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 The mandatory URL representing the media:
106 for plain file media - HTTP URL of this file,
108 for HLS - URL of the M3U8 media playlist,
109 for HDS - URL of the F4M manifest,
110 for DASH - URL of the MPD manifest,
111 for MSS - URL of the ISM manifest.
113 The URL of the manifest file in case of
115 for HLS - URL of the M3U8 master playlist,
116 for HDS - URL of the F4M manifest,
117 for DASH - URL of the MPD manifest,
118 for MSS - URL of the ISM manifest.
119 * ext Will be calculated from URL if missing
120 * format A human-readable description of the format
121 ("mp4 container with h264/opus").
122 Calculated from the format_id, width, height.
123 and format_note fields if missing.
124 * format_id A short description of the format
125 ("mp4_h264_opus" or "19").
126 Technically optional, but strongly recommended.
127 * format_note Additional info about the format
128 ("3D" or "DASH video")
129 * width Width of the video, if known
130 * height Height of the video, if known
131 * resolution Textual description of width and height
132 * tbr Average bitrate of audio and video in KBit/s
133 * abr Average audio bitrate in KBit/s
134 * acodec Name of the audio codec in use
135 * asr Audio sampling rate in Hertz
136 * vbr Average video bitrate in KBit/s
138 * vcodec Name of the video codec in use
139 * container Name of the container format
140 * filesize The number of bytes, if known in advance
141 * filesize_approx An estimate for the number of bytes
142 * player_url SWF Player URL (used for rtmpdump).
143 * protocol The protocol that will be used for the actual
144 download, lower-case.
145 "http", "https", "rtsp", "rtmp", "rtmpe",
146 "m3u8", "m3u8_native" or "http_dash_segments".
148 Base URL for fragments. Each fragment's path
149 value (if present) will be relative to
151 * fragments A list of fragments of a fragmented media.
152 Each fragment entry must contain either an url
153 or a path. If an url is present it should be
154 considered by a client. Otherwise both path and
155 fragment_base_url must be present. Here is
156 the list of all potential fields:
157 * "url" - fragment's URL
158 * "path" - fragment's path relative to
160 * "duration" (optional, int or float)
161 * "filesize" (optional, int)
162 * preference Order number of this format. If this field is
163 present and not None, the formats get sorted
164 by this field, regardless of all other values.
165 -1 for default (order by other properties),
166 -2 or smaller for less than default.
167 < -1000 to hide the format (if there is
168 another one which is strictly better)
169 * language Language code, e.g. "de" or "en-US".
170 * language_preference Is this in the language mentioned in
172 10 if it's what the URL is about,
173 -1 for default (don't know),
174 -10 otherwise, other values reserved for now.
175 * quality Order number of the video quality of this
176 format, irrespective of the file format.
177 -1 for default (order by other properties),
178 -2 or smaller for less than default.
179 * source_preference Order number for this video source
180 (quality takes higher priority)
181 -1 for default (order by other properties),
182 -2 or smaller for less than default.
183 * http_headers A dictionary of additional HTTP headers
184 to add to the request.
185 * stretched_ratio If given and not 1, indicates that the
186 video's pixels are not square.
187 width : height ratio as float.
188 * no_resume The server does not support resuming the
189 (HTTP or RTMP) download. Boolean.
190 * downloader_options A dictionary of downloader options as
191 described in FileDownloader
193 url: Final video URL.
194 ext: Video filename extension.
195 format: The video format, defaults to ext (used for --get-format)
196 player_url: SWF Player URL (used for rtmpdump).
198 The following fields are optional:
200 alt_title: A secondary title of the video.
201 display_id An alternative identifier for the video, not necessarily
202 unique, but available before title. Typically, id is
203 something like "4234987", title "Dancing naked mole rats",
204 and display_id "dancing-naked-mole-rats"
205 thumbnails: A list of dictionaries, with the following entries:
206 * "id" (optional, string) - Thumbnail format ID
208 * "preference" (optional, int) - quality of the image
209 * "width" (optional, int)
210 * "height" (optional, int)
211 * "resolution" (optional, string "{width}x{height"},
213 * "filesize" (optional, int)
214 thumbnail: Full URL to a video thumbnail image.
215 description: Full video description.
216 uploader: Full name of the video uploader.
217 license: License name the video is licensed under.
218 creator: The creator of the video.
219 release_date: The date (YYYYMMDD) when the video was released.
220 timestamp: UNIX timestamp of the moment the video became available.
221 upload_date: Video upload date (YYYYMMDD).
222 If not explicitly set, calculated from timestamp.
223 uploader_id: Nickname or id of the video uploader.
224 uploader_url: Full URL to a personal webpage of the video uploader.
225 channel: Full name of the channel the video is uploaded on.
226 Note that channel fields may or may not repeat uploader
227 fields. This depends on a particular extractor.
228 channel_id: Id of the channel.
229 channel_url: Full URL to a channel webpage.
230 location: Physical location where the video was filmed.
231 subtitles: The available subtitles as a dictionary in the format
232 {tag: subformats}. "tag" is usually a language code, and
233 "subformats" is a list sorted from lower to higher
234 preference, each element is a dictionary with the "ext"
236 * "data": The subtitles file contents
237 * "url": A URL pointing to the subtitles file
238 "ext" will be calculated from URL if missing
239 automatic_captions: Like 'subtitles', used by the YoutubeIE for
240 automatically generated captions
241 duration: Length of the video in seconds, as an integer or float.
242 view_count: How many users have watched the video on the platform.
243 like_count: Number of positive ratings of the video
244 dislike_count: Number of negative ratings of the video
245 repost_count: Number of reposts of the video
246 average_rating: Average rating give by users, the scale used depends on the webpage
247 comment_count: Number of comments on the video
248 comments: A list of comments, each with one or more of the following
249 properties (all but one of text or html optional):
250 * "author" - human-readable name of the comment author
251 * "author_id" - user ID of the comment author
253 * "html" - Comment as HTML
254 * "text" - Plain text of the comment
255 * "timestamp" - UNIX timestamp of comment
256 * "parent" - ID of the comment this one is replying to.
257 Set to "root" to indicate that this is a
258 comment to the original video.
259 age_limit: Age restriction for the video, as an integer (years)
260 webpage_url: The URL to the video webpage, if given to youtube-dl it
261 should allow to get the same result again. (It will be set
262 by YoutubeDL if it's missing)
263 categories: A list of categories that the video falls in, for example
265 tags: A list of tags assigned to the video, e.g. ["sweden", "pop music"]
266 is_live: True, False, or None (=unknown). Whether this video is a
267 live stream that goes on instead of a fixed-length video.
268 start_time: Time in seconds where the reproduction should start, as
269 specified in the URL.
270 end_time: Time in seconds where the reproduction should end, as
271 specified in the URL.
272 chapters: A list of dictionaries, with the following entries:
273 * "start_time" - The start time of the chapter in seconds
274 * "end_time" - The end time of the chapter in seconds
275 * "title" (optional, string)
277 The following fields should only be used when the video belongs to some logical
280 chapter: Name or title of the chapter the video belongs to.
281 chapter_number: Number of the chapter the video belongs to, as an integer.
282 chapter_id: Id of the chapter the video belongs to, as a unicode string.
284 The following fields should only be used when the video is an episode of some
285 series, programme or podcast:
287 series: Title of the series or programme the video episode belongs to.
288 season: Title of the season the video episode belongs to.
289 season_number: Number of the season the video episode belongs to, as an integer.
290 season_id: Id of the season the video episode belongs to, as a unicode string.
291 episode: Title of the video episode. Unlike mandatory video title field,
292 this field should denote the exact title of the video episode
293 without any kind of decoration.
294 episode_number: Number of the video episode within a season, as an integer.
295 episode_id: Id of the video episode, as a unicode string.
297 The following fields should only be used when the media is a track or a part of
300 track: Title of the track.
301 track_number: Number of the track within an album or a disc, as an integer.
302 track_id: Id of the track (useful in case of custom indexing, e.g. 6.iii),
304 artist: Artist(s) of the track.
305 genre: Genre(s) of the track.
306 album: Title of the album the track belongs to.
307 album_type: Type of the album (e.g. "Demo", "Full-length", "Split", "Compilation", etc).
308 album_artist: List of all artists appeared on the album (e.g.
309 "Ash Borer / Fell Voices" or "Various Artists", useful for splits
311 disc_number: Number of the disc or other physical medium the track belongs to,
313 release_year: Year (YYYY) when the album was released.
315 Unless mentioned otherwise, the fields should be Unicode strings.
317 Unless mentioned otherwise, None is equivalent to absence of information.
320 _type "playlist" indicates multiple videos.
321 There must be a key "entries", which is a list, an iterable, or a PagedList
322 object, each element of which is a valid dictionary by this specification.
324 Additionally, playlists can have "id", "title", "description", "uploader",
325 "uploader_id", "uploader_url" attributes with the same semantics as videos
329 _type "multi_video" indicates that there are multiple videos that
330 form a single show, for examples multiple acts of an opera or TV episode.
331 It must have an entries key like a playlist and contain all the keys
332 required for a video at the same time.
335 _type "url" indicates that the video must be extracted from another
336 location, possibly by a different extractor. Its only required key is:
337 "url" - the next URL to extract.
338 The key "ie_key" can be set to the class name (minus the trailing "IE",
339 e.g. "Youtube") if the extractor class is known in advance.
340 Additionally, the dictionary may have any properties of the resolved entity
341 known in advance, for example "title" if the title of the referred video is
345 _type "url_transparent" entities have the same specification as "url", but
346 indicate that the given additional information is more precise than the one
347 associated with the resolved URL.
348 This is useful when a site employs a video service that hosts the video and
349 its technical metadata, but that video service does not embed a useful
350 title, description etc.
353 Subclasses of this one should re-define the _real_initialize() and
354 _real_extract() methods and define a _VALID_URL regexp.
355 Probably, they should also be added to the list of extractors.
357 _GEO_BYPASS attribute may be set to False in order to disable
358 geo restriction bypass mechanisms for a particular extractor.
359 Though it won't disable explicit geo restriction bypass based on
360 country code provided with geo_bypass_country.
362 _GEO_COUNTRIES attribute may contain a list of presumably geo unrestricted
363 countries for this extractor. One of these countries will be used by
364 geo restriction bypass mechanism right away in order to bypass
365 geo restriction, of course, if the mechanism is not disabled.
367 _GEO_IP_BLOCKS attribute may contain a list of presumably geo unrestricted
368 IP blocks in CIDR notation for this extractor. One of these IP blocks
369 will be used by geo restriction bypass mechanism similarly
372 Finally, the _WORKING attribute should be set to False for broken IEs
373 in order to warn the users and skip the tests.
378 _x_forwarded_for_ip = None
380 _GEO_COUNTRIES = None
381 _GEO_IP_BLOCKS = None
384 def __init__(self, downloader=None):
385 """Constructor. Receives an optional downloader."""
387 self._x_forwarded_for_ip = None
388 self.set_downloader(downloader)
391 def suitable(cls, url):
392 """Receives a URL and returns True if suitable for this IE."""
394 # This does not use has/getattr intentionally - we want to know whether
395 # we have cached the regexp for *this* class, whereas getattr would also
396 # match the superclass
397 if '_VALID_URL_RE' not in cls.__dict__:
398 cls._VALID_URL_RE = re.compile(cls._VALID_URL)
399 return cls._VALID_URL_RE.match(url) is not None
402 def _match_id(cls, url):
403 if '_VALID_URL_RE' not in cls.__dict__:
404 cls._VALID_URL_RE = re.compile(cls._VALID_URL)
405 m = cls._VALID_URL_RE.match(url)
407 return compat_str(m.group('id'))
411 """Getter method for _WORKING."""
414 def initialize(self):
415 """Initializes an instance (authentication, etc)."""
416 self._initialize_geo_bypass({
417 'countries': self._GEO_COUNTRIES,
418 'ip_blocks': self._GEO_IP_BLOCKS,
421 self._real_initialize()
424 def _initialize_geo_bypass(self, geo_bypass_context):
426 Initialize geo restriction bypass mechanism.
428 This method is used to initialize geo bypass mechanism based on faking
429 X-Forwarded-For HTTP header. A random country from provided country list
430 is selected and a random IP belonging to this country is generated. This
431 IP will be passed as X-Forwarded-For HTTP header in all subsequent
434 This method will be used for initial geo bypass mechanism initialization
435 during the instance initialization with _GEO_COUNTRIES and
438 You may also manually call it from extractor's code if geo bypass
439 information is not available beforehand (e.g. obtained during
440 extraction) or due to some other reason. In this case you should pass
441 this information in geo bypass context passed as first argument. It may
442 contain following fields:
444 countries: List of geo unrestricted countries (similar
446 ip_blocks: List of geo unrestricted IP blocks in CIDR notation
447 (similar to _GEO_IP_BLOCKS)
450 if not self._x_forwarded_for_ip:
452 # Geo bypass mechanism is explicitly disabled by user
453 if not self._downloader.params.get('geo_bypass', True):
456 if not geo_bypass_context:
457 geo_bypass_context = {}
459 # Backward compatibility: previously _initialize_geo_bypass
460 # expected a list of countries, some 3rd party code may still use
462 if isinstance(geo_bypass_context, (list, tuple)):
463 geo_bypass_context = {
464 'countries': geo_bypass_context,
467 # The whole point of geo bypass mechanism is to fake IP
468 # as X-Forwarded-For HTTP header based on some IP block or
471 # Path 1: bypassing based on IP block in CIDR notation
473 # Explicit IP block specified by user, use it right away
474 # regardless of whether extractor is geo bypassable or not
475 ip_block = self._downloader.params.get('geo_bypass_ip_block', None)
477 # Otherwise use random IP block from geo bypass context but only
478 # if extractor is known as geo bypassable
480 ip_blocks = geo_bypass_context.get('ip_blocks')
481 if self._GEO_BYPASS and ip_blocks:
482 ip_block = random.choice(ip_blocks)
485 self._x_forwarded_for_ip = GeoUtils.random_ipv4(ip_block)
486 if self._downloader.params.get('verbose', False):
487 self._downloader.to_screen(
488 '[debug] Using fake IP %s as X-Forwarded-For.'
489 % self._x_forwarded_for_ip)
492 # Path 2: bypassing based on country code
494 # Explicit country code specified by user, use it right away
495 # regardless of whether extractor is geo bypassable or not
496 country = self._downloader.params.get('geo_bypass_country', None)
498 # Otherwise use random country code from geo bypass context but
499 # only if extractor is known as geo bypassable
501 countries = geo_bypass_context.get('countries')
502 if self._GEO_BYPASS and countries:
503 country = random.choice(countries)
506 self._x_forwarded_for_ip = GeoUtils.random_ipv4(country)
507 if self._downloader.params.get('verbose', False):
508 self._downloader.to_screen(
509 '[debug] Using fake IP %s (%s) as X-Forwarded-For.'
510 % (self._x_forwarded_for_ip, country.upper()))
512 def extract(self, url):
513 """Extracts URL information and returns it in list of dicts."""
518 ie_result = self._real_extract(url)
519 if self._x_forwarded_for_ip:
520 ie_result['__x_forwarded_for_ip'] = self._x_forwarded_for_ip
522 except GeoRestrictedError as e:
523 if self.__maybe_fake_ip_and_retry(e.countries):
526 except ExtractorError:
528 except compat_http_client.IncompleteRead as e:
529 raise ExtractorError('A network error has occurred.', cause=e, expected=True)
530 except (KeyError, StopIteration) as e:
531 raise ExtractorError('An extractor error has occurred.', cause=e)
533 def __maybe_fake_ip_and_retry(self, countries):
534 if (not self._downloader.params.get('geo_bypass_country', None) and
536 self._downloader.params.get('geo_bypass', True) and
537 not self._x_forwarded_for_ip and
539 country_code = random.choice(countries)
540 self._x_forwarded_for_ip = GeoUtils.random_ipv4(country_code)
541 if self._x_forwarded_for_ip:
543 'Video is geo restricted. Retrying extraction with fake IP %s (%s) as X-Forwarded-For.'
544 % (self._x_forwarded_for_ip, country_code.upper()))
548 def set_downloader(self, downloader):
549 """Sets the downloader for this IE."""
550 self._downloader = downloader
552 def _real_initialize(self):
553 """Real initialization process. Redefine in subclasses."""
556 def _real_extract(self, url):
557 """Real extraction process. Redefine in subclasses."""
562 """A string for getting the InfoExtractor with get_info_extractor"""
563 return compat_str(cls.__name__[:-2])
567 return compat_str(type(self).__name__[:-2])
570 def __can_accept_status_code(err, expected_status):
571 assert isinstance(err, compat_urllib_error.HTTPError)
572 if expected_status is None:
574 if isinstance(expected_status, compat_integer_types):
575 return err.code == expected_status
576 elif isinstance(expected_status, (list, tuple)):
577 return err.code in expected_status
578 elif callable(expected_status):
579 return expected_status(err.code) is True
583 def _request_webpage(self, url_or_request, video_id, note=None, errnote=None, fatal=True, data=None, headers={}, query={}, expected_status=None):
585 Return the response handle.
587 See _download_webpage docstring for arguments specification.
590 self.report_download_webpage(video_id)
591 elif note is not False:
593 self.to_screen('%s' % (note,))
595 self.to_screen('%s: %s' % (video_id, note))
597 # Some sites check X-Forwarded-For HTTP header in order to figure out
598 # the origin of the client behind proxy. This allows bypassing geo
599 # restriction by faking this header's value to IP that belongs to some
600 # geo unrestricted country. We will do so once we encounter any
601 # geo restriction error.
602 if self._x_forwarded_for_ip:
603 if 'X-Forwarded-For' not in headers:
604 headers['X-Forwarded-For'] = self._x_forwarded_for_ip
606 if isinstance(url_or_request, compat_urllib_request.Request):
607 url_or_request = update_Request(
608 url_or_request, data=data, headers=headers, query=query)
611 url_or_request = update_url_query(url_or_request, query)
612 if data is not None or headers:
613 url_or_request = sanitized_Request(url_or_request, data, headers)
615 return self._downloader.urlopen(url_or_request)
616 except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err:
617 if isinstance(err, compat_urllib_error.HTTPError):
618 if self.__can_accept_status_code(err, expected_status):
619 # Retain reference to error to prevent file object from
620 # being closed before it can be read. Works around the
621 # effects of <https://bugs.python.org/issue15002>
622 # introduced in Python 3.4.1.
629 errnote = 'Unable to download webpage'
631 errmsg = '%s: %s' % (errnote, error_to_compat_str(err))
633 raise ExtractorError(errmsg, sys.exc_info()[2], cause=err)
635 self._downloader.report_warning(errmsg)
638 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):
640 Return a tuple (page content as string, URL handle).
642 See _download_webpage docstring for arguments specification.
644 # Strip hashes from the URL (#1038)
645 if isinstance(url_or_request, (compat_str, str)):
646 url_or_request = url_or_request.partition('#')[0]
648 urlh = self._request_webpage(url_or_request, video_id, note, errnote, fatal, data=data, headers=headers, query=query, expected_status=expected_status)
652 content = self._webpage_read_content(urlh, url_or_request, video_id, note, errnote, fatal, encoding=encoding)
653 return (content, urlh)
656 def _guess_encoding_from_content(content_type, webpage_bytes):
657 m = re.match(r'[a-zA-Z0-9_.-]+/[a-zA-Z0-9_.-]+\s*;\s*charset=(.+)', content_type)
659 encoding = m.group(1)
661 m = re.search(br'<meta[^>]+charset=[\'"]?([^\'")]+)[ /\'">]',
662 webpage_bytes[:1024])
664 encoding = m.group(1).decode('ascii')
665 elif webpage_bytes.startswith(b'\xff\xfe'):
672 def __check_blocked(self, content):
673 first_block = content[:512]
674 if ('<title>Access to this site is blocked</title>' in content and
675 'Websense' in first_block):
676 msg = 'Access to this webpage has been blocked by Websense filtering software in your network.'
677 blocked_iframe = self._html_search_regex(
678 r'<iframe src="([^"]+)"', content,
679 'Websense information URL', default=None)
681 msg += ' Visit %s for more details' % blocked_iframe
682 raise ExtractorError(msg, expected=True)
683 if '<title>The URL you requested has been blocked</title>' in first_block:
685 'Access to this webpage has been blocked by Indian censorship. '
686 'Use a VPN or proxy server (with --proxy) to route around it.')
687 block_msg = self._html_search_regex(
688 r'</h1><p>(.*?)</p>',
689 content, 'block message', default=None)
691 msg += ' (Message: "%s")' % block_msg.replace('\n', ' ')
692 raise ExtractorError(msg, expected=True)
693 if ('<title>TTK :: Доступ к ресурсу ограничен</title>' in content and
694 'blocklist.rkn.gov.ru' in content):
695 raise ExtractorError(
696 'Access to this webpage has been blocked by decision of the Russian government. '
697 'Visit http://blocklist.rkn.gov.ru/ for a block reason.',
700 def _webpage_read_content(self, urlh, url_or_request, video_id, note=None, errnote=None, fatal=True, prefix=None, encoding=None):
701 content_type = urlh.headers.get('Content-Type', '')
702 webpage_bytes = urlh.read()
703 if prefix is not None:
704 webpage_bytes = prefix + webpage_bytes
706 encoding = self._guess_encoding_from_content(content_type, webpage_bytes)
707 if self._downloader.params.get('dump_intermediate_pages', False):
708 self.to_screen('Dumping request to ' + urlh.geturl())
709 dump = base64.b64encode(webpage_bytes).decode('ascii')
710 self._downloader.to_screen(dump)
711 if self._downloader.params.get('write_pages', False):
712 basen = '%s_%s' % (video_id, urlh.geturl())
714 h = '___' + hashlib.md5(basen.encode('utf-8')).hexdigest()
715 basen = basen[:240 - len(h)] + h
716 raw_filename = basen + '.dump'
717 filename = sanitize_filename(raw_filename, restricted=True)
718 self.to_screen('Saving request to ' + filename)
719 # Working around MAX_PATH limitation on Windows (see
720 # http://msdn.microsoft.com/en-us/library/windows/desktop/aa365247(v=vs.85).aspx)
721 if compat_os_name == 'nt':
722 absfilepath = os.path.abspath(filename)
723 if len(absfilepath) > 259:
724 filename = '\\\\?\\' + absfilepath
725 with open(filename, 'wb') as outf:
726 outf.write(webpage_bytes)
729 content = webpage_bytes.decode(encoding, 'replace')
731 content = webpage_bytes.decode('utf-8', 'replace')
733 self.__check_blocked(content)
737 def _download_webpage(
738 self, url_or_request, video_id, note=None, errnote=None,
739 fatal=True, tries=1, timeout=5, encoding=None, data=None,
740 headers={}, query={}, expected_status=None):
742 Return the data of the page as a string.
745 url_or_request -- plain text URL as a string or
746 a compat_urllib_request.Requestobject
747 video_id -- Video/playlist/item identifier (string)
750 note -- note printed before downloading (string)
751 errnote -- note printed in case of an error (string)
752 fatal -- flag denoting whether error should be considered fatal,
753 i.e. whether it should cause ExtractionError to be raised,
754 otherwise a warning will be reported and extraction continued
755 tries -- number of tries
756 timeout -- sleep interval between tries
757 encoding -- encoding for a page content decoding, guessed automatically
758 when not explicitly specified
759 data -- POST data (bytes)
760 headers -- HTTP headers (dict)
761 query -- URL query (dict)
762 expected_status -- allows to accept failed HTTP requests (non 2xx
763 status code) by explicitly specifying a set of accepted status
764 codes. Can be any of the following entities:
765 - an integer type specifying an exact failed status code to
767 - a list or a tuple of integer types specifying a list of
768 failed status codes to accept
769 - a callable accepting an actual failed status code and
770 returning True if it should be accepted
771 Note that this argument does not affect success status codes (2xx)
772 which are always accepted.
777 while success is False:
779 res = self._download_webpage_handle(
780 url_or_request, video_id, note, errnote, fatal,
781 encoding=encoding, data=data, headers=headers, query=query,
782 expected_status=expected_status)
784 except compat_http_client.IncompleteRead as e:
786 if try_count >= tries:
788 self._sleep(timeout, video_id)
795 def _download_xml_handle(
796 self, url_or_request, video_id, note='Downloading XML',
797 errnote='Unable to download XML', transform_source=None,
798 fatal=True, encoding=None, data=None, headers={}, query={},
799 expected_status=None):
801 Return a tuple (xml as an xml.etree.ElementTree.Element, URL handle).
803 See _download_webpage docstring for arguments specification.
805 res = self._download_webpage_handle(
806 url_or_request, video_id, note, errnote, fatal=fatal,
807 encoding=encoding, data=data, headers=headers, query=query,
808 expected_status=expected_status)
811 xml_string, urlh = res
812 return self._parse_xml(
813 xml_string, video_id, transform_source=transform_source,
817 self, url_or_request, video_id,
818 note='Downloading XML', errnote='Unable to download XML',
819 transform_source=None, fatal=True, encoding=None,
820 data=None, headers={}, query={}, expected_status=None):
822 Return the xml as an xml.etree.ElementTree.Element.
824 See _download_webpage docstring for arguments specification.
826 res = self._download_xml_handle(
827 url_or_request, video_id, note=note, errnote=errnote,
828 transform_source=transform_source, fatal=fatal, encoding=encoding,
829 data=data, headers=headers, query=query,
830 expected_status=expected_status)
831 return res if res is False else res[0]
833 def _parse_xml(self, xml_string, video_id, transform_source=None, fatal=True):
835 xml_string = transform_source(xml_string)
837 return compat_etree_fromstring(xml_string.encode('utf-8'))
838 except compat_xml_parse_error as ve:
839 errmsg = '%s: Failed to parse XML ' % video_id
841 raise ExtractorError(errmsg, cause=ve)
843 self.report_warning(errmsg + str(ve))
845 def _download_json_handle(
846 self, url_or_request, video_id, note='Downloading JSON metadata',
847 errnote='Unable to download JSON metadata', transform_source=None,
848 fatal=True, encoding=None, data=None, headers={}, query={},
849 expected_status=None):
851 Return a tuple (JSON object, URL handle).
853 See _download_webpage docstring for arguments specification.
855 res = self._download_webpage_handle(
856 url_or_request, video_id, note, errnote, fatal=fatal,
857 encoding=encoding, data=data, headers=headers, query=query,
858 expected_status=expected_status)
861 json_string, urlh = res
862 return self._parse_json(
863 json_string, video_id, transform_source=transform_source,
867 self, url_or_request, video_id, note='Downloading JSON metadata',
868 errnote='Unable to download JSON metadata', transform_source=None,
869 fatal=True, encoding=None, data=None, headers={}, query={},
870 expected_status=None):
872 Return the JSON object as a dict.
874 See _download_webpage docstring for arguments specification.
876 res = self._download_json_handle(
877 url_or_request, video_id, note=note, errnote=errnote,
878 transform_source=transform_source, fatal=fatal, encoding=encoding,
879 data=data, headers=headers, query=query,
880 expected_status=expected_status)
881 return res if res is False else res[0]
883 def _parse_json(self, json_string, video_id, transform_source=None, fatal=True):
885 json_string = transform_source(json_string)
887 return json.loads(json_string)
888 except ValueError as ve:
889 errmsg = '%s: Failed to parse JSON ' % video_id
891 raise ExtractorError(errmsg, cause=ve)
893 self.report_warning(errmsg + str(ve))
895 def report_warning(self, msg, video_id=None):
896 idstr = '' if video_id is None else '%s: ' % video_id
897 self._downloader.report_warning(
898 '[%s] %s%s' % (self.IE_NAME, idstr, msg))
900 def to_screen(self, msg):
901 """Print msg to screen, prefixing it with '[ie_name]'"""
902 self._downloader.to_screen('[%s] %s' % (self.IE_NAME, msg))
904 def report_extraction(self, id_or_name):
905 """Report information extraction."""
906 self.to_screen('%s: Extracting information' % id_or_name)
908 def report_download_webpage(self, video_id):
909 """Report webpage download."""
910 self.to_screen('%s: Downloading webpage' % video_id)
912 def report_age_confirmation(self):
913 """Report attempt to confirm age."""
914 self.to_screen('Confirming age')
916 def report_login(self):
917 """Report attempt to log in."""
918 self.to_screen('Logging in')
921 def raise_login_required(msg='This video is only available for registered users'):
922 raise ExtractorError(
923 '%s. Use --username and --password or --netrc to provide account credentials.' % msg,
927 def raise_geo_restricted(msg='This video is not available from your location due to geo restriction', countries=None):
928 raise GeoRestrictedError(msg, countries=countries)
930 # Methods for following #608
932 def url_result(url, ie=None, video_id=None, video_title=None):
933 """Returns a URL that points to a page that should be processed"""
934 # TODO: ie should be the class used for getting the info
935 video_info = {'_type': 'url',
938 if video_id is not None:
939 video_info['id'] = video_id
940 if video_title is not None:
941 video_info['title'] = video_title
944 def playlist_from_matches(self, matches, playlist_id=None, playlist_title=None, getter=None, ie=None):
946 self.url_result(self._proto_relative_url(getter(m) if getter else m), ie)
948 return self.playlist_result(
949 urls, playlist_id=playlist_id, playlist_title=playlist_title)
952 def playlist_result(entries, playlist_id=None, playlist_title=None, playlist_description=None):
953 """Returns a playlist"""
954 video_info = {'_type': 'playlist',
957 video_info['id'] = playlist_id
959 video_info['title'] = playlist_title
960 if playlist_description:
961 video_info['description'] = playlist_description
964 def _search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
966 Perform a regex search on the given string, using a single or a list of
967 patterns returning the first matching group.
968 In case of failure return a default value or raise a WARNING or a
969 RegexNotFoundError, depending on fatal, specifying the field name.
971 if isinstance(pattern, (str, compat_str, compiled_regex_type)):
972 mobj = re.search(pattern, string, flags)
975 mobj = re.search(p, string, flags)
979 if not self._downloader.params.get('no_color') and compat_os_name != 'nt' and sys.stderr.isatty():
980 _name = '\033[0;34m%s\033[0m' % name
986 # return the first matching group
987 return next(g for g in mobj.groups() if g is not None)
989 return mobj.group(group)
990 elif default is not NO_DEFAULT:
993 raise RegexNotFoundError('Unable to extract %s' % _name)
995 self._downloader.report_warning('unable to extract %s' % _name + bug_reports_message())
998 def _html_search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
1000 Like _search_regex, but strips HTML tags and unescapes entities.
1002 res = self._search_regex(pattern, string, name, default, fatal, flags, group)
1004 return clean_html(res).strip()
1008 def _get_netrc_login_info(self, netrc_machine=None):
1011 netrc_machine = netrc_machine or self._NETRC_MACHINE
1013 if self._downloader.params.get('usenetrc', False):
1015 info = netrc.netrc().authenticators(netrc_machine)
1016 if info is not None:
1020 raise netrc.NetrcParseError(
1021 'No authenticators for %s' % netrc_machine)
1022 except (IOError, netrc.NetrcParseError) as err:
1023 self._downloader.report_warning(
1024 'parsing .netrc: %s' % error_to_compat_str(err))
1026 return username, password
1028 def _get_login_info(self, username_option='username', password_option='password', netrc_machine=None):
1030 Get the login info as (username, password)
1031 First look for the manually specified credentials using username_option
1032 and password_option as keys in params dictionary. If no such credentials
1033 available look in the netrc file using the netrc_machine or _NETRC_MACHINE
1035 If there's no info available, return (None, None)
1037 if self._downloader is None:
1040 downloader_params = self._downloader.params
1042 # Attempt to use provided username and password or .netrc data
1043 if downloader_params.get(username_option) is not None:
1044 username = downloader_params[username_option]
1045 password = downloader_params[password_option]
1047 username, password = self._get_netrc_login_info(netrc_machine)
1049 return username, password
1051 def _get_tfa_info(self, note='two-factor verification code'):
1053 Get the two-factor authentication info
1054 TODO - asking the user will be required for sms/phone verify
1055 currently just uses the command line option
1056 If there's no info available, return None
1058 if self._downloader is None:
1060 downloader_params = self._downloader.params
1062 if downloader_params.get('twofactor') is not None:
1063 return downloader_params['twofactor']
1065 return compat_getpass('Type %s and press [Return]: ' % note)
1067 # Helper functions for extracting OpenGraph info
1069 def _og_regexes(prop):
1070 content_re = r'content=(?:"([^"]+?)"|\'([^\']+?)\'|\s*([^\s"\'=<>`]+?))'
1071 property_re = (r'(?:name|property)=(?:\'og[:-]%(prop)s\'|"og[:-]%(prop)s"|\s*og[:-]%(prop)s\b)'
1072 % {'prop': re.escape(prop)})
1073 template = r'<meta[^>]+?%s[^>]+?%s'
1075 template % (property_re, content_re),
1076 template % (content_re, property_re),
1080 def _meta_regex(prop):
1081 return r'''(?isx)<meta
1082 (?=[^>]+(?:itemprop|name|property|id|http-equiv)=(["\']?)%s\1)
1083 [^>]+?content=(["\'])(?P<content>.*?)\2''' % re.escape(prop)
1085 def _og_search_property(self, prop, html, name=None, **kargs):
1086 if not isinstance(prop, (list, tuple)):
1089 name = 'OpenGraph %s' % prop[0]
1092 og_regexes.extend(self._og_regexes(p))
1093 escaped = self._search_regex(og_regexes, html, name, flags=re.DOTALL, **kargs)
1096 return unescapeHTML(escaped)
1098 def _og_search_thumbnail(self, html, **kargs):
1099 return self._og_search_property('image', html, 'thumbnail URL', fatal=False, **kargs)
1101 def _og_search_description(self, html, **kargs):
1102 return self._og_search_property('description', html, fatal=False, **kargs)
1104 def _og_search_title(self, html, **kargs):
1105 return self._og_search_property('title', html, **kargs)
1107 def _og_search_video_url(self, html, name='video url', secure=True, **kargs):
1108 regexes = self._og_regexes('video') + self._og_regexes('video:url')
1110 regexes = self._og_regexes('video:secure_url') + regexes
1111 return self._html_search_regex(regexes, html, name, **kargs)
1113 def _og_search_url(self, html, **kargs):
1114 return self._og_search_property('url', html, **kargs)
1116 def _html_search_meta(self, name, html, display_name=None, fatal=False, **kwargs):
1117 if not isinstance(name, (list, tuple)):
1119 if display_name is None:
1120 display_name = name[0]
1121 return self._html_search_regex(
1122 [self._meta_regex(n) for n in name],
1123 html, display_name, fatal=fatal, group='content', **kwargs)
1125 def _dc_search_uploader(self, html):
1126 return self._html_search_meta('dc.creator', html, 'uploader')
1128 def _rta_search(self, html):
1129 # See http://www.rtalabel.org/index.php?content=howtofaq#single
1130 if re.search(r'(?ix)<meta\s+name="rating"\s+'
1131 r' content="RTA-5042-1996-1400-1577-RTA"',
1136 def _media_rating_search(self, html):
1137 # See http://www.tjg-designs.com/WP/metadata-code-examples-adding-metadata-to-your-web-pages/
1138 rating = self._html_search_meta('rating', html)
1150 return RATING_TABLE.get(rating.lower())
1152 def _family_friendly_search(self, html):
1153 # See http://schema.org/VideoObject
1154 family_friendly = self._html_search_meta(
1155 'isFamilyFriendly', html, default=None)
1157 if not family_friendly:
1166 return RATING_TABLE.get(family_friendly.lower())
1168 def _twitter_search_player(self, html):
1169 return self._html_search_meta('twitter:player', html,
1170 'twitter card player')
1172 def _search_json_ld(self, html, video_id, expected_type=None, **kwargs):
1173 json_ld = self._search_regex(
1174 JSON_LD_RE, html, 'JSON-LD', group='json_ld', **kwargs)
1175 default = kwargs.get('default', NO_DEFAULT)
1177 return default if default is not NO_DEFAULT else {}
1178 # JSON-LD may be malformed and thus `fatal` should be respected.
1179 # At the same time `default` may be passed that assumes `fatal=False`
1180 # for _search_regex. Let's simulate the same behavior here as well.
1181 fatal = kwargs.get('fatal', True) if default == NO_DEFAULT else False
1182 return self._json_ld(json_ld, video_id, fatal=fatal, expected_type=expected_type)
1184 def _json_ld(self, json_ld, video_id, fatal=True, expected_type=None):
1185 if isinstance(json_ld, compat_str):
1186 json_ld = self._parse_json(json_ld, video_id, fatal=fatal)
1190 if not isinstance(json_ld, (list, tuple, dict)):
1192 if isinstance(json_ld, dict):
1195 INTERACTION_TYPE_MAP = {
1196 'CommentAction': 'comment',
1197 'AgreeAction': 'like',
1198 'DisagreeAction': 'dislike',
1199 'LikeAction': 'like',
1200 'DislikeAction': 'dislike',
1201 'ListenAction': 'view',
1202 'WatchAction': 'view',
1203 'ViewAction': 'view',
1206 def extract_interaction_statistic(e):
1207 interaction_statistic = e.get('interactionStatistic')
1208 if not isinstance(interaction_statistic, list):
1210 for is_e in interaction_statistic:
1211 if not isinstance(is_e, dict):
1213 if is_e.get('@type') != 'InteractionCounter':
1215 interaction_type = is_e.get('interactionType')
1216 if not isinstance(interaction_type, compat_str):
1218 interaction_count = int_or_none(is_e.get('userInteractionCount'))
1219 if interaction_count is None:
1221 count_kind = INTERACTION_TYPE_MAP.get(interaction_type.split('/')[-1])
1224 count_key = '%s_count' % count_kind
1225 if info.get(count_key) is not None:
1227 info[count_key] = interaction_count
1229 def extract_video_object(e):
1230 assert e['@type'] == 'VideoObject'
1232 'url': url_or_none(e.get('contentUrl')),
1233 'title': unescapeHTML(e.get('name')),
1234 'description': unescapeHTML(e.get('description')),
1235 'thumbnail': url_or_none(e.get('thumbnailUrl') or e.get('thumbnailURL')),
1236 'duration': parse_duration(e.get('duration')),
1237 'timestamp': unified_timestamp(e.get('uploadDate')),
1238 'filesize': float_or_none(e.get('contentSize')),
1239 'tbr': int_or_none(e.get('bitrate')),
1240 'width': int_or_none(e.get('width')),
1241 'height': int_or_none(e.get('height')),
1242 'view_count': int_or_none(e.get('interactionCount')),
1244 extract_interaction_statistic(e)
1247 if isinstance(e.get('@context'), compat_str) and re.match(r'^https?://schema.org/?$', e.get('@context')):
1248 item_type = e.get('@type')
1249 if expected_type is not None and expected_type != item_type:
1251 if item_type in ('TVEpisode', 'Episode'):
1252 episode_name = unescapeHTML(e.get('name'))
1254 'episode': episode_name,
1255 'episode_number': int_or_none(e.get('episodeNumber')),
1256 'description': unescapeHTML(e.get('description')),
1258 if not info.get('title') and episode_name:
1259 info['title'] = episode_name
1260 part_of_season = e.get('partOfSeason')
1261 if isinstance(part_of_season, dict) and part_of_season.get('@type') in ('TVSeason', 'Season', 'CreativeWorkSeason'):
1263 'season': unescapeHTML(part_of_season.get('name')),
1264 'season_number': int_or_none(part_of_season.get('seasonNumber')),
1266 part_of_series = e.get('partOfSeries') or e.get('partOfTVSeries')
1267 if isinstance(part_of_series, dict) and part_of_series.get('@type') in ('TVSeries', 'Series', 'CreativeWorkSeries'):
1268 info['series'] = unescapeHTML(part_of_series.get('name'))
1269 elif item_type == 'Movie':
1271 'title': unescapeHTML(e.get('name')),
1272 'description': unescapeHTML(e.get('description')),
1273 'duration': parse_duration(e.get('duration')),
1274 'timestamp': unified_timestamp(e.get('dateCreated')),
1276 elif item_type in ('Article', 'NewsArticle'):
1278 'timestamp': parse_iso8601(e.get('datePublished')),
1279 'title': unescapeHTML(e.get('headline')),
1280 'description': unescapeHTML(e.get('articleBody')),
1282 elif item_type == 'VideoObject':
1283 extract_video_object(e)
1285 video = e.get('video')
1286 if isinstance(video, dict) and video.get('@type') == 'VideoObject':
1287 extract_video_object(video)
1289 return dict((k, v) for k, v in info.items() if v is not None)
1292 def _hidden_inputs(html):
1293 html = re.sub(r'<!--(?:(?!<!--).)*-->', '', html)
1295 for input in re.findall(r'(?i)(<input[^>]+>)', html):
1296 attrs = extract_attributes(input)
1299 if attrs.get('type') not in ('hidden', 'submit'):
1301 name = attrs.get('name') or attrs.get('id')
1302 value = attrs.get('value')
1303 if name and value is not None:
1304 hidden_inputs[name] = value
1305 return hidden_inputs
1307 def _form_hidden_inputs(self, form_id, html):
1308 form = self._search_regex(
1309 r'(?is)<form[^>]+?id=(["\'])%s\1[^>]*>(?P<form>.+?)</form>' % form_id,
1310 html, '%s form' % form_id, group='form')
1311 return self._hidden_inputs(form)
1313 def _sort_formats(self, formats, field_preference=None):
1315 raise ExtractorError('No video formats found')
1318 # Automatically determine tbr when missing based on abr and vbr (improves
1319 # formats sorting in some cases)
1320 if 'tbr' not in f and f.get('abr') is not None and f.get('vbr') is not None:
1321 f['tbr'] = f['abr'] + f['vbr']
1323 def _formats_key(f):
1324 # TODO remove the following workaround
1325 from ..utils import determine_ext
1326 if not f.get('ext') and 'url' in f:
1327 f['ext'] = determine_ext(f['url'])
1329 if isinstance(field_preference, (list, tuple)):
1332 if f.get(field) is not None
1333 else ('' if field == 'format_id' else -1)
1334 for field in field_preference)
1336 preference = f.get('preference')
1337 if preference is None:
1339 if f.get('ext') in ['f4f', 'f4m']: # Not yet supported
1342 protocol = f.get('protocol') or determine_protocol(f)
1343 proto_preference = 0 if protocol in ['http', 'https'] else (-0.5 if protocol == 'rtsp' else -0.1)
1345 if f.get('vcodec') == 'none': # audio only
1347 if self._downloader.params.get('prefer_free_formats'):
1348 ORDER = ['aac', 'mp3', 'm4a', 'webm', 'ogg', 'opus']
1350 ORDER = ['webm', 'opus', 'ogg', 'mp3', 'aac', 'm4a']
1353 audio_ext_preference = ORDER.index(f['ext'])
1355 audio_ext_preference = -1
1357 if f.get('acodec') == 'none': # video only
1359 if self._downloader.params.get('prefer_free_formats'):
1360 ORDER = ['flv', 'mp4', 'webm']
1362 ORDER = ['webm', 'flv', 'mp4']
1364 ext_preference = ORDER.index(f['ext'])
1367 audio_ext_preference = 0
1371 f.get('language_preference') if f.get('language_preference') is not None else -1,
1372 f.get('quality') if f.get('quality') is not None else -1,
1373 f.get('tbr') if f.get('tbr') is not None else -1,
1374 f.get('filesize') if f.get('filesize') is not None else -1,
1375 f.get('vbr') if f.get('vbr') is not None else -1,
1376 f.get('height') if f.get('height') is not None else -1,
1377 f.get('width') if f.get('width') is not None else -1,
1380 f.get('abr') if f.get('abr') is not None else -1,
1381 audio_ext_preference,
1382 f.get('fps') if f.get('fps') is not None else -1,
1383 f.get('filesize_approx') if f.get('filesize_approx') is not None else -1,
1384 f.get('source_preference') if f.get('source_preference') is not None else -1,
1385 f.get('format_id') if f.get('format_id') is not None else '',
1387 formats.sort(key=_formats_key)
1389 def _check_formats(self, formats, video_id):
1391 formats[:] = filter(
1392 lambda f: self._is_valid_url(
1394 item='%s video format' % f.get('format_id') if f.get('format_id') else 'video'),
1398 def _remove_duplicate_formats(formats):
1402 if f['url'] not in format_urls:
1403 format_urls.add(f['url'])
1404 unique_formats.append(f)
1405 formats[:] = unique_formats
1407 def _is_valid_url(self, url, video_id, item='video', headers={}):
1408 url = self._proto_relative_url(url, scheme='http:')
1409 # For now assume non HTTP(S) URLs always valid
1410 if not (url.startswith('http://') or url.startswith('https://')):
1413 self._request_webpage(url, video_id, 'Checking %s URL' % item, headers=headers)
1415 except ExtractorError as e:
1416 if isinstance(e.cause, compat_urllib_error.URLError):
1418 '%s: %s URL is invalid, skipping' % (video_id, item))
1422 def http_scheme(self):
1423 """ Either "http:" or "https:", depending on the user's preferences """
1426 if self._downloader.params.get('prefer_insecure', False)
1429 def _proto_relative_url(self, url, scheme=None):
1432 if url.startswith('//'):
1434 scheme = self.http_scheme()
1439 def _sleep(self, timeout, video_id, msg_template=None):
1440 if msg_template is None:
1441 msg_template = '%(video_id)s: Waiting for %(timeout)s seconds'
1442 msg = msg_template % {'video_id': video_id, 'timeout': timeout}
1446 def _extract_f4m_formats(self, manifest_url, video_id, preference=None, f4m_id=None,
1447 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1448 fatal=True, m3u8_id=None):
1449 manifest = self._download_xml(
1450 manifest_url, video_id, 'Downloading f4m manifest',
1451 'Unable to download f4m manifest',
1452 # Some manifests may be malformed, e.g. prosiebensat1 generated manifests
1453 # (see https://github.com/rg3/youtube-dl/issues/6215#issuecomment-121704244)
1454 transform_source=transform_source,
1457 if manifest is False:
1460 return self._parse_f4m_formats(
1461 manifest, manifest_url, video_id, preference=preference, f4m_id=f4m_id,
1462 transform_source=transform_source, fatal=fatal, m3u8_id=m3u8_id)
1464 def _parse_f4m_formats(self, manifest, manifest_url, video_id, preference=None, f4m_id=None,
1465 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1466 fatal=True, m3u8_id=None):
1467 # currently youtube-dl cannot decode the playerVerificationChallenge as Akamai uses Adobe Alchemy
1468 akamai_pv = manifest.find('{http://ns.adobe.com/f4m/1.0}pv-2.0')
1469 if akamai_pv is not None and ';' in akamai_pv.text:
1470 playerVerificationChallenge = akamai_pv.text.split(';')[0]
1471 if playerVerificationChallenge.strip() != '':
1475 manifest_version = '1.0'
1476 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/1.0}media')
1478 manifest_version = '2.0'
1479 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/2.0}media')
1480 # Remove unsupported DRM protected media from final formats
1481 # rendition (see https://github.com/rg3/youtube-dl/issues/8573).
1482 media_nodes = remove_encrypted_media(media_nodes)
1486 manifest_base_url = get_base_url(manifest)
1488 bootstrap_info = xpath_element(
1489 manifest, ['{http://ns.adobe.com/f4m/1.0}bootstrapInfo', '{http://ns.adobe.com/f4m/2.0}bootstrapInfo'],
1490 'bootstrap info', default=None)
1493 mime_type = xpath_text(
1494 manifest, ['{http://ns.adobe.com/f4m/1.0}mimeType', '{http://ns.adobe.com/f4m/2.0}mimeType'],
1495 'base URL', default=None)
1496 if mime_type and mime_type.startswith('audio/'):
1499 for i, media_el in enumerate(media_nodes):
1500 tbr = int_or_none(media_el.attrib.get('bitrate'))
1501 width = int_or_none(media_el.attrib.get('width'))
1502 height = int_or_none(media_el.attrib.get('height'))
1503 format_id = '-'.join(filter(None, [f4m_id, compat_str(i if tbr is None else tbr)]))
1504 # If <bootstrapInfo> is present, the specified f4m is a
1505 # stream-level manifest, and only set-level manifests may refer to
1506 # external resources. See section 11.4 and section 4 of F4M spec
1507 if bootstrap_info is None:
1509 # @href is introduced in 2.0, see section 11.6 of F4M spec
1510 if manifest_version == '2.0':
1511 media_url = media_el.attrib.get('href')
1512 if media_url is None:
1513 media_url = media_el.attrib.get('url')
1517 media_url if media_url.startswith('http://') or media_url.startswith('https://')
1518 else ((manifest_base_url or '/'.join(manifest_url.split('/')[:-1])) + '/' + media_url))
1519 # If media_url is itself a f4m manifest do the recursive extraction
1520 # since bitrates in parent manifest (this one) and media_url manifest
1521 # may differ leading to inability to resolve the format by requested
1522 # bitrate in f4m downloader
1523 ext = determine_ext(manifest_url)
1525 f4m_formats = self._extract_f4m_formats(
1526 manifest_url, video_id, preference=preference, f4m_id=f4m_id,
1527 transform_source=transform_source, fatal=fatal)
1528 # Sometimes stream-level manifest contains single media entry that
1529 # does not contain any quality metadata (e.g. http://matchtv.ru/#live-player).
1530 # At the same time parent's media entry in set-level manifest may
1531 # contain it. We will copy it from parent in such cases.
1532 if len(f4m_formats) == 1:
1535 'tbr': f.get('tbr') or tbr,
1536 'width': f.get('width') or width,
1537 'height': f.get('height') or height,
1538 'format_id': f.get('format_id') if not tbr else format_id,
1541 formats.extend(f4m_formats)
1544 formats.extend(self._extract_m3u8_formats(
1545 manifest_url, video_id, 'mp4', preference=preference,
1546 m3u8_id=m3u8_id, fatal=fatal))
1549 'format_id': format_id,
1550 'url': manifest_url,
1551 'manifest_url': manifest_url,
1552 'ext': 'flv' if bootstrap_info is not None else None,
1558 'preference': preference,
1562 def _m3u8_meta_format(self, m3u8_url, ext=None, preference=None, m3u8_id=None):
1564 'format_id': '-'.join(filter(None, [m3u8_id, 'meta'])),
1568 'preference': preference - 100 if preference else -100,
1569 'resolution': 'multiple',
1570 'format_note': 'Quality selection URL',
1573 def _extract_m3u8_formats(self, m3u8_url, video_id, ext=None,
1574 entry_protocol='m3u8', preference=None,
1575 m3u8_id=None, note=None, errnote=None,
1576 fatal=True, live=False):
1577 res = self._download_webpage_handle(
1579 note=note or 'Downloading m3u8 information',
1580 errnote=errnote or 'Failed to download m3u8 information',
1586 m3u8_doc, urlh = res
1587 m3u8_url = urlh.geturl()
1589 return self._parse_m3u8_formats(
1590 m3u8_doc, m3u8_url, ext=ext, entry_protocol=entry_protocol,
1591 preference=preference, m3u8_id=m3u8_id, live=live)
1593 def _parse_m3u8_formats(self, m3u8_doc, m3u8_url, ext=None,
1594 entry_protocol='m3u8', preference=None,
1595 m3u8_id=None, live=False):
1596 if '#EXT-X-FAXS-CM:' in m3u8_doc: # Adobe Flash Access
1599 if re.search(r'#EXT-X-SESSION-KEY:.*?URI="skd://', m3u8_doc): # Apple FairPlay
1604 format_url = lambda u: (
1606 if re.match(r'^https?://', u)
1607 else compat_urlparse.urljoin(m3u8_url, u))
1610 # 1. https://tools.ietf.org/html/draft-pantos-http-live-streaming-21
1611 # 2. https://github.com/rg3/youtube-dl/issues/12211
1612 # 3. https://github.com/rg3/youtube-dl/issues/18923
1614 # We should try extracting formats only from master playlists [1, 4.3.4],
1615 # i.e. playlists that describe available qualities. On the other hand
1616 # media playlists [1, 4.3.3] should be returned as is since they contain
1617 # just the media without qualities renditions.
1618 # Fortunately, master playlist can be easily distinguished from media
1619 # playlist based on particular tags availability. As of [1, 4.3.3, 4.3.4]
1620 # master playlist tags MUST NOT appear in a media playist and vice versa.
1621 # As of [1, 4.3.3.1] #EXT-X-TARGETDURATION tag is REQUIRED for every
1622 # media playlist and MUST NOT appear in master playlist thus we can
1623 # clearly detect media playlist with this criterion.
1625 if '#EXT-X-TARGETDURATION' in m3u8_doc: # media playlist, return as is
1628 'format_id': m3u8_id,
1630 'protocol': entry_protocol,
1631 'preference': preference,
1635 last_stream_inf = {}
1637 def extract_media(x_media_line):
1638 media = parse_m3u8_attributes(x_media_line)
1639 # As per [1, 4.3.4.1] TYPE, GROUP-ID and NAME are REQUIRED
1640 media_type, group_id, name = media.get('TYPE'), media.get('GROUP-ID'), media.get('NAME')
1641 if not (media_type and group_id and name):
1643 groups.setdefault(group_id, []).append(media)
1644 if media_type not in ('VIDEO', 'AUDIO'):
1646 media_url = media.get('URI')
1649 for v in (m3u8_id, group_id, name):
1653 'format_id': '-'.join(format_id),
1654 'url': format_url(media_url),
1655 'manifest_url': m3u8_url,
1656 'language': media.get('LANGUAGE'),
1658 'protocol': entry_protocol,
1659 'preference': preference,
1661 if media_type == 'AUDIO':
1662 f['vcodec'] = 'none'
1665 def build_stream_name():
1666 # Despite specification does not mention NAME attribute for
1667 # EXT-X-STREAM-INF tag it still sometimes may be present (see [1]
1668 # or vidio test in TestInfoExtractor.test_parse_m3u8_formats)
1669 # 1. http://www.vidio.com/watch/165683-dj_ambred-booyah-live-2015
1670 stream_name = last_stream_inf.get('NAME')
1673 # If there is no NAME in EXT-X-STREAM-INF it will be obtained
1674 # from corresponding rendition group
1675 stream_group_id = last_stream_inf.get('VIDEO')
1676 if not stream_group_id:
1678 stream_group = groups.get(stream_group_id)
1679 if not stream_group:
1680 return stream_group_id
1681 rendition = stream_group[0]
1682 return rendition.get('NAME') or stream_group_id
1684 # parse EXT-X-MEDIA tags before EXT-X-STREAM-INF in order to have the
1685 # chance to detect video only formats when EXT-X-STREAM-INF tags
1686 # precede EXT-X-MEDIA tags in HLS manifest such as [3].
1687 for line in m3u8_doc.splitlines():
1688 if line.startswith('#EXT-X-MEDIA:'):
1691 for line in m3u8_doc.splitlines():
1692 if line.startswith('#EXT-X-STREAM-INF:'):
1693 last_stream_inf = parse_m3u8_attributes(line)
1694 elif line.startswith('#') or not line.strip():
1697 tbr = float_or_none(
1698 last_stream_inf.get('AVERAGE-BANDWIDTH') or
1699 last_stream_inf.get('BANDWIDTH'), scale=1000)
1702 format_id.append(m3u8_id)
1703 stream_name = build_stream_name()
1704 # Bandwidth of live streams may differ over time thus making
1705 # format_id unpredictable. So it's better to keep provided
1708 format_id.append(stream_name if stream_name else '%d' % (tbr if tbr else len(formats)))
1709 manifest_url = format_url(line.strip())
1711 'format_id': '-'.join(format_id),
1712 'url': manifest_url,
1713 'manifest_url': m3u8_url,
1716 'fps': float_or_none(last_stream_inf.get('FRAME-RATE')),
1717 'protocol': entry_protocol,
1718 'preference': preference,
1720 resolution = last_stream_inf.get('RESOLUTION')
1722 mobj = re.search(r'(?P<width>\d+)[xX](?P<height>\d+)', resolution)
1724 f['width'] = int(mobj.group('width'))
1725 f['height'] = int(mobj.group('height'))
1726 # Unified Streaming Platform
1728 r'audio.*?(?:%3D|=)(\d+)(?:-video.*?(?:%3D|=)(\d+))?', f['url'])
1730 abr, vbr = mobj.groups()
1731 abr, vbr = float_or_none(abr, 1000), float_or_none(vbr, 1000)
1736 codecs = parse_codecs(last_stream_inf.get('CODECS'))
1738 audio_group_id = last_stream_inf.get('AUDIO')
1739 # As per [1, 4.3.4.1.1] any EXT-X-STREAM-INF tag which
1740 # references a rendition group MUST have a CODECS attribute.
1741 # However, this is not always respected, for example, [2]
1742 # contains EXT-X-STREAM-INF tag which references AUDIO
1743 # rendition group but does not have CODECS and despite
1744 # referencing an audio group it represents a complete
1745 # (with audio and video) format. So, for such cases we will
1746 # ignore references to rendition groups and treat them
1747 # as complete formats.
1748 if audio_group_id and codecs and f.get('vcodec') != 'none':
1749 audio_group = groups.get(audio_group_id)
1750 if audio_group and audio_group[0].get('URI'):
1751 # TODO: update acodec for audio only formats with
1753 f['acodec'] = 'none'
1755 last_stream_inf = {}
1759 def _xpath_ns(path, namespace=None):
1763 for c in path.split('/'):
1764 if not c or c == '.':
1767 out.append('{%s}%s' % (namespace, c))
1768 return '/'.join(out)
1770 def _extract_smil_formats(self, smil_url, video_id, fatal=True, f4m_params=None, transform_source=None):
1771 smil = self._download_smil(smil_url, video_id, fatal=fatal, transform_source=transform_source)
1777 namespace = self._parse_smil_namespace(smil)
1779 return self._parse_smil_formats(
1780 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
1782 def _extract_smil_info(self, smil_url, video_id, fatal=True, f4m_params=None):
1783 smil = self._download_smil(smil_url, video_id, fatal=fatal)
1786 return self._parse_smil(smil, smil_url, video_id, f4m_params=f4m_params)
1788 def _download_smil(self, smil_url, video_id, fatal=True, transform_source=None):
1789 return self._download_xml(
1790 smil_url, video_id, 'Downloading SMIL file',
1791 'Unable to download SMIL file', fatal=fatal, transform_source=transform_source)
1793 def _parse_smil(self, smil, smil_url, video_id, f4m_params=None):
1794 namespace = self._parse_smil_namespace(smil)
1796 formats = self._parse_smil_formats(
1797 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
1798 subtitles = self._parse_smil_subtitles(smil, namespace=namespace)
1800 video_id = os.path.splitext(url_basename(smil_url))[0]
1804 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
1805 name = meta.attrib.get('name')
1806 content = meta.attrib.get('content')
1807 if not name or not content:
1809 if not title and name == 'title':
1811 elif not description and name in ('description', 'abstract'):
1812 description = content
1813 elif not upload_date and name == 'date':
1814 upload_date = unified_strdate(content)
1817 'id': image.get('type'),
1818 'url': image.get('src'),
1819 'width': int_or_none(image.get('width')),
1820 'height': int_or_none(image.get('height')),
1821 } for image in smil.findall(self._xpath_ns('.//image', namespace)) if image.get('src')]
1825 'title': title or video_id,
1826 'description': description,
1827 'upload_date': upload_date,
1828 'thumbnails': thumbnails,
1830 'subtitles': subtitles,
1833 def _parse_smil_namespace(self, smil):
1834 return self._search_regex(
1835 r'(?i)^{([^}]+)?}smil$', smil.tag, 'namespace', default=None)
1837 def _parse_smil_formats(self, smil, smil_url, video_id, namespace=None, f4m_params=None, transform_rtmp_url=None):
1839 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
1840 b = meta.get('base') or meta.get('httpBase')
1851 media = smil.findall(self._xpath_ns('.//video', namespace)) + smil.findall(self._xpath_ns('.//audio', namespace))
1852 for medium in media:
1853 src = medium.get('src')
1854 if not src or src in srcs:
1858 bitrate = float_or_none(medium.get('system-bitrate') or medium.get('systemBitrate'), 1000)
1859 filesize = int_or_none(medium.get('size') or medium.get('fileSize'))
1860 width = int_or_none(medium.get('width'))
1861 height = int_or_none(medium.get('height'))
1862 proto = medium.get('proto')
1863 ext = medium.get('ext')
1864 src_ext = determine_ext(src)
1865 streamer = medium.get('streamer') or base
1867 if proto == 'rtmp' or streamer.startswith('rtmp'):
1873 'format_id': 'rtmp-%d' % (rtmp_count if bitrate is None else bitrate),
1875 'filesize': filesize,
1879 if transform_rtmp_url:
1880 streamer, src = transform_rtmp_url(streamer, src)
1881 formats[-1].update({
1887 src_url = src if src.startswith('http') else compat_urlparse.urljoin(base, src)
1888 src_url = src_url.strip()
1890 if proto == 'm3u8' or src_ext == 'm3u8':
1891 m3u8_formats = self._extract_m3u8_formats(
1892 src_url, video_id, ext or 'mp4', m3u8_id='hls', fatal=False)
1893 if len(m3u8_formats) == 1:
1895 m3u8_formats[0].update({
1896 'format_id': 'hls-%d' % (m3u8_count if bitrate is None else bitrate),
1901 formats.extend(m3u8_formats)
1902 elif src_ext == 'f4m':
1907 'plugin': 'flowplayer-3.2.0.1',
1909 f4m_url += '&' if '?' in f4m_url else '?'
1910 f4m_url += compat_urllib_parse_urlencode(f4m_params)
1911 formats.extend(self._extract_f4m_formats(f4m_url, video_id, f4m_id='hds', fatal=False))
1912 elif src_ext == 'mpd':
1913 formats.extend(self._extract_mpd_formats(
1914 src_url, video_id, mpd_id='dash', fatal=False))
1915 elif re.search(r'\.ism/[Mm]anifest', src_url):
1916 formats.extend(self._extract_ism_formats(
1917 src_url, video_id, ism_id='mss', fatal=False))
1918 elif src_url.startswith('http') and self._is_valid_url(src, video_id):
1922 'ext': ext or src_ext or 'flv',
1923 'format_id': 'http-%d' % (bitrate or http_count),
1925 'filesize': filesize,
1932 def _parse_smil_subtitles(self, smil, namespace=None, subtitles_lang='en'):
1935 for num, textstream in enumerate(smil.findall(self._xpath_ns('.//textstream', namespace))):
1936 src = textstream.get('src')
1937 if not src or src in urls:
1940 ext = textstream.get('ext') or mimetype2ext(textstream.get('type')) or determine_ext(src)
1941 lang = textstream.get('systemLanguage') or textstream.get('systemLanguageName') or textstream.get('lang') or subtitles_lang
1942 subtitles.setdefault(lang, []).append({
1948 def _extract_xspf_playlist(self, xspf_url, playlist_id, fatal=True):
1949 xspf = self._download_xml(
1950 xspf_url, playlist_id, 'Downloading xpsf playlist',
1951 'Unable to download xspf manifest', fatal=fatal)
1954 return self._parse_xspf(
1955 xspf, playlist_id, xspf_url=xspf_url,
1956 xspf_base_url=base_url(xspf_url))
1958 def _parse_xspf(self, xspf_doc, playlist_id, xspf_url=None, xspf_base_url=None):
1960 'xspf': 'http://xspf.org/ns/0/',
1961 's1': 'http://static.streamone.nl/player/ns/0',
1965 for track in xspf_doc.findall(xpath_with_ns('./xspf:trackList/xspf:track', NS_MAP)):
1967 track, xpath_with_ns('./xspf:title', NS_MAP), 'title', default=playlist_id)
1968 description = xpath_text(
1969 track, xpath_with_ns('./xspf:annotation', NS_MAP), 'description')
1970 thumbnail = xpath_text(
1971 track, xpath_with_ns('./xspf:image', NS_MAP), 'thumbnail')
1972 duration = float_or_none(
1973 xpath_text(track, xpath_with_ns('./xspf:duration', NS_MAP), 'duration'), 1000)
1976 for location in track.findall(xpath_with_ns('./xspf:location', NS_MAP)):
1977 format_url = urljoin(xspf_base_url, location.text)
1982 'manifest_url': xspf_url,
1983 'format_id': location.get(xpath_with_ns('s1:label', NS_MAP)),
1984 'width': int_or_none(location.get(xpath_with_ns('s1:width', NS_MAP))),
1985 'height': int_or_none(location.get(xpath_with_ns('s1:height', NS_MAP))),
1987 self._sort_formats(formats)
1992 'description': description,
1993 'thumbnail': thumbnail,
1994 'duration': duration,
1999 def _extract_mpd_formats(self, mpd_url, video_id, mpd_id=None, note=None, errnote=None, fatal=True, formats_dict={}):
2000 res = self._download_xml_handle(
2002 note=note or 'Downloading MPD manifest',
2003 errnote=errnote or 'Failed to download MPD manifest',
2008 mpd_base_url = base_url(urlh.geturl())
2010 return self._parse_mpd_formats(
2011 mpd_doc, mpd_id=mpd_id, mpd_base_url=mpd_base_url,
2012 formats_dict=formats_dict, mpd_url=mpd_url)
2014 def _parse_mpd_formats(self, mpd_doc, mpd_id=None, mpd_base_url='', formats_dict={}, mpd_url=None):
2016 Parse formats from MPD manifest.
2018 1. MPEG-DASH Standard, ISO/IEC 23009-1:2014(E),
2019 http://standards.iso.org/ittf/PubliclyAvailableStandards/c065274_ISO_IEC_23009-1_2014.zip
2020 2. https://en.wikipedia.org/wiki/Dynamic_Adaptive_Streaming_over_HTTP
2022 if mpd_doc.get('type') == 'dynamic':
2025 namespace = self._search_regex(r'(?i)^{([^}]+)?}MPD$', mpd_doc.tag, 'namespace', default=None)
2028 return self._xpath_ns(path, namespace)
2030 def is_drm_protected(element):
2031 return element.find(_add_ns('ContentProtection')) is not None
2033 def extract_multisegment_info(element, ms_parent_info):
2034 ms_info = ms_parent_info.copy()
2036 # As per [1, 5.3.9.2.2] SegmentList and SegmentTemplate share some
2037 # common attributes and elements. We will only extract relevant
2039 def extract_common(source):
2040 segment_timeline = source.find(_add_ns('SegmentTimeline'))
2041 if segment_timeline is not None:
2042 s_e = segment_timeline.findall(_add_ns('S'))
2044 ms_info['total_number'] = 0
2047 r = int(s.get('r', 0))
2048 ms_info['total_number'] += 1 + r
2049 ms_info['s'].append({
2050 't': int(s.get('t', 0)),
2051 # @d is mandatory (see [1, 5.3.9.6.2, Table 17, page 60])
2052 'd': int(s.attrib['d']),
2055 start_number = source.get('startNumber')
2057 ms_info['start_number'] = int(start_number)
2058 timescale = source.get('timescale')
2060 ms_info['timescale'] = int(timescale)
2061 segment_duration = source.get('duration')
2062 if segment_duration:
2063 ms_info['segment_duration'] = float(segment_duration)
2065 def extract_Initialization(source):
2066 initialization = source.find(_add_ns('Initialization'))
2067 if initialization is not None:
2068 ms_info['initialization_url'] = initialization.attrib['sourceURL']
2070 segment_list = element.find(_add_ns('SegmentList'))
2071 if segment_list is not None:
2072 extract_common(segment_list)
2073 extract_Initialization(segment_list)
2074 segment_urls_e = segment_list.findall(_add_ns('SegmentURL'))
2076 ms_info['segment_urls'] = [segment.attrib['media'] for segment in segment_urls_e]
2078 segment_template = element.find(_add_ns('SegmentTemplate'))
2079 if segment_template is not None:
2080 extract_common(segment_template)
2081 media = segment_template.get('media')
2083 ms_info['media'] = media
2084 initialization = segment_template.get('initialization')
2086 ms_info['initialization'] = initialization
2088 extract_Initialization(segment_template)
2091 mpd_duration = parse_duration(mpd_doc.get('mediaPresentationDuration'))
2093 for period in mpd_doc.findall(_add_ns('Period')):
2094 period_duration = parse_duration(period.get('duration')) or mpd_duration
2095 period_ms_info = extract_multisegment_info(period, {
2099 for adaptation_set in period.findall(_add_ns('AdaptationSet')):
2100 if is_drm_protected(adaptation_set):
2102 adaption_set_ms_info = extract_multisegment_info(adaptation_set, period_ms_info)
2103 for representation in adaptation_set.findall(_add_ns('Representation')):
2104 if is_drm_protected(representation):
2106 representation_attrib = adaptation_set.attrib.copy()
2107 representation_attrib.update(representation.attrib)
2108 # According to [1, 5.3.7.2, Table 9, page 41], @mimeType is mandatory
2109 mime_type = representation_attrib['mimeType']
2110 content_type = mime_type.split('/')[0]
2111 if content_type == 'text':
2112 # TODO implement WebVTT downloading
2114 elif content_type in ('video', 'audio'):
2116 for element in (representation, adaptation_set, period, mpd_doc):
2117 base_url_e = element.find(_add_ns('BaseURL'))
2118 if base_url_e is not None:
2119 base_url = base_url_e.text + base_url
2120 if re.match(r'^https?://', base_url):
2122 if mpd_base_url and not re.match(r'^https?://', base_url):
2123 if not mpd_base_url.endswith('/') and not base_url.startswith('/'):
2125 base_url = mpd_base_url + base_url
2126 representation_id = representation_attrib.get('id')
2127 lang = representation_attrib.get('lang')
2128 url_el = representation.find(_add_ns('BaseURL'))
2129 filesize = int_or_none(url_el.attrib.get('{http://youtube.com/yt/2012/10/10}contentLength') if url_el is not None else None)
2130 bandwidth = int_or_none(representation_attrib.get('bandwidth'))
2132 'format_id': '%s-%s' % (mpd_id, representation_id) if mpd_id else representation_id,
2134 'manifest_url': mpd_url,
2135 'ext': mimetype2ext(mime_type),
2136 'width': int_or_none(representation_attrib.get('width')),
2137 'height': int_or_none(representation_attrib.get('height')),
2138 'tbr': float_or_none(bandwidth, 1000),
2139 'asr': int_or_none(representation_attrib.get('audioSamplingRate')),
2140 'fps': int_or_none(representation_attrib.get('frameRate')),
2141 'language': lang if lang not in ('mul', 'und', 'zxx', 'mis') else None,
2142 'format_note': 'DASH %s' % content_type,
2143 'filesize': filesize,
2144 'container': mimetype2ext(mime_type) + '_dash',
2146 f.update(parse_codecs(representation_attrib.get('codecs')))
2147 representation_ms_info = extract_multisegment_info(representation, adaption_set_ms_info)
2149 def prepare_template(template_name, identifiers):
2150 tmpl = representation_ms_info[template_name]
2151 # First of, % characters outside $...$ templates
2152 # must be escaped by doubling for proper processing
2153 # by % operator string formatting used further (see
2154 # https://github.com/rg3/youtube-dl/issues/16867).
2160 in_template = not in_template
2161 elif c == '%' and not in_template:
2163 # Next, $...$ templates are translated to their
2164 # %(...) counterparts to be used with % operator
2165 t = t.replace('$RepresentationID$', representation_id)
2166 t = re.sub(r'\$(%s)\$' % '|'.join(identifiers), r'%(\1)d', t)
2167 t = re.sub(r'\$(%s)%%([^$]+)\$' % '|'.join(identifiers), r'%(\1)\2', t)
2168 t.replace('$$', '$')
2171 # @initialization is a regular template like @media one
2172 # so it should be handled just the same way (see
2173 # https://github.com/rg3/youtube-dl/issues/11605)
2174 if 'initialization' in representation_ms_info:
2175 initialization_template = prepare_template(
2177 # As per [1, 5.3.9.4.2, Table 15, page 54] $Number$ and
2178 # $Time$ shall not be included for @initialization thus
2179 # only $Bandwidth$ remains
2181 representation_ms_info['initialization_url'] = initialization_template % {
2182 'Bandwidth': bandwidth,
2185 def location_key(location):
2186 return 'url' if re.match(r'^https?://', location) else 'path'
2188 if 'segment_urls' not in representation_ms_info and 'media' in representation_ms_info:
2190 media_template = prepare_template('media', ('Number', 'Bandwidth', 'Time'))
2191 media_location_key = location_key(media_template)
2193 # As per [1, 5.3.9.4.4, Table 16, page 55] $Number$ and $Time$
2194 # can't be used at the same time
2195 if '%(Number' in media_template and 's' not in representation_ms_info:
2196 segment_duration = None
2197 if 'total_number' not in representation_ms_info and 'segment_duration' in representation_ms_info:
2198 segment_duration = float_or_none(representation_ms_info['segment_duration'], representation_ms_info['timescale'])
2199 representation_ms_info['total_number'] = int(math.ceil(float(period_duration) / segment_duration))
2200 representation_ms_info['fragments'] = [{
2201 media_location_key: media_template % {
2202 'Number': segment_number,
2203 'Bandwidth': bandwidth,
2205 'duration': segment_duration,
2206 } for segment_number in range(
2207 representation_ms_info['start_number'],
2208 representation_ms_info['total_number'] + representation_ms_info['start_number'])]
2210 # $Number*$ or $Time$ in media template with S list available
2211 # Example $Number*$: http://www.svtplay.se/klipp/9023742/stopptid-om-bjorn-borg
2212 # Example $Time$: https://play.arkena.com/embed/avp/v2/player/media/b41dda37-d8e7-4d3f-b1b5-9a9db578bdfe/1/129411
2213 representation_ms_info['fragments'] = []
2216 segment_number = representation_ms_info['start_number']
2218 def add_segment_url():
2219 segment_url = media_template % {
2220 'Time': segment_time,
2221 'Bandwidth': bandwidth,
2222 'Number': segment_number,
2224 representation_ms_info['fragments'].append({
2225 media_location_key: segment_url,
2226 'duration': float_or_none(segment_d, representation_ms_info['timescale']),
2229 for num, s in enumerate(representation_ms_info['s']):
2230 segment_time = s.get('t') or segment_time
2234 for r in range(s.get('r', 0)):
2235 segment_time += segment_d
2238 segment_time += segment_d
2239 elif 'segment_urls' in representation_ms_info and 's' in representation_ms_info:
2241 # Example: https://www.youtube.com/watch?v=iXZV5uAYMJI
2242 # or any YouTube dashsegments video
2245 timescale = representation_ms_info['timescale']
2246 for s in representation_ms_info['s']:
2247 duration = float_or_none(s['d'], timescale)
2248 for r in range(s.get('r', 0) + 1):
2249 segment_uri = representation_ms_info['segment_urls'][segment_index]
2251 location_key(segment_uri): segment_uri,
2252 'duration': duration,
2255 representation_ms_info['fragments'] = fragments
2256 elif 'segment_urls' in representation_ms_info:
2257 # Segment URLs with no SegmentTimeline
2258 # Example: https://www.seznam.cz/zpravy/clanek/cesko-zasahne-vitr-o-sile-vichrice-muze-byt-i-zivotu-nebezpecny-39091
2259 # https://github.com/rg3/youtube-dl/pull/14844
2261 segment_duration = float_or_none(
2262 representation_ms_info['segment_duration'],
2263 representation_ms_info['timescale']) if 'segment_duration' in representation_ms_info else None
2264 for segment_url in representation_ms_info['segment_urls']:
2266 location_key(segment_url): segment_url,
2268 if segment_duration:
2269 fragment['duration'] = segment_duration
2270 fragments.append(fragment)
2271 representation_ms_info['fragments'] = fragments
2272 # NB: MPD manifest may contain direct URLs to unfragmented media.
2273 # No fragments key is present in this case.
2274 if 'fragments' in representation_ms_info:
2276 'fragment_base_url': base_url,
2278 'protocol': 'http_dash_segments',
2280 if 'initialization_url' in representation_ms_info:
2281 initialization_url = representation_ms_info['initialization_url']
2282 if not f.get('url'):
2283 f['url'] = initialization_url
2284 f['fragments'].append({location_key(initialization_url): initialization_url})
2285 f['fragments'].extend(representation_ms_info['fragments'])
2286 # According to [1, 5.3.5.2, Table 7, page 35] @id of Representation
2287 # is not necessarily unique within a Period thus formats with
2288 # the same `format_id` are quite possible. There are numerous examples
2289 # of such manifests (see https://github.com/rg3/youtube-dl/issues/15111,
2290 # https://github.com/rg3/youtube-dl/issues/13919)
2291 full_info = formats_dict.get(representation_id, {}).copy()
2293 formats.append(full_info)
2295 self.report_warning('Unknown MIME type %s in DASH manifest' % mime_type)
2298 def _extract_ism_formats(self, ism_url, video_id, ism_id=None, note=None, errnote=None, fatal=True):
2299 res = self._download_xml_handle(
2301 note=note or 'Downloading ISM manifest',
2302 errnote=errnote or 'Failed to download ISM manifest',
2308 return self._parse_ism_formats(ism_doc, urlh.geturl(), ism_id)
2310 def _parse_ism_formats(self, ism_doc, ism_url, ism_id=None):
2312 Parse formats from ISM manifest.
2314 1. [MS-SSTR]: Smooth Streaming Protocol,
2315 https://msdn.microsoft.com/en-us/library/ff469518.aspx
2317 if ism_doc.get('IsLive') == 'TRUE' or ism_doc.find('Protection') is not None:
2320 duration = int(ism_doc.attrib['Duration'])
2321 timescale = int_or_none(ism_doc.get('TimeScale')) or 10000000
2324 for stream in ism_doc.findall('StreamIndex'):
2325 stream_type = stream.get('Type')
2326 if stream_type not in ('video', 'audio'):
2328 url_pattern = stream.attrib['Url']
2329 stream_timescale = int_or_none(stream.get('TimeScale')) or timescale
2330 stream_name = stream.get('Name')
2331 for track in stream.findall('QualityLevel'):
2332 fourcc = track.get('FourCC', 'AACL' if track.get('AudioTag') == '255' else None)
2333 # TODO: add support for WVC1 and WMAP
2334 if fourcc not in ('H264', 'AVC1', 'AACL'):
2335 self.report_warning('%s is not a supported codec' % fourcc)
2337 tbr = int(track.attrib['Bitrate']) // 1000
2338 # [1] does not mention Width and Height attributes. However,
2339 # they're often present while MaxWidth and MaxHeight are
2340 # missing, so should be used as fallbacks
2341 width = int_or_none(track.get('MaxWidth') or track.get('Width'))
2342 height = int_or_none(track.get('MaxHeight') or track.get('Height'))
2343 sampling_rate = int_or_none(track.get('SamplingRate'))
2345 track_url_pattern = re.sub(r'{[Bb]itrate}', track.attrib['Bitrate'], url_pattern)
2346 track_url_pattern = compat_urlparse.urljoin(ism_url, track_url_pattern)
2352 stream_fragments = stream.findall('c')
2353 for stream_fragment_index, stream_fragment in enumerate(stream_fragments):
2354 fragment_ctx['time'] = int_or_none(stream_fragment.get('t')) or fragment_ctx['time']
2355 fragment_repeat = int_or_none(stream_fragment.get('r')) or 1
2356 fragment_ctx['duration'] = int_or_none(stream_fragment.get('d'))
2357 if not fragment_ctx['duration']:
2359 next_fragment_time = int(stream_fragment[stream_fragment_index + 1].attrib['t'])
2361 next_fragment_time = duration
2362 fragment_ctx['duration'] = (next_fragment_time - fragment_ctx['time']) / fragment_repeat
2363 for _ in range(fragment_repeat):
2365 'url': re.sub(r'{start[ _]time}', compat_str(fragment_ctx['time']), track_url_pattern),
2366 'duration': fragment_ctx['duration'] / stream_timescale,
2368 fragment_ctx['time'] += fragment_ctx['duration']
2372 format_id.append(ism_id)
2374 format_id.append(stream_name)
2375 format_id.append(compat_str(tbr))
2378 'format_id': '-'.join(format_id),
2380 'manifest_url': ism_url,
2381 'ext': 'ismv' if stream_type == 'video' else 'isma',
2385 'asr': sampling_rate,
2386 'vcodec': 'none' if stream_type == 'audio' else fourcc,
2387 'acodec': 'none' if stream_type == 'video' else fourcc,
2389 'fragments': fragments,
2390 '_download_params': {
2391 'duration': duration,
2392 'timescale': stream_timescale,
2393 'width': width or 0,
2394 'height': height or 0,
2396 'codec_private_data': track.get('CodecPrivateData'),
2397 'sampling_rate': sampling_rate,
2398 'channels': int_or_none(track.get('Channels', 2)),
2399 'bits_per_sample': int_or_none(track.get('BitsPerSample', 16)),
2400 'nal_unit_length_field': int_or_none(track.get('NALUnitLengthField', 4)),
2405 def _parse_html5_media_entries(self, base_url, webpage, video_id, m3u8_id=None, m3u8_entry_protocol='m3u8', mpd_id=None, preference=None):
2406 def absolute_url(item_url):
2407 return urljoin(base_url, item_url)
2409 def parse_content_type(content_type):
2410 if not content_type:
2412 ctr = re.search(r'(?P<mimetype>[^/]+/[^;]+)(?:;\s*codecs="?(?P<codecs>[^"]+))?', content_type)
2414 mimetype, codecs = ctr.groups()
2415 f = parse_codecs(codecs)
2416 f['ext'] = mimetype2ext(mimetype)
2420 def _media_formats(src, cur_media_type, type_info={}):
2421 full_url = absolute_url(src)
2422 ext = type_info.get('ext') or determine_ext(full_url)
2424 is_plain_url = False
2425 formats = self._extract_m3u8_formats(
2426 full_url, video_id, ext='mp4',
2427 entry_protocol=m3u8_entry_protocol, m3u8_id=m3u8_id,
2428 preference=preference, fatal=False)
2430 is_plain_url = False
2431 formats = self._extract_mpd_formats(
2432 full_url, video_id, mpd_id=mpd_id, fatal=False)
2437 'vcodec': 'none' if cur_media_type == 'audio' else None,
2439 return is_plain_url, formats
2442 # amp-video and amp-audio are very similar to their HTML5 counterparts
2443 # so we wll include them right here (see
2444 # https://www.ampproject.org/docs/reference/components/amp-video)
2445 media_tags = [(media_tag, media_type, '')
2446 for media_tag, media_type
2447 in re.findall(r'(?s)(<(?:amp-)?(video|audio)[^>]*/>)', webpage)]
2448 media_tags.extend(re.findall(
2449 # We only allow video|audio followed by a whitespace or '>'.
2450 # Allowing more characters may end up in significant slow down (see
2451 # https://github.com/rg3/youtube-dl/issues/11979, example URL:
2452 # http://www.porntrex.com/maps/videositemap.xml).
2453 r'(?s)(<(?P<tag>(?:amp-)?(?:video|audio))(?:\s+[^>]*)?>)(.*?)</(?P=tag)>', webpage))
2454 for media_tag, media_type, media_content in media_tags:
2459 media_attributes = extract_attributes(media_tag)
2460 src = media_attributes.get('src')
2462 _, formats = _media_formats(src, media_type)
2463 media_info['formats'].extend(formats)
2464 media_info['thumbnail'] = absolute_url(media_attributes.get('poster'))
2466 for source_tag in re.findall(r'<source[^>]+>', media_content):
2467 source_attributes = extract_attributes(source_tag)
2468 src = source_attributes.get('src')
2471 f = parse_content_type(source_attributes.get('type'))
2472 is_plain_url, formats = _media_formats(src, media_type, f)
2474 # res attribute is not standard but seen several times
2477 'height': int_or_none(source_attributes.get('res')),
2478 'format_id': source_attributes.get('label'),
2480 f.update(formats[0])
2481 media_info['formats'].append(f)
2483 media_info['formats'].extend(formats)
2484 for track_tag in re.findall(r'<track[^>]+>', media_content):
2485 track_attributes = extract_attributes(track_tag)
2486 kind = track_attributes.get('kind')
2487 if not kind or kind in ('subtitles', 'captions'):
2488 src = track_attributes.get('src')
2491 lang = track_attributes.get('srclang') or track_attributes.get('lang') or track_attributes.get('label')
2492 media_info['subtitles'].setdefault(lang, []).append({
2493 'url': absolute_url(src),
2495 for f in media_info['formats']:
2496 f.setdefault('http_headers', {})['Referer'] = base_url
2497 if media_info['formats'] or media_info['subtitles']:
2498 entries.append(media_info)
2501 def _extract_akamai_formats(self, manifest_url, video_id, hosts={}):
2503 hdcore_sign = 'hdcore=3.7.0'
2504 f4m_url = re.sub(r'(https?://[^/]+)/i/', r'\1/z/', manifest_url).replace('/master.m3u8', '/manifest.f4m')
2505 hds_host = hosts.get('hds')
2507 f4m_url = re.sub(r'(https?://)[^/]+', r'\1' + hds_host, f4m_url)
2508 if 'hdcore=' not in f4m_url:
2509 f4m_url += ('&' if '?' in f4m_url else '?') + hdcore_sign
2510 f4m_formats = self._extract_f4m_formats(
2511 f4m_url, video_id, f4m_id='hds', fatal=False)
2512 for entry in f4m_formats:
2513 entry.update({'extra_param_to_segment_url': hdcore_sign})
2514 formats.extend(f4m_formats)
2515 m3u8_url = re.sub(r'(https?://[^/]+)/z/', r'\1/i/', manifest_url).replace('/manifest.f4m', '/master.m3u8')
2516 hls_host = hosts.get('hls')
2518 m3u8_url = re.sub(r'(https?://)[^/]+', r'\1' + hls_host, m3u8_url)
2519 formats.extend(self._extract_m3u8_formats(
2520 m3u8_url, video_id, 'mp4', 'm3u8_native',
2521 m3u8_id='hls', fatal=False))
2524 def _extract_wowza_formats(self, url, video_id, m3u8_entry_protocol='m3u8_native', skip_protocols=[]):
2525 query = compat_urlparse.urlparse(url).query
2526 url = re.sub(r'/(?:manifest|playlist|jwplayer)\.(?:m3u8|f4m|mpd|smil)', '', url)
2528 r'(?:(?:http|rtmp|rtsp)(?P<s>s)?:)?(?P<url>//[^?]+)', url)
2529 url_base = mobj.group('url')
2530 http_base_url = '%s%s:%s' % ('http', mobj.group('s') or '', url_base)
2533 def manifest_url(manifest):
2534 m_url = '%s/%s' % (http_base_url, manifest)
2536 m_url += '?%s' % query
2539 if 'm3u8' not in skip_protocols:
2540 formats.extend(self._extract_m3u8_formats(
2541 manifest_url('playlist.m3u8'), video_id, 'mp4',
2542 m3u8_entry_protocol, m3u8_id='hls', fatal=False))
2543 if 'f4m' not in skip_protocols:
2544 formats.extend(self._extract_f4m_formats(
2545 manifest_url('manifest.f4m'),
2546 video_id, f4m_id='hds', fatal=False))
2547 if 'dash' not in skip_protocols:
2548 formats.extend(self._extract_mpd_formats(
2549 manifest_url('manifest.mpd'),
2550 video_id, mpd_id='dash', fatal=False))
2551 if re.search(r'(?:/smil:|\.smil)', url_base):
2552 if 'smil' not in skip_protocols:
2553 rtmp_formats = self._extract_smil_formats(
2554 manifest_url('jwplayer.smil'),
2555 video_id, fatal=False)
2556 for rtmp_format in rtmp_formats:
2557 rtsp_format = rtmp_format.copy()
2558 rtsp_format['url'] = '%s/%s' % (rtmp_format['url'], rtmp_format['play_path'])
2559 del rtsp_format['play_path']
2560 del rtsp_format['ext']
2561 rtsp_format.update({
2562 'url': rtsp_format['url'].replace('rtmp://', 'rtsp://'),
2563 'format_id': rtmp_format['format_id'].replace('rtmp', 'rtsp'),
2566 formats.extend([rtmp_format, rtsp_format])
2568 for protocol in ('rtmp', 'rtsp'):
2569 if protocol not in skip_protocols:
2571 'url': '%s:%s' % (protocol, url_base),
2572 'format_id': protocol,
2573 'protocol': protocol,
2577 def _find_jwplayer_data(self, webpage, video_id=None, transform_source=js_to_json):
2579 r'(?s)jwplayer\((?P<quote>[\'"])[^\'" ]+(?P=quote)\)(?!</script>).*?\.setup\s*\((?P<options>[^)]+)\)',
2583 jwplayer_data = self._parse_json(mobj.group('options'),
2585 transform_source=transform_source)
2586 except ExtractorError:
2589 if isinstance(jwplayer_data, dict):
2590 return jwplayer_data
2592 def _extract_jwplayer_data(self, webpage, video_id, *args, **kwargs):
2593 jwplayer_data = self._find_jwplayer_data(
2594 webpage, video_id, transform_source=js_to_json)
2595 return self._parse_jwplayer_data(
2596 jwplayer_data, video_id, *args, **kwargs)
2598 def _parse_jwplayer_data(self, jwplayer_data, video_id=None, require_title=True,
2599 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
2600 # JWPlayer backward compatibility: flattened playlists
2601 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/api/config.js#L81-L96
2602 if 'playlist' not in jwplayer_data:
2603 jwplayer_data = {'playlist': [jwplayer_data]}
2607 # JWPlayer backward compatibility: single playlist item
2608 # https://github.com/jwplayer/jwplayer/blob/v7.7.0/src/js/playlist/playlist.js#L10
2609 if not isinstance(jwplayer_data['playlist'], list):
2610 jwplayer_data['playlist'] = [jwplayer_data['playlist']]
2612 for video_data in jwplayer_data['playlist']:
2613 # JWPlayer backward compatibility: flattened sources
2614 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/playlist/item.js#L29-L35
2615 if 'sources' not in video_data:
2616 video_data['sources'] = [video_data]
2618 this_video_id = video_id or video_data['mediaid']
2620 formats = self._parse_jwplayer_formats(
2621 video_data['sources'], video_id=this_video_id, m3u8_id=m3u8_id,
2622 mpd_id=mpd_id, rtmp_params=rtmp_params, base_url=base_url)
2625 tracks = video_data.get('tracks')
2626 if tracks and isinstance(tracks, list):
2627 for track in tracks:
2628 if not isinstance(track, dict):
2630 track_kind = track.get('kind')
2631 if not track_kind or not isinstance(track_kind, compat_str):
2633 if track_kind.lower() not in ('captions', 'subtitles'):
2635 track_url = urljoin(base_url, track.get('file'))
2638 subtitles.setdefault(track.get('label') or 'en', []).append({
2639 'url': self._proto_relative_url(track_url)
2643 'id': this_video_id,
2644 'title': unescapeHTML(video_data['title'] if require_title else video_data.get('title')),
2645 'description': video_data.get('description'),
2646 'thumbnail': urljoin(base_url, self._proto_relative_url(video_data.get('image'))),
2647 'timestamp': int_or_none(video_data.get('pubdate')),
2648 'duration': float_or_none(jwplayer_data.get('duration') or video_data.get('duration')),
2649 'subtitles': subtitles,
2651 # https://github.com/jwplayer/jwplayer/blob/master/src/js/utils/validator.js#L32
2652 if len(formats) == 1 and re.search(r'^(?:http|//).*(?:youtube\.com|youtu\.be)/.+', formats[0]['url']):
2654 '_type': 'url_transparent',
2655 'url': formats[0]['url'],
2658 self._sort_formats(formats)
2659 entry['formats'] = formats
2660 entries.append(entry)
2661 if len(entries) == 1:
2664 return self.playlist_result(entries)
2666 def _parse_jwplayer_formats(self, jwplayer_sources_data, video_id=None,
2667 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
2670 for source in jwplayer_sources_data:
2671 if not isinstance(source, dict):
2673 source_url = urljoin(
2674 base_url, self._proto_relative_url(source.get('file')))
2675 if not source_url or source_url in urls:
2677 urls.append(source_url)
2678 source_type = source.get('type') or ''
2679 ext = mimetype2ext(source_type) or determine_ext(source_url)
2680 if source_type == 'hls' or ext == 'm3u8':
2681 formats.extend(self._extract_m3u8_formats(
2682 source_url, video_id, 'mp4', entry_protocol='m3u8_native',
2683 m3u8_id=m3u8_id, fatal=False))
2684 elif source_type == 'dash' or ext == 'mpd':
2685 formats.extend(self._extract_mpd_formats(
2686 source_url, video_id, mpd_id=mpd_id, fatal=False))
2688 formats.extend(self._extract_smil_formats(
2689 source_url, video_id, fatal=False))
2690 # https://github.com/jwplayer/jwplayer/blob/master/src/js/providers/default.js#L67
2691 elif source_type.startswith('audio') or ext in (
2692 'oga', 'aac', 'mp3', 'mpeg', 'vorbis'):
2699 height = int_or_none(source.get('height'))
2701 # Often no height is provided but there is a label in
2702 # format like "1080p", "720p SD", or 1080.
2703 height = int_or_none(self._search_regex(
2704 r'^(\d{3,4})[pP]?(?:\b|$)', compat_str(source.get('label') or ''),
2705 'height', default=None))
2708 'width': int_or_none(source.get('width')),
2710 'tbr': int_or_none(source.get('bitrate')),
2713 if source_url.startswith('rtmp'):
2714 a_format['ext'] = 'flv'
2715 # See com/longtailvideo/jwplayer/media/RTMPMediaProvider.as
2716 # of jwplayer.flash.swf
2717 rtmp_url_parts = re.split(
2718 r'((?:mp4|mp3|flv):)', source_url, 1)
2719 if len(rtmp_url_parts) == 3:
2720 rtmp_url, prefix, play_path = rtmp_url_parts
2723 'play_path': prefix + play_path,
2726 a_format.update(rtmp_params)
2727 formats.append(a_format)
2730 def _live_title(self, name):
2731 """ Generate the title for a live video """
2732 now = datetime.datetime.now()
2733 now_str = now.strftime('%Y-%m-%d %H:%M')
2734 return name + ' ' + now_str
2736 def _int(self, v, name, fatal=False, **kwargs):
2737 res = int_or_none(v, **kwargs)
2738 if 'get_attr' in kwargs:
2739 print(getattr(v, kwargs['get_attr']))
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 _float(self, v, name, fatal=False, **kwargs):
2749 res = float_or_none(v, **kwargs)
2751 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
2753 raise ExtractorError(msg)
2755 self._downloader.report_warning(msg)
2758 def _set_cookie(self, domain, name, value, expire_time=None, port=None,
2759 path='/', secure=False, discard=False, rest={}, **kwargs):
2760 cookie = compat_cookiejar.Cookie(
2761 0, name, value, port, port is not None, domain, True,
2762 domain.startswith('.'), path, True, secure, expire_time,
2763 discard, None, None, rest)
2764 self._downloader.cookiejar.set_cookie(cookie)
2766 def _get_cookies(self, url):
2767 """ Return a compat_cookies.SimpleCookie with the cookies for the url """
2768 req = sanitized_Request(url)
2769 self._downloader.cookiejar.add_cookie_header(req)
2770 return compat_cookies.SimpleCookie(req.get_header('Cookie'))
2772 def get_testcases(self, include_onlymatching=False):
2773 t = getattr(self, '_TEST', None)
2775 assert not hasattr(self, '_TESTS'), \
2776 '%s has _TEST and _TESTS' % type(self).__name__
2779 tests = getattr(self, '_TESTS', [])
2781 if not include_onlymatching and t.get('only_matching', False):
2783 t['name'] = type(self).__name__[:-len('IE')]
2786 def is_suitable(self, age_limit):
2787 """ Test whether the extractor is generally suitable for the given
2788 age limit (i.e. pornographic sites are not, all others usually are) """
2790 any_restricted = False
2791 for tc in self.get_testcases(include_onlymatching=False):
2792 if tc.get('playlist', []):
2793 tc = tc['playlist'][0]
2794 is_restricted = age_restricted(
2795 tc.get('info_dict', {}).get('age_limit'), age_limit)
2796 if not is_restricted:
2798 any_restricted = any_restricted or is_restricted
2799 return not any_restricted
2801 def extract_subtitles(self, *args, **kwargs):
2802 if (self._downloader.params.get('writesubtitles', False) or
2803 self._downloader.params.get('listsubtitles')):
2804 return self._get_subtitles(*args, **kwargs)
2807 def _get_subtitles(self, *args, **kwargs):
2808 raise NotImplementedError('This method must be implemented by subclasses')
2811 def _merge_subtitle_items(subtitle_list1, subtitle_list2):
2812 """ Merge subtitle items for one language. Items with duplicated URLs
2813 will be dropped. """
2814 list1_urls = set([item['url'] for item in subtitle_list1])
2815 ret = list(subtitle_list1)
2816 ret.extend([item for item in subtitle_list2 if item['url'] not in list1_urls])
2820 def _merge_subtitles(cls, subtitle_dict1, subtitle_dict2):
2821 """ Merge two subtitle dictionaries, language by language. """
2822 ret = dict(subtitle_dict1)
2823 for lang in subtitle_dict2:
2824 ret[lang] = cls._merge_subtitle_items(subtitle_dict1.get(lang, []), subtitle_dict2[lang])
2827 def extract_automatic_captions(self, *args, **kwargs):
2828 if (self._downloader.params.get('writeautomaticsub', False) or
2829 self._downloader.params.get('listsubtitles')):
2830 return self._get_automatic_captions(*args, **kwargs)
2833 def _get_automatic_captions(self, *args, **kwargs):
2834 raise NotImplementedError('This method must be implemented by subclasses')
2836 def mark_watched(self, *args, **kwargs):
2837 if (self._downloader.params.get('mark_watched', False) and
2838 (self._get_login_info()[0] is not None or
2839 self._downloader.params.get('cookiefile') is not None)):
2840 self._mark_watched(*args, **kwargs)
2842 def _mark_watched(self, *args, **kwargs):
2843 raise NotImplementedError('This method must be implemented by subclasses')
2845 def geo_verification_headers(self):
2847 geo_verification_proxy = self._downloader.params.get('geo_verification_proxy')
2848 if geo_verification_proxy:
2849 headers['Ytdl-request-proxy'] = geo_verification_proxy
2852 def _generic_id(self, url):
2853 return compat_urllib_parse_unquote(os.path.splitext(url.rstrip('/').split('/')[-1])[0])
2855 def _generic_title(self, url):
2856 return compat_urllib_parse_unquote(os.path.splitext(url_basename(url))[0])
2859 class SearchInfoExtractor(InfoExtractor):
2861 Base class for paged search queries extractors.
2862 They accept URLs in the format _SEARCH_KEY(|all|[0-9]):{query}
2863 Instances should define _SEARCH_KEY and _MAX_RESULTS.
2867 def _make_valid_url(cls):
2868 return r'%s(?P<prefix>|[1-9][0-9]*|all):(?P<query>[\s\S]+)' % cls._SEARCH_KEY
2871 def suitable(cls, url):
2872 return re.match(cls._make_valid_url(), url) is not None
2874 def _real_extract(self, query):
2875 mobj = re.match(self._make_valid_url(), query)
2877 raise ExtractorError('Invalid search query "%s"' % query)
2879 prefix = mobj.group('prefix')
2880 query = mobj.group('query')
2882 return self._get_n_results(query, 1)
2883 elif prefix == 'all':
2884 return self._get_n_results(query, self._MAX_RESULTS)
2888 raise ExtractorError('invalid download number %s for query "%s"' % (n, query))
2889 elif n > self._MAX_RESULTS:
2890 self._downloader.report_warning('%s returns max %i results (you requested %i)' % (self._SEARCH_KEY, self._MAX_RESULTS, n))
2891 n = self._MAX_RESULTS
2892 return self._get_n_results(query, n)
2894 def _get_n_results(self, query, n):
2895 """Get a specified number of results for a query"""
2896 raise NotImplementedError('This method must be implemented by subclasses')
2899 def SEARCH_KEY(self):
2900 return self._SEARCH_KEY