2 from __future__ import unicode_literals
17 from ..compat import (
20 compat_etree_fromstring,
27 compat_urllib_parse_unquote,
28 compat_urllib_parse_urlencode,
29 compat_urllib_request,
31 compat_xml_parse_error,
33 from ..downloader.f4m import (
35 remove_encrypted_media,
61 parse_m3u8_attributes,
78 class InfoExtractor(object):
79 """Information Extractor class.
81 Information extractors are the classes that, given a URL, extract
82 information about the video (or videos) the URL refers to. This
83 information includes the real video URL, the video title, author and
84 others. The information is stored in a dictionary which is then
85 passed to the YoutubeDL. The YoutubeDL processes this
86 information possibly downloading the video to the file system, among
87 other possible outcomes.
89 The type field determines the type of the result.
90 By far the most common value (and the default if _type is missing) is
91 "video", which indicates a single video.
93 For a video, the dictionaries must include the following fields:
96 title: Video title, unescaped.
98 Additionally, it must contain either a formats entry or a url one:
100 formats: A list of dictionaries for each format available, ordered
101 from worst to best quality.
104 * url Mandatory. The URL of the video file
106 The URL of the manifest file in case of
107 fragmented media (DASH, hls, hds)
108 * ext Will be calculated from URL if missing
109 * format A human-readable description of the format
110 ("mp4 container with h264/opus").
111 Calculated from the format_id, width, height.
112 and format_note fields if missing.
113 * format_id A short description of the format
114 ("mp4_h264_opus" or "19").
115 Technically optional, but strongly recommended.
116 * format_note Additional info about the format
117 ("3D" or "DASH video")
118 * width Width of the video, if known
119 * height Height of the video, if known
120 * resolution Textual description of width and height
121 * tbr Average bitrate of audio and video in KBit/s
122 * abr Average audio bitrate in KBit/s
123 * acodec Name of the audio codec in use
124 * asr Audio sampling rate in Hertz
125 * vbr Average video bitrate in KBit/s
127 * vcodec Name of the video codec in use
128 * container Name of the container format
129 * filesize The number of bytes, if known in advance
130 * filesize_approx An estimate for the number of bytes
131 * player_url SWF Player URL (used for rtmpdump).
132 * protocol The protocol that will be used for the actual
133 download, lower-case.
134 "http", "https", "rtsp", "rtmp", "rtmpe",
135 "m3u8", "m3u8_native" or "http_dash_segments".
137 Base URL for fragments. Each fragment's path
138 value (if present) will be relative to
140 * fragments A list of fragments of a fragmented media.
141 Each fragment entry must contain either an url
142 or a path. If an url is present it should be
143 considered by a client. Otherwise both path and
144 fragment_base_url must be present. Here is
145 the list of all potential fields:
146 * "url" - fragment's URL
147 * "path" - fragment's path relative to
149 * "duration" (optional, int or float)
150 * "filesize" (optional, int)
151 * preference Order number of this format. If this field is
152 present and not None, the formats get sorted
153 by this field, regardless of all other values.
154 -1 for default (order by other properties),
155 -2 or smaller for less than default.
156 < -1000 to hide the format (if there is
157 another one which is strictly better)
158 * language Language code, e.g. "de" or "en-US".
159 * language_preference Is this in the language mentioned in
161 10 if it's what the URL is about,
162 -1 for default (don't know),
163 -10 otherwise, other values reserved for now.
164 * quality Order number of the video quality of this
165 format, irrespective of the file format.
166 -1 for default (order by other properties),
167 -2 or smaller for less than default.
168 * source_preference Order number for this video source
169 (quality takes higher priority)
170 -1 for default (order by other properties),
171 -2 or smaller for less than default.
172 * http_headers A dictionary of additional HTTP headers
173 to add to the request.
174 * stretched_ratio If given and not 1, indicates that the
175 video's pixels are not square.
176 width : height ratio as float.
177 * no_resume The server does not support resuming the
178 (HTTP or RTMP) download. Boolean.
179 * downloader_options A dictionary of downloader options as
180 described in FileDownloader
182 url: Final video URL.
183 ext: Video filename extension.
184 format: The video format, defaults to ext (used for --get-format)
185 player_url: SWF Player URL (used for rtmpdump).
187 The following fields are optional:
189 alt_title: A secondary title of the video.
190 display_id An alternative identifier for the video, not necessarily
191 unique, but available before title. Typically, id is
192 something like "4234987", title "Dancing naked mole rats",
193 and display_id "dancing-naked-mole-rats"
194 thumbnails: A list of dictionaries, with the following entries:
195 * "id" (optional, string) - Thumbnail format ID
197 * "preference" (optional, int) - quality of the image
198 * "width" (optional, int)
199 * "height" (optional, int)
200 * "resolution" (optional, string "{width}x{height"},
202 * "filesize" (optional, int)
203 thumbnail: Full URL to a video thumbnail image.
204 description: Full video description.
205 uploader: Full name of the video uploader.
206 license: License name the video is licensed under.
207 creator: The creator of the video.
208 release_date: The date (YYYYMMDD) when the video was released.
209 timestamp: UNIX timestamp of the moment the video became available.
210 upload_date: Video upload date (YYYYMMDD).
211 If not explicitly set, calculated from timestamp.
212 uploader_id: Nickname or id of the video uploader.
213 uploader_url: Full URL to a personal webpage of the video uploader.
214 location: Physical location where the video was filmed.
215 subtitles: The available subtitles as a dictionary in the format
216 {tag: subformats}. "tag" is usually a language code, and
217 "subformats" is a list sorted from lower to higher
218 preference, each element is a dictionary with the "ext"
220 * "data": The subtitles file contents
221 * "url": A URL pointing to the subtitles file
222 "ext" will be calculated from URL if missing
223 automatic_captions: Like 'subtitles', used by the YoutubeIE for
224 automatically generated captions
225 duration: Length of the video in seconds, as an integer or float.
226 view_count: How many users have watched the video on the platform.
227 like_count: Number of positive ratings of the video
228 dislike_count: Number of negative ratings of the video
229 repost_count: Number of reposts of the video
230 average_rating: Average rating give by users, the scale used depends on the webpage
231 comment_count: Number of comments on the video
232 comments: A list of comments, each with one or more of the following
233 properties (all but one of text or html optional):
234 * "author" - human-readable name of the comment author
235 * "author_id" - user ID of the comment author
237 * "html" - Comment as HTML
238 * "text" - Plain text of the comment
239 * "timestamp" - UNIX timestamp of comment
240 * "parent" - ID of the comment this one is replying to.
241 Set to "root" to indicate that this is a
242 comment to the original video.
243 age_limit: Age restriction for the video, as an integer (years)
244 webpage_url: The URL to the video webpage, if given to youtube-dl it
245 should allow to get the same result again. (It will be set
246 by YoutubeDL if it's missing)
247 categories: A list of categories that the video falls in, for example
249 tags: A list of tags assigned to the video, e.g. ["sweden", "pop music"]
250 is_live: True, False, or None (=unknown). Whether this video is a
251 live stream that goes on instead of a fixed-length video.
252 start_time: Time in seconds where the reproduction should start, as
253 specified in the URL.
254 end_time: Time in seconds where the reproduction should end, as
255 specified in the URL.
256 chapters: A list of dictionaries, with the following entries:
257 * "start_time" - The start time of the chapter in seconds
258 * "end_time" - The end time of the chapter in seconds
259 * "title" (optional, string)
261 The following fields should only be used when the video belongs to some logical
264 chapter: Name or title of the chapter the video belongs to.
265 chapter_number: Number of the chapter the video belongs to, as an integer.
266 chapter_id: Id of the chapter the video belongs to, as a unicode string.
268 The following fields should only be used when the video is an episode of some
269 series, programme or podcast:
271 series: Title of the series or programme the video episode belongs to.
272 season: Title of the season the video episode belongs to.
273 season_number: Number of the season the video episode belongs to, as an integer.
274 season_id: Id of the season the video episode belongs to, as a unicode string.
275 episode: Title of the video episode. Unlike mandatory video title field,
276 this field should denote the exact title of the video episode
277 without any kind of decoration.
278 episode_number: Number of the video episode within a season, as an integer.
279 episode_id: Id of the video episode, as a unicode string.
281 The following fields should only be used when the media is a track or a part of
284 track: Title of the track.
285 track_number: Number of the track within an album or a disc, as an integer.
286 track_id: Id of the track (useful in case of custom indexing, e.g. 6.iii),
288 artist: Artist(s) of the track.
289 genre: Genre(s) of the track.
290 album: Title of the album the track belongs to.
291 album_type: Type of the album (e.g. "Demo", "Full-length", "Split", "Compilation", etc).
292 album_artist: List of all artists appeared on the album (e.g.
293 "Ash Borer / Fell Voices" or "Various Artists", useful for splits
295 disc_number: Number of the disc or other physical medium the track belongs to,
297 release_year: Year (YYYY) when the album was released.
299 Unless mentioned otherwise, the fields should be Unicode strings.
301 Unless mentioned otherwise, None is equivalent to absence of information.
304 _type "playlist" indicates multiple videos.
305 There must be a key "entries", which is a list, an iterable, or a PagedList
306 object, each element of which is a valid dictionary by this specification.
308 Additionally, playlists can have "id", "title", "description", "uploader",
309 "uploader_id", "uploader_url" attributes with the same semantics as videos
313 _type "multi_video" indicates that there are multiple videos that
314 form a single show, for examples multiple acts of an opera or TV episode.
315 It must have an entries key like a playlist and contain all the keys
316 required for a video at the same time.
319 _type "url" indicates that the video must be extracted from another
320 location, possibly by a different extractor. Its only required key is:
321 "url" - the next URL to extract.
322 The key "ie_key" can be set to the class name (minus the trailing "IE",
323 e.g. "Youtube") if the extractor class is known in advance.
324 Additionally, the dictionary may have any properties of the resolved entity
325 known in advance, for example "title" if the title of the referred video is
329 _type "url_transparent" entities have the same specification as "url", but
330 indicate that the given additional information is more precise than the one
331 associated with the resolved URL.
332 This is useful when a site employs a video service that hosts the video and
333 its technical metadata, but that video service does not embed a useful
334 title, description etc.
337 Subclasses of this one should re-define the _real_initialize() and
338 _real_extract() methods and define a _VALID_URL regexp.
339 Probably, they should also be added to the list of extractors.
341 _GEO_BYPASS attribute may be set to False in order to disable
342 geo restriction bypass mechanisms for a particular extractor.
343 Though it won't disable explicit geo restriction bypass based on
344 country code provided with geo_bypass_country.
346 _GEO_COUNTRIES attribute may contain a list of presumably geo unrestricted
347 countries for this extractor. One of these countries will be used by
348 geo restriction bypass mechanism right away in order to bypass
349 geo restriction, of course, if the mechanism is not disabled.
351 _GEO_IP_BLOCKS attribute may contain a list of presumably geo unrestricted
352 IP blocks in CIDR notation for this extractor. One of these IP blocks
353 will be used by geo restriction bypass mechanism similarly
356 Finally, the _WORKING attribute should be set to False for broken IEs
357 in order to warn the users and skip the tests.
362 _x_forwarded_for_ip = None
364 _GEO_COUNTRIES = None
365 _GEO_IP_BLOCKS = None
368 def __init__(self, downloader=None):
369 """Constructor. Receives an optional downloader."""
371 self._x_forwarded_for_ip = None
372 self.set_downloader(downloader)
375 def suitable(cls, url):
376 """Receives a URL and returns True if suitable for this IE."""
378 # This does not use has/getattr intentionally - we want to know whether
379 # we have cached the regexp for *this* class, whereas getattr would also
380 # match the superclass
381 if '_VALID_URL_RE' not in cls.__dict__:
382 cls._VALID_URL_RE = re.compile(cls._VALID_URL)
383 return cls._VALID_URL_RE.match(url) is not None
386 def _match_id(cls, url):
387 if '_VALID_URL_RE' not in cls.__dict__:
388 cls._VALID_URL_RE = re.compile(cls._VALID_URL)
389 m = cls._VALID_URL_RE.match(url)
391 return compat_str(m.group('id'))
395 """Getter method for _WORKING."""
398 def initialize(self):
399 """Initializes an instance (authentication, etc)."""
400 self._initialize_geo_bypass({
401 'countries': self._GEO_COUNTRIES,
402 'ip_blocks': self._GEO_IP_BLOCKS,
405 self._real_initialize()
408 def _initialize_geo_bypass(self, geo_bypass_context):
410 Initialize geo restriction bypass mechanism.
412 This method is used to initialize geo bypass mechanism based on faking
413 X-Forwarded-For HTTP header. A random country from provided country list
414 is selected and a random IP belonging to this country is generated. This
415 IP will be passed as X-Forwarded-For HTTP header in all subsequent
418 This method will be used for initial geo bypass mechanism initialization
419 during the instance initialization with _GEO_COUNTRIES and
422 You may also manually call it from extractor's code if geo bypass
423 information is not available beforehand (e.g. obtained during
424 extraction) or due to some other reason. In this case you should pass
425 this information in geo bypass context passed as first argument. It may
426 contain following fields:
428 countries: List of geo unrestricted countries (similar
430 ip_blocks: List of geo unrestricted IP blocks in CIDR notation
431 (similar to _GEO_IP_BLOCKS)
434 if not self._x_forwarded_for_ip:
436 # Geo bypass mechanism is explicitly disabled by user
437 if not self._downloader.params.get('geo_bypass', True):
440 if not geo_bypass_context:
441 geo_bypass_context = {}
443 # Backward compatibility: previously _initialize_geo_bypass
444 # expected a list of countries, some 3rd party code may still use
446 if isinstance(geo_bypass_context, (list, tuple)):
447 geo_bypass_context = {
448 'countries': geo_bypass_context,
451 # The whole point of geo bypass mechanism is to fake IP
452 # as X-Forwarded-For HTTP header based on some IP block or
455 # Path 1: bypassing based on IP block in CIDR notation
457 # Explicit IP block specified by user, use it right away
458 # regardless of whether extractor is geo bypassable or not
459 ip_block = self._downloader.params.get('geo_bypass_ip_block', None)
461 # Otherwise use random IP block from geo bypass context but only
462 # if extractor is known as geo bypassable
464 ip_blocks = geo_bypass_context.get('ip_blocks')
465 if self._GEO_BYPASS and ip_blocks:
466 ip_block = random.choice(ip_blocks)
469 self._x_forwarded_for_ip = GeoUtils.random_ipv4(ip_block)
470 if self._downloader.params.get('verbose', False):
471 self._downloader.to_screen(
472 '[debug] Using fake IP %s as X-Forwarded-For.'
473 % self._x_forwarded_for_ip)
476 # Path 2: bypassing based on country code
478 # Explicit country code specified by user, use it right away
479 # regardless of whether extractor is geo bypassable or not
480 country = self._downloader.params.get('geo_bypass_country', None)
482 # Otherwise use random country code from geo bypass context but
483 # only if extractor is known as geo bypassable
485 countries = geo_bypass_context.get('countries')
486 if self._GEO_BYPASS and countries:
487 country = random.choice(countries)
490 self._x_forwarded_for_ip = GeoUtils.random_ipv4(country)
491 if self._downloader.params.get('verbose', False):
492 self._downloader.to_screen(
493 '[debug] Using fake IP %s (%s) as X-Forwarded-For.'
494 % (self._x_forwarded_for_ip, country.upper()))
496 def extract(self, url):
497 """Extracts URL information and returns it in list of dicts."""
502 ie_result = self._real_extract(url)
503 if self._x_forwarded_for_ip:
504 ie_result['__x_forwarded_for_ip'] = self._x_forwarded_for_ip
506 except GeoRestrictedError as e:
507 if self.__maybe_fake_ip_and_retry(e.countries):
510 except ExtractorError:
512 except compat_http_client.IncompleteRead as e:
513 raise ExtractorError('A network error has occurred.', cause=e, expected=True)
514 except (KeyError, StopIteration) as e:
515 raise ExtractorError('An extractor error has occurred.', cause=e)
517 def __maybe_fake_ip_and_retry(self, countries):
518 if (not self._downloader.params.get('geo_bypass_country', None) and
520 self._downloader.params.get('geo_bypass', True) and
521 not self._x_forwarded_for_ip and
523 country_code = random.choice(countries)
524 self._x_forwarded_for_ip = GeoUtils.random_ipv4(country_code)
525 if self._x_forwarded_for_ip:
527 'Video is geo restricted. Retrying extraction with fake IP %s (%s) as X-Forwarded-For.'
528 % (self._x_forwarded_for_ip, country_code.upper()))
532 def set_downloader(self, downloader):
533 """Sets the downloader for this IE."""
534 self._downloader = downloader
536 def _real_initialize(self):
537 """Real initialization process. Redefine in subclasses."""
540 def _real_extract(self, url):
541 """Real extraction process. Redefine in subclasses."""
546 """A string for getting the InfoExtractor with get_info_extractor"""
547 return compat_str(cls.__name__[:-2])
551 return compat_str(type(self).__name__[:-2])
554 def __can_accept_status_code(err, expected_status):
555 assert isinstance(err, compat_urllib_error.HTTPError)
556 if expected_status is None:
558 if isinstance(expected_status, compat_integer_types):
559 return err.code == expected_status
560 elif isinstance(expected_status, (list, tuple)):
561 return err.code in expected_status
562 elif callable(expected_status):
563 return expected_status(err.code) is True
567 def _request_webpage(self, url_or_request, video_id, note=None, errnote=None, fatal=True, data=None, headers={}, query={}, expected_status=None):
569 Return the response handle.
571 See _download_webpage docstring for arguments specification.
574 self.report_download_webpage(video_id)
575 elif note is not False:
577 self.to_screen('%s' % (note,))
579 self.to_screen('%s: %s' % (video_id, note))
581 # Some sites check X-Forwarded-For HTTP header in order to figure out
582 # the origin of the client behind proxy. This allows bypassing geo
583 # restriction by faking this header's value to IP that belongs to some
584 # geo unrestricted country. We will do so once we encounter any
585 # geo restriction error.
586 if self._x_forwarded_for_ip:
587 if 'X-Forwarded-For' not in headers:
588 headers['X-Forwarded-For'] = self._x_forwarded_for_ip
590 if isinstance(url_or_request, compat_urllib_request.Request):
591 url_or_request = update_Request(
592 url_or_request, data=data, headers=headers, query=query)
595 url_or_request = update_url_query(url_or_request, query)
596 if data is not None or headers:
597 url_or_request = sanitized_Request(url_or_request, data, headers)
599 return self._downloader.urlopen(url_or_request)
600 except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err:
601 if isinstance(err, compat_urllib_error.HTTPError):
602 if self.__can_accept_status_code(err, expected_status):
608 errnote = 'Unable to download webpage'
610 errmsg = '%s: %s' % (errnote, error_to_compat_str(err))
612 raise ExtractorError(errmsg, sys.exc_info()[2], cause=err)
614 self._downloader.report_warning(errmsg)
617 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):
619 Return a tuple (page content as string, URL handle).
621 See _download_webpage docstring for arguments specification.
623 # Strip hashes from the URL (#1038)
624 if isinstance(url_or_request, (compat_str, str)):
625 url_or_request = url_or_request.partition('#')[0]
627 urlh = self._request_webpage(url_or_request, video_id, note, errnote, fatal, data=data, headers=headers, query=query, expected_status=expected_status)
631 content = self._webpage_read_content(urlh, url_or_request, video_id, note, errnote, fatal, encoding=encoding)
632 return (content, urlh)
635 def _guess_encoding_from_content(content_type, webpage_bytes):
636 m = re.match(r'[a-zA-Z0-9_.-]+/[a-zA-Z0-9_.-]+\s*;\s*charset=(.+)', content_type)
638 encoding = m.group(1)
640 m = re.search(br'<meta[^>]+charset=[\'"]?([^\'")]+)[ /\'">]',
641 webpage_bytes[:1024])
643 encoding = m.group(1).decode('ascii')
644 elif webpage_bytes.startswith(b'\xff\xfe'):
651 def __check_blocked(self, content):
652 first_block = content[:512]
653 if ('<title>Access to this site is blocked</title>' in content and
654 'Websense' in first_block):
655 msg = 'Access to this webpage has been blocked by Websense filtering software in your network.'
656 blocked_iframe = self._html_search_regex(
657 r'<iframe src="([^"]+)"', content,
658 'Websense information URL', default=None)
660 msg += ' Visit %s for more details' % blocked_iframe
661 raise ExtractorError(msg, expected=True)
662 if '<title>The URL you requested has been blocked</title>' in first_block:
664 'Access to this webpage has been blocked by Indian censorship. '
665 'Use a VPN or proxy server (with --proxy) to route around it.')
666 block_msg = self._html_search_regex(
667 r'</h1><p>(.*?)</p>',
668 content, 'block message', default=None)
670 msg += ' (Message: "%s")' % block_msg.replace('\n', ' ')
671 raise ExtractorError(msg, expected=True)
672 if ('<title>TTK :: Доступ к ресурсу ограничен</title>' in content and
673 'blocklist.rkn.gov.ru' in content):
674 raise ExtractorError(
675 'Access to this webpage has been blocked by decision of the Russian government. '
676 'Visit http://blocklist.rkn.gov.ru/ for a block reason.',
679 def _webpage_read_content(self, urlh, url_or_request, video_id, note=None, errnote=None, fatal=True, prefix=None, encoding=None):
680 content_type = urlh.headers.get('Content-Type', '')
681 webpage_bytes = urlh.read()
682 if prefix is not None:
683 webpage_bytes = prefix + webpage_bytes
685 encoding = self._guess_encoding_from_content(content_type, webpage_bytes)
686 if self._downloader.params.get('dump_intermediate_pages', False):
687 self.to_screen('Dumping request to ' + urlh.geturl())
688 dump = base64.b64encode(webpage_bytes).decode('ascii')
689 self._downloader.to_screen(dump)
690 if self._downloader.params.get('write_pages', False):
691 basen = '%s_%s' % (video_id, urlh.geturl())
693 h = '___' + hashlib.md5(basen.encode('utf-8')).hexdigest()
694 basen = basen[:240 - len(h)] + h
695 raw_filename = basen + '.dump'
696 filename = sanitize_filename(raw_filename, restricted=True)
697 self.to_screen('Saving request to ' + filename)
698 # Working around MAX_PATH limitation on Windows (see
699 # http://msdn.microsoft.com/en-us/library/windows/desktop/aa365247(v=vs.85).aspx)
700 if compat_os_name == 'nt':
701 absfilepath = os.path.abspath(filename)
702 if len(absfilepath) > 259:
703 filename = '\\\\?\\' + absfilepath
704 with open(filename, 'wb') as outf:
705 outf.write(webpage_bytes)
708 content = webpage_bytes.decode(encoding, 'replace')
710 content = webpage_bytes.decode('utf-8', 'replace')
712 self.__check_blocked(content)
716 def _download_webpage(
717 self, url_or_request, video_id, note=None, errnote=None,
718 fatal=True, tries=1, timeout=5, encoding=None, data=None,
719 headers={}, query={}, expected_status=None):
721 Return the data of the page as a string.
724 url_or_request -- plain text URL as a string or
725 a compat_urllib_request.Requestobject
726 video_id -- Video/playlist/item identifier (string)
729 note -- note printed before downloading (string)
730 errnote -- note printed in case of an error (string)
731 fatal -- flag denoting whether error should be considered fatal,
732 i.e. whether it should cause ExtractionError to be raised,
733 otherwise a warning will be reported and extraction continued
734 tries -- number of tries
735 timeout -- sleep interval between tries
736 encoding -- encoding for a page content decoding, guessed automatically
737 when not explicitly specified
738 data -- POST data (bytes)
739 headers -- HTTP headers (dict)
740 query -- URL query (dict)
741 expected_status -- allows to accept failed HTTP requests (non 2xx
742 status code) by explicitly specifying a set of accepted status
743 codes. Can be any of the following entities:
744 - an integer type specifying an exact failed status code to
746 - a list or a tuple of integer types specifying a list of
747 failed status codes to accept
748 - a callable accepting an actual failed status code and
749 returning True if it should be accepted
750 Note that this argument does not affect success status codes (2xx)
751 which are always accepted.
756 while success is False:
758 res = self._download_webpage_handle(
759 url_or_request, video_id, note, errnote, fatal,
760 encoding=encoding, data=data, headers=headers, query=query,
761 expected_status=expected_status)
763 except compat_http_client.IncompleteRead as e:
765 if try_count >= tries:
767 self._sleep(timeout, video_id)
774 def _download_xml_handle(
775 self, url_or_request, video_id, note='Downloading XML',
776 errnote='Unable to download XML', transform_source=None,
777 fatal=True, encoding=None, data=None, headers={}, query={},
778 expected_status=None):
780 Return a tuple (xml as an xml.etree.ElementTree.Element, URL handle).
782 See _download_webpage docstring for arguments specification.
784 res = self._download_webpage_handle(
785 url_or_request, video_id, note, errnote, fatal=fatal,
786 encoding=encoding, data=data, headers=headers, query=query,
787 expected_status=expected_status)
790 xml_string, urlh = res
791 return self._parse_xml(
792 xml_string, video_id, transform_source=transform_source,
796 self, url_or_request, video_id,
797 note='Downloading XML', errnote='Unable to download XML',
798 transform_source=None, fatal=True, encoding=None,
799 data=None, headers={}, query={}, expected_status=None):
801 Return the xml as an xml.etree.ElementTree.Element.
803 See _download_webpage docstring for arguments specification.
805 res = self._download_xml_handle(
806 url_or_request, video_id, note=note, errnote=errnote,
807 transform_source=transform_source, fatal=fatal, encoding=encoding,
808 data=data, headers=headers, query=query,
809 expected_status=expected_status)
810 return res if res is False else res[0]
812 def _parse_xml(self, xml_string, video_id, transform_source=None, fatal=True):
814 xml_string = transform_source(xml_string)
816 return compat_etree_fromstring(xml_string.encode('utf-8'))
817 except compat_xml_parse_error as ve:
818 errmsg = '%s: Failed to parse XML ' % video_id
820 raise ExtractorError(errmsg, cause=ve)
822 self.report_warning(errmsg + str(ve))
824 def _download_json_handle(
825 self, url_or_request, video_id, note='Downloading JSON metadata',
826 errnote='Unable to download JSON metadata', transform_source=None,
827 fatal=True, encoding=None, data=None, headers={}, query={},
828 expected_status=None):
830 Return a tuple (JSON object, URL handle).
832 See _download_webpage docstring for arguments specification.
834 res = self._download_webpage_handle(
835 url_or_request, video_id, note, errnote, fatal=fatal,
836 encoding=encoding, data=data, headers=headers, query=query,
837 expected_status=expected_status)
840 json_string, urlh = res
841 return self._parse_json(
842 json_string, video_id, transform_source=transform_source,
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 the JSON object as a dict.
853 See _download_webpage docstring for arguments specification.
855 res = self._download_json_handle(
856 url_or_request, video_id, note=note, errnote=errnote,
857 transform_source=transform_source, fatal=fatal, encoding=encoding,
858 data=data, headers=headers, query=query,
859 expected_status=expected_status)
860 return res if res is False else res[0]
862 def _parse_json(self, json_string, video_id, transform_source=None, fatal=True):
864 json_string = transform_source(json_string)
866 return json.loads(json_string)
867 except ValueError as ve:
868 errmsg = '%s: Failed to parse JSON ' % video_id
870 raise ExtractorError(errmsg, cause=ve)
872 self.report_warning(errmsg + str(ve))
874 def report_warning(self, msg, video_id=None):
875 idstr = '' if video_id is None else '%s: ' % video_id
876 self._downloader.report_warning(
877 '[%s] %s%s' % (self.IE_NAME, idstr, msg))
879 def to_screen(self, msg):
880 """Print msg to screen, prefixing it with '[ie_name]'"""
881 self._downloader.to_screen('[%s] %s' % (self.IE_NAME, msg))
883 def report_extraction(self, id_or_name):
884 """Report information extraction."""
885 self.to_screen('%s: Extracting information' % id_or_name)
887 def report_download_webpage(self, video_id):
888 """Report webpage download."""
889 self.to_screen('%s: Downloading webpage' % video_id)
891 def report_age_confirmation(self):
892 """Report attempt to confirm age."""
893 self.to_screen('Confirming age')
895 def report_login(self):
896 """Report attempt to log in."""
897 self.to_screen('Logging in')
900 def raise_login_required(msg='This video is only available for registered users'):
901 raise ExtractorError(
902 '%s. Use --username and --password or --netrc to provide account credentials.' % msg,
906 def raise_geo_restricted(msg='This video is not available from your location due to geo restriction', countries=None):
907 raise GeoRestrictedError(msg, countries=countries)
909 # Methods for following #608
911 def url_result(url, ie=None, video_id=None, video_title=None):
912 """Returns a URL that points to a page that should be processed"""
913 # TODO: ie should be the class used for getting the info
914 video_info = {'_type': 'url',
917 if video_id is not None:
918 video_info['id'] = video_id
919 if video_title is not None:
920 video_info['title'] = video_title
923 def playlist_from_matches(self, matches, playlist_id=None, playlist_title=None, getter=None, ie=None):
925 self.url_result(self._proto_relative_url(getter(m) if getter else m), ie)
927 return self.playlist_result(
928 urls, playlist_id=playlist_id, playlist_title=playlist_title)
931 def playlist_result(entries, playlist_id=None, playlist_title=None, playlist_description=None):
932 """Returns a playlist"""
933 video_info = {'_type': 'playlist',
936 video_info['id'] = playlist_id
938 video_info['title'] = playlist_title
939 if playlist_description:
940 video_info['description'] = playlist_description
943 def _search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
945 Perform a regex search on the given string, using a single or a list of
946 patterns returning the first matching group.
947 In case of failure return a default value or raise a WARNING or a
948 RegexNotFoundError, depending on fatal, specifying the field name.
950 if isinstance(pattern, (str, compat_str, compiled_regex_type)):
951 mobj = re.search(pattern, string, flags)
954 mobj = re.search(p, string, flags)
958 if not self._downloader.params.get('no_color') and compat_os_name != 'nt' and sys.stderr.isatty():
959 _name = '\033[0;34m%s\033[0m' % name
965 # return the first matching group
966 return next(g for g in mobj.groups() if g is not None)
968 return mobj.group(group)
969 elif default is not NO_DEFAULT:
972 raise RegexNotFoundError('Unable to extract %s' % _name)
974 self._downloader.report_warning('unable to extract %s' % _name + bug_reports_message())
977 def _html_search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
979 Like _search_regex, but strips HTML tags and unescapes entities.
981 res = self._search_regex(pattern, string, name, default, fatal, flags, group)
983 return clean_html(res).strip()
987 def _get_netrc_login_info(self, netrc_machine=None):
990 netrc_machine = netrc_machine or self._NETRC_MACHINE
992 if self._downloader.params.get('usenetrc', False):
994 info = netrc.netrc().authenticators(netrc_machine)
999 raise netrc.NetrcParseError(
1000 'No authenticators for %s' % netrc_machine)
1001 except (IOError, netrc.NetrcParseError) as err:
1002 self._downloader.report_warning(
1003 'parsing .netrc: %s' % error_to_compat_str(err))
1005 return username, password
1007 def _get_login_info(self, username_option='username', password_option='password', netrc_machine=None):
1009 Get the login info as (username, password)
1010 First look for the manually specified credentials using username_option
1011 and password_option as keys in params dictionary. If no such credentials
1012 available look in the netrc file using the netrc_machine or _NETRC_MACHINE
1014 If there's no info available, return (None, None)
1016 if self._downloader is None:
1019 downloader_params = self._downloader.params
1021 # Attempt to use provided username and password or .netrc data
1022 if downloader_params.get(username_option) is not None:
1023 username = downloader_params[username_option]
1024 password = downloader_params[password_option]
1026 username, password = self._get_netrc_login_info(netrc_machine)
1028 return username, password
1030 def _get_tfa_info(self, note='two-factor verification code'):
1032 Get the two-factor authentication info
1033 TODO - asking the user will be required for sms/phone verify
1034 currently just uses the command line option
1035 If there's no info available, return None
1037 if self._downloader is None:
1039 downloader_params = self._downloader.params
1041 if downloader_params.get('twofactor') is not None:
1042 return downloader_params['twofactor']
1044 return compat_getpass('Type %s and press [Return]: ' % note)
1046 # Helper functions for extracting OpenGraph info
1048 def _og_regexes(prop):
1049 content_re = r'content=(?:"([^"]+?)"|\'([^\']+?)\'|\s*([^\s"\'=<>`]+?))'
1050 property_re = (r'(?:name|property)=(?:\'og:%(prop)s\'|"og:%(prop)s"|\s*og:%(prop)s\b)'
1051 % {'prop': re.escape(prop)})
1052 template = r'<meta[^>]+?%s[^>]+?%s'
1054 template % (property_re, content_re),
1055 template % (content_re, property_re),
1059 def _meta_regex(prop):
1060 return r'''(?isx)<meta
1061 (?=[^>]+(?:itemprop|name|property|id|http-equiv)=(["\']?)%s\1)
1062 [^>]+?content=(["\'])(?P<content>.*?)\2''' % re.escape(prop)
1064 def _og_search_property(self, prop, html, name=None, **kargs):
1065 if not isinstance(prop, (list, tuple)):
1068 name = 'OpenGraph %s' % prop[0]
1071 og_regexes.extend(self._og_regexes(p))
1072 escaped = self._search_regex(og_regexes, html, name, flags=re.DOTALL, **kargs)
1075 return unescapeHTML(escaped)
1077 def _og_search_thumbnail(self, html, **kargs):
1078 return self._og_search_property('image', html, 'thumbnail URL', fatal=False, **kargs)
1080 def _og_search_description(self, html, **kargs):
1081 return self._og_search_property('description', html, fatal=False, **kargs)
1083 def _og_search_title(self, html, **kargs):
1084 return self._og_search_property('title', html, **kargs)
1086 def _og_search_video_url(self, html, name='video url', secure=True, **kargs):
1087 regexes = self._og_regexes('video') + self._og_regexes('video:url')
1089 regexes = self._og_regexes('video:secure_url') + regexes
1090 return self._html_search_regex(regexes, html, name, **kargs)
1092 def _og_search_url(self, html, **kargs):
1093 return self._og_search_property('url', html, **kargs)
1095 def _html_search_meta(self, name, html, display_name=None, fatal=False, **kwargs):
1096 if not isinstance(name, (list, tuple)):
1098 if display_name is None:
1099 display_name = name[0]
1100 return self._html_search_regex(
1101 [self._meta_regex(n) for n in name],
1102 html, display_name, fatal=fatal, group='content', **kwargs)
1104 def _dc_search_uploader(self, html):
1105 return self._html_search_meta('dc.creator', html, 'uploader')
1107 def _rta_search(self, html):
1108 # See http://www.rtalabel.org/index.php?content=howtofaq#single
1109 if re.search(r'(?ix)<meta\s+name="rating"\s+'
1110 r' content="RTA-5042-1996-1400-1577-RTA"',
1115 def _media_rating_search(self, html):
1116 # See http://www.tjg-designs.com/WP/metadata-code-examples-adding-metadata-to-your-web-pages/
1117 rating = self._html_search_meta('rating', html)
1129 return RATING_TABLE.get(rating.lower())
1131 def _family_friendly_search(self, html):
1132 # See http://schema.org/VideoObject
1133 family_friendly = self._html_search_meta(
1134 'isFamilyFriendly', html, default=None)
1136 if not family_friendly:
1145 return RATING_TABLE.get(family_friendly.lower())
1147 def _twitter_search_player(self, html):
1148 return self._html_search_meta('twitter:player', html,
1149 'twitter card player')
1151 def _search_json_ld(self, html, video_id, expected_type=None, **kwargs):
1152 json_ld = self._search_regex(
1153 JSON_LD_RE, html, 'JSON-LD', group='json_ld', **kwargs)
1154 default = kwargs.get('default', NO_DEFAULT)
1156 return default if default is not NO_DEFAULT else {}
1157 # JSON-LD may be malformed and thus `fatal` should be respected.
1158 # At the same time `default` may be passed that assumes `fatal=False`
1159 # for _search_regex. Let's simulate the same behavior here as well.
1160 fatal = kwargs.get('fatal', True) if default == NO_DEFAULT else False
1161 return self._json_ld(json_ld, video_id, fatal=fatal, expected_type=expected_type)
1163 def _json_ld(self, json_ld, video_id, fatal=True, expected_type=None):
1164 if isinstance(json_ld, compat_str):
1165 json_ld = self._parse_json(json_ld, video_id, fatal=fatal)
1169 if not isinstance(json_ld, (list, tuple, dict)):
1171 if isinstance(json_ld, dict):
1174 INTERACTION_TYPE_MAP = {
1175 'CommentAction': 'comment',
1176 'AgreeAction': 'like',
1177 'DisagreeAction': 'dislike',
1178 'LikeAction': 'like',
1179 'DislikeAction': 'dislike',
1180 'ListenAction': 'view',
1181 'WatchAction': 'view',
1182 'ViewAction': 'view',
1185 def extract_interaction_statistic(e):
1186 interaction_statistic = e.get('interactionStatistic')
1187 if not isinstance(interaction_statistic, list):
1189 for is_e in interaction_statistic:
1190 if not isinstance(is_e, dict):
1192 if is_e.get('@type') != 'InteractionCounter':
1194 interaction_type = is_e.get('interactionType')
1195 if not isinstance(interaction_type, compat_str):
1197 interaction_count = int_or_none(is_e.get('userInteractionCount'))
1198 if interaction_count is None:
1200 count_kind = INTERACTION_TYPE_MAP.get(interaction_type.split('/')[-1])
1203 count_key = '%s_count' % count_kind
1204 if info.get(count_key) is not None:
1206 info[count_key] = interaction_count
1208 def extract_video_object(e):
1209 assert e['@type'] == 'VideoObject'
1211 'url': e.get('contentUrl'),
1212 'title': unescapeHTML(e.get('name')),
1213 'description': unescapeHTML(e.get('description')),
1214 'thumbnail': e.get('thumbnailUrl') or e.get('thumbnailURL'),
1215 'duration': parse_duration(e.get('duration')),
1216 'timestamp': unified_timestamp(e.get('uploadDate')),
1217 'filesize': float_or_none(e.get('contentSize')),
1218 'tbr': int_or_none(e.get('bitrate')),
1219 'width': int_or_none(e.get('width')),
1220 'height': int_or_none(e.get('height')),
1221 'view_count': int_or_none(e.get('interactionCount')),
1223 extract_interaction_statistic(e)
1226 if isinstance(e.get('@context'), compat_str) and re.match(r'^https?://schema.org/?$', e.get('@context')):
1227 item_type = e.get('@type')
1228 if expected_type is not None and expected_type != item_type:
1230 if item_type in ('TVEpisode', 'Episode'):
1232 'episode': unescapeHTML(e.get('name')),
1233 'episode_number': int_or_none(e.get('episodeNumber')),
1234 'description': unescapeHTML(e.get('description')),
1236 part_of_season = e.get('partOfSeason')
1237 if isinstance(part_of_season, dict) and part_of_season.get('@type') in ('TVSeason', 'Season', 'CreativeWorkSeason'):
1238 info['season_number'] = int_or_none(part_of_season.get('seasonNumber'))
1239 part_of_series = e.get('partOfSeries') or e.get('partOfTVSeries')
1240 if isinstance(part_of_series, dict) and part_of_series.get('@type') in ('TVSeries', 'Series', 'CreativeWorkSeries'):
1241 info['series'] = unescapeHTML(part_of_series.get('name'))
1242 elif item_type in ('Article', 'NewsArticle'):
1244 'timestamp': parse_iso8601(e.get('datePublished')),
1245 'title': unescapeHTML(e.get('headline')),
1246 'description': unescapeHTML(e.get('articleBody')),
1248 elif item_type == 'VideoObject':
1249 extract_video_object(e)
1251 video = e.get('video')
1252 if isinstance(video, dict) and video.get('@type') == 'VideoObject':
1253 extract_video_object(video)
1255 return dict((k, v) for k, v in info.items() if v is not None)
1258 def _hidden_inputs(html):
1259 html = re.sub(r'<!--(?:(?!<!--).)*-->', '', html)
1261 for input in re.findall(r'(?i)(<input[^>]+>)', html):
1262 attrs = extract_attributes(input)
1265 if attrs.get('type') not in ('hidden', 'submit'):
1267 name = attrs.get('name') or attrs.get('id')
1268 value = attrs.get('value')
1269 if name and value is not None:
1270 hidden_inputs[name] = value
1271 return hidden_inputs
1273 def _form_hidden_inputs(self, form_id, html):
1274 form = self._search_regex(
1275 r'(?is)<form[^>]+?id=(["\'])%s\1[^>]*>(?P<form>.+?)</form>' % form_id,
1276 html, '%s form' % form_id, group='form')
1277 return self._hidden_inputs(form)
1279 def _sort_formats(self, formats, field_preference=None):
1281 raise ExtractorError('No video formats found')
1284 # Automatically determine tbr when missing based on abr and vbr (improves
1285 # formats sorting in some cases)
1286 if 'tbr' not in f and f.get('abr') is not None and f.get('vbr') is not None:
1287 f['tbr'] = f['abr'] + f['vbr']
1289 def _formats_key(f):
1290 # TODO remove the following workaround
1291 from ..utils import determine_ext
1292 if not f.get('ext') and 'url' in f:
1293 f['ext'] = determine_ext(f['url'])
1295 if isinstance(field_preference, (list, tuple)):
1298 if f.get(field) is not None
1299 else ('' if field == 'format_id' else -1)
1300 for field in field_preference)
1302 preference = f.get('preference')
1303 if preference is None:
1305 if f.get('ext') in ['f4f', 'f4m']: # Not yet supported
1308 protocol = f.get('protocol') or determine_protocol(f)
1309 proto_preference = 0 if protocol in ['http', 'https'] else (-0.5 if protocol == 'rtsp' else -0.1)
1311 if f.get('vcodec') == 'none': # audio only
1313 if self._downloader.params.get('prefer_free_formats'):
1314 ORDER = ['aac', 'mp3', 'm4a', 'webm', 'ogg', 'opus']
1316 ORDER = ['webm', 'opus', 'ogg', 'mp3', 'aac', 'm4a']
1319 audio_ext_preference = ORDER.index(f['ext'])
1321 audio_ext_preference = -1
1323 if f.get('acodec') == 'none': # video only
1325 if self._downloader.params.get('prefer_free_formats'):
1326 ORDER = ['flv', 'mp4', 'webm']
1328 ORDER = ['webm', 'flv', 'mp4']
1330 ext_preference = ORDER.index(f['ext'])
1333 audio_ext_preference = 0
1337 f.get('language_preference') if f.get('language_preference') is not None else -1,
1338 f.get('quality') if f.get('quality') is not None else -1,
1339 f.get('tbr') if f.get('tbr') is not None else -1,
1340 f.get('filesize') if f.get('filesize') is not None else -1,
1341 f.get('vbr') if f.get('vbr') is not None else -1,
1342 f.get('height') if f.get('height') is not None else -1,
1343 f.get('width') if f.get('width') is not None else -1,
1346 f.get('abr') if f.get('abr') is not None else -1,
1347 audio_ext_preference,
1348 f.get('fps') if f.get('fps') is not None else -1,
1349 f.get('filesize_approx') if f.get('filesize_approx') is not None else -1,
1350 f.get('source_preference') if f.get('source_preference') is not None else -1,
1351 f.get('format_id') if f.get('format_id') is not None else '',
1353 formats.sort(key=_formats_key)
1355 def _check_formats(self, formats, video_id):
1357 formats[:] = filter(
1358 lambda f: self._is_valid_url(
1360 item='%s video format' % f.get('format_id') if f.get('format_id') else 'video'),
1364 def _remove_duplicate_formats(formats):
1368 if f['url'] not in format_urls:
1369 format_urls.add(f['url'])
1370 unique_formats.append(f)
1371 formats[:] = unique_formats
1373 def _is_valid_url(self, url, video_id, item='video', headers={}):
1374 url = self._proto_relative_url(url, scheme='http:')
1375 # For now assume non HTTP(S) URLs always valid
1376 if not (url.startswith('http://') or url.startswith('https://')):
1379 self._request_webpage(url, video_id, 'Checking %s URL' % item, headers=headers)
1381 except ExtractorError as e:
1382 if isinstance(e.cause, compat_urllib_error.URLError):
1384 '%s: %s URL is invalid, skipping' % (video_id, item))
1388 def http_scheme(self):
1389 """ Either "http:" or "https:", depending on the user's preferences """
1392 if self._downloader.params.get('prefer_insecure', False)
1395 def _proto_relative_url(self, url, scheme=None):
1398 if url.startswith('//'):
1400 scheme = self.http_scheme()
1405 def _sleep(self, timeout, video_id, msg_template=None):
1406 if msg_template is None:
1407 msg_template = '%(video_id)s: Waiting for %(timeout)s seconds'
1408 msg = msg_template % {'video_id': video_id, 'timeout': timeout}
1412 def _extract_f4m_formats(self, manifest_url, video_id, preference=None, f4m_id=None,
1413 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1414 fatal=True, m3u8_id=None):
1415 manifest = self._download_xml(
1416 manifest_url, video_id, 'Downloading f4m manifest',
1417 'Unable to download f4m manifest',
1418 # Some manifests may be malformed, e.g. prosiebensat1 generated manifests
1419 # (see https://github.com/rg3/youtube-dl/issues/6215#issuecomment-121704244)
1420 transform_source=transform_source,
1423 if manifest is False:
1426 return self._parse_f4m_formats(
1427 manifest, manifest_url, video_id, preference=preference, f4m_id=f4m_id,
1428 transform_source=transform_source, fatal=fatal, m3u8_id=m3u8_id)
1430 def _parse_f4m_formats(self, manifest, manifest_url, video_id, preference=None, f4m_id=None,
1431 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1432 fatal=True, m3u8_id=None):
1433 # currently youtube-dl cannot decode the playerVerificationChallenge as Akamai uses Adobe Alchemy
1434 akamai_pv = manifest.find('{http://ns.adobe.com/f4m/1.0}pv-2.0')
1435 if akamai_pv is not None and ';' in akamai_pv.text:
1436 playerVerificationChallenge = akamai_pv.text.split(';')[0]
1437 if playerVerificationChallenge.strip() != '':
1441 manifest_version = '1.0'
1442 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/1.0}media')
1444 manifest_version = '2.0'
1445 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/2.0}media')
1446 # Remove unsupported DRM protected media from final formats
1447 # rendition (see https://github.com/rg3/youtube-dl/issues/8573).
1448 media_nodes = remove_encrypted_media(media_nodes)
1452 manifest_base_url = get_base_url(manifest)
1454 bootstrap_info = xpath_element(
1455 manifest, ['{http://ns.adobe.com/f4m/1.0}bootstrapInfo', '{http://ns.adobe.com/f4m/2.0}bootstrapInfo'],
1456 'bootstrap info', default=None)
1459 mime_type = xpath_text(
1460 manifest, ['{http://ns.adobe.com/f4m/1.0}mimeType', '{http://ns.adobe.com/f4m/2.0}mimeType'],
1461 'base URL', default=None)
1462 if mime_type and mime_type.startswith('audio/'):
1465 for i, media_el in enumerate(media_nodes):
1466 tbr = int_or_none(media_el.attrib.get('bitrate'))
1467 width = int_or_none(media_el.attrib.get('width'))
1468 height = int_or_none(media_el.attrib.get('height'))
1469 format_id = '-'.join(filter(None, [f4m_id, compat_str(i if tbr is None else tbr)]))
1470 # If <bootstrapInfo> is present, the specified f4m is a
1471 # stream-level manifest, and only set-level manifests may refer to
1472 # external resources. See section 11.4 and section 4 of F4M spec
1473 if bootstrap_info is None:
1475 # @href is introduced in 2.0, see section 11.6 of F4M spec
1476 if manifest_version == '2.0':
1477 media_url = media_el.attrib.get('href')
1478 if media_url is None:
1479 media_url = media_el.attrib.get('url')
1483 media_url if media_url.startswith('http://') or media_url.startswith('https://')
1484 else ((manifest_base_url or '/'.join(manifest_url.split('/')[:-1])) + '/' + media_url))
1485 # If media_url is itself a f4m manifest do the recursive extraction
1486 # since bitrates in parent manifest (this one) and media_url manifest
1487 # may differ leading to inability to resolve the format by requested
1488 # bitrate in f4m downloader
1489 ext = determine_ext(manifest_url)
1491 f4m_formats = self._extract_f4m_formats(
1492 manifest_url, video_id, preference=preference, f4m_id=f4m_id,
1493 transform_source=transform_source, fatal=fatal)
1494 # Sometimes stream-level manifest contains single media entry that
1495 # does not contain any quality metadata (e.g. http://matchtv.ru/#live-player).
1496 # At the same time parent's media entry in set-level manifest may
1497 # contain it. We will copy it from parent in such cases.
1498 if len(f4m_formats) == 1:
1501 'tbr': f.get('tbr') or tbr,
1502 'width': f.get('width') or width,
1503 'height': f.get('height') or height,
1504 'format_id': f.get('format_id') if not tbr else format_id,
1507 formats.extend(f4m_formats)
1510 formats.extend(self._extract_m3u8_formats(
1511 manifest_url, video_id, 'mp4', preference=preference,
1512 m3u8_id=m3u8_id, fatal=fatal))
1515 'format_id': format_id,
1516 'url': manifest_url,
1517 'manifest_url': manifest_url,
1518 'ext': 'flv' if bootstrap_info is not None else None,
1524 'preference': preference,
1528 def _m3u8_meta_format(self, m3u8_url, ext=None, preference=None, m3u8_id=None):
1530 'format_id': '-'.join(filter(None, [m3u8_id, 'meta'])),
1534 'preference': preference - 100 if preference else -100,
1535 'resolution': 'multiple',
1536 'format_note': 'Quality selection URL',
1539 def _extract_m3u8_formats(self, m3u8_url, video_id, ext=None,
1540 entry_protocol='m3u8', preference=None,
1541 m3u8_id=None, note=None, errnote=None,
1542 fatal=True, live=False):
1543 res = self._download_webpage_handle(
1545 note=note or 'Downloading m3u8 information',
1546 errnote=errnote or 'Failed to download m3u8 information',
1552 m3u8_doc, urlh = res
1553 m3u8_url = urlh.geturl()
1555 return self._parse_m3u8_formats(
1556 m3u8_doc, m3u8_url, ext=ext, entry_protocol=entry_protocol,
1557 preference=preference, m3u8_id=m3u8_id, live=live)
1559 def _parse_m3u8_formats(self, m3u8_doc, m3u8_url, ext=None,
1560 entry_protocol='m3u8', preference=None,
1561 m3u8_id=None, live=False):
1562 if '#EXT-X-FAXS-CM:' in m3u8_doc: # Adobe Flash Access
1565 if re.search(r'#EXT-X-SESSION-KEY:.*?URI="skd://', m3u8_doc): # Apple FairPlay
1570 format_url = lambda u: (
1572 if re.match(r'^https?://', u)
1573 else compat_urlparse.urljoin(m3u8_url, u))
1576 # 1. https://tools.ietf.org/html/draft-pantos-http-live-streaming-21
1577 # 2. https://github.com/rg3/youtube-dl/issues/12211
1579 # We should try extracting formats only from master playlists [1, 4.3.4],
1580 # i.e. playlists that describe available qualities. On the other hand
1581 # media playlists [1, 4.3.3] should be returned as is since they contain
1582 # just the media without qualities renditions.
1583 # Fortunately, master playlist can be easily distinguished from media
1584 # playlist based on particular tags availability. As of [1, 4.3.3, 4.3.4]
1585 # master playlist tags MUST NOT appear in a media playist and vice versa.
1586 # As of [1, 4.3.3.1] #EXT-X-TARGETDURATION tag is REQUIRED for every
1587 # media playlist and MUST NOT appear in master playlist thus we can
1588 # clearly detect media playlist with this criterion.
1590 if '#EXT-X-TARGETDURATION' in m3u8_doc: # media playlist, return as is
1593 'format_id': m3u8_id,
1595 'protocol': entry_protocol,
1596 'preference': preference,
1600 last_stream_inf = {}
1602 def extract_media(x_media_line):
1603 media = parse_m3u8_attributes(x_media_line)
1604 # As per [1, 4.3.4.1] TYPE, GROUP-ID and NAME are REQUIRED
1605 media_type, group_id, name = media.get('TYPE'), media.get('GROUP-ID'), media.get('NAME')
1606 if not (media_type and group_id and name):
1608 groups.setdefault(group_id, []).append(media)
1609 if media_type not in ('VIDEO', 'AUDIO'):
1611 media_url = media.get('URI')
1614 for v in (m3u8_id, group_id, name):
1618 'format_id': '-'.join(format_id),
1619 'url': format_url(media_url),
1620 'manifest_url': m3u8_url,
1621 'language': media.get('LANGUAGE'),
1623 'protocol': entry_protocol,
1624 'preference': preference,
1626 if media_type == 'AUDIO':
1627 f['vcodec'] = 'none'
1630 def build_stream_name():
1631 # Despite specification does not mention NAME attribute for
1632 # EXT-X-STREAM-INF tag it still sometimes may be present (see [1]
1633 # or vidio test in TestInfoExtractor.test_parse_m3u8_formats)
1634 # 1. http://www.vidio.com/watch/165683-dj_ambred-booyah-live-2015
1635 stream_name = last_stream_inf.get('NAME')
1638 # If there is no NAME in EXT-X-STREAM-INF it will be obtained
1639 # from corresponding rendition group
1640 stream_group_id = last_stream_inf.get('VIDEO')
1641 if not stream_group_id:
1643 stream_group = groups.get(stream_group_id)
1644 if not stream_group:
1645 return stream_group_id
1646 rendition = stream_group[0]
1647 return rendition.get('NAME') or stream_group_id
1649 for line in m3u8_doc.splitlines():
1650 if line.startswith('#EXT-X-STREAM-INF:'):
1651 last_stream_inf = parse_m3u8_attributes(line)
1652 elif line.startswith('#EXT-X-MEDIA:'):
1654 elif line.startswith('#') or not line.strip():
1657 tbr = float_or_none(
1658 last_stream_inf.get('AVERAGE-BANDWIDTH') or
1659 last_stream_inf.get('BANDWIDTH'), scale=1000)
1662 format_id.append(m3u8_id)
1663 stream_name = build_stream_name()
1664 # Bandwidth of live streams may differ over time thus making
1665 # format_id unpredictable. So it's better to keep provided
1668 format_id.append(stream_name if stream_name else '%d' % (tbr if tbr else len(formats)))
1669 manifest_url = format_url(line.strip())
1671 'format_id': '-'.join(format_id),
1672 'url': manifest_url,
1673 'manifest_url': m3u8_url,
1676 'fps': float_or_none(last_stream_inf.get('FRAME-RATE')),
1677 'protocol': entry_protocol,
1678 'preference': preference,
1680 resolution = last_stream_inf.get('RESOLUTION')
1682 mobj = re.search(r'(?P<width>\d+)[xX](?P<height>\d+)', resolution)
1684 f['width'] = int(mobj.group('width'))
1685 f['height'] = int(mobj.group('height'))
1686 # Unified Streaming Platform
1688 r'audio.*?(?:%3D|=)(\d+)(?:-video.*?(?:%3D|=)(\d+))?', f['url'])
1690 abr, vbr = mobj.groups()
1691 abr, vbr = float_or_none(abr, 1000), float_or_none(vbr, 1000)
1696 codecs = parse_codecs(last_stream_inf.get('CODECS'))
1698 audio_group_id = last_stream_inf.get('AUDIO')
1699 # As per [1, 4.3.4.1.1] any EXT-X-STREAM-INF tag which
1700 # references a rendition group MUST have a CODECS attribute.
1701 # However, this is not always respected, for example, [2]
1702 # contains EXT-X-STREAM-INF tag which references AUDIO
1703 # rendition group but does not have CODECS and despite
1704 # referencing audio group an audio group, it represents
1705 # a complete (with audio and video) format. So, for such cases
1706 # we will ignore references to rendition groups and treat them
1707 # as complete formats.
1708 if audio_group_id and codecs and f.get('vcodec') != 'none':
1709 audio_group = groups.get(audio_group_id)
1710 if audio_group and audio_group[0].get('URI'):
1711 # TODO: update acodec for audio only formats with
1713 f['acodec'] = 'none'
1715 last_stream_inf = {}
1719 def _xpath_ns(path, namespace=None):
1723 for c in path.split('/'):
1724 if not c or c == '.':
1727 out.append('{%s}%s' % (namespace, c))
1728 return '/'.join(out)
1730 def _extract_smil_formats(self, smil_url, video_id, fatal=True, f4m_params=None, transform_source=None):
1731 smil = self._download_smil(smil_url, video_id, fatal=fatal, transform_source=transform_source)
1737 namespace = self._parse_smil_namespace(smil)
1739 return self._parse_smil_formats(
1740 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
1742 def _extract_smil_info(self, smil_url, video_id, fatal=True, f4m_params=None):
1743 smil = self._download_smil(smil_url, video_id, fatal=fatal)
1746 return self._parse_smil(smil, smil_url, video_id, f4m_params=f4m_params)
1748 def _download_smil(self, smil_url, video_id, fatal=True, transform_source=None):
1749 return self._download_xml(
1750 smil_url, video_id, 'Downloading SMIL file',
1751 'Unable to download SMIL file', fatal=fatal, transform_source=transform_source)
1753 def _parse_smil(self, smil, smil_url, video_id, f4m_params=None):
1754 namespace = self._parse_smil_namespace(smil)
1756 formats = self._parse_smil_formats(
1757 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
1758 subtitles = self._parse_smil_subtitles(smil, namespace=namespace)
1760 video_id = os.path.splitext(url_basename(smil_url))[0]
1764 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
1765 name = meta.attrib.get('name')
1766 content = meta.attrib.get('content')
1767 if not name or not content:
1769 if not title and name == 'title':
1771 elif not description and name in ('description', 'abstract'):
1772 description = content
1773 elif not upload_date and name == 'date':
1774 upload_date = unified_strdate(content)
1777 'id': image.get('type'),
1778 'url': image.get('src'),
1779 'width': int_or_none(image.get('width')),
1780 'height': int_or_none(image.get('height')),
1781 } for image in smil.findall(self._xpath_ns('.//image', namespace)) if image.get('src')]
1785 'title': title or video_id,
1786 'description': description,
1787 'upload_date': upload_date,
1788 'thumbnails': thumbnails,
1790 'subtitles': subtitles,
1793 def _parse_smil_namespace(self, smil):
1794 return self._search_regex(
1795 r'(?i)^{([^}]+)?}smil$', smil.tag, 'namespace', default=None)
1797 def _parse_smil_formats(self, smil, smil_url, video_id, namespace=None, f4m_params=None, transform_rtmp_url=None):
1799 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
1800 b = meta.get('base') or meta.get('httpBase')
1811 media = smil.findall(self._xpath_ns('.//video', namespace)) + smil.findall(self._xpath_ns('.//audio', namespace))
1812 for medium in media:
1813 src = medium.get('src')
1814 if not src or src in srcs:
1818 bitrate = float_or_none(medium.get('system-bitrate') or medium.get('systemBitrate'), 1000)
1819 filesize = int_or_none(medium.get('size') or medium.get('fileSize'))
1820 width = int_or_none(medium.get('width'))
1821 height = int_or_none(medium.get('height'))
1822 proto = medium.get('proto')
1823 ext = medium.get('ext')
1824 src_ext = determine_ext(src)
1825 streamer = medium.get('streamer') or base
1827 if proto == 'rtmp' or streamer.startswith('rtmp'):
1833 'format_id': 'rtmp-%d' % (rtmp_count if bitrate is None else bitrate),
1835 'filesize': filesize,
1839 if transform_rtmp_url:
1840 streamer, src = transform_rtmp_url(streamer, src)
1841 formats[-1].update({
1847 src_url = src if src.startswith('http') else compat_urlparse.urljoin(base, src)
1848 src_url = src_url.strip()
1850 if proto == 'm3u8' or src_ext == 'm3u8':
1851 m3u8_formats = self._extract_m3u8_formats(
1852 src_url, video_id, ext or 'mp4', m3u8_id='hls', fatal=False)
1853 if len(m3u8_formats) == 1:
1855 m3u8_formats[0].update({
1856 'format_id': 'hls-%d' % (m3u8_count if bitrate is None else bitrate),
1861 formats.extend(m3u8_formats)
1864 if src_ext == 'f4m':
1869 'plugin': 'flowplayer-3.2.0.1',
1871 f4m_url += '&' if '?' in f4m_url else '?'
1872 f4m_url += compat_urllib_parse_urlencode(f4m_params)
1873 formats.extend(self._extract_f4m_formats(f4m_url, video_id, f4m_id='hds', fatal=False))
1876 if src_url.startswith('http') and self._is_valid_url(src, video_id):
1880 'ext': ext or src_ext or 'flv',
1881 'format_id': 'http-%d' % (bitrate or http_count),
1883 'filesize': filesize,
1891 def _parse_smil_subtitles(self, smil, namespace=None, subtitles_lang='en'):
1894 for num, textstream in enumerate(smil.findall(self._xpath_ns('.//textstream', namespace))):
1895 src = textstream.get('src')
1896 if not src or src in urls:
1899 ext = textstream.get('ext') or mimetype2ext(textstream.get('type')) or determine_ext(src)
1900 lang = textstream.get('systemLanguage') or textstream.get('systemLanguageName') or textstream.get('lang') or subtitles_lang
1901 subtitles.setdefault(lang, []).append({
1907 def _extract_xspf_playlist(self, xspf_url, playlist_id, fatal=True):
1908 xspf = self._download_xml(
1909 xspf_url, playlist_id, 'Downloading xpsf playlist',
1910 'Unable to download xspf manifest', fatal=fatal)
1913 return self._parse_xspf(
1914 xspf, playlist_id, xspf_url=xspf_url,
1915 xspf_base_url=base_url(xspf_url))
1917 def _parse_xspf(self, xspf_doc, playlist_id, xspf_url=None, xspf_base_url=None):
1919 'xspf': 'http://xspf.org/ns/0/',
1920 's1': 'http://static.streamone.nl/player/ns/0',
1924 for track in xspf_doc.findall(xpath_with_ns('./xspf:trackList/xspf:track', NS_MAP)):
1926 track, xpath_with_ns('./xspf:title', NS_MAP), 'title', default=playlist_id)
1927 description = xpath_text(
1928 track, xpath_with_ns('./xspf:annotation', NS_MAP), 'description')
1929 thumbnail = xpath_text(
1930 track, xpath_with_ns('./xspf:image', NS_MAP), 'thumbnail')
1931 duration = float_or_none(
1932 xpath_text(track, xpath_with_ns('./xspf:duration', NS_MAP), 'duration'), 1000)
1935 for location in track.findall(xpath_with_ns('./xspf:location', NS_MAP)):
1936 format_url = urljoin(xspf_base_url, location.text)
1941 'manifest_url': xspf_url,
1942 'format_id': location.get(xpath_with_ns('s1:label', NS_MAP)),
1943 'width': int_or_none(location.get(xpath_with_ns('s1:width', NS_MAP))),
1944 'height': int_or_none(location.get(xpath_with_ns('s1:height', NS_MAP))),
1946 self._sort_formats(formats)
1951 'description': description,
1952 'thumbnail': thumbnail,
1953 'duration': duration,
1958 def _extract_mpd_formats(self, mpd_url, video_id, mpd_id=None, note=None, errnote=None, fatal=True, formats_dict={}):
1959 res = self._download_xml_handle(
1961 note=note or 'Downloading MPD manifest',
1962 errnote=errnote or 'Failed to download MPD manifest',
1967 mpd_base_url = base_url(urlh.geturl())
1969 return self._parse_mpd_formats(
1970 mpd_doc, mpd_id=mpd_id, mpd_base_url=mpd_base_url,
1971 formats_dict=formats_dict, mpd_url=mpd_url)
1973 def _parse_mpd_formats(self, mpd_doc, mpd_id=None, mpd_base_url='', formats_dict={}, mpd_url=None):
1975 Parse formats from MPD manifest.
1977 1. MPEG-DASH Standard, ISO/IEC 23009-1:2014(E),
1978 http://standards.iso.org/ittf/PubliclyAvailableStandards/c065274_ISO_IEC_23009-1_2014.zip
1979 2. https://en.wikipedia.org/wiki/Dynamic_Adaptive_Streaming_over_HTTP
1981 if mpd_doc.get('type') == 'dynamic':
1984 namespace = self._search_regex(r'(?i)^{([^}]+)?}MPD$', mpd_doc.tag, 'namespace', default=None)
1987 return self._xpath_ns(path, namespace)
1989 def is_drm_protected(element):
1990 return element.find(_add_ns('ContentProtection')) is not None
1992 def extract_multisegment_info(element, ms_parent_info):
1993 ms_info = ms_parent_info.copy()
1995 # As per [1, 5.3.9.2.2] SegmentList and SegmentTemplate share some
1996 # common attributes and elements. We will only extract relevant
1998 def extract_common(source):
1999 segment_timeline = source.find(_add_ns('SegmentTimeline'))
2000 if segment_timeline is not None:
2001 s_e = segment_timeline.findall(_add_ns('S'))
2003 ms_info['total_number'] = 0
2006 r = int(s.get('r', 0))
2007 ms_info['total_number'] += 1 + r
2008 ms_info['s'].append({
2009 't': int(s.get('t', 0)),
2010 # @d is mandatory (see [1, 5.3.9.6.2, Table 17, page 60])
2011 'd': int(s.attrib['d']),
2014 start_number = source.get('startNumber')
2016 ms_info['start_number'] = int(start_number)
2017 timescale = source.get('timescale')
2019 ms_info['timescale'] = int(timescale)
2020 segment_duration = source.get('duration')
2021 if segment_duration:
2022 ms_info['segment_duration'] = float(segment_duration)
2024 def extract_Initialization(source):
2025 initialization = source.find(_add_ns('Initialization'))
2026 if initialization is not None:
2027 ms_info['initialization_url'] = initialization.attrib['sourceURL']
2029 segment_list = element.find(_add_ns('SegmentList'))
2030 if segment_list is not None:
2031 extract_common(segment_list)
2032 extract_Initialization(segment_list)
2033 segment_urls_e = segment_list.findall(_add_ns('SegmentURL'))
2035 ms_info['segment_urls'] = [segment.attrib['media'] for segment in segment_urls_e]
2037 segment_template = element.find(_add_ns('SegmentTemplate'))
2038 if segment_template is not None:
2039 extract_common(segment_template)
2040 media = segment_template.get('media')
2042 ms_info['media'] = media
2043 initialization = segment_template.get('initialization')
2045 ms_info['initialization'] = initialization
2047 extract_Initialization(segment_template)
2050 mpd_duration = parse_duration(mpd_doc.get('mediaPresentationDuration'))
2052 for period in mpd_doc.findall(_add_ns('Period')):
2053 period_duration = parse_duration(period.get('duration')) or mpd_duration
2054 period_ms_info = extract_multisegment_info(period, {
2058 for adaptation_set in period.findall(_add_ns('AdaptationSet')):
2059 if is_drm_protected(adaptation_set):
2061 adaption_set_ms_info = extract_multisegment_info(adaptation_set, period_ms_info)
2062 for representation in adaptation_set.findall(_add_ns('Representation')):
2063 if is_drm_protected(representation):
2065 representation_attrib = adaptation_set.attrib.copy()
2066 representation_attrib.update(representation.attrib)
2067 # According to [1, 5.3.7.2, Table 9, page 41], @mimeType is mandatory
2068 mime_type = representation_attrib['mimeType']
2069 content_type = mime_type.split('/')[0]
2070 if content_type == 'text':
2071 # TODO implement WebVTT downloading
2073 elif content_type in ('video', 'audio'):
2075 for element in (representation, adaptation_set, period, mpd_doc):
2076 base_url_e = element.find(_add_ns('BaseURL'))
2077 if base_url_e is not None:
2078 base_url = base_url_e.text + base_url
2079 if re.match(r'^https?://', base_url):
2081 if mpd_base_url and not re.match(r'^https?://', base_url):
2082 if not mpd_base_url.endswith('/') and not base_url.startswith('/'):
2084 base_url = mpd_base_url + base_url
2085 representation_id = representation_attrib.get('id')
2086 lang = representation_attrib.get('lang')
2087 url_el = representation.find(_add_ns('BaseURL'))
2088 filesize = int_or_none(url_el.attrib.get('{http://youtube.com/yt/2012/10/10}contentLength') if url_el is not None else None)
2089 bandwidth = int_or_none(representation_attrib.get('bandwidth'))
2091 'format_id': '%s-%s' % (mpd_id, representation_id) if mpd_id else representation_id,
2093 'manifest_url': mpd_url,
2094 'ext': mimetype2ext(mime_type),
2095 'width': int_or_none(representation_attrib.get('width')),
2096 'height': int_or_none(representation_attrib.get('height')),
2097 'tbr': float_or_none(bandwidth, 1000),
2098 'asr': int_or_none(representation_attrib.get('audioSamplingRate')),
2099 'fps': int_or_none(representation_attrib.get('frameRate')),
2100 'language': lang if lang not in ('mul', 'und', 'zxx', 'mis') else None,
2101 'format_note': 'DASH %s' % content_type,
2102 'filesize': filesize,
2103 'container': mimetype2ext(mime_type) + '_dash',
2105 f.update(parse_codecs(representation_attrib.get('codecs')))
2106 representation_ms_info = extract_multisegment_info(representation, adaption_set_ms_info)
2108 def prepare_template(template_name, identifiers):
2109 tmpl = representation_ms_info[template_name]
2110 # First of, % characters outside $...$ templates
2111 # must be escaped by doubling for proper processing
2112 # by % operator string formatting used further (see
2113 # https://github.com/rg3/youtube-dl/issues/16867).
2119 in_template = not in_template
2120 elif c == '%' and not in_template:
2122 # Next, $...$ templates are translated to their
2123 # %(...) counterparts to be used with % operator
2124 t = t.replace('$RepresentationID$', representation_id)
2125 t = re.sub(r'\$(%s)\$' % '|'.join(identifiers), r'%(\1)d', t)
2126 t = re.sub(r'\$(%s)%%([^$]+)\$' % '|'.join(identifiers), r'%(\1)\2', t)
2127 t.replace('$$', '$')
2130 # @initialization is a regular template like @media one
2131 # so it should be handled just the same way (see
2132 # https://github.com/rg3/youtube-dl/issues/11605)
2133 if 'initialization' in representation_ms_info:
2134 initialization_template = prepare_template(
2136 # As per [1, 5.3.9.4.2, Table 15, page 54] $Number$ and
2137 # $Time$ shall not be included for @initialization thus
2138 # only $Bandwidth$ remains
2140 representation_ms_info['initialization_url'] = initialization_template % {
2141 'Bandwidth': bandwidth,
2144 def location_key(location):
2145 return 'url' if re.match(r'^https?://', location) else 'path'
2147 if 'segment_urls' not in representation_ms_info and 'media' in representation_ms_info:
2149 media_template = prepare_template('media', ('Number', 'Bandwidth', 'Time'))
2150 media_location_key = location_key(media_template)
2152 # As per [1, 5.3.9.4.4, Table 16, page 55] $Number$ and $Time$
2153 # can't be used at the same time
2154 if '%(Number' in media_template and 's' not in representation_ms_info:
2155 segment_duration = None
2156 if 'total_number' not in representation_ms_info and 'segment_duration' in representation_ms_info:
2157 segment_duration = float_or_none(representation_ms_info['segment_duration'], representation_ms_info['timescale'])
2158 representation_ms_info['total_number'] = int(math.ceil(float(period_duration) / segment_duration))
2159 representation_ms_info['fragments'] = [{
2160 media_location_key: media_template % {
2161 'Number': segment_number,
2162 'Bandwidth': bandwidth,
2164 'duration': segment_duration,
2165 } for segment_number in range(
2166 representation_ms_info['start_number'],
2167 representation_ms_info['total_number'] + representation_ms_info['start_number'])]
2169 # $Number*$ or $Time$ in media template with S list available
2170 # Example $Number*$: http://www.svtplay.se/klipp/9023742/stopptid-om-bjorn-borg
2171 # Example $Time$: https://play.arkena.com/embed/avp/v2/player/media/b41dda37-d8e7-4d3f-b1b5-9a9db578bdfe/1/129411
2172 representation_ms_info['fragments'] = []
2175 segment_number = representation_ms_info['start_number']
2177 def add_segment_url():
2178 segment_url = media_template % {
2179 'Time': segment_time,
2180 'Bandwidth': bandwidth,
2181 'Number': segment_number,
2183 representation_ms_info['fragments'].append({
2184 media_location_key: segment_url,
2185 'duration': float_or_none(segment_d, representation_ms_info['timescale']),
2188 for num, s in enumerate(representation_ms_info['s']):
2189 segment_time = s.get('t') or segment_time
2193 for r in range(s.get('r', 0)):
2194 segment_time += segment_d
2197 segment_time += segment_d
2198 elif 'segment_urls' in representation_ms_info and 's' in representation_ms_info:
2200 # Example: https://www.youtube.com/watch?v=iXZV5uAYMJI
2201 # or any YouTube dashsegments video
2204 timescale = representation_ms_info['timescale']
2205 for s in representation_ms_info['s']:
2206 duration = float_or_none(s['d'], timescale)
2207 for r in range(s.get('r', 0) + 1):
2208 segment_uri = representation_ms_info['segment_urls'][segment_index]
2210 location_key(segment_uri): segment_uri,
2211 'duration': duration,
2214 representation_ms_info['fragments'] = fragments
2215 elif 'segment_urls' in representation_ms_info:
2216 # Segment URLs with no SegmentTimeline
2217 # Example: https://www.seznam.cz/zpravy/clanek/cesko-zasahne-vitr-o-sile-vichrice-muze-byt-i-zivotu-nebezpecny-39091
2218 # https://github.com/rg3/youtube-dl/pull/14844
2220 segment_duration = float_or_none(
2221 representation_ms_info['segment_duration'],
2222 representation_ms_info['timescale']) if 'segment_duration' in representation_ms_info else None
2223 for segment_url in representation_ms_info['segment_urls']:
2225 location_key(segment_url): segment_url,
2227 if segment_duration:
2228 fragment['duration'] = segment_duration
2229 fragments.append(fragment)
2230 representation_ms_info['fragments'] = fragments
2231 # NB: MPD manifest may contain direct URLs to unfragmented media.
2232 # No fragments key is present in this case.
2233 if 'fragments' in representation_ms_info:
2235 'fragment_base_url': base_url,
2237 'protocol': 'http_dash_segments',
2239 if 'initialization_url' in representation_ms_info:
2240 initialization_url = representation_ms_info['initialization_url']
2241 if not f.get('url'):
2242 f['url'] = initialization_url
2243 f['fragments'].append({location_key(initialization_url): initialization_url})
2244 f['fragments'].extend(representation_ms_info['fragments'])
2245 # According to [1, 5.3.5.2, Table 7, page 35] @id of Representation
2246 # is not necessarily unique within a Period thus formats with
2247 # the same `format_id` are quite possible. There are numerous examples
2248 # of such manifests (see https://github.com/rg3/youtube-dl/issues/15111,
2249 # https://github.com/rg3/youtube-dl/issues/13919)
2250 full_info = formats_dict.get(representation_id, {}).copy()
2252 formats.append(full_info)
2254 self.report_warning('Unknown MIME type %s in DASH manifest' % mime_type)
2257 def _extract_ism_formats(self, ism_url, video_id, ism_id=None, note=None, errnote=None, fatal=True):
2258 res = self._download_xml_handle(
2260 note=note or 'Downloading ISM manifest',
2261 errnote=errnote or 'Failed to download ISM manifest',
2267 return self._parse_ism_formats(ism_doc, urlh.geturl(), ism_id)
2269 def _parse_ism_formats(self, ism_doc, ism_url, ism_id=None):
2271 Parse formats from ISM manifest.
2273 1. [MS-SSTR]: Smooth Streaming Protocol,
2274 https://msdn.microsoft.com/en-us/library/ff469518.aspx
2276 if ism_doc.get('IsLive') == 'TRUE' or ism_doc.find('Protection') is not None:
2279 duration = int(ism_doc.attrib['Duration'])
2280 timescale = int_or_none(ism_doc.get('TimeScale')) or 10000000
2283 for stream in ism_doc.findall('StreamIndex'):
2284 stream_type = stream.get('Type')
2285 if stream_type not in ('video', 'audio'):
2287 url_pattern = stream.attrib['Url']
2288 stream_timescale = int_or_none(stream.get('TimeScale')) or timescale
2289 stream_name = stream.get('Name')
2290 for track in stream.findall('QualityLevel'):
2291 fourcc = track.get('FourCC', 'AACL' if track.get('AudioTag') == '255' else None)
2292 # TODO: add support for WVC1 and WMAP
2293 if fourcc not in ('H264', 'AVC1', 'AACL'):
2294 self.report_warning('%s is not a supported codec' % fourcc)
2296 tbr = int(track.attrib['Bitrate']) // 1000
2297 # [1] does not mention Width and Height attributes. However,
2298 # they're often present while MaxWidth and MaxHeight are
2299 # missing, so should be used as fallbacks
2300 width = int_or_none(track.get('MaxWidth') or track.get('Width'))
2301 height = int_or_none(track.get('MaxHeight') or track.get('Height'))
2302 sampling_rate = int_or_none(track.get('SamplingRate'))
2304 track_url_pattern = re.sub(r'{[Bb]itrate}', track.attrib['Bitrate'], url_pattern)
2305 track_url_pattern = compat_urlparse.urljoin(ism_url, track_url_pattern)
2311 stream_fragments = stream.findall('c')
2312 for stream_fragment_index, stream_fragment in enumerate(stream_fragments):
2313 fragment_ctx['time'] = int_or_none(stream_fragment.get('t')) or fragment_ctx['time']
2314 fragment_repeat = int_or_none(stream_fragment.get('r')) or 1
2315 fragment_ctx['duration'] = int_or_none(stream_fragment.get('d'))
2316 if not fragment_ctx['duration']:
2318 next_fragment_time = int(stream_fragment[stream_fragment_index + 1].attrib['t'])
2320 next_fragment_time = duration
2321 fragment_ctx['duration'] = (next_fragment_time - fragment_ctx['time']) / fragment_repeat
2322 for _ in range(fragment_repeat):
2324 'url': re.sub(r'{start[ _]time}', compat_str(fragment_ctx['time']), track_url_pattern),
2325 'duration': fragment_ctx['duration'] / stream_timescale,
2327 fragment_ctx['time'] += fragment_ctx['duration']
2331 format_id.append(ism_id)
2333 format_id.append(stream_name)
2334 format_id.append(compat_str(tbr))
2337 'format_id': '-'.join(format_id),
2339 'manifest_url': ism_url,
2340 'ext': 'ismv' if stream_type == 'video' else 'isma',
2344 'asr': sampling_rate,
2345 'vcodec': 'none' if stream_type == 'audio' else fourcc,
2346 'acodec': 'none' if stream_type == 'video' else fourcc,
2348 'fragments': fragments,
2349 '_download_params': {
2350 'duration': duration,
2351 'timescale': stream_timescale,
2352 'width': width or 0,
2353 'height': height or 0,
2355 'codec_private_data': track.get('CodecPrivateData'),
2356 'sampling_rate': sampling_rate,
2357 'channels': int_or_none(track.get('Channels', 2)),
2358 'bits_per_sample': int_or_none(track.get('BitsPerSample', 16)),
2359 'nal_unit_length_field': int_or_none(track.get('NALUnitLengthField', 4)),
2364 def _parse_html5_media_entries(self, base_url, webpage, video_id, m3u8_id=None, m3u8_entry_protocol='m3u8', mpd_id=None, preference=None):
2365 def absolute_url(item_url):
2366 return urljoin(base_url, item_url)
2368 def parse_content_type(content_type):
2369 if not content_type:
2371 ctr = re.search(r'(?P<mimetype>[^/]+/[^;]+)(?:;\s*codecs="?(?P<codecs>[^"]+))?', content_type)
2373 mimetype, codecs = ctr.groups()
2374 f = parse_codecs(codecs)
2375 f['ext'] = mimetype2ext(mimetype)
2379 def _media_formats(src, cur_media_type, type_info={}):
2380 full_url = absolute_url(src)
2381 ext = type_info.get('ext') or determine_ext(full_url)
2383 is_plain_url = False
2384 formats = self._extract_m3u8_formats(
2385 full_url, video_id, ext='mp4',
2386 entry_protocol=m3u8_entry_protocol, m3u8_id=m3u8_id,
2387 preference=preference, fatal=False)
2389 is_plain_url = False
2390 formats = self._extract_mpd_formats(
2391 full_url, video_id, mpd_id=mpd_id, fatal=False)
2396 'vcodec': 'none' if cur_media_type == 'audio' else None,
2398 return is_plain_url, formats
2401 # amp-video and amp-audio are very similar to their HTML5 counterparts
2402 # so we wll include them right here (see
2403 # https://www.ampproject.org/docs/reference/components/amp-video)
2404 media_tags = [(media_tag, media_type, '')
2405 for media_tag, media_type
2406 in re.findall(r'(?s)(<(?:amp-)?(video|audio)[^>]*/>)', webpage)]
2407 media_tags.extend(re.findall(
2408 # We only allow video|audio followed by a whitespace or '>'.
2409 # Allowing more characters may end up in significant slow down (see
2410 # https://github.com/rg3/youtube-dl/issues/11979, example URL:
2411 # http://www.porntrex.com/maps/videositemap.xml).
2412 r'(?s)(<(?P<tag>(?:amp-)?(?:video|audio))(?:\s+[^>]*)?>)(.*?)</(?P=tag)>', webpage))
2413 for media_tag, media_type, media_content in media_tags:
2418 media_attributes = extract_attributes(media_tag)
2419 src = media_attributes.get('src')
2421 _, formats = _media_formats(src, media_type)
2422 media_info['formats'].extend(formats)
2423 media_info['thumbnail'] = absolute_url(media_attributes.get('poster'))
2425 for source_tag in re.findall(r'<source[^>]+>', media_content):
2426 source_attributes = extract_attributes(source_tag)
2427 src = source_attributes.get('src')
2430 f = parse_content_type(source_attributes.get('type'))
2431 is_plain_url, formats = _media_formats(src, media_type, f)
2433 # res attribute is not standard but seen several times
2436 'height': int_or_none(source_attributes.get('res')),
2437 'format_id': source_attributes.get('label'),
2439 f.update(formats[0])
2440 media_info['formats'].append(f)
2442 media_info['formats'].extend(formats)
2443 for track_tag in re.findall(r'<track[^>]+>', media_content):
2444 track_attributes = extract_attributes(track_tag)
2445 kind = track_attributes.get('kind')
2446 if not kind or kind in ('subtitles', 'captions'):
2447 src = track_attributes.get('src')
2450 lang = track_attributes.get('srclang') or track_attributes.get('lang') or track_attributes.get('label')
2451 media_info['subtitles'].setdefault(lang, []).append({
2452 'url': absolute_url(src),
2454 for f in media_info['formats']:
2455 f.setdefault('http_headers', {})['Referer'] = base_url
2456 if media_info['formats'] or media_info['subtitles']:
2457 entries.append(media_info)
2460 def _extract_akamai_formats(self, manifest_url, video_id, hosts={}):
2462 hdcore_sign = 'hdcore=3.7.0'
2463 f4m_url = re.sub(r'(https?://[^/]+)/i/', r'\1/z/', manifest_url).replace('/master.m3u8', '/manifest.f4m')
2464 hds_host = hosts.get('hds')
2466 f4m_url = re.sub(r'(https?://)[^/]+', r'\1' + hds_host, f4m_url)
2467 if 'hdcore=' not in f4m_url:
2468 f4m_url += ('&' if '?' in f4m_url else '?') + hdcore_sign
2469 f4m_formats = self._extract_f4m_formats(
2470 f4m_url, video_id, f4m_id='hds', fatal=False)
2471 for entry in f4m_formats:
2472 entry.update({'extra_param_to_segment_url': hdcore_sign})
2473 formats.extend(f4m_formats)
2474 m3u8_url = re.sub(r'(https?://[^/]+)/z/', r'\1/i/', manifest_url).replace('/manifest.f4m', '/master.m3u8')
2475 hls_host = hosts.get('hls')
2477 m3u8_url = re.sub(r'(https?://)[^/]+', r'\1' + hls_host, m3u8_url)
2478 formats.extend(self._extract_m3u8_formats(
2479 m3u8_url, video_id, 'mp4', 'm3u8_native',
2480 m3u8_id='hls', fatal=False))
2483 def _extract_wowza_formats(self, url, video_id, m3u8_entry_protocol='m3u8_native', skip_protocols=[]):
2484 query = compat_urlparse.urlparse(url).query
2485 url = re.sub(r'/(?:manifest|playlist|jwplayer)\.(?:m3u8|f4m|mpd|smil)', '', url)
2487 r'(?:(?:http|rtmp|rtsp)(?P<s>s)?:)?(?P<url>//[^?]+)', url)
2488 url_base = mobj.group('url')
2489 http_base_url = '%s%s:%s' % ('http', mobj.group('s') or '', url_base)
2492 def manifest_url(manifest):
2493 m_url = '%s/%s' % (http_base_url, manifest)
2495 m_url += '?%s' % query
2498 if 'm3u8' not in skip_protocols:
2499 formats.extend(self._extract_m3u8_formats(
2500 manifest_url('playlist.m3u8'), video_id, 'mp4',
2501 m3u8_entry_protocol, m3u8_id='hls', fatal=False))
2502 if 'f4m' not in skip_protocols:
2503 formats.extend(self._extract_f4m_formats(
2504 manifest_url('manifest.f4m'),
2505 video_id, f4m_id='hds', fatal=False))
2506 if 'dash' not in skip_protocols:
2507 formats.extend(self._extract_mpd_formats(
2508 manifest_url('manifest.mpd'),
2509 video_id, mpd_id='dash', fatal=False))
2510 if re.search(r'(?:/smil:|\.smil)', url_base):
2511 if 'smil' not in skip_protocols:
2512 rtmp_formats = self._extract_smil_formats(
2513 manifest_url('jwplayer.smil'),
2514 video_id, fatal=False)
2515 for rtmp_format in rtmp_formats:
2516 rtsp_format = rtmp_format.copy()
2517 rtsp_format['url'] = '%s/%s' % (rtmp_format['url'], rtmp_format['play_path'])
2518 del rtsp_format['play_path']
2519 del rtsp_format['ext']
2520 rtsp_format.update({
2521 'url': rtsp_format['url'].replace('rtmp://', 'rtsp://'),
2522 'format_id': rtmp_format['format_id'].replace('rtmp', 'rtsp'),
2525 formats.extend([rtmp_format, rtsp_format])
2527 for protocol in ('rtmp', 'rtsp'):
2528 if protocol not in skip_protocols:
2530 'url': '%s:%s' % (protocol, url_base),
2531 'format_id': protocol,
2532 'protocol': protocol,
2536 def _find_jwplayer_data(self, webpage, video_id=None, transform_source=js_to_json):
2538 r'(?s)jwplayer\((?P<quote>[\'"])[^\'" ]+(?P=quote)\)(?!</script>).*?\.setup\s*\((?P<options>[^)]+)\)',
2542 jwplayer_data = self._parse_json(mobj.group('options'),
2544 transform_source=transform_source)
2545 except ExtractorError:
2548 if isinstance(jwplayer_data, dict):
2549 return jwplayer_data
2551 def _extract_jwplayer_data(self, webpage, video_id, *args, **kwargs):
2552 jwplayer_data = self._find_jwplayer_data(
2553 webpage, video_id, transform_source=js_to_json)
2554 return self._parse_jwplayer_data(
2555 jwplayer_data, video_id, *args, **kwargs)
2557 def _parse_jwplayer_data(self, jwplayer_data, video_id=None, require_title=True,
2558 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
2559 # JWPlayer backward compatibility: flattened playlists
2560 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/api/config.js#L81-L96
2561 if 'playlist' not in jwplayer_data:
2562 jwplayer_data = {'playlist': [jwplayer_data]}
2566 # JWPlayer backward compatibility: single playlist item
2567 # https://github.com/jwplayer/jwplayer/blob/v7.7.0/src/js/playlist/playlist.js#L10
2568 if not isinstance(jwplayer_data['playlist'], list):
2569 jwplayer_data['playlist'] = [jwplayer_data['playlist']]
2571 for video_data in jwplayer_data['playlist']:
2572 # JWPlayer backward compatibility: flattened sources
2573 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/playlist/item.js#L29-L35
2574 if 'sources' not in video_data:
2575 video_data['sources'] = [video_data]
2577 this_video_id = video_id or video_data['mediaid']
2579 formats = self._parse_jwplayer_formats(
2580 video_data['sources'], video_id=this_video_id, m3u8_id=m3u8_id,
2581 mpd_id=mpd_id, rtmp_params=rtmp_params, base_url=base_url)
2584 tracks = video_data.get('tracks')
2585 if tracks and isinstance(tracks, list):
2586 for track in tracks:
2587 if not isinstance(track, dict):
2589 track_kind = track.get('kind')
2590 if not track_kind or not isinstance(track_kind, compat_str):
2592 if track_kind.lower() not in ('captions', 'subtitles'):
2594 track_url = urljoin(base_url, track.get('file'))
2597 subtitles.setdefault(track.get('label') or 'en', []).append({
2598 'url': self._proto_relative_url(track_url)
2602 'id': this_video_id,
2603 'title': unescapeHTML(video_data['title'] if require_title else video_data.get('title')),
2604 'description': video_data.get('description'),
2605 'thumbnail': self._proto_relative_url(video_data.get('image')),
2606 'timestamp': int_or_none(video_data.get('pubdate')),
2607 'duration': float_or_none(jwplayer_data.get('duration') or video_data.get('duration')),
2608 'subtitles': subtitles,
2610 # https://github.com/jwplayer/jwplayer/blob/master/src/js/utils/validator.js#L32
2611 if len(formats) == 1 and re.search(r'^(?:http|//).*(?:youtube\.com|youtu\.be)/.+', formats[0]['url']):
2613 '_type': 'url_transparent',
2614 'url': formats[0]['url'],
2617 self._sort_formats(formats)
2618 entry['formats'] = formats
2619 entries.append(entry)
2620 if len(entries) == 1:
2623 return self.playlist_result(entries)
2625 def _parse_jwplayer_formats(self, jwplayer_sources_data, video_id=None,
2626 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
2629 for source in jwplayer_sources_data:
2630 if not isinstance(source, dict):
2632 source_url = self._proto_relative_url(source.get('file'))
2636 source_url = compat_urlparse.urljoin(base_url, source_url)
2637 if source_url in urls:
2639 urls.append(source_url)
2640 source_type = source.get('type') or ''
2641 ext = mimetype2ext(source_type) or determine_ext(source_url)
2642 if source_type == 'hls' or ext == 'm3u8':
2643 formats.extend(self._extract_m3u8_formats(
2644 source_url, video_id, 'mp4', entry_protocol='m3u8_native',
2645 m3u8_id=m3u8_id, fatal=False))
2646 elif source_type == 'dash' or ext == 'mpd':
2647 formats.extend(self._extract_mpd_formats(
2648 source_url, video_id, mpd_id=mpd_id, fatal=False))
2650 formats.extend(self._extract_smil_formats(
2651 source_url, video_id, fatal=False))
2652 # https://github.com/jwplayer/jwplayer/blob/master/src/js/providers/default.js#L67
2653 elif source_type.startswith('audio') or ext in (
2654 'oga', 'aac', 'mp3', 'mpeg', 'vorbis'):
2661 height = int_or_none(source.get('height'))
2663 # Often no height is provided but there is a label in
2664 # format like "1080p", "720p SD", or 1080.
2665 height = int_or_none(self._search_regex(
2666 r'^(\d{3,4})[pP]?(?:\b|$)', compat_str(source.get('label') or ''),
2667 'height', default=None))
2670 'width': int_or_none(source.get('width')),
2672 'tbr': int_or_none(source.get('bitrate')),
2675 if source_url.startswith('rtmp'):
2676 a_format['ext'] = 'flv'
2677 # See com/longtailvideo/jwplayer/media/RTMPMediaProvider.as
2678 # of jwplayer.flash.swf
2679 rtmp_url_parts = re.split(
2680 r'((?:mp4|mp3|flv):)', source_url, 1)
2681 if len(rtmp_url_parts) == 3:
2682 rtmp_url, prefix, play_path = rtmp_url_parts
2685 'play_path': prefix + play_path,
2688 a_format.update(rtmp_params)
2689 formats.append(a_format)
2692 def _live_title(self, name):
2693 """ Generate the title for a live video """
2694 now = datetime.datetime.now()
2695 now_str = now.strftime('%Y-%m-%d %H:%M')
2696 return name + ' ' + now_str
2698 def _int(self, v, name, fatal=False, **kwargs):
2699 res = int_or_none(v, **kwargs)
2700 if 'get_attr' in kwargs:
2701 print(getattr(v, kwargs['get_attr']))
2703 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
2705 raise ExtractorError(msg)
2707 self._downloader.report_warning(msg)
2710 def _float(self, v, name, fatal=False, **kwargs):
2711 res = float_or_none(v, **kwargs)
2713 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
2715 raise ExtractorError(msg)
2717 self._downloader.report_warning(msg)
2720 def _set_cookie(self, domain, name, value, expire_time=None, port=None,
2721 path='/', secure=False, discard=False, rest={}, **kwargs):
2722 cookie = compat_cookiejar.Cookie(
2723 0, name, value, port, port is not None, domain, True,
2724 domain.startswith('.'), path, True, secure, expire_time,
2725 discard, None, None, rest)
2726 self._downloader.cookiejar.set_cookie(cookie)
2728 def _get_cookies(self, url):
2729 """ Return a compat_cookies.SimpleCookie with the cookies for the url """
2730 req = sanitized_Request(url)
2731 self._downloader.cookiejar.add_cookie_header(req)
2732 return compat_cookies.SimpleCookie(req.get_header('Cookie'))
2734 def get_testcases(self, include_onlymatching=False):
2735 t = getattr(self, '_TEST', None)
2737 assert not hasattr(self, '_TESTS'), \
2738 '%s has _TEST and _TESTS' % type(self).__name__
2741 tests = getattr(self, '_TESTS', [])
2743 if not include_onlymatching and t.get('only_matching', False):
2745 t['name'] = type(self).__name__[:-len('IE')]
2748 def is_suitable(self, age_limit):
2749 """ Test whether the extractor is generally suitable for the given
2750 age limit (i.e. pornographic sites are not, all others usually are) """
2752 any_restricted = False
2753 for tc in self.get_testcases(include_onlymatching=False):
2754 if tc.get('playlist', []):
2755 tc = tc['playlist'][0]
2756 is_restricted = age_restricted(
2757 tc.get('info_dict', {}).get('age_limit'), age_limit)
2758 if not is_restricted:
2760 any_restricted = any_restricted or is_restricted
2761 return not any_restricted
2763 def extract_subtitles(self, *args, **kwargs):
2764 if (self._downloader.params.get('writesubtitles', False) or
2765 self._downloader.params.get('listsubtitles')):
2766 return self._get_subtitles(*args, **kwargs)
2769 def _get_subtitles(self, *args, **kwargs):
2770 raise NotImplementedError('This method must be implemented by subclasses')
2773 def _merge_subtitle_items(subtitle_list1, subtitle_list2):
2774 """ Merge subtitle items for one language. Items with duplicated URLs
2775 will be dropped. """
2776 list1_urls = set([item['url'] for item in subtitle_list1])
2777 ret = list(subtitle_list1)
2778 ret.extend([item for item in subtitle_list2 if item['url'] not in list1_urls])
2782 def _merge_subtitles(cls, subtitle_dict1, subtitle_dict2):
2783 """ Merge two subtitle dictionaries, language by language. """
2784 ret = dict(subtitle_dict1)
2785 for lang in subtitle_dict2:
2786 ret[lang] = cls._merge_subtitle_items(subtitle_dict1.get(lang, []), subtitle_dict2[lang])
2789 def extract_automatic_captions(self, *args, **kwargs):
2790 if (self._downloader.params.get('writeautomaticsub', False) or
2791 self._downloader.params.get('listsubtitles')):
2792 return self._get_automatic_captions(*args, **kwargs)
2795 def _get_automatic_captions(self, *args, **kwargs):
2796 raise NotImplementedError('This method must be implemented by subclasses')
2798 def mark_watched(self, *args, **kwargs):
2799 if (self._downloader.params.get('mark_watched', False) and
2800 (self._get_login_info()[0] is not None or
2801 self._downloader.params.get('cookiefile') is not None)):
2802 self._mark_watched(*args, **kwargs)
2804 def _mark_watched(self, *args, **kwargs):
2805 raise NotImplementedError('This method must be implemented by subclasses')
2807 def geo_verification_headers(self):
2809 geo_verification_proxy = self._downloader.params.get('geo_verification_proxy')
2810 if geo_verification_proxy:
2811 headers['Ytdl-request-proxy'] = geo_verification_proxy
2814 def _generic_id(self, url):
2815 return compat_urllib_parse_unquote(os.path.splitext(url.rstrip('/').split('/')[-1])[0])
2817 def _generic_title(self, url):
2818 return compat_urllib_parse_unquote(os.path.splitext(url_basename(url))[0])
2821 class SearchInfoExtractor(InfoExtractor):
2823 Base class for paged search queries extractors.
2824 They accept URLs in the format _SEARCH_KEY(|all|[0-9]):{query}
2825 Instances should define _SEARCH_KEY and _MAX_RESULTS.
2829 def _make_valid_url(cls):
2830 return r'%s(?P<prefix>|[1-9][0-9]*|all):(?P<query>[\s\S]+)' % cls._SEARCH_KEY
2833 def suitable(cls, url):
2834 return re.match(cls._make_valid_url(), url) is not None
2836 def _real_extract(self, query):
2837 mobj = re.match(self._make_valid_url(), query)
2839 raise ExtractorError('Invalid search query "%s"' % query)
2841 prefix = mobj.group('prefix')
2842 query = mobj.group('query')
2844 return self._get_n_results(query, 1)
2845 elif prefix == 'all':
2846 return self._get_n_results(query, self._MAX_RESULTS)
2850 raise ExtractorError('invalid download number %s for query "%s"' % (n, query))
2851 elif n > self._MAX_RESULTS:
2852 self._downloader.report_warning('%s returns max %i results (you requested %i)' % (self._SEARCH_KEY, self._MAX_RESULTS, n))
2853 n = self._MAX_RESULTS
2854 return self._get_n_results(query, n)
2856 def _get_n_results(self, query, n):
2857 """Get a specified number of results for a query"""
2858 raise NotImplementedError('This method must be implemented by subclasses')
2861 def SEARCH_KEY(self):
2862 return self._SEARCH_KEY