2 from __future__ import unicode_literals
17 from ..compat import (
20 compat_etree_fromstring,
26 compat_urllib_parse_unquote,
27 compat_urllib_parse_urlencode,
28 compat_urllib_request,
30 compat_xml_parse_error,
32 from ..downloader.f4m import (
34 remove_encrypted_media,
59 parse_m3u8_attributes,
76 class InfoExtractor(object):
77 """Information Extractor class.
79 Information extractors are the classes that, given a URL, extract
80 information about the video (or videos) the URL refers to. This
81 information includes the real video URL, the video title, author and
82 others. The information is stored in a dictionary which is then
83 passed to the YoutubeDL. The YoutubeDL processes this
84 information possibly downloading the video to the file system, among
85 other possible outcomes.
87 The type field determines the type of the result.
88 By far the most common value (and the default if _type is missing) is
89 "video", which indicates a single video.
91 For a video, the dictionaries must include the following fields:
94 title: Video title, unescaped.
96 Additionally, it must contain either a formats entry or a url one:
98 formats: A list of dictionaries for each format available, ordered
99 from worst to best quality.
102 * url Mandatory. The URL of the video file
104 The URL of the manifest file in case of
105 fragmented media (DASH, hls, hds)
106 * ext Will be calculated from URL if missing
107 * format A human-readable description of the format
108 ("mp4 container with h264/opus").
109 Calculated from the format_id, width, height.
110 and format_note fields if missing.
111 * format_id A short description of the format
112 ("mp4_h264_opus" or "19").
113 Technically optional, but strongly recommended.
114 * format_note Additional info about the format
115 ("3D" or "DASH video")
116 * width Width of the video, if known
117 * height Height of the video, if known
118 * resolution Textual description of width and height
119 * tbr Average bitrate of audio and video in KBit/s
120 * abr Average audio bitrate in KBit/s
121 * acodec Name of the audio codec in use
122 * asr Audio sampling rate in Hertz
123 * vbr Average video bitrate in KBit/s
125 * vcodec Name of the video codec in use
126 * container Name of the container format
127 * filesize The number of bytes, if known in advance
128 * filesize_approx An estimate for the number of bytes
129 * player_url SWF Player URL (used for rtmpdump).
130 * protocol The protocol that will be used for the actual
131 download, lower-case.
132 "http", "https", "rtsp", "rtmp", "rtmpe",
133 "m3u8", "m3u8_native" or "http_dash_segments".
135 Base URL for fragments. Each fragment's path
136 value (if present) will be relative to
138 * fragments A list of fragments of a fragmented media.
139 Each fragment entry must contain either an url
140 or a path. If an url is present it should be
141 considered by a client. Otherwise both path and
142 fragment_base_url must be present. Here is
143 the list of all potential fields:
144 * "url" - fragment's URL
145 * "path" - fragment's path relative to
147 * "duration" (optional, int or float)
148 * "filesize" (optional, int)
149 * preference Order number of this format. If this field is
150 present and not None, the formats get sorted
151 by this field, regardless of all other values.
152 -1 for default (order by other properties),
153 -2 or smaller for less than default.
154 < -1000 to hide the format (if there is
155 another one which is strictly better)
156 * language Language code, e.g. "de" or "en-US".
157 * language_preference Is this in the language mentioned in
159 10 if it's what the URL is about,
160 -1 for default (don't know),
161 -10 otherwise, other values reserved for now.
162 * quality Order number of the video quality of this
163 format, irrespective of the file format.
164 -1 for default (order by other properties),
165 -2 or smaller for less than default.
166 * source_preference Order number for this video source
167 (quality takes higher priority)
168 -1 for default (order by other properties),
169 -2 or smaller for less than default.
170 * http_headers A dictionary of additional HTTP headers
171 to add to the request.
172 * stretched_ratio If given and not 1, indicates that the
173 video's pixels are not square.
174 width : height ratio as float.
175 * no_resume The server does not support resuming the
176 (HTTP or RTMP) download. Boolean.
178 url: Final video URL.
179 ext: Video filename extension.
180 format: The video format, defaults to ext (used for --get-format)
181 player_url: SWF Player URL (used for rtmpdump).
183 The following fields are optional:
185 alt_title: A secondary title of the video.
186 display_id An alternative identifier for the video, not necessarily
187 unique, but available before title. Typically, id is
188 something like "4234987", title "Dancing naked mole rats",
189 and display_id "dancing-naked-mole-rats"
190 thumbnails: A list of dictionaries, with the following entries:
191 * "id" (optional, string) - Thumbnail format ID
193 * "preference" (optional, int) - quality of the image
194 * "width" (optional, int)
195 * "height" (optional, int)
196 * "resolution" (optional, string "{width}x{height"},
198 * "filesize" (optional, int)
199 thumbnail: Full URL to a video thumbnail image.
200 description: Full video description.
201 uploader: Full name of the video uploader.
202 license: License name the video is licensed under.
203 creator: The creator of the video.
204 release_date: The date (YYYYMMDD) when the video was released.
205 timestamp: UNIX timestamp of the moment the video became available.
206 upload_date: Video upload date (YYYYMMDD).
207 If not explicitly set, calculated from timestamp.
208 uploader_id: Nickname or id of the video uploader.
209 uploader_url: Full URL to a personal webpage of the video uploader.
210 location: Physical location where the video was filmed.
211 subtitles: The available subtitles as a dictionary in the format
212 {tag: subformats}. "tag" is usually a language code, and
213 "subformats" is a list sorted from lower to higher
214 preference, each element is a dictionary with the "ext"
216 * "data": The subtitles file contents
217 * "url": A URL pointing to the subtitles file
218 "ext" will be calculated from URL if missing
219 automatic_captions: Like 'subtitles', used by the YoutubeIE for
220 automatically generated captions
221 duration: Length of the video in seconds, as an integer or float.
222 view_count: How many users have watched the video on the platform.
223 like_count: Number of positive ratings of the video
224 dislike_count: Number of negative ratings of the video
225 repost_count: Number of reposts of the video
226 average_rating: Average rating give by users, the scale used depends on the webpage
227 comment_count: Number of comments on the video
228 comments: A list of comments, each with one or more of the following
229 properties (all but one of text or html optional):
230 * "author" - human-readable name of the comment author
231 * "author_id" - user ID of the comment author
233 * "html" - Comment as HTML
234 * "text" - Plain text of the comment
235 * "timestamp" - UNIX timestamp of comment
236 * "parent" - ID of the comment this one is replying to.
237 Set to "root" to indicate that this is a
238 comment to the original video.
239 age_limit: Age restriction for the video, as an integer (years)
240 webpage_url: The URL to the video webpage, if given to youtube-dl it
241 should allow to get the same result again. (It will be set
242 by YoutubeDL if it's missing)
243 categories: A list of categories that the video falls in, for example
245 tags: A list of tags assigned to the video, e.g. ["sweden", "pop music"]
246 is_live: True, False, or None (=unknown). Whether this video is a
247 live stream that goes on instead of a fixed-length video.
248 start_time: Time in seconds where the reproduction should start, as
249 specified in the URL.
250 end_time: Time in seconds where the reproduction should end, as
251 specified in the URL.
252 chapters: A list of dictionaries, with the following entries:
253 * "start_time" - The start time of the chapter in seconds
254 * "end_time" - The end time of the chapter in seconds
255 * "title" (optional, string)
257 The following fields should only be used when the video belongs to some logical
260 chapter: Name or title of the chapter the video belongs to.
261 chapter_number: Number of the chapter the video belongs to, as an integer.
262 chapter_id: Id of the chapter the video belongs to, as a unicode string.
264 The following fields should only be used when the video is an episode of some
265 series, programme or podcast:
267 series: Title of the series or programme the video episode belongs to.
268 season: Title of the season the video episode belongs to.
269 season_number: Number of the season the video episode belongs to, as an integer.
270 season_id: Id of the season the video episode belongs to, as a unicode string.
271 episode: Title of the video episode. Unlike mandatory video title field,
272 this field should denote the exact title of the video episode
273 without any kind of decoration.
274 episode_number: Number of the video episode within a season, as an integer.
275 episode_id: Id of the video episode, as a unicode string.
277 The following fields should only be used when the media is a track or a part of
280 track: Title of the track.
281 track_number: Number of the track within an album or a disc, as an integer.
282 track_id: Id of the track (useful in case of custom indexing, e.g. 6.iii),
284 artist: Artist(s) of the track.
285 genre: Genre(s) of the track.
286 album: Title of the album the track belongs to.
287 album_type: Type of the album (e.g. "Demo", "Full-length", "Split", "Compilation", etc).
288 album_artist: List of all artists appeared on the album (e.g.
289 "Ash Borer / Fell Voices" or "Various Artists", useful for splits
291 disc_number: Number of the disc or other physical medium the track belongs to,
293 release_year: Year (YYYY) when the album was released.
295 Unless mentioned otherwise, the fields should be Unicode strings.
297 Unless mentioned otherwise, None is equivalent to absence of information.
300 _type "playlist" indicates multiple videos.
301 There must be a key "entries", which is a list, an iterable, or a PagedList
302 object, each element of which is a valid dictionary by this specification.
304 Additionally, playlists can have "id", "title", "description", "uploader",
305 "uploader_id", "uploader_url" attributes with the same semantics as videos
309 _type "multi_video" indicates that there are multiple videos that
310 form a single show, for examples multiple acts of an opera or TV episode.
311 It must have an entries key like a playlist and contain all the keys
312 required for a video at the same time.
315 _type "url" indicates that the video must be extracted from another
316 location, possibly by a different extractor. Its only required key is:
317 "url" - the next URL to extract.
318 The key "ie_key" can be set to the class name (minus the trailing "IE",
319 e.g. "Youtube") if the extractor class is known in advance.
320 Additionally, the dictionary may have any properties of the resolved entity
321 known in advance, for example "title" if the title of the referred video is
325 _type "url_transparent" entities have the same specification as "url", but
326 indicate that the given additional information is more precise than the one
327 associated with the resolved URL.
328 This is useful when a site employs a video service that hosts the video and
329 its technical metadata, but that video service does not embed a useful
330 title, description etc.
333 Subclasses of this one should re-define the _real_initialize() and
334 _real_extract() methods and define a _VALID_URL regexp.
335 Probably, they should also be added to the list of extractors.
337 _GEO_BYPASS attribute may be set to False in order to disable
338 geo restriction bypass mechanisms for a particular extractor.
339 Though it won't disable explicit geo restriction bypass based on
340 country code provided with geo_bypass_country. (experimental)
342 _GEO_COUNTRIES attribute may contain a list of presumably geo unrestricted
343 countries for this extractor. One of these countries will be used by
344 geo restriction bypass mechanism right away in order to bypass
345 geo restriction, of course, if the mechanism is not disabled. (experimental)
347 NB: both these geo attributes are experimental and may change in future
348 or be completely removed.
350 Finally, the _WORKING attribute should be set to False for broken IEs
351 in order to warn the users and skip the tests.
356 _x_forwarded_for_ip = None
358 _GEO_COUNTRIES = None
361 def __init__(self, downloader=None):
362 """Constructor. Receives an optional downloader."""
364 self._x_forwarded_for_ip = None
365 self.set_downloader(downloader)
368 def suitable(cls, url):
369 """Receives a URL and returns True if suitable for this IE."""
371 # This does not use has/getattr intentionally - we want to know whether
372 # we have cached the regexp for *this* class, whereas getattr would also
373 # match the superclass
374 if '_VALID_URL_RE' not in cls.__dict__:
375 cls._VALID_URL_RE = re.compile(cls._VALID_URL)
376 return cls._VALID_URL_RE.match(url) is not None
379 def _match_id(cls, url):
380 if '_VALID_URL_RE' not in cls.__dict__:
381 cls._VALID_URL_RE = re.compile(cls._VALID_URL)
382 m = cls._VALID_URL_RE.match(url)
384 return compat_str(m.group('id'))
388 """Getter method for _WORKING."""
391 def initialize(self):
392 """Initializes an instance (authentication, etc)."""
393 self._initialize_geo_bypass(self._GEO_COUNTRIES)
395 self._real_initialize()
398 def _initialize_geo_bypass(self, countries):
400 Initialize geo restriction bypass mechanism.
402 This method is used to initialize geo bypass mechanism based on faking
403 X-Forwarded-For HTTP header. A random country from provided country list
404 is selected and a random IP belonging to this country is generated. This
405 IP will be passed as X-Forwarded-For HTTP header in all subsequent
408 This method will be used for initial geo bypass mechanism initialization
409 during the instance initialization with _GEO_COUNTRIES.
411 You may also manually call it from extractor's code if geo countries
412 information is not available beforehand (e.g. obtained during
413 extraction) or due to some another reason.
415 if not self._x_forwarded_for_ip:
416 country_code = self._downloader.params.get('geo_bypass_country', None)
417 # If there is no explicit country for geo bypass specified and
418 # the extractor is known to be geo restricted let's fake IP
419 # as X-Forwarded-For right away.
420 if (not country_code and
422 self._downloader.params.get('geo_bypass', True) and
424 country_code = random.choice(countries)
426 self._x_forwarded_for_ip = GeoUtils.random_ipv4(country_code)
427 if self._downloader.params.get('verbose', False):
428 self._downloader.to_screen(
429 '[debug] Using fake IP %s (%s) as X-Forwarded-For.'
430 % (self._x_forwarded_for_ip, country_code.upper()))
432 def extract(self, url):
433 """Extracts URL information and returns it in list of dicts."""
438 ie_result = self._real_extract(url)
439 if self._x_forwarded_for_ip:
440 ie_result['__x_forwarded_for_ip'] = self._x_forwarded_for_ip
442 except GeoRestrictedError as e:
443 if self.__maybe_fake_ip_and_retry(e.countries):
446 except ExtractorError:
448 except compat_http_client.IncompleteRead as e:
449 raise ExtractorError('A network error has occurred.', cause=e, expected=True)
450 except (KeyError, StopIteration) as e:
451 raise ExtractorError('An extractor error has occurred.', cause=e)
453 def __maybe_fake_ip_and_retry(self, countries):
454 if (not self._downloader.params.get('geo_bypass_country', None) and
456 self._downloader.params.get('geo_bypass', True) and
457 not self._x_forwarded_for_ip and
459 country_code = random.choice(countries)
460 self._x_forwarded_for_ip = GeoUtils.random_ipv4(country_code)
461 if self._x_forwarded_for_ip:
463 'Video is geo restricted. Retrying extraction with fake IP %s (%s) as X-Forwarded-For.'
464 % (self._x_forwarded_for_ip, country_code.upper()))
468 def set_downloader(self, downloader):
469 """Sets the downloader for this IE."""
470 self._downloader = downloader
472 def _real_initialize(self):
473 """Real initialization process. Redefine in subclasses."""
476 def _real_extract(self, url):
477 """Real extraction process. Redefine in subclasses."""
482 """A string for getting the InfoExtractor with get_info_extractor"""
483 return compat_str(cls.__name__[:-2])
487 return compat_str(type(self).__name__[:-2])
489 def _request_webpage(self, url_or_request, video_id, note=None, errnote=None, fatal=True, data=None, headers={}, query={}):
490 """ Returns the response handle """
492 self.report_download_webpage(video_id)
493 elif note is not False:
495 self.to_screen('%s' % (note,))
497 self.to_screen('%s: %s' % (video_id, note))
498 if isinstance(url_or_request, compat_urllib_request.Request):
499 url_or_request = update_Request(
500 url_or_request, data=data, headers=headers, query=query)
503 url_or_request = update_url_query(url_or_request, query)
504 if data is not None or headers:
505 url_or_request = sanitized_Request(url_or_request, data, headers)
507 return self._downloader.urlopen(url_or_request)
508 except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err:
512 errnote = 'Unable to download webpage'
514 errmsg = '%s: %s' % (errnote, error_to_compat_str(err))
516 raise ExtractorError(errmsg, sys.exc_info()[2], cause=err)
518 self._downloader.report_warning(errmsg)
521 def _download_webpage_handle(self, url_or_request, video_id, note=None, errnote=None, fatal=True, encoding=None, data=None, headers={}, query={}):
522 """ Returns a tuple (page content as string, URL handle) """
523 # Strip hashes from the URL (#1038)
524 if isinstance(url_or_request, (compat_str, str)):
525 url_or_request = url_or_request.partition('#')[0]
527 # Some sites check X-Forwarded-For HTTP header in order to figure out
528 # the origin of the client behind proxy. This allows bypassing geo
529 # restriction by faking this header's value to IP that belongs to some
530 # geo unrestricted country. We will do so once we encounter any
531 # geo restriction error.
532 if self._x_forwarded_for_ip:
533 if 'X-Forwarded-For' not in headers:
534 headers['X-Forwarded-For'] = self._x_forwarded_for_ip
536 urlh = self._request_webpage(url_or_request, video_id, note, errnote, fatal, data=data, headers=headers, query=query)
540 content = self._webpage_read_content(urlh, url_or_request, video_id, note, errnote, fatal, encoding=encoding)
541 return (content, urlh)
544 def _guess_encoding_from_content(content_type, webpage_bytes):
545 m = re.match(r'[a-zA-Z0-9_.-]+/[a-zA-Z0-9_.-]+\s*;\s*charset=(.+)', content_type)
547 encoding = m.group(1)
549 m = re.search(br'<meta[^>]+charset=[\'"]?([^\'")]+)[ /\'">]',
550 webpage_bytes[:1024])
552 encoding = m.group(1).decode('ascii')
553 elif webpage_bytes.startswith(b'\xff\xfe'):
560 def __check_blocked(self, content):
561 first_block = content[:512]
562 if ('<title>Access to this site is blocked</title>' in content and
563 'Websense' in first_block):
564 msg = 'Access to this webpage has been blocked by Websense filtering software in your network.'
565 blocked_iframe = self._html_search_regex(
566 r'<iframe src="([^"]+)"', content,
567 'Websense information URL', default=None)
569 msg += ' Visit %s for more details' % blocked_iframe
570 raise ExtractorError(msg, expected=True)
571 if '<title>The URL you requested has been blocked</title>' in first_block:
573 'Access to this webpage has been blocked by Indian censorship. '
574 'Use a VPN or proxy server (with --proxy) to route around it.')
575 block_msg = self._html_search_regex(
576 r'</h1><p>(.*?)</p>',
577 content, 'block message', default=None)
579 msg += ' (Message: "%s")' % block_msg.replace('\n', ' ')
580 raise ExtractorError(msg, expected=True)
581 if ('<title>TTK :: Доступ к ресурсу ограничен</title>' in content and
582 'blocklist.rkn.gov.ru' in content):
583 raise ExtractorError(
584 'Access to this webpage has been blocked by decision of the Russian government. '
585 'Visit http://blocklist.rkn.gov.ru/ for a block reason.',
588 def _webpage_read_content(self, urlh, url_or_request, video_id, note=None, errnote=None, fatal=True, prefix=None, encoding=None):
589 content_type = urlh.headers.get('Content-Type', '')
590 webpage_bytes = urlh.read()
591 if prefix is not None:
592 webpage_bytes = prefix + webpage_bytes
594 encoding = self._guess_encoding_from_content(content_type, webpage_bytes)
595 if self._downloader.params.get('dump_intermediate_pages', False):
596 self.to_screen('Dumping request to ' + urlh.geturl())
597 dump = base64.b64encode(webpage_bytes).decode('ascii')
598 self._downloader.to_screen(dump)
599 if self._downloader.params.get('write_pages', False):
600 basen = '%s_%s' % (video_id, urlh.geturl())
602 h = '___' + hashlib.md5(basen.encode('utf-8')).hexdigest()
603 basen = basen[:240 - len(h)] + h
604 raw_filename = basen + '.dump'
605 filename = sanitize_filename(raw_filename, restricted=True)
606 self.to_screen('Saving request to ' + filename)
607 # Working around MAX_PATH limitation on Windows (see
608 # http://msdn.microsoft.com/en-us/library/windows/desktop/aa365247(v=vs.85).aspx)
609 if compat_os_name == 'nt':
610 absfilepath = os.path.abspath(filename)
611 if len(absfilepath) > 259:
612 filename = '\\\\?\\' + absfilepath
613 with open(filename, 'wb') as outf:
614 outf.write(webpage_bytes)
617 content = webpage_bytes.decode(encoding, 'replace')
619 content = webpage_bytes.decode('utf-8', 'replace')
621 self.__check_blocked(content)
625 def _download_webpage(self, url_or_request, video_id, note=None, errnote=None, fatal=True, tries=1, timeout=5, encoding=None, data=None, headers={}, query={}):
626 """ Returns the data of the page as a string """
629 while success is False:
631 res = self._download_webpage_handle(url_or_request, video_id, note, errnote, fatal, encoding=encoding, data=data, headers=headers, query=query)
633 except compat_http_client.IncompleteRead as e:
635 if try_count >= tries:
637 self._sleep(timeout, video_id)
644 def _download_xml(self, url_or_request, video_id,
645 note='Downloading XML', errnote='Unable to download XML',
646 transform_source=None, fatal=True, encoding=None,
647 data=None, headers={}, query={}):
648 """Return the xml as an xml.etree.ElementTree.Element"""
649 xml_string = self._download_webpage(
650 url_or_request, video_id, note, errnote, fatal=fatal,
651 encoding=encoding, data=data, headers=headers, query=query)
652 if xml_string is False:
654 return self._parse_xml(
655 xml_string, video_id, transform_source=transform_source,
658 def _parse_xml(self, xml_string, video_id, transform_source=None, fatal=True):
660 xml_string = transform_source(xml_string)
662 return compat_etree_fromstring(xml_string.encode('utf-8'))
663 except compat_xml_parse_error as ve:
664 errmsg = '%s: Failed to parse XML ' % video_id
666 raise ExtractorError(errmsg, cause=ve)
668 self.report_warning(errmsg + str(ve))
670 def _download_json(self, url_or_request, video_id,
671 note='Downloading JSON metadata',
672 errnote='Unable to download JSON metadata',
673 transform_source=None,
674 fatal=True, encoding=None, data=None, headers={}, query={}):
675 json_string = self._download_webpage(
676 url_or_request, video_id, note, errnote, fatal=fatal,
677 encoding=encoding, data=data, headers=headers, query=query)
678 if (not fatal) and json_string is False:
680 return self._parse_json(
681 json_string, video_id, transform_source=transform_source, fatal=fatal)
683 def _parse_json(self, json_string, video_id, transform_source=None, fatal=True):
685 json_string = transform_source(json_string)
687 return json.loads(json_string)
688 except ValueError as ve:
689 errmsg = '%s: Failed to parse JSON ' % video_id
691 raise ExtractorError(errmsg, cause=ve)
693 self.report_warning(errmsg + str(ve))
695 def report_warning(self, msg, video_id=None):
696 idstr = '' if video_id is None else '%s: ' % video_id
697 self._downloader.report_warning(
698 '[%s] %s%s' % (self.IE_NAME, idstr, msg))
700 def to_screen(self, msg):
701 """Print msg to screen, prefixing it with '[ie_name]'"""
702 self._downloader.to_screen('[%s] %s' % (self.IE_NAME, msg))
704 def report_extraction(self, id_or_name):
705 """Report information extraction."""
706 self.to_screen('%s: Extracting information' % id_or_name)
708 def report_download_webpage(self, video_id):
709 """Report webpage download."""
710 self.to_screen('%s: Downloading webpage' % video_id)
712 def report_age_confirmation(self):
713 """Report attempt to confirm age."""
714 self.to_screen('Confirming age')
716 def report_login(self):
717 """Report attempt to log in."""
718 self.to_screen('Logging in')
721 def raise_login_required(msg='This video is only available for registered users'):
722 raise ExtractorError(
723 '%s. Use --username and --password or --netrc to provide account credentials.' % msg,
727 def raise_geo_restricted(msg='This video is not available from your location due to geo restriction', countries=None):
728 raise GeoRestrictedError(msg, countries=countries)
730 # Methods for following #608
732 def url_result(url, ie=None, video_id=None, video_title=None):
733 """Returns a URL that points to a page that should be processed"""
734 # TODO: ie should be the class used for getting the info
735 video_info = {'_type': 'url',
738 if video_id is not None:
739 video_info['id'] = video_id
740 if video_title is not None:
741 video_info['title'] = video_title
744 def playlist_from_matches(self, matches, playlist_id=None, playlist_title=None, getter=None, ie=None):
746 self.url_result(self._proto_relative_url(getter(m) if getter else m), ie)
748 return self.playlist_result(
749 urls, playlist_id=playlist_id, playlist_title=playlist_title)
752 def playlist_result(entries, playlist_id=None, playlist_title=None, playlist_description=None):
753 """Returns a playlist"""
754 video_info = {'_type': 'playlist',
757 video_info['id'] = playlist_id
759 video_info['title'] = playlist_title
760 if playlist_description:
761 video_info['description'] = playlist_description
764 def _search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
766 Perform a regex search on the given string, using a single or a list of
767 patterns returning the first matching group.
768 In case of failure return a default value or raise a WARNING or a
769 RegexNotFoundError, depending on fatal, specifying the field name.
771 if isinstance(pattern, (str, compat_str, compiled_regex_type)):
772 mobj = re.search(pattern, string, flags)
775 mobj = re.search(p, string, flags)
779 if not self._downloader.params.get('no_color') and compat_os_name != 'nt' and sys.stderr.isatty():
780 _name = '\033[0;34m%s\033[0m' % name
786 # return the first matching group
787 return next(g for g in mobj.groups() if g is not None)
789 return mobj.group(group)
790 elif default is not NO_DEFAULT:
793 raise RegexNotFoundError('Unable to extract %s' % _name)
795 self._downloader.report_warning('unable to extract %s' % _name + bug_reports_message())
798 def _html_search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
800 Like _search_regex, but strips HTML tags and unescapes entities.
802 res = self._search_regex(pattern, string, name, default, fatal, flags, group)
804 return clean_html(res).strip()
808 def _get_netrc_login_info(self, netrc_machine=None):
811 netrc_machine = netrc_machine or self._NETRC_MACHINE
813 if self._downloader.params.get('usenetrc', False):
815 info = netrc.netrc().authenticators(netrc_machine)
820 raise netrc.NetrcParseError(
821 'No authenticators for %s' % netrc_machine)
822 except (IOError, netrc.NetrcParseError) as err:
823 self._downloader.report_warning(
824 'parsing .netrc: %s' % error_to_compat_str(err))
826 return username, password
828 def _get_login_info(self, username_option='username', password_option='password', netrc_machine=None):
830 Get the login info as (username, password)
831 First look for the manually specified credentials using username_option
832 and password_option as keys in params dictionary. If no such credentials
833 available look in the netrc file using the netrc_machine or _NETRC_MACHINE
835 If there's no info available, return (None, None)
837 if self._downloader is None:
840 downloader_params = self._downloader.params
842 # Attempt to use provided username and password or .netrc data
843 if downloader_params.get(username_option) is not None:
844 username = downloader_params[username_option]
845 password = downloader_params[password_option]
847 username, password = self._get_netrc_login_info(netrc_machine)
849 return username, password
851 def _get_tfa_info(self, note='two-factor verification code'):
853 Get the two-factor authentication info
854 TODO - asking the user will be required for sms/phone verify
855 currently just uses the command line option
856 If there's no info available, return None
858 if self._downloader is None:
860 downloader_params = self._downloader.params
862 if downloader_params.get('twofactor') is not None:
863 return downloader_params['twofactor']
865 return compat_getpass('Type %s and press [Return]: ' % note)
867 # Helper functions for extracting OpenGraph info
869 def _og_regexes(prop):
870 content_re = r'content=(?:"([^"]+?)"|\'([^\']+?)\'|\s*([^\s"\'=<>`]+?))'
871 property_re = (r'(?:name|property)=(?:\'og:%(prop)s\'|"og:%(prop)s"|\s*og:%(prop)s\b)'
872 % {'prop': re.escape(prop)})
873 template = r'<meta[^>]+?%s[^>]+?%s'
875 template % (property_re, content_re),
876 template % (content_re, property_re),
880 def _meta_regex(prop):
881 return r'''(?isx)<meta
882 (?=[^>]+(?:itemprop|name|property|id|http-equiv)=(["\']?)%s\1)
883 [^>]+?content=(["\'])(?P<content>.*?)\2''' % re.escape(prop)
885 def _og_search_property(self, prop, html, name=None, **kargs):
886 if not isinstance(prop, (list, tuple)):
889 name = 'OpenGraph %s' % prop[0]
892 og_regexes.extend(self._og_regexes(p))
893 escaped = self._search_regex(og_regexes, html, name, flags=re.DOTALL, **kargs)
896 return unescapeHTML(escaped)
898 def _og_search_thumbnail(self, html, **kargs):
899 return self._og_search_property('image', html, 'thumbnail URL', fatal=False, **kargs)
901 def _og_search_description(self, html, **kargs):
902 return self._og_search_property('description', html, fatal=False, **kargs)
904 def _og_search_title(self, html, **kargs):
905 return self._og_search_property('title', html, **kargs)
907 def _og_search_video_url(self, html, name='video url', secure=True, **kargs):
908 regexes = self._og_regexes('video') + self._og_regexes('video:url')
910 regexes = self._og_regexes('video:secure_url') + regexes
911 return self._html_search_regex(regexes, html, name, **kargs)
913 def _og_search_url(self, html, **kargs):
914 return self._og_search_property('url', html, **kargs)
916 def _html_search_meta(self, name, html, display_name=None, fatal=False, **kwargs):
917 if not isinstance(name, (list, tuple)):
919 if display_name is None:
920 display_name = name[0]
921 return self._html_search_regex(
922 [self._meta_regex(n) for n in name],
923 html, display_name, fatal=fatal, group='content', **kwargs)
925 def _dc_search_uploader(self, html):
926 return self._html_search_meta('dc.creator', html, 'uploader')
928 def _rta_search(self, html):
929 # See http://www.rtalabel.org/index.php?content=howtofaq#single
930 if re.search(r'(?ix)<meta\s+name="rating"\s+'
931 r' content="RTA-5042-1996-1400-1577-RTA"',
936 def _media_rating_search(self, html):
937 # See http://www.tjg-designs.com/WP/metadata-code-examples-adding-metadata-to-your-web-pages/
938 rating = self._html_search_meta('rating', html)
950 return RATING_TABLE.get(rating.lower())
952 def _family_friendly_search(self, html):
953 # See http://schema.org/VideoObject
954 family_friendly = self._html_search_meta(
955 'isFamilyFriendly', html, default=None)
957 if not family_friendly:
966 return RATING_TABLE.get(family_friendly.lower())
968 def _twitter_search_player(self, html):
969 return self._html_search_meta('twitter:player', html,
970 'twitter card player')
972 def _search_json_ld(self, html, video_id, expected_type=None, **kwargs):
973 json_ld = self._search_regex(
974 r'(?s)<script[^>]+type=(["\'])application/ld\+json\1[^>]*>(?P<json_ld>.+?)</script>',
975 html, 'JSON-LD', group='json_ld', **kwargs)
976 default = kwargs.get('default', NO_DEFAULT)
978 return default if default is not NO_DEFAULT else {}
979 # JSON-LD may be malformed and thus `fatal` should be respected.
980 # At the same time `default` may be passed that assumes `fatal=False`
981 # for _search_regex. Let's simulate the same behavior here as well.
982 fatal = kwargs.get('fatal', True) if default == NO_DEFAULT else False
983 return self._json_ld(json_ld, video_id, fatal=fatal, expected_type=expected_type)
985 def _json_ld(self, json_ld, video_id, fatal=True, expected_type=None):
986 if isinstance(json_ld, compat_str):
987 json_ld = self._parse_json(json_ld, video_id, fatal=fatal)
991 if not isinstance(json_ld, (list, tuple, dict)):
993 if isinstance(json_ld, dict):
996 def extract_video_object(e):
997 assert e['@type'] == 'VideoObject'
999 'url': e.get('contentUrl'),
1000 'title': unescapeHTML(e.get('name')),
1001 'description': unescapeHTML(e.get('description')),
1002 'thumbnail': e.get('thumbnailUrl') or e.get('thumbnailURL'),
1003 'duration': parse_duration(e.get('duration')),
1004 'timestamp': unified_timestamp(e.get('uploadDate')),
1005 'filesize': float_or_none(e.get('contentSize')),
1006 'tbr': int_or_none(e.get('bitrate')),
1007 'width': int_or_none(e.get('width')),
1008 'height': int_or_none(e.get('height')),
1009 'view_count': int_or_none(e.get('interactionCount')),
1013 if e.get('@context') == 'http://schema.org':
1014 item_type = e.get('@type')
1015 if expected_type is not None and expected_type != item_type:
1017 if item_type in ('TVEpisode', 'Episode'):
1019 'episode': unescapeHTML(e.get('name')),
1020 'episode_number': int_or_none(e.get('episodeNumber')),
1021 'description': unescapeHTML(e.get('description')),
1023 part_of_season = e.get('partOfSeason')
1024 if isinstance(part_of_season, dict) and part_of_season.get('@type') in ('TVSeason', 'Season', 'CreativeWorkSeason'):
1025 info['season_number'] = int_or_none(part_of_season.get('seasonNumber'))
1026 part_of_series = e.get('partOfSeries') or e.get('partOfTVSeries')
1027 if isinstance(part_of_series, dict) and part_of_series.get('@type') in ('TVSeries', 'Series', 'CreativeWorkSeries'):
1028 info['series'] = unescapeHTML(part_of_series.get('name'))
1029 elif item_type == 'Article':
1031 'timestamp': parse_iso8601(e.get('datePublished')),
1032 'title': unescapeHTML(e.get('headline')),
1033 'description': unescapeHTML(e.get('articleBody')),
1035 elif item_type == 'VideoObject':
1036 extract_video_object(e)
1038 video = e.get('video')
1039 if isinstance(video, dict) and video.get('@type') == 'VideoObject':
1040 extract_video_object(video)
1042 return dict((k, v) for k, v in info.items() if v is not None)
1045 def _hidden_inputs(html):
1046 html = re.sub(r'<!--(?:(?!<!--).)*-->', '', html)
1048 for input in re.findall(r'(?i)(<input[^>]+>)', html):
1049 attrs = extract_attributes(input)
1052 if attrs.get('type') not in ('hidden', 'submit'):
1054 name = attrs.get('name') or attrs.get('id')
1055 value = attrs.get('value')
1056 if name and value is not None:
1057 hidden_inputs[name] = value
1058 return hidden_inputs
1060 def _form_hidden_inputs(self, form_id, html):
1061 form = self._search_regex(
1062 r'(?is)<form[^>]+?id=(["\'])%s\1[^>]*>(?P<form>.+?)</form>' % form_id,
1063 html, '%s form' % form_id, group='form')
1064 return self._hidden_inputs(form)
1066 def _sort_formats(self, formats, field_preference=None):
1068 raise ExtractorError('No video formats found')
1071 # Automatically determine tbr when missing based on abr and vbr (improves
1072 # formats sorting in some cases)
1073 if 'tbr' not in f and f.get('abr') is not None and f.get('vbr') is not None:
1074 f['tbr'] = f['abr'] + f['vbr']
1076 def _formats_key(f):
1077 # TODO remove the following workaround
1078 from ..utils import determine_ext
1079 if not f.get('ext') and 'url' in f:
1080 f['ext'] = determine_ext(f['url'])
1082 if isinstance(field_preference, (list, tuple)):
1085 if f.get(field) is not None
1086 else ('' if field == 'format_id' else -1)
1087 for field in field_preference)
1089 preference = f.get('preference')
1090 if preference is None:
1092 if f.get('ext') in ['f4f', 'f4m']: # Not yet supported
1095 protocol = f.get('protocol') or determine_protocol(f)
1096 proto_preference = 0 if protocol in ['http', 'https'] else (-0.5 if protocol == 'rtsp' else -0.1)
1098 if f.get('vcodec') == 'none': # audio only
1100 if self._downloader.params.get('prefer_free_formats'):
1101 ORDER = ['aac', 'mp3', 'm4a', 'webm', 'ogg', 'opus']
1103 ORDER = ['webm', 'opus', 'ogg', 'mp3', 'aac', 'm4a']
1106 audio_ext_preference = ORDER.index(f['ext'])
1108 audio_ext_preference = -1
1110 if f.get('acodec') == 'none': # video only
1112 if self._downloader.params.get('prefer_free_formats'):
1113 ORDER = ['flv', 'mp4', 'webm']
1115 ORDER = ['webm', 'flv', 'mp4']
1117 ext_preference = ORDER.index(f['ext'])
1120 audio_ext_preference = 0
1124 f.get('language_preference') if f.get('language_preference') is not None else -1,
1125 f.get('quality') if f.get('quality') is not None else -1,
1126 f.get('tbr') if f.get('tbr') is not None else -1,
1127 f.get('filesize') if f.get('filesize') is not None else -1,
1128 f.get('vbr') if f.get('vbr') is not None else -1,
1129 f.get('height') if f.get('height') is not None else -1,
1130 f.get('width') if f.get('width') is not None else -1,
1133 f.get('abr') if f.get('abr') is not None else -1,
1134 audio_ext_preference,
1135 f.get('fps') if f.get('fps') is not None else -1,
1136 f.get('filesize_approx') if f.get('filesize_approx') is not None else -1,
1137 f.get('source_preference') if f.get('source_preference') is not None else -1,
1138 f.get('format_id') if f.get('format_id') is not None else '',
1140 formats.sort(key=_formats_key)
1142 def _check_formats(self, formats, video_id):
1144 formats[:] = filter(
1145 lambda f: self._is_valid_url(
1147 item='%s video format' % f.get('format_id') if f.get('format_id') else 'video'),
1151 def _remove_duplicate_formats(formats):
1155 if f['url'] not in format_urls:
1156 format_urls.add(f['url'])
1157 unique_formats.append(f)
1158 formats[:] = unique_formats
1160 def _is_valid_url(self, url, video_id, item='video', headers={}):
1161 url = self._proto_relative_url(url, scheme='http:')
1162 # For now assume non HTTP(S) URLs always valid
1163 if not (url.startswith('http://') or url.startswith('https://')):
1166 self._request_webpage(url, video_id, 'Checking %s URL' % item, headers=headers)
1168 except ExtractorError as e:
1169 if isinstance(e.cause, compat_urllib_error.URLError):
1171 '%s: %s URL is invalid, skipping' % (video_id, item))
1175 def http_scheme(self):
1176 """ Either "http:" or "https:", depending on the user's preferences """
1179 if self._downloader.params.get('prefer_insecure', False)
1182 def _proto_relative_url(self, url, scheme=None):
1185 if url.startswith('//'):
1187 scheme = self.http_scheme()
1192 def _sleep(self, timeout, video_id, msg_template=None):
1193 if msg_template is None:
1194 msg_template = '%(video_id)s: Waiting for %(timeout)s seconds'
1195 msg = msg_template % {'video_id': video_id, 'timeout': timeout}
1199 def _extract_f4m_formats(self, manifest_url, video_id, preference=None, f4m_id=None,
1200 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1201 fatal=True, m3u8_id=None):
1202 manifest = self._download_xml(
1203 manifest_url, video_id, 'Downloading f4m manifest',
1204 'Unable to download f4m manifest',
1205 # Some manifests may be malformed, e.g. prosiebensat1 generated manifests
1206 # (see https://github.com/rg3/youtube-dl/issues/6215#issuecomment-121704244)
1207 transform_source=transform_source,
1210 if manifest is False:
1213 return self._parse_f4m_formats(
1214 manifest, manifest_url, video_id, preference=preference, f4m_id=f4m_id,
1215 transform_source=transform_source, fatal=fatal, m3u8_id=m3u8_id)
1217 def _parse_f4m_formats(self, manifest, manifest_url, video_id, preference=None, f4m_id=None,
1218 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1219 fatal=True, m3u8_id=None):
1220 # currently youtube-dl cannot decode the playerVerificationChallenge as Akamai uses Adobe Alchemy
1221 akamai_pv = manifest.find('{http://ns.adobe.com/f4m/1.0}pv-2.0')
1222 if akamai_pv is not None and ';' in akamai_pv.text:
1223 playerVerificationChallenge = akamai_pv.text.split(';')[0]
1224 if playerVerificationChallenge.strip() != '':
1228 manifest_version = '1.0'
1229 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/1.0}media')
1231 manifest_version = '2.0'
1232 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/2.0}media')
1233 # Remove unsupported DRM protected media from final formats
1234 # rendition (see https://github.com/rg3/youtube-dl/issues/8573).
1235 media_nodes = remove_encrypted_media(media_nodes)
1239 manifest_base_url = get_base_url(manifest)
1241 bootstrap_info = xpath_element(
1242 manifest, ['{http://ns.adobe.com/f4m/1.0}bootstrapInfo', '{http://ns.adobe.com/f4m/2.0}bootstrapInfo'],
1243 'bootstrap info', default=None)
1246 mime_type = xpath_text(
1247 manifest, ['{http://ns.adobe.com/f4m/1.0}mimeType', '{http://ns.adobe.com/f4m/2.0}mimeType'],
1248 'base URL', default=None)
1249 if mime_type and mime_type.startswith('audio/'):
1252 for i, media_el in enumerate(media_nodes):
1253 tbr = int_or_none(media_el.attrib.get('bitrate'))
1254 width = int_or_none(media_el.attrib.get('width'))
1255 height = int_or_none(media_el.attrib.get('height'))
1256 format_id = '-'.join(filter(None, [f4m_id, compat_str(i if tbr is None else tbr)]))
1257 # If <bootstrapInfo> is present, the specified f4m is a
1258 # stream-level manifest, and only set-level manifests may refer to
1259 # external resources. See section 11.4 and section 4 of F4M spec
1260 if bootstrap_info is None:
1262 # @href is introduced in 2.0, see section 11.6 of F4M spec
1263 if manifest_version == '2.0':
1264 media_url = media_el.attrib.get('href')
1265 if media_url is None:
1266 media_url = media_el.attrib.get('url')
1270 media_url if media_url.startswith('http://') or media_url.startswith('https://')
1271 else ((manifest_base_url or '/'.join(manifest_url.split('/')[:-1])) + '/' + media_url))
1272 # If media_url is itself a f4m manifest do the recursive extraction
1273 # since bitrates in parent manifest (this one) and media_url manifest
1274 # may differ leading to inability to resolve the format by requested
1275 # bitrate in f4m downloader
1276 ext = determine_ext(manifest_url)
1278 f4m_formats = self._extract_f4m_formats(
1279 manifest_url, video_id, preference=preference, f4m_id=f4m_id,
1280 transform_source=transform_source, fatal=fatal)
1281 # Sometimes stream-level manifest contains single media entry that
1282 # does not contain any quality metadata (e.g. http://matchtv.ru/#live-player).
1283 # At the same time parent's media entry in set-level manifest may
1284 # contain it. We will copy it from parent in such cases.
1285 if len(f4m_formats) == 1:
1288 'tbr': f.get('tbr') or tbr,
1289 'width': f.get('width') or width,
1290 'height': f.get('height') or height,
1291 'format_id': f.get('format_id') if not tbr else format_id,
1294 formats.extend(f4m_formats)
1297 formats.extend(self._extract_m3u8_formats(
1298 manifest_url, video_id, 'mp4', preference=preference,
1299 m3u8_id=m3u8_id, fatal=fatal))
1302 'format_id': format_id,
1303 'url': manifest_url,
1304 'manifest_url': manifest_url,
1305 'ext': 'flv' if bootstrap_info is not None else None,
1311 'preference': preference,
1315 def _m3u8_meta_format(self, m3u8_url, ext=None, preference=None, m3u8_id=None):
1317 'format_id': '-'.join(filter(None, [m3u8_id, 'meta'])),
1321 'preference': preference - 100 if preference else -100,
1322 'resolution': 'multiple',
1323 'format_note': 'Quality selection URL',
1326 def _extract_m3u8_formats(self, m3u8_url, video_id, ext=None,
1327 entry_protocol='m3u8', preference=None,
1328 m3u8_id=None, note=None, errnote=None,
1329 fatal=True, live=False):
1330 res = self._download_webpage_handle(
1332 note=note or 'Downloading m3u8 information',
1333 errnote=errnote or 'Failed to download m3u8 information',
1339 m3u8_doc, urlh = res
1340 m3u8_url = urlh.geturl()
1342 return self._parse_m3u8_formats(
1343 m3u8_doc, m3u8_url, ext=ext, entry_protocol=entry_protocol,
1344 preference=preference, m3u8_id=m3u8_id, live=live)
1346 def _parse_m3u8_formats(self, m3u8_doc, m3u8_url, ext=None,
1347 entry_protocol='m3u8', preference=None,
1348 m3u8_id=None, live=False):
1349 if '#EXT-X-FAXS-CM:' in m3u8_doc: # Adobe Flash Access
1352 if re.search(r'#EXT-X-SESSION-KEY:.*?URI="skd://', m3u8_doc): # Apple FairPlay
1357 format_url = lambda u: (
1359 if re.match(r'^https?://', u)
1360 else compat_urlparse.urljoin(m3u8_url, u))
1363 # 1. https://tools.ietf.org/html/draft-pantos-http-live-streaming-21
1364 # 2. https://github.com/rg3/youtube-dl/issues/12211
1366 # We should try extracting formats only from master playlists [1, 4.3.4],
1367 # i.e. playlists that describe available qualities. On the other hand
1368 # media playlists [1, 4.3.3] should be returned as is since they contain
1369 # just the media without qualities renditions.
1370 # Fortunately, master playlist can be easily distinguished from media
1371 # playlist based on particular tags availability. As of [1, 4.3.3, 4.3.4]
1372 # master playlist tags MUST NOT appear in a media playist and vice versa.
1373 # As of [1, 4.3.3.1] #EXT-X-TARGETDURATION tag is REQUIRED for every
1374 # media playlist and MUST NOT appear in master playlist thus we can
1375 # clearly detect media playlist with this criterion.
1377 if '#EXT-X-TARGETDURATION' in m3u8_doc: # media playlist, return as is
1380 'format_id': m3u8_id,
1382 'protocol': entry_protocol,
1383 'preference': preference,
1387 last_stream_inf = {}
1389 def extract_media(x_media_line):
1390 media = parse_m3u8_attributes(x_media_line)
1391 # As per [1, 4.3.4.1] TYPE, GROUP-ID and NAME are REQUIRED
1392 media_type, group_id, name = media.get('TYPE'), media.get('GROUP-ID'), media.get('NAME')
1393 if not (media_type and group_id and name):
1395 groups.setdefault(group_id, []).append(media)
1396 if media_type not in ('VIDEO', 'AUDIO'):
1398 media_url = media.get('URI')
1401 for v in (m3u8_id, group_id, name):
1405 'format_id': '-'.join(format_id),
1406 'url': format_url(media_url),
1407 'manifest_url': m3u8_url,
1408 'language': media.get('LANGUAGE'),
1410 'protocol': entry_protocol,
1411 'preference': preference,
1413 if media_type == 'AUDIO':
1414 f['vcodec'] = 'none'
1417 def build_stream_name():
1418 # Despite specification does not mention NAME attribute for
1419 # EXT-X-STREAM-INF tag it still sometimes may be present (see [1]
1420 # or vidio test in TestInfoExtractor.test_parse_m3u8_formats)
1421 # 1. http://www.vidio.com/watch/165683-dj_ambred-booyah-live-2015
1422 stream_name = last_stream_inf.get('NAME')
1425 # If there is no NAME in EXT-X-STREAM-INF it will be obtained
1426 # from corresponding rendition group
1427 stream_group_id = last_stream_inf.get('VIDEO')
1428 if not stream_group_id:
1430 stream_group = groups.get(stream_group_id)
1431 if not stream_group:
1432 return stream_group_id
1433 rendition = stream_group[0]
1434 return rendition.get('NAME') or stream_group_id
1436 for line in m3u8_doc.splitlines():
1437 if line.startswith('#EXT-X-STREAM-INF:'):
1438 last_stream_inf = parse_m3u8_attributes(line)
1439 elif line.startswith('#EXT-X-MEDIA:'):
1441 elif line.startswith('#') or not line.strip():
1444 tbr = float_or_none(
1445 last_stream_inf.get('AVERAGE-BANDWIDTH') or
1446 last_stream_inf.get('BANDWIDTH'), scale=1000)
1449 format_id.append(m3u8_id)
1450 stream_name = build_stream_name()
1451 # Bandwidth of live streams may differ over time thus making
1452 # format_id unpredictable. So it's better to keep provided
1455 format_id.append(stream_name if stream_name else '%d' % (tbr if tbr else len(formats)))
1456 manifest_url = format_url(line.strip())
1458 'format_id': '-'.join(format_id),
1459 'url': manifest_url,
1460 'manifest_url': m3u8_url,
1463 'fps': float_or_none(last_stream_inf.get('FRAME-RATE')),
1464 'protocol': entry_protocol,
1465 'preference': preference,
1467 resolution = last_stream_inf.get('RESOLUTION')
1469 mobj = re.search(r'(?P<width>\d+)[xX](?P<height>\d+)', resolution)
1471 f['width'] = int(mobj.group('width'))
1472 f['height'] = int(mobj.group('height'))
1473 # Unified Streaming Platform
1475 r'audio.*?(?:%3D|=)(\d+)(?:-video.*?(?:%3D|=)(\d+))?', f['url'])
1477 abr, vbr = mobj.groups()
1478 abr, vbr = float_or_none(abr, 1000), float_or_none(vbr, 1000)
1483 codecs = parse_codecs(last_stream_inf.get('CODECS'))
1485 audio_group_id = last_stream_inf.get('AUDIO')
1486 # As per [1, 4.3.4.1.1] any EXT-X-STREAM-INF tag which
1487 # references a rendition group MUST have a CODECS attribute.
1488 # However, this is not always respected, for example, [2]
1489 # contains EXT-X-STREAM-INF tag which references AUDIO
1490 # rendition group but does not have CODECS and despite
1491 # referencing audio group an audio group, it represents
1492 # a complete (with audio and video) format. So, for such cases
1493 # we will ignore references to rendition groups and treat them
1494 # as complete formats.
1495 if audio_group_id and codecs and f.get('vcodec') != 'none':
1496 audio_group = groups.get(audio_group_id)
1497 if audio_group and audio_group[0].get('URI'):
1498 # TODO: update acodec for audio only formats with
1500 f['acodec'] = 'none'
1502 last_stream_inf = {}
1506 def _xpath_ns(path, namespace=None):
1510 for c in path.split('/'):
1511 if not c or c == '.':
1514 out.append('{%s}%s' % (namespace, c))
1515 return '/'.join(out)
1517 def _extract_smil_formats(self, smil_url, video_id, fatal=True, f4m_params=None, transform_source=None):
1518 smil = self._download_smil(smil_url, video_id, fatal=fatal, transform_source=transform_source)
1524 namespace = self._parse_smil_namespace(smil)
1526 return self._parse_smil_formats(
1527 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
1529 def _extract_smil_info(self, smil_url, video_id, fatal=True, f4m_params=None):
1530 smil = self._download_smil(smil_url, video_id, fatal=fatal)
1533 return self._parse_smil(smil, smil_url, video_id, f4m_params=f4m_params)
1535 def _download_smil(self, smil_url, video_id, fatal=True, transform_source=None):
1536 return self._download_xml(
1537 smil_url, video_id, 'Downloading SMIL file',
1538 'Unable to download SMIL file', fatal=fatal, transform_source=transform_source)
1540 def _parse_smil(self, smil, smil_url, video_id, f4m_params=None):
1541 namespace = self._parse_smil_namespace(smil)
1543 formats = self._parse_smil_formats(
1544 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
1545 subtitles = self._parse_smil_subtitles(smil, namespace=namespace)
1547 video_id = os.path.splitext(url_basename(smil_url))[0]
1551 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
1552 name = meta.attrib.get('name')
1553 content = meta.attrib.get('content')
1554 if not name or not content:
1556 if not title and name == 'title':
1558 elif not description and name in ('description', 'abstract'):
1559 description = content
1560 elif not upload_date and name == 'date':
1561 upload_date = unified_strdate(content)
1564 'id': image.get('type'),
1565 'url': image.get('src'),
1566 'width': int_or_none(image.get('width')),
1567 'height': int_or_none(image.get('height')),
1568 } for image in smil.findall(self._xpath_ns('.//image', namespace)) if image.get('src')]
1572 'title': title or video_id,
1573 'description': description,
1574 'upload_date': upload_date,
1575 'thumbnails': thumbnails,
1577 'subtitles': subtitles,
1580 def _parse_smil_namespace(self, smil):
1581 return self._search_regex(
1582 r'(?i)^{([^}]+)?}smil$', smil.tag, 'namespace', default=None)
1584 def _parse_smil_formats(self, smil, smil_url, video_id, namespace=None, f4m_params=None, transform_rtmp_url=None):
1586 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
1587 b = meta.get('base') or meta.get('httpBase')
1598 media = smil.findall(self._xpath_ns('.//video', namespace)) + smil.findall(self._xpath_ns('.//audio', namespace))
1599 for medium in media:
1600 src = medium.get('src')
1601 if not src or src in srcs:
1605 bitrate = float_or_none(medium.get('system-bitrate') or medium.get('systemBitrate'), 1000)
1606 filesize = int_or_none(medium.get('size') or medium.get('fileSize'))
1607 width = int_or_none(medium.get('width'))
1608 height = int_or_none(medium.get('height'))
1609 proto = medium.get('proto')
1610 ext = medium.get('ext')
1611 src_ext = determine_ext(src)
1612 streamer = medium.get('streamer') or base
1614 if proto == 'rtmp' or streamer.startswith('rtmp'):
1620 'format_id': 'rtmp-%d' % (rtmp_count if bitrate is None else bitrate),
1622 'filesize': filesize,
1626 if transform_rtmp_url:
1627 streamer, src = transform_rtmp_url(streamer, src)
1628 formats[-1].update({
1634 src_url = src if src.startswith('http') else compat_urlparse.urljoin(base, src)
1635 src_url = src_url.strip()
1637 if proto == 'm3u8' or src_ext == 'm3u8':
1638 m3u8_formats = self._extract_m3u8_formats(
1639 src_url, video_id, ext or 'mp4', m3u8_id='hls', fatal=False)
1640 if len(m3u8_formats) == 1:
1642 m3u8_formats[0].update({
1643 'format_id': 'hls-%d' % (m3u8_count if bitrate is None else bitrate),
1648 formats.extend(m3u8_formats)
1651 if src_ext == 'f4m':
1656 'plugin': 'flowplayer-3.2.0.1',
1658 f4m_url += '&' if '?' in f4m_url else '?'
1659 f4m_url += compat_urllib_parse_urlencode(f4m_params)
1660 formats.extend(self._extract_f4m_formats(f4m_url, video_id, f4m_id='hds', fatal=False))
1663 if src_url.startswith('http') and self._is_valid_url(src, video_id):
1667 'ext': ext or src_ext or 'flv',
1668 'format_id': 'http-%d' % (bitrate or http_count),
1670 'filesize': filesize,
1678 def _parse_smil_subtitles(self, smil, namespace=None, subtitles_lang='en'):
1681 for num, textstream in enumerate(smil.findall(self._xpath_ns('.//textstream', namespace))):
1682 src = textstream.get('src')
1683 if not src or src in urls:
1686 ext = textstream.get('ext') or mimetype2ext(textstream.get('type')) or determine_ext(src)
1687 lang = textstream.get('systemLanguage') or textstream.get('systemLanguageName') or textstream.get('lang') or subtitles_lang
1688 subtitles.setdefault(lang, []).append({
1694 def _extract_xspf_playlist(self, playlist_url, playlist_id, fatal=True):
1695 xspf = self._download_xml(
1696 playlist_url, playlist_id, 'Downloading xpsf playlist',
1697 'Unable to download xspf manifest', fatal=fatal)
1700 return self._parse_xspf(xspf, playlist_id)
1702 def _parse_xspf(self, playlist, playlist_id):
1704 'xspf': 'http://xspf.org/ns/0/',
1705 's1': 'http://static.streamone.nl/player/ns/0',
1709 for track in playlist.findall(xpath_with_ns('./xspf:trackList/xspf:track', NS_MAP)):
1711 track, xpath_with_ns('./xspf:title', NS_MAP), 'title', default=playlist_id)
1712 description = xpath_text(
1713 track, xpath_with_ns('./xspf:annotation', NS_MAP), 'description')
1714 thumbnail = xpath_text(
1715 track, xpath_with_ns('./xspf:image', NS_MAP), 'thumbnail')
1716 duration = float_or_none(
1717 xpath_text(track, xpath_with_ns('./xspf:duration', NS_MAP), 'duration'), 1000)
1720 'url': location.text,
1721 'format_id': location.get(xpath_with_ns('s1:label', NS_MAP)),
1722 'width': int_or_none(location.get(xpath_with_ns('s1:width', NS_MAP))),
1723 'height': int_or_none(location.get(xpath_with_ns('s1:height', NS_MAP))),
1724 } for location in track.findall(xpath_with_ns('./xspf:location', NS_MAP))]
1725 self._sort_formats(formats)
1730 'description': description,
1731 'thumbnail': thumbnail,
1732 'duration': duration,
1737 def _extract_mpd_formats(self, mpd_url, video_id, mpd_id=None, note=None, errnote=None, fatal=True, formats_dict={}):
1738 res = self._download_webpage_handle(
1740 note=note or 'Downloading MPD manifest',
1741 errnote=errnote or 'Failed to download MPD manifest',
1746 mpd_base_url = base_url(urlh.geturl())
1748 return self._parse_mpd_formats(
1749 compat_etree_fromstring(mpd.encode('utf-8')), mpd_id, mpd_base_url,
1750 formats_dict=formats_dict, mpd_url=mpd_url)
1752 def _parse_mpd_formats(self, mpd_doc, mpd_id=None, mpd_base_url='', formats_dict={}, mpd_url=None):
1754 Parse formats from MPD manifest.
1756 1. MPEG-DASH Standard, ISO/IEC 23009-1:2014(E),
1757 http://standards.iso.org/ittf/PubliclyAvailableStandards/c065274_ISO_IEC_23009-1_2014.zip
1758 2. https://en.wikipedia.org/wiki/Dynamic_Adaptive_Streaming_over_HTTP
1760 if mpd_doc.get('type') == 'dynamic':
1763 namespace = self._search_regex(r'(?i)^{([^}]+)?}MPD$', mpd_doc.tag, 'namespace', default=None)
1766 return self._xpath_ns(path, namespace)
1768 def is_drm_protected(element):
1769 return element.find(_add_ns('ContentProtection')) is not None
1771 def extract_multisegment_info(element, ms_parent_info):
1772 ms_info = ms_parent_info.copy()
1774 # As per [1, 5.3.9.2.2] SegmentList and SegmentTemplate share some
1775 # common attributes and elements. We will only extract relevant
1777 def extract_common(source):
1778 segment_timeline = source.find(_add_ns('SegmentTimeline'))
1779 if segment_timeline is not None:
1780 s_e = segment_timeline.findall(_add_ns('S'))
1782 ms_info['total_number'] = 0
1785 r = int(s.get('r', 0))
1786 ms_info['total_number'] += 1 + r
1787 ms_info['s'].append({
1788 't': int(s.get('t', 0)),
1789 # @d is mandatory (see [1, 5.3.9.6.2, Table 17, page 60])
1790 'd': int(s.attrib['d']),
1793 start_number = source.get('startNumber')
1795 ms_info['start_number'] = int(start_number)
1796 timescale = source.get('timescale')
1798 ms_info['timescale'] = int(timescale)
1799 segment_duration = source.get('duration')
1800 if segment_duration:
1801 ms_info['segment_duration'] = float(segment_duration)
1803 def extract_Initialization(source):
1804 initialization = source.find(_add_ns('Initialization'))
1805 if initialization is not None:
1806 ms_info['initialization_url'] = initialization.attrib['sourceURL']
1808 segment_list = element.find(_add_ns('SegmentList'))
1809 if segment_list is not None:
1810 extract_common(segment_list)
1811 extract_Initialization(segment_list)
1812 segment_urls_e = segment_list.findall(_add_ns('SegmentURL'))
1814 ms_info['segment_urls'] = [segment.attrib['media'] for segment in segment_urls_e]
1816 segment_template = element.find(_add_ns('SegmentTemplate'))
1817 if segment_template is not None:
1818 extract_common(segment_template)
1819 media = segment_template.get('media')
1821 ms_info['media'] = media
1822 initialization = segment_template.get('initialization')
1824 ms_info['initialization'] = initialization
1826 extract_Initialization(segment_template)
1829 mpd_duration = parse_duration(mpd_doc.get('mediaPresentationDuration'))
1831 for period in mpd_doc.findall(_add_ns('Period')):
1832 period_duration = parse_duration(period.get('duration')) or mpd_duration
1833 period_ms_info = extract_multisegment_info(period, {
1837 for adaptation_set in period.findall(_add_ns('AdaptationSet')):
1838 if is_drm_protected(adaptation_set):
1840 adaption_set_ms_info = extract_multisegment_info(adaptation_set, period_ms_info)
1841 for representation in adaptation_set.findall(_add_ns('Representation')):
1842 if is_drm_protected(representation):
1844 representation_attrib = adaptation_set.attrib.copy()
1845 representation_attrib.update(representation.attrib)
1846 # According to [1, 5.3.7.2, Table 9, page 41], @mimeType is mandatory
1847 mime_type = representation_attrib['mimeType']
1848 content_type = mime_type.split('/')[0]
1849 if content_type == 'text':
1850 # TODO implement WebVTT downloading
1852 elif content_type in ('video', 'audio'):
1854 for element in (representation, adaptation_set, period, mpd_doc):
1855 base_url_e = element.find(_add_ns('BaseURL'))
1856 if base_url_e is not None:
1857 base_url = base_url_e.text + base_url
1858 if re.match(r'^https?://', base_url):
1860 if mpd_base_url and not re.match(r'^https?://', base_url):
1861 if not mpd_base_url.endswith('/') and not base_url.startswith('/'):
1863 base_url = mpd_base_url + base_url
1864 representation_id = representation_attrib.get('id')
1865 lang = representation_attrib.get('lang')
1866 url_el = representation.find(_add_ns('BaseURL'))
1867 filesize = int_or_none(url_el.attrib.get('{http://youtube.com/yt/2012/10/10}contentLength') if url_el is not None else None)
1868 bandwidth = int_or_none(representation_attrib.get('bandwidth'))
1870 'format_id': '%s-%s' % (mpd_id, representation_id) if mpd_id else representation_id,
1872 'manifest_url': mpd_url,
1873 'ext': mimetype2ext(mime_type),
1874 'width': int_or_none(representation_attrib.get('width')),
1875 'height': int_or_none(representation_attrib.get('height')),
1876 'tbr': float_or_none(bandwidth, 1000),
1877 'asr': int_or_none(representation_attrib.get('audioSamplingRate')),
1878 'fps': int_or_none(representation_attrib.get('frameRate')),
1879 'language': lang if lang not in ('mul', 'und', 'zxx', 'mis') else None,
1880 'format_note': 'DASH %s' % content_type,
1881 'filesize': filesize,
1883 f.update(parse_codecs(representation_attrib.get('codecs')))
1884 representation_ms_info = extract_multisegment_info(representation, adaption_set_ms_info)
1886 def prepare_template(template_name, identifiers):
1887 t = representation_ms_info[template_name]
1888 t = t.replace('$RepresentationID$', representation_id)
1889 t = re.sub(r'\$(%s)\$' % '|'.join(identifiers), r'%(\1)d', t)
1890 t = re.sub(r'\$(%s)%%([^$]+)\$' % '|'.join(identifiers), r'%(\1)\2', t)
1891 t.replace('$$', '$')
1894 # @initialization is a regular template like @media one
1895 # so it should be handled just the same way (see
1896 # https://github.com/rg3/youtube-dl/issues/11605)
1897 if 'initialization' in representation_ms_info:
1898 initialization_template = prepare_template(
1900 # As per [1, 5.3.9.4.2, Table 15, page 54] $Number$ and
1901 # $Time$ shall not be included for @initialization thus
1902 # only $Bandwidth$ remains
1904 representation_ms_info['initialization_url'] = initialization_template % {
1905 'Bandwidth': bandwidth,
1908 def location_key(location):
1909 return 'url' if re.match(r'^https?://', location) else 'path'
1911 if 'segment_urls' not in representation_ms_info and 'media' in representation_ms_info:
1913 media_template = prepare_template('media', ('Number', 'Bandwidth', 'Time'))
1914 media_location_key = location_key(media_template)
1916 # As per [1, 5.3.9.4.4, Table 16, page 55] $Number$ and $Time$
1917 # can't be used at the same time
1918 if '%(Number' in media_template and 's' not in representation_ms_info:
1919 segment_duration = None
1920 if 'total_number' not in representation_ms_info and 'segment_duration' in representation_ms_info:
1921 segment_duration = float_or_none(representation_ms_info['segment_duration'], representation_ms_info['timescale'])
1922 representation_ms_info['total_number'] = int(math.ceil(float(period_duration) / segment_duration))
1923 representation_ms_info['fragments'] = [{
1924 media_location_key: media_template % {
1925 'Number': segment_number,
1926 'Bandwidth': bandwidth,
1928 'duration': segment_duration,
1929 } for segment_number in range(
1930 representation_ms_info['start_number'],
1931 representation_ms_info['total_number'] + representation_ms_info['start_number'])]
1933 # $Number*$ or $Time$ in media template with S list available
1934 # Example $Number*$: http://www.svtplay.se/klipp/9023742/stopptid-om-bjorn-borg
1935 # Example $Time$: https://play.arkena.com/embed/avp/v2/player/media/b41dda37-d8e7-4d3f-b1b5-9a9db578bdfe/1/129411
1936 representation_ms_info['fragments'] = []
1939 segment_number = representation_ms_info['start_number']
1941 def add_segment_url():
1942 segment_url = media_template % {
1943 'Time': segment_time,
1944 'Bandwidth': bandwidth,
1945 'Number': segment_number,
1947 representation_ms_info['fragments'].append({
1948 media_location_key: segment_url,
1949 'duration': float_or_none(segment_d, representation_ms_info['timescale']),
1952 for num, s in enumerate(representation_ms_info['s']):
1953 segment_time = s.get('t') or segment_time
1957 for r in range(s.get('r', 0)):
1958 segment_time += segment_d
1961 segment_time += segment_d
1962 elif 'segment_urls' in representation_ms_info and 's' in representation_ms_info:
1964 # Example: https://www.youtube.com/watch?v=iXZV5uAYMJI
1965 # or any YouTube dashsegments video
1968 timescale = representation_ms_info['timescale']
1969 for s in representation_ms_info['s']:
1970 duration = float_or_none(s['d'], timescale)
1971 for r in range(s.get('r', 0) + 1):
1972 segment_uri = representation_ms_info['segment_urls'][segment_index]
1974 location_key(segment_uri): segment_uri,
1975 'duration': duration,
1978 representation_ms_info['fragments'] = fragments
1979 elif 'segment_urls' in representation_ms_info:
1980 # Segment URLs with no SegmentTimeline
1981 # Example: https://www.seznam.cz/zpravy/clanek/cesko-zasahne-vitr-o-sile-vichrice-muze-byt-i-zivotu-nebezpecny-39091
1982 # https://github.com/rg3/youtube-dl/pull/14844
1984 segment_duration = float_or_none(
1985 representation_ms_info['segment_duration'],
1986 representation_ms_info['timescale']) if 'segment_duration' in representation_ms_info else None
1987 for segment_url in representation_ms_info['segment_urls']:
1989 location_key(segment_url): segment_url,
1991 if segment_duration:
1992 fragment['duration'] = segment_duration
1993 fragments.append(fragment)
1994 representation_ms_info['fragments'] = fragments
1995 # NB: MPD manifest may contain direct URLs to unfragmented media.
1996 # No fragments key is present in this case.
1997 if 'fragments' in representation_ms_info:
1999 'fragment_base_url': base_url,
2001 'protocol': 'http_dash_segments',
2003 if 'initialization_url' in representation_ms_info:
2004 initialization_url = representation_ms_info['initialization_url']
2005 if not f.get('url'):
2006 f['url'] = initialization_url
2007 f['fragments'].append({location_key(initialization_url): initialization_url})
2008 f['fragments'].extend(representation_ms_info['fragments'])
2010 existing_format = next(
2011 fo for fo in formats
2012 if fo['format_id'] == representation_id)
2013 except StopIteration:
2014 full_info = formats_dict.get(representation_id, {}).copy()
2016 formats.append(full_info)
2018 existing_format.update(f)
2020 self.report_warning('Unknown MIME type %s in DASH manifest' % mime_type)
2023 def _extract_ism_formats(self, ism_url, video_id, ism_id=None, note=None, errnote=None, fatal=True):
2024 res = self._download_webpage_handle(
2026 note=note or 'Downloading ISM manifest',
2027 errnote=errnote or 'Failed to download ISM manifest',
2033 return self._parse_ism_formats(
2034 compat_etree_fromstring(ism.encode('utf-8')), urlh.geturl(), ism_id)
2036 def _parse_ism_formats(self, ism_doc, ism_url, ism_id=None):
2038 Parse formats from ISM manifest.
2040 1. [MS-SSTR]: Smooth Streaming Protocol,
2041 https://msdn.microsoft.com/en-us/library/ff469518.aspx
2043 if ism_doc.get('IsLive') == 'TRUE' or ism_doc.find('Protection') is not None:
2046 duration = int(ism_doc.attrib['Duration'])
2047 timescale = int_or_none(ism_doc.get('TimeScale')) or 10000000
2050 for stream in ism_doc.findall('StreamIndex'):
2051 stream_type = stream.get('Type')
2052 if stream_type not in ('video', 'audio'):
2054 url_pattern = stream.attrib['Url']
2055 stream_timescale = int_or_none(stream.get('TimeScale')) or timescale
2056 stream_name = stream.get('Name')
2057 for track in stream.findall('QualityLevel'):
2058 fourcc = track.get('FourCC')
2059 # TODO: add support for WVC1 and WMAP
2060 if fourcc not in ('H264', 'AVC1', 'AACL'):
2061 self.report_warning('%s is not a supported codec' % fourcc)
2063 tbr = int(track.attrib['Bitrate']) // 1000
2064 # [1] does not mention Width and Height attributes. However,
2065 # they're often present while MaxWidth and MaxHeight are
2066 # missing, so should be used as fallbacks
2067 width = int_or_none(track.get('MaxWidth') or track.get('Width'))
2068 height = int_or_none(track.get('MaxHeight') or track.get('Height'))
2069 sampling_rate = int_or_none(track.get('SamplingRate'))
2071 track_url_pattern = re.sub(r'{[Bb]itrate}', track.attrib['Bitrate'], url_pattern)
2072 track_url_pattern = compat_urlparse.urljoin(ism_url, track_url_pattern)
2078 stream_fragments = stream.findall('c')
2079 for stream_fragment_index, stream_fragment in enumerate(stream_fragments):
2080 fragment_ctx['time'] = int_or_none(stream_fragment.get('t')) or fragment_ctx['time']
2081 fragment_repeat = int_or_none(stream_fragment.get('r')) or 1
2082 fragment_ctx['duration'] = int_or_none(stream_fragment.get('d'))
2083 if not fragment_ctx['duration']:
2085 next_fragment_time = int(stream_fragment[stream_fragment_index + 1].attrib['t'])
2087 next_fragment_time = duration
2088 fragment_ctx['duration'] = (next_fragment_time - fragment_ctx['time']) / fragment_repeat
2089 for _ in range(fragment_repeat):
2091 'url': re.sub(r'{start[ _]time}', compat_str(fragment_ctx['time']), track_url_pattern),
2092 'duration': fragment_ctx['duration'] / stream_timescale,
2094 fragment_ctx['time'] += fragment_ctx['duration']
2098 format_id.append(ism_id)
2100 format_id.append(stream_name)
2101 format_id.append(compat_str(tbr))
2104 'format_id': '-'.join(format_id),
2106 'manifest_url': ism_url,
2107 'ext': 'ismv' if stream_type == 'video' else 'isma',
2111 'asr': sampling_rate,
2112 'vcodec': 'none' if stream_type == 'audio' else fourcc,
2113 'acodec': 'none' if stream_type == 'video' else fourcc,
2115 'fragments': fragments,
2116 '_download_params': {
2117 'duration': duration,
2118 'timescale': stream_timescale,
2119 'width': width or 0,
2120 'height': height or 0,
2122 'codec_private_data': track.get('CodecPrivateData'),
2123 'sampling_rate': sampling_rate,
2124 'channels': int_or_none(track.get('Channels', 2)),
2125 'bits_per_sample': int_or_none(track.get('BitsPerSample', 16)),
2126 'nal_unit_length_field': int_or_none(track.get('NALUnitLengthField', 4)),
2131 def _parse_html5_media_entries(self, base_url, webpage, video_id, m3u8_id=None, m3u8_entry_protocol='m3u8', mpd_id=None, preference=None):
2132 def absolute_url(video_url):
2133 return compat_urlparse.urljoin(base_url, video_url)
2135 def parse_content_type(content_type):
2136 if not content_type:
2138 ctr = re.search(r'(?P<mimetype>[^/]+/[^;]+)(?:;\s*codecs="?(?P<codecs>[^"]+))?', content_type)
2140 mimetype, codecs = ctr.groups()
2141 f = parse_codecs(codecs)
2142 f['ext'] = mimetype2ext(mimetype)
2146 def _media_formats(src, cur_media_type, type_info={}):
2147 full_url = absolute_url(src)
2148 ext = type_info.get('ext') or determine_ext(full_url)
2150 is_plain_url = False
2151 formats = self._extract_m3u8_formats(
2152 full_url, video_id, ext='mp4',
2153 entry_protocol=m3u8_entry_protocol, m3u8_id=m3u8_id,
2154 preference=preference, fatal=False)
2156 is_plain_url = False
2157 formats = self._extract_mpd_formats(
2158 full_url, video_id, mpd_id=mpd_id, fatal=False)
2163 'vcodec': 'none' if cur_media_type == 'audio' else None,
2165 return is_plain_url, formats
2168 # amp-video and amp-audio are very similar to their HTML5 counterparts
2169 # so we wll include them right here (see
2170 # https://www.ampproject.org/docs/reference/components/amp-video)
2171 media_tags = [(media_tag, media_type, '')
2172 for media_tag, media_type
2173 in re.findall(r'(?s)(<(?:amp-)?(video|audio)[^>]*/>)', webpage)]
2174 media_tags.extend(re.findall(
2175 # We only allow video|audio followed by a whitespace or '>'.
2176 # Allowing more characters may end up in significant slow down (see
2177 # https://github.com/rg3/youtube-dl/issues/11979, example URL:
2178 # http://www.porntrex.com/maps/videositemap.xml).
2179 r'(?s)(<(?P<tag>(?:amp-)?(?:video|audio))(?:\s+[^>]*)?>)(.*?)</(?P=tag)>', webpage))
2180 for media_tag, media_type, media_content in media_tags:
2185 media_attributes = extract_attributes(media_tag)
2186 src = media_attributes.get('src')
2188 _, formats = _media_formats(src, media_type)
2189 media_info['formats'].extend(formats)
2190 media_info['thumbnail'] = media_attributes.get('poster')
2192 for source_tag in re.findall(r'<source[^>]+>', media_content):
2193 source_attributes = extract_attributes(source_tag)
2194 src = source_attributes.get('src')
2197 f = parse_content_type(source_attributes.get('type'))
2198 is_plain_url, formats = _media_formats(src, media_type, f)
2200 # res attribute is not standard but seen several times
2203 'height': int_or_none(source_attributes.get('res')),
2204 'format_id': source_attributes.get('label'),
2206 f.update(formats[0])
2207 media_info['formats'].append(f)
2209 media_info['formats'].extend(formats)
2210 for track_tag in re.findall(r'<track[^>]+>', media_content):
2211 track_attributes = extract_attributes(track_tag)
2212 kind = track_attributes.get('kind')
2213 if not kind or kind in ('subtitles', 'captions'):
2214 src = track_attributes.get('src')
2217 lang = track_attributes.get('srclang') or track_attributes.get('lang') or track_attributes.get('label')
2218 media_info['subtitles'].setdefault(lang, []).append({
2219 'url': absolute_url(src),
2221 if media_info['formats'] or media_info['subtitles']:
2222 entries.append(media_info)
2225 def _extract_akamai_formats(self, manifest_url, video_id, hosts={}):
2227 hdcore_sign = 'hdcore=3.7.0'
2228 f4m_url = re.sub(r'(https?://[^/]+)/i/', r'\1/z/', manifest_url).replace('/master.m3u8', '/manifest.f4m')
2229 hds_host = hosts.get('hds')
2231 f4m_url = re.sub(r'(https?://)[^/]+', r'\1' + hds_host, f4m_url)
2232 if 'hdcore=' not in f4m_url:
2233 f4m_url += ('&' if '?' in f4m_url else '?') + hdcore_sign
2234 f4m_formats = self._extract_f4m_formats(
2235 f4m_url, video_id, f4m_id='hds', fatal=False)
2236 for entry in f4m_formats:
2237 entry.update({'extra_param_to_segment_url': hdcore_sign})
2238 formats.extend(f4m_formats)
2239 m3u8_url = re.sub(r'(https?://[^/]+)/z/', r'\1/i/', manifest_url).replace('/manifest.f4m', '/master.m3u8')
2240 hls_host = hosts.get('hls')
2242 m3u8_url = re.sub(r'(https?://)[^/]+', r'\1' + hls_host, m3u8_url)
2243 formats.extend(self._extract_m3u8_formats(
2244 m3u8_url, video_id, 'mp4', 'm3u8_native',
2245 m3u8_id='hls', fatal=False))
2248 def _extract_wowza_formats(self, url, video_id, m3u8_entry_protocol='m3u8_native', skip_protocols=[]):
2249 query = compat_urlparse.urlparse(url).query
2250 url = re.sub(r'/(?:manifest|playlist|jwplayer)\.(?:m3u8|f4m|mpd|smil)', '', url)
2251 url_base = self._search_regex(
2252 r'(?:(?:https?|rtmp|rtsp):)?(//[^?]+)', url, 'format url')
2253 http_base_url = '%s:%s' % ('http', url_base)
2256 def manifest_url(manifest):
2257 m_url = '%s/%s' % (http_base_url, manifest)
2259 m_url += '?%s' % query
2262 if 'm3u8' not in skip_protocols:
2263 formats.extend(self._extract_m3u8_formats(
2264 manifest_url('playlist.m3u8'), video_id, 'mp4',
2265 m3u8_entry_protocol, m3u8_id='hls', fatal=False))
2266 if 'f4m' not in skip_protocols:
2267 formats.extend(self._extract_f4m_formats(
2268 manifest_url('manifest.f4m'),
2269 video_id, f4m_id='hds', fatal=False))
2270 if 'dash' not in skip_protocols:
2271 formats.extend(self._extract_mpd_formats(
2272 manifest_url('manifest.mpd'),
2273 video_id, mpd_id='dash', fatal=False))
2274 if re.search(r'(?:/smil:|\.smil)', url_base):
2275 if 'smil' not in skip_protocols:
2276 rtmp_formats = self._extract_smil_formats(
2277 manifest_url('jwplayer.smil'),
2278 video_id, fatal=False)
2279 for rtmp_format in rtmp_formats:
2280 rtsp_format = rtmp_format.copy()
2281 rtsp_format['url'] = '%s/%s' % (rtmp_format['url'], rtmp_format['play_path'])
2282 del rtsp_format['play_path']
2283 del rtsp_format['ext']
2284 rtsp_format.update({
2285 'url': rtsp_format['url'].replace('rtmp://', 'rtsp://'),
2286 'format_id': rtmp_format['format_id'].replace('rtmp', 'rtsp'),
2289 formats.extend([rtmp_format, rtsp_format])
2291 for protocol in ('rtmp', 'rtsp'):
2292 if protocol not in skip_protocols:
2294 'url': '%s:%s' % (protocol, url_base),
2295 'format_id': protocol,
2296 'protocol': protocol,
2300 def _find_jwplayer_data(self, webpage, video_id=None, transform_source=js_to_json):
2302 r'(?s)jwplayer\((?P<quote>[\'"])[^\'" ]+(?P=quote)\)(?!</script>).*?\.setup\s*\((?P<options>[^)]+)\)',
2306 jwplayer_data = self._parse_json(mobj.group('options'),
2308 transform_source=transform_source)
2309 except ExtractorError:
2312 if isinstance(jwplayer_data, dict):
2313 return jwplayer_data
2315 def _extract_jwplayer_data(self, webpage, video_id, *args, **kwargs):
2316 jwplayer_data = self._find_jwplayer_data(
2317 webpage, video_id, transform_source=js_to_json)
2318 return self._parse_jwplayer_data(
2319 jwplayer_data, video_id, *args, **kwargs)
2321 def _parse_jwplayer_data(self, jwplayer_data, video_id=None, require_title=True,
2322 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
2323 # JWPlayer backward compatibility: flattened playlists
2324 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/api/config.js#L81-L96
2325 if 'playlist' not in jwplayer_data:
2326 jwplayer_data = {'playlist': [jwplayer_data]}
2330 # JWPlayer backward compatibility: single playlist item
2331 # https://github.com/jwplayer/jwplayer/blob/v7.7.0/src/js/playlist/playlist.js#L10
2332 if not isinstance(jwplayer_data['playlist'], list):
2333 jwplayer_data['playlist'] = [jwplayer_data['playlist']]
2335 for video_data in jwplayer_data['playlist']:
2336 # JWPlayer backward compatibility: flattened sources
2337 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/playlist/item.js#L29-L35
2338 if 'sources' not in video_data:
2339 video_data['sources'] = [video_data]
2341 this_video_id = video_id or video_data['mediaid']
2343 formats = self._parse_jwplayer_formats(
2344 video_data['sources'], video_id=this_video_id, m3u8_id=m3u8_id,
2345 mpd_id=mpd_id, rtmp_params=rtmp_params, base_url=base_url)
2348 tracks = video_data.get('tracks')
2349 if tracks and isinstance(tracks, list):
2350 for track in tracks:
2351 if not isinstance(track, dict):
2353 if track.get('kind') != 'captions':
2355 track_url = urljoin(base_url, track.get('file'))
2358 subtitles.setdefault(track.get('label') or 'en', []).append({
2359 'url': self._proto_relative_url(track_url)
2363 'id': this_video_id,
2364 'title': unescapeHTML(video_data['title'] if require_title else video_data.get('title')),
2365 'description': video_data.get('description'),
2366 'thumbnail': self._proto_relative_url(video_data.get('image')),
2367 'timestamp': int_or_none(video_data.get('pubdate')),
2368 'duration': float_or_none(jwplayer_data.get('duration') or video_data.get('duration')),
2369 'subtitles': subtitles,
2371 # https://github.com/jwplayer/jwplayer/blob/master/src/js/utils/validator.js#L32
2372 if len(formats) == 1 and re.search(r'^(?:http|//).*(?:youtube\.com|youtu\.be)/.+', formats[0]['url']):
2374 '_type': 'url_transparent',
2375 'url': formats[0]['url'],
2378 self._sort_formats(formats)
2379 entry['formats'] = formats
2380 entries.append(entry)
2381 if len(entries) == 1:
2384 return self.playlist_result(entries)
2386 def _parse_jwplayer_formats(self, jwplayer_sources_data, video_id=None,
2387 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
2390 for source in jwplayer_sources_data:
2391 if not isinstance(source, dict):
2393 source_url = self._proto_relative_url(source.get('file'))
2397 source_url = compat_urlparse.urljoin(base_url, source_url)
2398 if source_url in urls:
2400 urls.append(source_url)
2401 source_type = source.get('type') or ''
2402 ext = mimetype2ext(source_type) or determine_ext(source_url)
2403 if source_type == 'hls' or ext == 'm3u8':
2404 formats.extend(self._extract_m3u8_formats(
2405 source_url, video_id, 'mp4', entry_protocol='m3u8_native',
2406 m3u8_id=m3u8_id, fatal=False))
2408 formats.extend(self._extract_mpd_formats(
2409 source_url, video_id, mpd_id=mpd_id, fatal=False))
2411 formats.extend(self._extract_smil_formats(
2412 source_url, video_id, fatal=False))
2413 # https://github.com/jwplayer/jwplayer/blob/master/src/js/providers/default.js#L67
2414 elif source_type.startswith('audio') or ext in (
2415 'oga', 'aac', 'mp3', 'mpeg', 'vorbis'):
2422 height = int_or_none(source.get('height'))
2424 # Often no height is provided but there is a label in
2425 # format like "1080p", "720p SD", or 1080.
2426 height = int_or_none(self._search_regex(
2427 r'^(\d{3,4})[pP]?(?:\b|$)', compat_str(source.get('label') or ''),
2428 'height', default=None))
2431 'width': int_or_none(source.get('width')),
2433 'tbr': int_or_none(source.get('bitrate')),
2436 if source_url.startswith('rtmp'):
2437 a_format['ext'] = 'flv'
2438 # See com/longtailvideo/jwplayer/media/RTMPMediaProvider.as
2439 # of jwplayer.flash.swf
2440 rtmp_url_parts = re.split(
2441 r'((?:mp4|mp3|flv):)', source_url, 1)
2442 if len(rtmp_url_parts) == 3:
2443 rtmp_url, prefix, play_path = rtmp_url_parts
2446 'play_path': prefix + play_path,
2449 a_format.update(rtmp_params)
2450 formats.append(a_format)
2453 def _live_title(self, name):
2454 """ Generate the title for a live video """
2455 now = datetime.datetime.now()
2456 now_str = now.strftime('%Y-%m-%d %H:%M')
2457 return name + ' ' + now_str
2459 def _int(self, v, name, fatal=False, **kwargs):
2460 res = int_or_none(v, **kwargs)
2461 if 'get_attr' in kwargs:
2462 print(getattr(v, kwargs['get_attr']))
2464 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
2466 raise ExtractorError(msg)
2468 self._downloader.report_warning(msg)
2471 def _float(self, v, name, fatal=False, **kwargs):
2472 res = float_or_none(v, **kwargs)
2474 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
2476 raise ExtractorError(msg)
2478 self._downloader.report_warning(msg)
2481 def _set_cookie(self, domain, name, value, expire_time=None, port=None,
2482 path='/', secure=False, discard=False, rest={}, **kwargs):
2483 cookie = compat_cookiejar.Cookie(
2484 0, name, value, port, port is not None, domain, True,
2485 domain.startswith('.'), path, True, secure, expire_time,
2486 discard, None, None, rest)
2487 self._downloader.cookiejar.set_cookie(cookie)
2489 def _get_cookies(self, url):
2490 """ Return a compat_cookies.SimpleCookie with the cookies for the url """
2491 req = sanitized_Request(url)
2492 self._downloader.cookiejar.add_cookie_header(req)
2493 return compat_cookies.SimpleCookie(req.get_header('Cookie'))
2495 def get_testcases(self, include_onlymatching=False):
2496 t = getattr(self, '_TEST', None)
2498 assert not hasattr(self, '_TESTS'), \
2499 '%s has _TEST and _TESTS' % type(self).__name__
2502 tests = getattr(self, '_TESTS', [])
2504 if not include_onlymatching and t.get('only_matching', False):
2506 t['name'] = type(self).__name__[:-len('IE')]
2509 def is_suitable(self, age_limit):
2510 """ Test whether the extractor is generally suitable for the given
2511 age limit (i.e. pornographic sites are not, all others usually are) """
2513 any_restricted = False
2514 for tc in self.get_testcases(include_onlymatching=False):
2515 if tc.get('playlist', []):
2516 tc = tc['playlist'][0]
2517 is_restricted = age_restricted(
2518 tc.get('info_dict', {}).get('age_limit'), age_limit)
2519 if not is_restricted:
2521 any_restricted = any_restricted or is_restricted
2522 return not any_restricted
2524 def extract_subtitles(self, *args, **kwargs):
2525 if (self._downloader.params.get('writesubtitles', False) or
2526 self._downloader.params.get('listsubtitles')):
2527 return self._get_subtitles(*args, **kwargs)
2530 def _get_subtitles(self, *args, **kwargs):
2531 raise NotImplementedError('This method must be implemented by subclasses')
2534 def _merge_subtitle_items(subtitle_list1, subtitle_list2):
2535 """ Merge subtitle items for one language. Items with duplicated URLs
2536 will be dropped. """
2537 list1_urls = set([item['url'] for item in subtitle_list1])
2538 ret = list(subtitle_list1)
2539 ret.extend([item for item in subtitle_list2 if item['url'] not in list1_urls])
2543 def _merge_subtitles(cls, subtitle_dict1, subtitle_dict2):
2544 """ Merge two subtitle dictionaries, language by language. """
2545 ret = dict(subtitle_dict1)
2546 for lang in subtitle_dict2:
2547 ret[lang] = cls._merge_subtitle_items(subtitle_dict1.get(lang, []), subtitle_dict2[lang])
2550 def extract_automatic_captions(self, *args, **kwargs):
2551 if (self._downloader.params.get('writeautomaticsub', False) or
2552 self._downloader.params.get('listsubtitles')):
2553 return self._get_automatic_captions(*args, **kwargs)
2556 def _get_automatic_captions(self, *args, **kwargs):
2557 raise NotImplementedError('This method must be implemented by subclasses')
2559 def mark_watched(self, *args, **kwargs):
2560 if (self._downloader.params.get('mark_watched', False) and
2561 (self._get_login_info()[0] is not None or
2562 self._downloader.params.get('cookiefile') is not None)):
2563 self._mark_watched(*args, **kwargs)
2565 def _mark_watched(self, *args, **kwargs):
2566 raise NotImplementedError('This method must be implemented by subclasses')
2568 def geo_verification_headers(self):
2570 geo_verification_proxy = self._downloader.params.get('geo_verification_proxy')
2571 if geo_verification_proxy:
2572 headers['Ytdl-request-proxy'] = geo_verification_proxy
2575 def _generic_id(self, url):
2576 return compat_urllib_parse_unquote(os.path.splitext(url.rstrip('/').split('/')[-1])[0])
2578 def _generic_title(self, url):
2579 return compat_urllib_parse_unquote(os.path.splitext(url_basename(url))[0])
2582 class SearchInfoExtractor(InfoExtractor):
2584 Base class for paged search queries extractors.
2585 They accept URLs in the format _SEARCH_KEY(|all|[0-9]):{query}
2586 Instances should define _SEARCH_KEY and _MAX_RESULTS.
2590 def _make_valid_url(cls):
2591 return r'%s(?P<prefix>|[1-9][0-9]*|all):(?P<query>[\s\S]+)' % cls._SEARCH_KEY
2594 def suitable(cls, url):
2595 return re.match(cls._make_valid_url(), url) is not None
2597 def _real_extract(self, query):
2598 mobj = re.match(self._make_valid_url(), query)
2600 raise ExtractorError('Invalid search query "%s"' % query)
2602 prefix = mobj.group('prefix')
2603 query = mobj.group('query')
2605 return self._get_n_results(query, 1)
2606 elif prefix == 'all':
2607 return self._get_n_results(query, self._MAX_RESULTS)
2611 raise ExtractorError('invalid download number %s for query "%s"' % (n, query))
2612 elif n > self._MAX_RESULTS:
2613 self._downloader.report_warning('%s returns max %i results (you requested %i)' % (self._SEARCH_KEY, self._MAX_RESULTS, n))
2614 n = self._MAX_RESULTS
2615 return self._get_n_results(query, n)
2617 def _get_n_results(self, query, n):
2618 """Get a specified number of results for a query"""
2619 raise NotImplementedError('This method must be implemented by subclasses')
2622 def SEARCH_KEY(self):
2623 return self._SEARCH_KEY