_duration'],
2253 representation_ms_info['timescale']) if 'segment_duration' in representation_ms_info else None
2254 for segment_url in representation_ms_info['segment_urls']:
2256 location_key(segment_url): segment_url,
2258 if segment_duration:
2259 fragment['duration'] = segment_duration
2260 fragments.append(fragment)
2261 representation_ms_info['fragments'] = fragments
2262 # NB: MPD manifest may contain direct URLs to unfragmented media.
2263 # No fragments key is present in this case.
2264 if 'fragments' in representation_ms_info:
2266 'fragment_base_url': base_url,
2268 'protocol': 'http_dash_segments',
2270 if 'initialization_url' in representation_ms_info:
2271 initialization_url = representation_ms_info['initialization_url']
2272 if not f.get('url'):
2273 f['url'] = initialization_url
2274 f['fragments'].append({location_key(initialization_url): initialization_url})
2275 f['fragments'].extend(representation_ms_info['fragments'])
2276 # According to [1, 5.3.5.2, Table 7, page 35] @id of Representation
2277 # is not necessarily unique within a Period thus formats with
2278 # the same `format_id` are quite possible. There are numerous examples
2279 # of such manifests (see https://github.com/rg3/youtube-dl/issues/15111,
2280 # https://github.com/rg3/youtube-dl/issues/13919)
2281 full_info = formats_dict.get(representation_id, {}).copy()
2283 formats.append(full_info)
2285 self.report_warning('Unknown MIME type %s in DASH manifest' % mime_type)
2288 def _extract_ism_formats(self, ism_url, video_id, ism_id=None, note=None, errnote=None, fatal=True):
2289 res = self._download_xml_handle(
2291 note=note or 'Downloading ISM manifest',
2292 errnote=errnote or 'Failed to download ISM manifest',
2298 return self._parse_ism_formats(ism_doc, urlh.geturl(), ism_id)
2300 def _parse_ism_formats(self, ism_doc, ism_url, ism_id=None):
2302 Parse formats from ISM manifest.
2304 1. [MS-SSTR]: Smooth Streaming Protocol,
2305 https://msdn.microsoft.com/en-us/library/ff469518.aspx
2307 if ism_doc.get('IsLive') == 'TRUE' or ism_doc.find('Protection') is not None:
2310 duration = int(ism_doc.attrib['Duration'])
2311 timescale = int_or_none(ism_doc.get('TimeScale')) or 10000000
2314 for stream in ism_doc.findall('StreamIndex'):
2315 stream_type = stream.get('Type')
2316 if stream_type not in ('video', 'audio'):
2318 url_pattern = stream.attrib['Url']
2319 stream_timescale = int_or_none(stream.get('TimeScale')) or timescale
2320 stream_name = stream.get('Name')
2321 for track in stream.findall('QualityLevel'):
2322 fourcc = track.get('FourCC', 'AACL' if track.get('AudioTag') == '255' else None)
2323 # TODO: add support for WVC1 and WMAP
2324 if fourcc not in ('H264', 'AVC1', 'AACL'):
2325 self.report_warning('%s is not a supported codec' % fourcc)
2327 tbr = int(track.attrib['Bitrate']) // 1000
2328 # [1] does not mention Width and Height attributes. However,
2329 # they're often present while MaxWidth and MaxHeight are
2330 # missing, so should be used as fallbacks
2331 width = int_or_none(track.get('MaxWidth') or track.get('Width'))
2332 height = int_or_none(track.get('MaxHeight') or track.get('Height'))
2333 sampling_rate = int_or_none(track.get('SamplingRate'))
2335 track_url_pattern = re.sub(r'{[Bb]itrate}', track.attrib['Bitrate'], url_pattern)
2336 track_url_pattern = compat_urlparse.urljoin(ism_url, track_url_pattern)
2342 stream_fragments = stream.findall('c')
2343 for stream_fragment_index, stream_fragment in enumerate(stream_fragments):
2344 fragment_ctx['time'] = int_or_none(stream_fragment.get('t')) or fragment_ctx['time']
2345 fragment_repeat = int_or_none(stream_fragment.get('r')) or 1
2346 fragment_ctx['duration'] = int_or_none(stream_fragment.get('d'))
2347 if not fragment_ctx['duration']:
2349 next_fragment_time = int(stream_fragment[stream_fragment_index + 1].attrib['t'])
2351 next_fragment_time = duration
2352 fragment_ctx['duration'] = (next_fragment_time - fragment_ctx['time']) / fragment_repeat
2353 for _ in range(fragment_repeat):
2355 'url': re.sub(r'{start[ _]time}', compat_str(fragment_ctx['time']), track_url_pattern),
2356 'duration': fragment_ctx['duration'] / stream_timescale,
2358 fragment_ctx['time'] += fragment_ctx['duration']
2362 format_id.append(ism_id)
2364 format_id.append(stream_name)
2365 format_id.append(compat_str(tbr))
2368 'format_id': '-'.join(format_id),
2370 'manifest_url': ism_url,
2371 'ext': 'ismv' if stream_type == 'video' else 'isma',
2375 'asr': sampling_rate,
2376 'vcodec': 'none' if stream_type == 'audio' else fourcc,
2377 'acodec': 'none' if stream_type == 'video' else fourcc,
2379 'fragments': fragments,
2380 '_download_params': {
2381 'duration': duration,
2382 'timescale': stream_timescale,
2383 'width': width or 0,
2384 'height': height or 0,
2386 'codec_private_data': track.get('CodecPrivateData'),
2387 'sampling_rate': sampling_rate,
2388 'channels': int_or_none(track.get('Channels', 2)),
2389 'bits_per_sample': int_or_none(track.get('BitsPerSample', 16)),
2390 'nal_unit_length_field': int_or_none(track.get('NALUnitLengthField', 4)),
2395 def _parse_html5_media_entries(self, base_url, webpage, video_id, m3u8_id=None, m3u8_entry_protocol='m3u8', mpd_id=None, preference=None):
2396 def absolute_url(item_url):
2397 return urljoin(base_url, item_url)
2399 def parse_content_type(content_type):
2400 if not content_type:
2402 ctr = re.search(r'(?P<mimetype>[^/]+/[^;]+)(?:;\s*codecs="?(?P<codecs>[^"]+))?', content_type)
2404 mimetype, codecs = ctr.groups()
2405 f = parse_codecs(codecs)
2406 f['ext'] = mimetype2ext(mimetype)
2410 def _media_formats(src, cur_media_type, type_info={}):
2411 full_url = absolute_url(src)
2412 ext = type_info.get('ext') or determine_ext(full_url)
2414 is_plain_url = False
2415 formats = self._extract_m3u8_formats(
2416 full_url, video_id, ext='mp4',
2417 entry_protocol=m3u8_entry_protocol, m3u8_id=m3u8_id,
2418 preference=preference, fatal=False)
2420 is_plain_url = False
2421 formats = self._extract_mpd_formats(
2422 full_url, video_id, mpd_id=mpd_id, fatal=False)
2427 'vcodec': 'none' if cur_media_type == 'audio' else None,
2429 return is_plain_url, formats
2432 # amp-video and amp-audio are very similar to their HTML5 counterparts
2433 # so we wll include them right here (see
2434 # https://www.ampproject.org/docs/reference/components/amp-video)
2435 media_tags = [(media_tag, media_type, '')
2436 for media_tag, media_type
2437 in re.findall(r'(?s)(<(?:amp-)?(video|audio)[^>]*/>)', webpage)]
2438 media_tags.extend(re.findall(
2439 # We only allow video|audio followed by a whitespace or '>'.
2440 # Allowing more characters may end up in significant slow down (see
2441 # https://github.com/rg3/youtube-dl/issues/11979, example URL:
2442 # http://www.porntrex.com/maps/videositemap.xml).
2443 r'(?s)(<(?P<tag>(?:amp-)?(?:video|audio))(?:\s+[^>]*)?>)(.*?)</(?P=tag)>', webpage))
2444 for media_tag, media_type, media_content in media_tags:
2449 media_attributes = extract_attributes(media_tag)
2450 src = media_attributes.get('src')
2452 _, formats = _media_formats(src, media_type)
2453 media_info['formats'].extend(formats)
2454 media_info['thumbnail'] = absolute_url(media_attributes.get('poster'))
2456 for source_tag in re.findall(r'<source[^>]+>', media_content):
2457 source_attributes = extract_attributes(source_tag)
2458 src = source_attributes.get('src')
2461 f = parse_content_type(source_attributes.get('type'))
2462 is_plain_url, formats = _media_formats(src, media_type, f)
2464 # res attribute is not standard but seen several times
2467 'height': int_or_none(source_attributes.get('res')),
2468 'format_id': source_attributes.get('label'),
2470 f.update(formats[0])
2471 media_info['formats'].append(f)
2473 media_info['formats'].extend(formats)
2474 for track_tag in re.findall(r'<track[^>]+>', media_content):
2475 track_attributes = extract_attributes(track_tag)
2476 kind = track_attributes.get('kind')
2477 if not kind or kind in ('subtitles', 'captions'):
2478 src = track_attributes.get('src')
2481 lang = track_attributes.get('srclang') or track_attributes.get('lang') or track_attributes.get('label')
2482 media_info['subtitles'].setdefault(lang, []).append({
2483 'url': absolute_url(src),
2485 for f in media_info['formats']:
2486 f.setdefault('http_headers', {})['Referer'] = base_url
2487 if media_info['formats'] or media_info['subtitles']:
2488 entries.append(media_info)
2491 def _extract_akamai_formats(self, manifest_url, video_id, hosts={}):
2493 hdcore_sign = 'hdcore=3.7.0'
2494 f4m_url = re.sub(r'(https?://[^/]+)/i/', r'\1/z/', manifest_url).replace('/master.m3u8', '/manifest.f4m')
2495 hds_host = hosts.get('hds')
2497 f4m_url = re.sub(r'(https?://)[^/]+', r'\1' + hds_host, f4m_url)
2498 if 'hdcore=' not in f4m_url:
2499 f4m_url += ('&' if '?' in f4m_url else '?') + hdcore_sign
2500 f4m_formats = self._extract_f4m_formats(
2501 f4m_url, video_id, f4m_id='hds', fatal=False)
2502 for entry in f4m_formats:
2503 entry.update({'extra_param_to_segment_url': hdcore_sign})
2504 formats.extend(f4m_formats)
2505 m3u8_url = re.sub(r'(https?://[^/]+)/z/', r'\1/i/', manifest_url).replace('/manifest.f4m', '/master.m3u8')
2506 hls_host = hosts.get('hls')
2508 m3u8_url = re.sub(r'(https?://)[^/]+', r'\1' + hls_host, m3u8_url)
2509 formats.extend(self._extract_m3u8_formats(
2510 m3u8_url, video_id, 'mp4', 'm3u8_native',
2511 m3u8_id='hls', fatal=False))
2514 def _extract_wowza_formats(self, url, video_id, m3u8_entry_protocol='m3u8_native', skip_protocols=[]):
2515 query = compat_urlparse.urlparse(url).query
2516 url = re.sub(r'/(?:manifest|playlist|jwplayer)\.(?:m3u8|f4m|mpd|smil)', '', url)
2518 r'(?:(?:http|rtmp|rtsp)(?P<s>s)?:)?(?P<url>//[^?]+)', url)
2519 url_base = mobj.group('url')
2520 http_base_url = '%s%s:%s' % ('http', mobj.group('s') or '', url_base)
2523 def manifest_url(manifest):
2524 m_url = '%s/%s' % (http_base_url, manifest)
2526 m_url += '?%s' % query
2529 if 'm3u8' not in skip_protocols:
2530 formats.extend(self._extract_m3u8_formats(
2531 manifest_url('playlist.m3u8'), video_id, 'mp4',
2532 m3u8_entry_protocol, m3u8_id='hls', fatal=False))
2533 if 'f4m' not in skip_protocols:
2534 formats.extend(self._extract_f4m_formats(
2535 manifest_url('manifest.f4m'),
2536 video_id, f4m_id='hds', fatal=False))
2537 if 'dash' not in skip_protocols:
2538 formats.extend(self._extract_mpd_formats(
2539 manifest_url('manifest.mpd'),
2540 video_id, mpd_id='dash', fatal=False))
2541 if re.search(r'(?:/smil:|\.smil)', url_base):
2542 if 'smil' not in skip_protocols:
2543 rtmp_formats = self._extract_smil_formats(
2544 manifest_url('jwplayer.smil'),
2545 video_id, fatal=False)
2546 for rtmp_format in rtmp_formats:
2547 rtsp_format = rtmp_format.copy()
2548 rtsp_format['url'] = '%s/%s' % (rtmp_format['url'], rtmp_format['play_path'])
2549 del rtsp_format['play_path']
2550 del rtsp_format['ext']
2551 rtsp_format.update({
2552 'url': rtsp_format['url'].replace('rtmp://', 'rtsp://'),
2553 'format_id': rtmp_format['format_id'].replace('rtmp', 'rtsp'),
2556 formats.extend([rtmp_format, rtsp_format])
2558 for protocol in ('rtmp', 'rtsp'):
2559 if protocol not in skip_protocols:
2561 'url': '%s:%s' % (protocol, url_base),
2562 'format_id': protocol,
2563 'protocol': protocol,
2567 def _find_jwplayer_data(self, webpage, video_id=None, transform_source=js_to_json):
2569 r'(?s)jwplayer\((?P<quote>[\'"])[^\'" ]+(?P=quote)\)(?!</script>).*?\.setup\s*\((?P<options>[^)]+)\)',
2573 jwplayer_data = self._parse_json(mobj.group('options'),
2575 transform_source=transform_source)
2576 except ExtractorError:
2579 if isinstance(jwplayer_data, dict):
2580 return jwplayer_data
2582 def _extract_jwplayer_data(self, webpage, video_id, *args, **kwargs):
2583 jwplayer_data = self._find_jwplayer_data(
2584 webpage, video_id, transform_source=js_to_json)
2585 return self._parse_jwplayer_data(
2586 jwplayer_data, video_id, *args, **kwargs)
2588 def _parse_jwplayer_data(self, jwplayer_data, video_id=None, require_title=True,
2589 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
2590 # JWPlayer backward compatibility: flattened playlists
2591 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/api/config.js#L81-L96
2592 if 'playlist' not in jwplayer_data:
2593 jwplayer_data = {'playlist': [jwplayer_data]}
2597 # JWPlayer backward compatibility: single playlist item
2598 # https://github.com/jwplayer/jwplayer/blob/v7.7.0/src/js/playlist/playlist.js#L10
2599 if not isinstance(jwplayer_data['playlist'], list):
2600 jwplayer_data['playlist'] = [jwplayer_data['playlist']]
2602 for video_data in jwplayer_data['playlist']:
2603 # JWPlayer backward compatibility: flattened sources
2604 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/playlist/item.js#L29-L35
2605 if 'sources' not in video_data:
2606 video_data['sources'] = [video_data]
2608 this_video_id = video_id or video_data['mediaid']
2610 formats = self._parse_jwplayer_formats(
2611 video_data['sources'], video_id=this_video_id, m3u8_id=m3u8_id,
2612 mpd_id=mpd_id, rtmp_params=rtmp_params, base_url=base_url)
2615 tracks = video_data.get('tracks')
2616 if tracks and isinstance(tracks, list):
2617 for track in tracks:
2618 if not isinstance(track, dict):
2620 track_kind = track.get('kind')
2621 if not track_kind or not isinstance(track_kind, compat_str):
2623 if track_kind.lower() not in ('captions', 'subtitles'):
2625 track_url = urljoin(base_url, track.get('file'))
2628 subtitles.setdefault(track.get('label') or 'en', []).append({
2629 'url': self._proto_relative_url(track_url)
2633 'id': this_video_id,
2634 'title': unescapeHTML(video_data['title'] if require_title else video_data.get('title')),
2635 'description': video_data.get('description'),
2636 'thumbnail': urljoin(base_url, self._proto_relative_url(video_data.get('image'))),
2637 'timestamp': int_or_none(video_data.get('pubdate')),
2638 'duration': float_or_none(jwplayer_data.get('duration') or video_data.get('duration')),
2639 'subtitles': subtitles,
2641 # https://github.com/jwplayer/jwplayer/blob/master/src/js/utils/validator.js#L32
2642 if len(formats) == 1 and re.search(r'^(?:http|//).*(?:youtube\.com|youtu\.be)/.+', formats[0]['url']):
2644 '_type': 'url_transparent',
2645 'url': formats[0]['url'],
2648 self._sort_formats(formats)
2649 entry['formats'] = formats
2650 entries.append(entry)
2651 if len(entries) == 1:
2654 return self.playlist_result(entries)
2656 def _parse_jwplayer_formats(self, jwplayer_sources_data, video_id=None,
2657 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
2660 for source in jwplayer_sources_data:
2661 if not isinstance(source, dict):
2663 source_url = urljoin(
2664 base_url, self._proto_relative_url(source.get('file')))
2665 if not source_url or source_url in urls:
2667 urls.append(source_url)
2668 source_type = source.get('type') or ''
2669 ext = mimetype2ext(source_type) or determine_ext(source_url)
2670 if source_type == 'hls' or ext == 'm3u8':
2671 formats.extend(self._extract_m3u8_formats(
2672 source_url, video_id, 'mp4', entry_protocol='m3u8_native',
2673 m3u8_id=m3u8_id, fatal=False))
2674 elif source_type == 'dash' or ext == 'mpd':
2675 formats.extend(self._extract_mpd_formats(
2676 source_url, video_id, mpd_id=mpd_id, fatal=False))
2678 formats.extend(self._extract_smil_formats(
2679 source_url, video_id, fatal=False))
2680 # https://github.com/jwplayer/jwplayer/blob/master/src/js/providers/default.js#L67
2681 elif source_type.startswith('audio') or ext in (
2682 'oga', 'aac', 'mp3', 'mpeg', 'vorbis'):
2689 height = int_or_none(source.get('height'))
2691 # Often no height is provided but there is a label in
2692 # format like "1080p", "720p SD", or 1080.
2693 height = int_or_none(self._search_regex(
2694 r'^(\d{3,4})[pP]?(?:\b|$)', compat_str(source.get('label') or ''),
2695 'height', default=None))
2698 'width': int_or_none(source.get('width')),
2700 'tbr': int_or_none(source.get('bitrate')),
2703 if source_url.startswith('rtmp'):
2704 a_format['ext'] = 'flv'
2705 # See com/longtailvideo/jwplayer/media/RTMPMediaProvider.as
2706 # of jwplayer.flash.swf
2707 rtmp_url_parts = re.split(
2708 r'((?:mp4|mp3|flv):)', source_url, 1)
2709 if len(rtmp_url_parts) == 3:
2710 rtmp_url, prefix, play_path = rtmp_url_parts
2713 'play_path': prefix + play_path,
2716 a_format.update(rtmp_params)
2717 formats.append(a_format)
2720 def _live_title(self, name):
2721 """ Generate the title for a live video """
2722 now = datetime.datetime.now()
2723 now_str = now.strftime('%Y-%m-%d %H:%M')
2724 return name + ' ' + now_str
2726 def _int(self, v, name, fatal=False, **kwargs):
2727 res = int_or_none(v, **kwargs)
2728 if 'get_attr' in kwargs:
2729 print(getattr(v, kwargs['get_attr']))
2731 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
2733 raise ExtractorError(msg)
2735 self._downloader.report_warning(msg)
2738 def _float(self, v, name, fatal=False, **kwargs):
2739 res = float_or_none(v, **kwargs)
2741 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
2743 raise ExtractorError(msg)
2745 self._downloader.report_warning(msg)
2748 def _set_cookie(self, domain, name, value, expire_time=None, port=None,
2749 path='/', secure=False, discard=False, rest={}, **kwargs):
2750 cookie = compat_cookiejar.Cookie(
2751 0, name, value, port, port is not None, domain, True,
2752 domain.startswith('.'), path, True, secure, expire_time,
2753 discard, None, None, rest)
2754 self._downloader.cookiejar.set_cookie(cookie)
2756 def _get_cookies(self, url):
2757 """ Return a compat_cookies.SimpleCookie with the cookies for the url """
2758 req = sanitized_Request(url)
2759 self._downloader.cookiejar.add_cookie_header(req)
2760 return compat_cookies.SimpleCookie(req.get_header('Cookie'))
2762 def get_testcases(self, include_onlymatching=False):
2763 t = getattr(self, '_TEST', None)
2765 assert not hasattr(self, '_TESTS'), \
2766 '%s has _TEST and _TESTS' % type(self).__name__
2769 tests = getattr(self, '_TESTS', [])
2771 if not include_onlymatching and t.get('only_matching', False):
2773 t['name'] = type(self).__name__[:-len('IE')]
2776 def is_suitable(self, age_limit):
2777 """ Test whether the extractor is generally suitable for the given
2778 age limit (i.e. pornographic sites are not, all others usually are) """
2780 any_restricted = False
2781 for tc in self.get_testcases(include_onlymatching=False):
2782 if tc.get('playlist', []):
2783 tc = tc['playlist'][0]
2784 is_restricted = age_restricted(
2785 tc.get('info_dict', {}).get('age_limit'), age_limit)
2786 if not is_restricted:
2788 any_restricted = any_restricted or is_restricted
2789 return not any_restricted
2791 def extract_subtitles(self, *args, **kwargs):
2792 if (self._downloader.params.get('writesubtitles', False) or
2793 self._downloader.params.get('listsubtitles')):
2794 return self._get_subtitles(*args, **kwargs)
2797 def _get_subtitles(self, *args, **kwargs):
2798 raise NotImplementedError('This method must be implemented by subclasses')
2801 def _merge_subtitle_items(subtitle_list1, subtitle_list2):
2802 """ Merge subtitle items for one language. Items with duplicated URLs
2803 will be dropped. """
2804 list1_urls = set([item['url'] for item in subtitle_list1])
2805 ret = list(subtitle_list1)
2806 ret.extend([item for item in subtitle_list2 if item['url'] not in list1_urls])
2810 def _merge_subtitles(cls, subtitle_dict1, subtitle_dict2):
2811 """ Merge two subtitle dictionaries, language by language. """
2812 ret = dict(subtitle_dict1)
2813 for lang in subtitle_dict2:
2814 ret[lang] = cls._merge_subtitle_items(subtitle_dict1.get(lang, []), subtitle_dict2[lang])
2817 def extract_automatic_captions(self, *args, **kwargs):
2818 if (self._downloader.params.get('writeautomaticsub', False) or
2819 self._downloader.params.get('listsubtitles')):
2820 return self._get_automatic_captions(*args, **kwargs)
2823 def _get_automatic_captions(self, *args, **kwargs):
2824 raise NotImplementedError('This method must be implemented by subclasses')
2826 def mark_watched(self, *args, **kwargs):
2827 if (self._downloader.params.get('mark_watched', False) and
2828 (self._get_login_info()[0] is not None or
2829 self._downloader.params.get('cookiefile') is not None)):
2830 self._mark_watched(*args, **kwargs)
2832 def _mark_watched(self, *args, **kwargs):
2833 raise NotImplementedError('This method must be implemented by subclasses')
2835 def geo_verification_headers(self):
2837 geo_verification_proxy = self._downloader.params.get('geo_verification_proxy')
2838 if geo_verification_proxy:
2839 headers['Ytdl-request-proxy'] = geo_verification_proxy
2842 def _generic_id(self, url):
2843 return compat_urllib_parse_unquote(os.path.splitext(url.rstrip('/').split('/')[-1])[0])
2845 def _generic_title(self, url):
2846 return compat_urllib_parse_unquote(os.path.splitext(url_basename(url))[0])
2849 class SearchInfoExtractor(InfoExtractor):
2851 Base class for paged search queries extractors.
2852 They accept URLs in the format _SEARCH_KEY(|all|[0-9]):{query}
2853 Instances should define _SEARCH_KEY and _MAX_RESULTS.
2857 def _make_valid_url(cls):
2858 return r'%s(?P<prefix>|[1-9][0-9]*|all):(?P<query>[\s\S]+)' % cls._SEARCH_KEY
2861 def suitable(cls, url):
2862 return re.match(cls._make_valid_url(), url) is not None
2864 def _real_extract(self, query):
2865 mobj = re.match(self._make_valid_url(), query)
2867 raise ExtractorError('Invalid search query "%s"' % query)
2869 prefix = mobj.group('prefix')
2870 query = mobj.group('query')
2872 return self._get_n_results(query, 1)
2873 elif prefix == 'all':
2874 return self._get_n_results(query, self._MAX_RESULTS)
2878 raise ExtractorError('invalid download number %s for query "%s"' % (n, query))
2879 elif n > self._MAX_RESULTS:
2880 self._downloader.report_warning('%s returns max %i results (you requested %i)' % (self._SEARCH_KEY, self._MAX_RESULTS, n))
2881 n = self._MAX_RESULTS
2882 return self._get_n_results(query, n)
2884 def _get_n_results(self, query, n):
2885 """Get a specified number of results for a query"""
2886 raise NotImplementedError('This method must be implemented by subclasses')
2889 def SEARCH_KEY(self):
2890 return self._SEARCH_KEY