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,
31 from ..downloader.f4m import remove_encrypted_media
55 parse_m3u8_attributes,
72 class InfoExtractor(object):
73 """Information Extractor class.
75 Information extractors are the classes that, given a URL, extract
76 information about the video (or videos) the URL refers to. This
77 information includes the real video URL, the video title, author and
78 others. The information is stored in a dictionary which is then
79 passed to the YoutubeDL. The YoutubeDL processes this
80 information possibly downloading the video to the file system, among
81 other possible outcomes.
83 The type field determines the type of the result.
84 By far the most common value (and the default if _type is missing) is
85 "video", which indicates a single video.
87 For a video, the dictionaries must include the following fields:
90 title: Video title, unescaped.
92 Additionally, it must contain either a formats entry or a url one:
94 formats: A list of dictionaries for each format available, ordered
95 from worst to best quality.
98 * url Mandatory. The URL of the video file
100 The URL of the manifest file in case of
101 fragmented media (DASH, hls, hds)
102 * ext Will be calculated from URL if missing
103 * format A human-readable description of the format
104 ("mp4 container with h264/opus").
105 Calculated from the format_id, width, height.
106 and format_note fields if missing.
107 * format_id A short description of the format
108 ("mp4_h264_opus" or "19").
109 Technically optional, but strongly recommended.
110 * format_note Additional info about the format
111 ("3D" or "DASH video")
112 * width Width of the video, if known
113 * height Height of the video, if known
114 * resolution Textual description of width and height
115 * tbr Average bitrate of audio and video in KBit/s
116 * abr Average audio bitrate in KBit/s
117 * acodec Name of the audio codec in use
118 * asr Audio sampling rate in Hertz
119 * vbr Average video bitrate in KBit/s
121 * vcodec Name of the video codec in use
122 * container Name of the container format
123 * filesize The number of bytes, if known in advance
124 * filesize_approx An estimate for the number of bytes
125 * player_url SWF Player URL (used for rtmpdump).
126 * protocol The protocol that will be used for the actual
127 download, lower-case.
128 "http", "https", "rtsp", "rtmp", "rtmpe",
129 "m3u8", "m3u8_native" or "http_dash_segments".
131 Base URL for fragments. Each fragment's path
132 value (if present) will be relative to
134 * fragments A list of fragments of a fragmented media.
135 Each fragment entry must contain either an url
136 or a path. If an url is present it should be
137 considered by a client. Otherwise both path and
138 fragment_base_url must be present. Here is
139 the list of all potential fields:
140 * "url" - fragment's URL
141 * "path" - fragment's path relative to
143 * "duration" (optional, int or float)
144 * "filesize" (optional, int)
145 * preference Order number of this format. If this field is
146 present and not None, the formats get sorted
147 by this field, regardless of all other values.
148 -1 for default (order by other properties),
149 -2 or smaller for less than default.
150 < -1000 to hide the format (if there is
151 another one which is strictly better)
152 * language Language code, e.g. "de" or "en-US".
153 * language_preference Is this in the language mentioned in
155 10 if it's what the URL is about,
156 -1 for default (don't know),
157 -10 otherwise, other values reserved for now.
158 * quality Order number of the video quality of this
159 format, irrespective of the file format.
160 -1 for default (order by other properties),
161 -2 or smaller for less than default.
162 * source_preference Order number for this video source
163 (quality takes higher priority)
164 -1 for default (order by other properties),
165 -2 or smaller for less than default.
166 * http_headers A dictionary of additional HTTP headers
167 to add to the request.
168 * stretched_ratio If given and not 1, indicates that the
169 video's pixels are not square.
170 width : height ratio as float.
171 * no_resume The server does not support resuming the
172 (HTTP or RTMP) download. Boolean.
174 url: Final video URL.
175 ext: Video filename extension.
176 format: The video format, defaults to ext (used for --get-format)
177 player_url: SWF Player URL (used for rtmpdump).
179 The following fields are optional:
181 alt_title: A secondary title of the video.
182 display_id An alternative identifier for the video, not necessarily
183 unique, but available before title. Typically, id is
184 something like "4234987", title "Dancing naked mole rats",
185 and display_id "dancing-naked-mole-rats"
186 thumbnails: A list of dictionaries, with the following entries:
187 * "id" (optional, string) - Thumbnail format ID
189 * "preference" (optional, int) - quality of the image
190 * "width" (optional, int)
191 * "height" (optional, int)
192 * "resolution" (optional, string "{width}x{height"},
194 * "filesize" (optional, int)
195 thumbnail: Full URL to a video thumbnail image.
196 description: Full video description.
197 uploader: Full name of the video uploader.
198 license: License name the video is licensed under.
199 creator: The creator of the video.
200 release_date: The date (YYYYMMDD) when the video was released.
201 timestamp: UNIX timestamp of the moment the video became available.
202 upload_date: Video upload date (YYYYMMDD).
203 If not explicitly set, calculated from timestamp.
204 uploader_id: Nickname or id of the video uploader.
205 uploader_url: Full URL to a personal webpage of the video uploader.
206 location: Physical location where the video was filmed.
207 subtitles: The available subtitles as a dictionary in the format
208 {tag: subformats}. "tag" is usually a language code, and
209 "subformats" is a list sorted from lower to higher
210 preference, each element is a dictionary with the "ext"
212 * "data": The subtitles file contents
213 * "url": A URL pointing to the subtitles file
214 "ext" will be calculated from URL if missing
215 automatic_captions: Like 'subtitles', used by the YoutubeIE for
216 automatically generated captions
217 duration: Length of the video in seconds, as an integer or float.
218 view_count: How many users have watched the video on the platform.
219 like_count: Number of positive ratings of the video
220 dislike_count: Number of negative ratings of the video
221 repost_count: Number of reposts of the video
222 average_rating: Average rating give by users, the scale used depends on the webpage
223 comment_count: Number of comments on the video
224 comments: A list of comments, each with one or more of the following
225 properties (all but one of text or html optional):
226 * "author" - human-readable name of the comment author
227 * "author_id" - user ID of the comment author
229 * "html" - Comment as HTML
230 * "text" - Plain text of the comment
231 * "timestamp" - UNIX timestamp of comment
232 * "parent" - ID of the comment this one is replying to.
233 Set to "root" to indicate that this is a
234 comment to the original video.
235 age_limit: Age restriction for the video, as an integer (years)
236 webpage_url: The URL to the video webpage, if given to youtube-dl it
237 should allow to get the same result again. (It will be set
238 by YoutubeDL if it's missing)
239 categories: A list of categories that the video falls in, for example
241 tags: A list of tags assigned to the video, e.g. ["sweden", "pop music"]
242 is_live: True, False, or None (=unknown). Whether this video is a
243 live stream that goes on instead of a fixed-length video.
244 start_time: Time in seconds where the reproduction should start, as
245 specified in the URL.
246 end_time: Time in seconds where the reproduction should end, as
247 specified in the URL.
248 chapters: A list of dictionaries, with the following entries:
249 * "start_time" - The start time of the chapter in seconds
250 * "end_time" - The end time of the chapter in seconds
251 * "title" (optional, string)
253 The following fields should only be used when the video belongs to some logical
256 chapter: Name or title of the chapter the video belongs to.
257 chapter_number: Number of the chapter the video belongs to, as an integer.
258 chapter_id: Id of the chapter the video belongs to, as a unicode string.
260 The following fields should only be used when the video is an episode of some
261 series, programme or podcast:
263 series: Title of the series or programme the video episode belongs to.
264 season: Title of the season the video episode belongs to.
265 season_number: Number of the season the video episode belongs to, as an integer.
266 season_id: Id of the season the video episode belongs to, as a unicode string.
267 episode: Title of the video episode. Unlike mandatory video title field,
268 this field should denote the exact title of the video episode
269 without any kind of decoration.
270 episode_number: Number of the video episode within a season, as an integer.
271 episode_id: Id of the video episode, as a unicode string.
273 The following fields should only be used when the media is a track or a part of
276 track: Title of the track.
277 track_number: Number of the track within an album or a disc, as an integer.
278 track_id: Id of the track (useful in case of custom indexing, e.g. 6.iii),
280 artist: Artist(s) of the track.
281 genre: Genre(s) of the track.
282 album: Title of the album the track belongs to.
283 album_type: Type of the album (e.g. "Demo", "Full-length", "Split", "Compilation", etc).
284 album_artist: List of all artists appeared on the album (e.g.
285 "Ash Borer / Fell Voices" or "Various Artists", useful for splits
287 disc_number: Number of the disc or other physical medium the track belongs to,
289 release_year: Year (YYYY) when the album was released.
291 Unless mentioned otherwise, the fields should be Unicode strings.
293 Unless mentioned otherwise, None is equivalent to absence of information.
296 _type "playlist" indicates multiple videos.
297 There must be a key "entries", which is a list, an iterable, or a PagedList
298 object, each element of which is a valid dictionary by this specification.
300 Additionally, playlists can have "title", "description" and "id" attributes
301 with the same semantics as videos (see above).
304 _type "multi_video" indicates that there are multiple videos that
305 form a single show, for examples multiple acts of an opera or TV episode.
306 It must have an entries key like a playlist and contain all the keys
307 required for a video at the same time.
310 _type "url" indicates that the video must be extracted from another
311 location, possibly by a different extractor. Its only required key is:
312 "url" - the next URL to extract.
313 The key "ie_key" can be set to the class name (minus the trailing "IE",
314 e.g. "Youtube") if the extractor class is known in advance.
315 Additionally, the dictionary may have any properties of the resolved entity
316 known in advance, for example "title" if the title of the referred video is
320 _type "url_transparent" entities have the same specification as "url", but
321 indicate that the given additional information is more precise than the one
322 associated with the resolved URL.
323 This is useful when a site employs a video service that hosts the video and
324 its technical metadata, but that video service does not embed a useful
325 title, description etc.
328 Subclasses of this one should re-define the _real_initialize() and
329 _real_extract() methods and define a _VALID_URL regexp.
330 Probably, they should also be added to the list of extractors.
332 _GEO_BYPASS attribute may be set to False in order to disable
333 geo restriction bypass mechanisms for a particular extractor.
334 Though it won't disable explicit geo restriction bypass based on
335 country code provided with geo_bypass_country. (experimental)
337 _GEO_COUNTRIES attribute may contain a list of presumably geo unrestricted
338 countries for this extractor. One of these countries will be used by
339 geo restriction bypass mechanism right away in order to bypass
340 geo restriction, of course, if the mechanism is not disabled. (experimental)
342 NB: both these geo attributes are experimental and may change in future
343 or be completely removed.
345 Finally, the _WORKING attribute should be set to False for broken IEs
346 in order to warn the users and skip the tests.
351 _x_forwarded_for_ip = None
353 _GEO_COUNTRIES = None
356 def __init__(self, downloader=None):
357 """Constructor. Receives an optional downloader."""
359 self._x_forwarded_for_ip = None
360 self.set_downloader(downloader)
363 def suitable(cls, url):
364 """Receives a URL and returns True if suitable for this IE."""
366 # This does not use has/getattr intentionally - we want to know whether
367 # we have cached the regexp for *this* class, whereas getattr would also
368 # match the superclass
369 if '_VALID_URL_RE' not in cls.__dict__:
370 cls._VALID_URL_RE = re.compile(cls._VALID_URL)
371 return cls._VALID_URL_RE.match(url) is not None
374 def _match_id(cls, url):
375 if '_VALID_URL_RE' not in cls.__dict__:
376 cls._VALID_URL_RE = re.compile(cls._VALID_URL)
377 m = cls._VALID_URL_RE.match(url)
379 return compat_str(m.group('id'))
383 """Getter method for _WORKING."""
386 def initialize(self):
387 """Initializes an instance (authentication, etc)."""
388 self._initialize_geo_bypass(self._GEO_COUNTRIES)
390 self._real_initialize()
393 def _initialize_geo_bypass(self, countries):
395 Initialize geo restriction bypass mechanism.
397 This method is used to initialize geo bypass mechanism based on faking
398 X-Forwarded-For HTTP header. A random country from provided country list
399 is selected and a random IP belonging to this country is generated. This
400 IP will be passed as X-Forwarded-For HTTP header in all subsequent
403 This method will be used for initial geo bypass mechanism initialization
404 during the instance initialization with _GEO_COUNTRIES.
406 You may also manually call it from extractor's code if geo countries
407 information is not available beforehand (e.g. obtained during
408 extraction) or due to some another reason.
410 if not self._x_forwarded_for_ip:
411 country_code = self._downloader.params.get('geo_bypass_country', None)
412 # If there is no explicit country for geo bypass specified and
413 # the extractor is known to be geo restricted let's fake IP
414 # as X-Forwarded-For right away.
415 if (not country_code and
417 self._downloader.params.get('geo_bypass', True) and
419 country_code = random.choice(countries)
421 self._x_forwarded_for_ip = GeoUtils.random_ipv4(country_code)
422 if self._downloader.params.get('verbose', False):
423 self._downloader.to_screen(
424 '[debug] Using fake IP %s (%s) as X-Forwarded-For.'
425 % (self._x_forwarded_for_ip, country_code.upper()))
427 def extract(self, url):
428 """Extracts URL information and returns it in list of dicts."""
433 ie_result = self._real_extract(url)
434 if self._x_forwarded_for_ip:
435 ie_result['__x_forwarded_for_ip'] = self._x_forwarded_for_ip
437 except GeoRestrictedError as e:
438 if self.__maybe_fake_ip_and_retry(e.countries):
441 except ExtractorError:
443 except compat_http_client.IncompleteRead as e:
444 raise ExtractorError('A network error has occurred.', cause=e, expected=True)
445 except (KeyError, StopIteration) as e:
446 raise ExtractorError('An extractor error has occurred.', cause=e)
448 def __maybe_fake_ip_and_retry(self, countries):
449 if (not self._downloader.params.get('geo_bypass_country', None) and
451 self._downloader.params.get('geo_bypass', True) and
452 not self._x_forwarded_for_ip and
454 country_code = random.choice(countries)
455 self._x_forwarded_for_ip = GeoUtils.random_ipv4(country_code)
456 if self._x_forwarded_for_ip:
458 'Video is geo restricted. Retrying extraction with fake IP %s (%s) as X-Forwarded-For.'
459 % (self._x_forwarded_for_ip, country_code.upper()))
463 def set_downloader(self, downloader):
464 """Sets the downloader for this IE."""
465 self._downloader = downloader
467 def _real_initialize(self):
468 """Real initialization process. Redefine in subclasses."""
471 def _real_extract(self, url):
472 """Real extraction process. Redefine in subclasses."""
477 """A string for getting the InfoExtractor with get_info_extractor"""
478 return compat_str(cls.__name__[:-2])
482 return compat_str(type(self).__name__[:-2])
484 def _request_webpage(self, url_or_request, video_id, note=None, errnote=None, fatal=True, data=None, headers={}, query={}):
485 """ Returns the response handle """
487 self.report_download_webpage(video_id)
488 elif note is not False:
490 self.to_screen('%s' % (note,))
492 self.to_screen('%s: %s' % (video_id, note))
493 if isinstance(url_or_request, compat_urllib_request.Request):
494 url_or_request = update_Request(
495 url_or_request, data=data, headers=headers, query=query)
498 url_or_request = update_url_query(url_or_request, query)
499 if data is not None or headers:
500 url_or_request = sanitized_Request(url_or_request, data, headers)
502 return self._downloader.urlopen(url_or_request)
503 except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err:
507 errnote = 'Unable to download webpage'
509 errmsg = '%s: %s' % (errnote, error_to_compat_str(err))
511 raise ExtractorError(errmsg, sys.exc_info()[2], cause=err)
513 self._downloader.report_warning(errmsg)
516 def _download_webpage_handle(self, url_or_request, video_id, note=None, errnote=None, fatal=True, encoding=None, data=None, headers={}, query={}):
517 """ Returns a tuple (page content as string, URL handle) """
518 # Strip hashes from the URL (#1038)
519 if isinstance(url_or_request, (compat_str, str)):
520 url_or_request = url_or_request.partition('#')[0]
522 # Some sites check X-Forwarded-For HTTP header in order to figure out
523 # the origin of the client behind proxy. This allows bypassing geo
524 # restriction by faking this header's value to IP that belongs to some
525 # geo unrestricted country. We will do so once we encounter any
526 # geo restriction error.
527 if self._x_forwarded_for_ip:
528 if 'X-Forwarded-For' not in headers:
529 headers['X-Forwarded-For'] = self._x_forwarded_for_ip
531 urlh = self._request_webpage(url_or_request, video_id, note, errnote, fatal, data=data, headers=headers, query=query)
535 content = self._webpage_read_content(urlh, url_or_request, video_id, note, errnote, fatal, encoding=encoding)
536 return (content, urlh)
539 def _guess_encoding_from_content(content_type, webpage_bytes):
540 m = re.match(r'[a-zA-Z0-9_.-]+/[a-zA-Z0-9_.-]+\s*;\s*charset=(.+)', content_type)
542 encoding = m.group(1)
544 m = re.search(br'<meta[^>]+charset=[\'"]?([^\'")]+)[ /\'">]',
545 webpage_bytes[:1024])
547 encoding = m.group(1).decode('ascii')
548 elif webpage_bytes.startswith(b'\xff\xfe'):
555 def __check_blocked(self, content):
556 first_block = content[:512]
557 if ('<title>Access to this site is blocked</title>' in content and
558 'Websense' in first_block):
559 msg = 'Access to this webpage has been blocked by Websense filtering software in your network.'
560 blocked_iframe = self._html_search_regex(
561 r'<iframe src="([^"]+)"', content,
562 'Websense information URL', default=None)
564 msg += ' Visit %s for more details' % blocked_iframe
565 raise ExtractorError(msg, expected=True)
566 if '<title>The URL you requested has been blocked</title>' in first_block:
568 'Access to this webpage has been blocked by Indian censorship. '
569 'Use a VPN or proxy server (with --proxy) to route around it.')
570 block_msg = self._html_search_regex(
571 r'</h1><p>(.*?)</p>',
572 content, 'block message', default=None)
574 msg += ' (Message: "%s")' % block_msg.replace('\n', ' ')
575 raise ExtractorError(msg, expected=True)
576 if ('<title>TTK :: Доступ к ресурсу ограничен</title>' in content and
577 'blocklist.rkn.gov.ru' in content):
578 raise ExtractorError(
579 'Access to this webpage has been blocked by decision of the Russian government. '
580 'Visit http://blocklist.rkn.gov.ru/ for a block reason.',
583 def _webpage_read_content(self, urlh, url_or_request, video_id, note=None, errnote=None, fatal=True, prefix=None, encoding=None):
584 content_type = urlh.headers.get('Content-Type', '')
585 webpage_bytes = urlh.read()
586 if prefix is not None:
587 webpage_bytes = prefix + webpage_bytes
589 encoding = self._guess_encoding_from_content(content_type, webpage_bytes)
590 if self._downloader.params.get('dump_intermediate_pages', False):
592 url = url_or_request.get_full_url()
593 except AttributeError:
595 self.to_screen('Dumping request to ' + url)
596 dump = base64.b64encode(webpage_bytes).decode('ascii')
597 self._downloader.to_screen(dump)
598 if self._downloader.params.get('write_pages', False):
600 url = url_or_request.get_full_url()
601 except AttributeError:
603 basen = '%s_%s' % (video_id, url)
605 h = '___' + hashlib.md5(basen.encode('utf-8')).hexdigest()
606 basen = basen[:240 - len(h)] + h
607 raw_filename = basen + '.dump'
608 filename = sanitize_filename(raw_filename, restricted=True)
609 self.to_screen('Saving request to ' + filename)
610 # Working around MAX_PATH limitation on Windows (see
611 # http://msdn.microsoft.com/en-us/library/windows/desktop/aa365247(v=vs.85).aspx)
612 if compat_os_name == 'nt':
613 absfilepath = os.path.abspath(filename)
614 if len(absfilepath) > 259:
615 filename = '\\\\?\\' + absfilepath
616 with open(filename, 'wb') as outf:
617 outf.write(webpage_bytes)
620 content = webpage_bytes.decode(encoding, 'replace')
622 content = webpage_bytes.decode('utf-8', 'replace')
624 self.__check_blocked(content)
628 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={}):
629 """ Returns the data of the page as a string """
632 while success is False:
634 res = self._download_webpage_handle(url_or_request, video_id, note, errnote, fatal, encoding=encoding, data=data, headers=headers, query=query)
636 except compat_http_client.IncompleteRead as e:
638 if try_count >= tries:
640 self._sleep(timeout, video_id)
647 def _download_xml(self, url_or_request, video_id,
648 note='Downloading XML', errnote='Unable to download XML',
649 transform_source=None, fatal=True, encoding=None, data=None, headers={}, query={}):
650 """Return the xml as an xml.etree.ElementTree.Element"""
651 xml_string = self._download_webpage(
652 url_or_request, video_id, note, errnote, fatal=fatal, encoding=encoding, data=data, headers=headers, query=query)
653 if xml_string is False:
656 xml_string = transform_source(xml_string)
657 return compat_etree_fromstring(xml_string.encode('utf-8'))
659 def _download_json(self, url_or_request, video_id,
660 note='Downloading JSON metadata',
661 errnote='Unable to download JSON metadata',
662 transform_source=None,
663 fatal=True, encoding=None, data=None, headers={}, query={}):
664 json_string = self._download_webpage(
665 url_or_request, video_id, note, errnote, fatal=fatal,
666 encoding=encoding, data=data, headers=headers, query=query)
667 if (not fatal) and json_string is False:
669 return self._parse_json(
670 json_string, video_id, transform_source=transform_source, fatal=fatal)
672 def _parse_json(self, json_string, video_id, transform_source=None, fatal=True):
674 json_string = transform_source(json_string)
676 return json.loads(json_string)
677 except ValueError as ve:
678 errmsg = '%s: Failed to parse JSON ' % video_id
680 raise ExtractorError(errmsg, cause=ve)
682 self.report_warning(errmsg + str(ve))
684 def report_warning(self, msg, video_id=None):
685 idstr = '' if video_id is None else '%s: ' % video_id
686 self._downloader.report_warning(
687 '[%s] %s%s' % (self.IE_NAME, idstr, msg))
689 def to_screen(self, msg):
690 """Print msg to screen, prefixing it with '[ie_name]'"""
691 self._downloader.to_screen('[%s] %s' % (self.IE_NAME, msg))
693 def report_extraction(self, id_or_name):
694 """Report information extraction."""
695 self.to_screen('%s: Extracting information' % id_or_name)
697 def report_download_webpage(self, video_id):
698 """Report webpage download."""
699 self.to_screen('%s: Downloading webpage' % video_id)
701 def report_age_confirmation(self):
702 """Report attempt to confirm age."""
703 self.to_screen('Confirming age')
705 def report_login(self):
706 """Report attempt to log in."""
707 self.to_screen('Logging in')
710 def raise_login_required(msg='This video is only available for registered users'):
711 raise ExtractorError(
712 '%s. Use --username and --password or --netrc to provide account credentials.' % msg,
716 def raise_geo_restricted(msg='This video is not available from your location due to geo restriction', countries=None):
717 raise GeoRestrictedError(msg, countries=countries)
719 # Methods for following #608
721 def url_result(url, ie=None, video_id=None, video_title=None):
722 """Returns a URL that points to a page that should be processed"""
723 # TODO: ie should be the class used for getting the info
724 video_info = {'_type': 'url',
727 if video_id is not None:
728 video_info['id'] = video_id
729 if video_title is not None:
730 video_info['title'] = video_title
733 def playlist_from_matches(self, matches, playlist_id=None, playlist_title=None, getter=None, ie=None):
735 self.url_result(self._proto_relative_url(getter(m) if getter else m), ie)
737 return self.playlist_result(
738 urls, playlist_id=playlist_id, playlist_title=playlist_title)
741 def playlist_result(entries, playlist_id=None, playlist_title=None, playlist_description=None):
742 """Returns a playlist"""
743 video_info = {'_type': 'playlist',
746 video_info['id'] = playlist_id
748 video_info['title'] = playlist_title
749 if playlist_description:
750 video_info['description'] = playlist_description
753 def _search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
755 Perform a regex search on the given string, using a single or a list of
756 patterns returning the first matching group.
757 In case of failure return a default value or raise a WARNING or a
758 RegexNotFoundError, depending on fatal, specifying the field name.
760 if isinstance(pattern, (str, compat_str, compiled_regex_type)):
761 mobj = re.search(pattern, string, flags)
764 mobj = re.search(p, string, flags)
768 if not self._downloader.params.get('no_color') and compat_os_name != 'nt' and sys.stderr.isatty():
769 _name = '\033[0;34m%s\033[0m' % name
775 # return the first matching group
776 return next(g for g in mobj.groups() if g is not None)
778 return mobj.group(group)
779 elif default is not NO_DEFAULT:
782 raise RegexNotFoundError('Unable to extract %s' % _name)
784 self._downloader.report_warning('unable to extract %s' % _name + bug_reports_message())
787 def _html_search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
789 Like _search_regex, but strips HTML tags and unescapes entities.
791 res = self._search_regex(pattern, string, name, default, fatal, flags, group)
793 return clean_html(res).strip()
797 def _get_netrc_login_info(self, netrc_machine=None):
800 netrc_machine = netrc_machine or self._NETRC_MACHINE
802 if self._downloader.params.get('usenetrc', False):
804 info = netrc.netrc().authenticators(netrc_machine)
809 raise netrc.NetrcParseError(
810 'No authenticators for %s' % netrc_machine)
811 except (IOError, netrc.NetrcParseError) as err:
812 self._downloader.report_warning(
813 'parsing .netrc: %s' % error_to_compat_str(err))
815 return username, password
817 def _get_login_info(self, username_option='username', password_option='password', netrc_machine=None):
819 Get the login info as (username, password)
820 First look for the manually specified credentials using username_option
821 and password_option as keys in params dictionary. If no such credentials
822 available look in the netrc file using the netrc_machine or _NETRC_MACHINE
824 If there's no info available, return (None, None)
826 if self._downloader is None:
829 downloader_params = self._downloader.params
831 # Attempt to use provided username and password or .netrc data
832 if downloader_params.get(username_option) is not None:
833 username = downloader_params[username_option]
834 password = downloader_params[password_option]
836 username, password = self._get_netrc_login_info(netrc_machine)
838 return username, password
840 def _get_tfa_info(self, note='two-factor verification code'):
842 Get the two-factor authentication info
843 TODO - asking the user will be required for sms/phone verify
844 currently just uses the command line option
845 If there's no info available, return None
847 if self._downloader is None:
849 downloader_params = self._downloader.params
851 if downloader_params.get('twofactor') is not None:
852 return downloader_params['twofactor']
854 return compat_getpass('Type %s and press [Return]: ' % note)
856 # Helper functions for extracting OpenGraph info
858 def _og_regexes(prop):
859 content_re = r'content=(?:"([^"]+?)"|\'([^\']+?)\'|\s*([^\s"\'=<>`]+?))'
860 property_re = (r'(?:name|property)=(?:\'og:%(prop)s\'|"og:%(prop)s"|\s*og:%(prop)s\b)'
861 % {'prop': re.escape(prop)})
862 template = r'<meta[^>]+?%s[^>]+?%s'
864 template % (property_re, content_re),
865 template % (content_re, property_re),
869 def _meta_regex(prop):
870 return r'''(?isx)<meta
871 (?=[^>]+(?:itemprop|name|property|id|http-equiv)=(["\']?)%s\1)
872 [^>]+?content=(["\'])(?P<content>.*?)\2''' % re.escape(prop)
874 def _og_search_property(self, prop, html, name=None, **kargs):
875 if not isinstance(prop, (list, tuple)):
878 name = 'OpenGraph %s' % prop[0]
881 og_regexes.extend(self._og_regexes(p))
882 escaped = self._search_regex(og_regexes, html, name, flags=re.DOTALL, **kargs)
885 return unescapeHTML(escaped)
887 def _og_search_thumbnail(self, html, **kargs):
888 return self._og_search_property('image', html, 'thumbnail URL', fatal=False, **kargs)
890 def _og_search_description(self, html, **kargs):
891 return self._og_search_property('description', html, fatal=False, **kargs)
893 def _og_search_title(self, html, **kargs):
894 return self._og_search_property('title', html, **kargs)
896 def _og_search_video_url(self, html, name='video url', secure=True, **kargs):
897 regexes = self._og_regexes('video') + self._og_regexes('video:url')
899 regexes = self._og_regexes('video:secure_url') + regexes
900 return self._html_search_regex(regexes, html, name, **kargs)
902 def _og_search_url(self, html, **kargs):
903 return self._og_search_property('url', html, **kargs)
905 def _html_search_meta(self, name, html, display_name=None, fatal=False, **kwargs):
906 if not isinstance(name, (list, tuple)):
908 if display_name is None:
909 display_name = name[0]
910 return self._html_search_regex(
911 [self._meta_regex(n) for n in name],
912 html, display_name, fatal=fatal, group='content', **kwargs)
914 def _dc_search_uploader(self, html):
915 return self._html_search_meta('dc.creator', html, 'uploader')
917 def _rta_search(self, html):
918 # See http://www.rtalabel.org/index.php?content=howtofaq#single
919 if re.search(r'(?ix)<meta\s+name="rating"\s+'
920 r' content="RTA-5042-1996-1400-1577-RTA"',
925 def _media_rating_search(self, html):
926 # See http://www.tjg-designs.com/WP/metadata-code-examples-adding-metadata-to-your-web-pages/
927 rating = self._html_search_meta('rating', html)
939 return RATING_TABLE.get(rating.lower())
941 def _family_friendly_search(self, html):
942 # See http://schema.org/VideoObject
943 family_friendly = self._html_search_meta(
944 'isFamilyFriendly', html, default=None)
946 if not family_friendly:
955 return RATING_TABLE.get(family_friendly.lower())
957 def _twitter_search_player(self, html):
958 return self._html_search_meta('twitter:player', html,
959 'twitter card player')
961 def _search_json_ld(self, html, video_id, expected_type=None, **kwargs):
962 json_ld = self._search_regex(
963 r'(?s)<script[^>]+type=(["\'])application/ld\+json\1[^>]*>(?P<json_ld>.+?)</script>',
964 html, 'JSON-LD', group='json_ld', **kwargs)
965 default = kwargs.get('default', NO_DEFAULT)
967 return default if default is not NO_DEFAULT else {}
968 # JSON-LD may be malformed and thus `fatal` should be respected.
969 # At the same time `default` may be passed that assumes `fatal=False`
970 # for _search_regex. Let's simulate the same behavior here as well.
971 fatal = kwargs.get('fatal', True) if default == NO_DEFAULT else False
972 return self._json_ld(json_ld, video_id, fatal=fatal, expected_type=expected_type)
974 def _json_ld(self, json_ld, video_id, fatal=True, expected_type=None):
975 if isinstance(json_ld, compat_str):
976 json_ld = self._parse_json(json_ld, video_id, fatal=fatal)
980 if not isinstance(json_ld, (list, tuple, dict)):
982 if isinstance(json_ld, dict):
985 def extract_video_object(e):
986 assert e['@type'] == 'VideoObject'
988 'url': e.get('contentUrl'),
989 'title': unescapeHTML(e.get('name')),
990 'description': unescapeHTML(e.get('description')),
991 'thumbnail': e.get('thumbnailUrl') or e.get('thumbnailURL'),
992 'duration': parse_duration(e.get('duration')),
993 'timestamp': unified_timestamp(e.get('uploadDate')),
994 'filesize': float_or_none(e.get('contentSize')),
995 'tbr': int_or_none(e.get('bitrate')),
996 'width': int_or_none(e.get('width')),
997 'height': int_or_none(e.get('height')),
998 'view_count': int_or_none(e.get('interactionCount')),
1002 if e.get('@context') == 'http://schema.org':
1003 item_type = e.get('@type')
1004 if expected_type is not None and expected_type != item_type:
1006 if item_type in ('TVEpisode', 'Episode'):
1008 'episode': unescapeHTML(e.get('name')),
1009 'episode_number': int_or_none(e.get('episodeNumber')),
1010 'description': unescapeHTML(e.get('description')),
1012 part_of_season = e.get('partOfSeason')
1013 if isinstance(part_of_season, dict) and part_of_season.get('@type') in ('TVSeason', 'Season', 'CreativeWorkSeason'):
1014 info['season_number'] = int_or_none(part_of_season.get('seasonNumber'))
1015 part_of_series = e.get('partOfSeries') or e.get('partOfTVSeries')
1016 if isinstance(part_of_series, dict) and part_of_series.get('@type') in ('TVSeries', 'Series', 'CreativeWorkSeries'):
1017 info['series'] = unescapeHTML(part_of_series.get('name'))
1018 elif item_type == 'Article':
1020 'timestamp': parse_iso8601(e.get('datePublished')),
1021 'title': unescapeHTML(e.get('headline')),
1022 'description': unescapeHTML(e.get('articleBody')),
1024 elif item_type == 'VideoObject':
1025 extract_video_object(e)
1027 video = e.get('video')
1028 if isinstance(video, dict) and video.get('@type') == 'VideoObject':
1029 extract_video_object(video)
1031 return dict((k, v) for k, v in info.items() if v is not None)
1034 def _hidden_inputs(html):
1035 html = re.sub(r'<!--(?:(?!<!--).)*-->', '', html)
1037 for input in re.findall(r'(?i)(<input[^>]+>)', html):
1038 attrs = extract_attributes(input)
1041 if attrs.get('type') not in ('hidden', 'submit'):
1043 name = attrs.get('name') or attrs.get('id')
1044 value = attrs.get('value')
1045 if name and value is not None:
1046 hidden_inputs[name] = value
1047 return hidden_inputs
1049 def _form_hidden_inputs(self, form_id, html):
1050 form = self._search_regex(
1051 r'(?is)<form[^>]+?id=(["\'])%s\1[^>]*>(?P<form>.+?)</form>' % form_id,
1052 html, '%s form' % form_id, group='form')
1053 return self._hidden_inputs(form)
1055 def _sort_formats(self, formats, field_preference=None):
1057 raise ExtractorError('No video formats found')
1060 # Automatically determine tbr when missing based on abr and vbr (improves
1061 # formats sorting in some cases)
1062 if 'tbr' not in f and f.get('abr') is not None and f.get('vbr') is not None:
1063 f['tbr'] = f['abr'] + f['vbr']
1065 def _formats_key(f):
1066 # TODO remove the following workaround
1067 from ..utils import determine_ext
1068 if not f.get('ext') and 'url' in f:
1069 f['ext'] = determine_ext(f['url'])
1071 if isinstance(field_preference, (list, tuple)):
1074 if f.get(field) is not None
1075 else ('' if field == 'format_id' else -1)
1076 for field in field_preference)
1078 preference = f.get('preference')
1079 if preference is None:
1081 if f.get('ext') in ['f4f', 'f4m']: # Not yet supported
1084 protocol = f.get('protocol') or determine_protocol(f)
1085 proto_preference = 0 if protocol in ['http', 'https'] else (-0.5 if protocol == 'rtsp' else -0.1)
1087 if f.get('vcodec') == 'none': # audio only
1089 if self._downloader.params.get('prefer_free_formats'):
1090 ORDER = ['aac', 'mp3', 'm4a', 'webm', 'ogg', 'opus']
1092 ORDER = ['webm', 'opus', 'ogg', 'mp3', 'aac', 'm4a']
1095 audio_ext_preference = ORDER.index(f['ext'])
1097 audio_ext_preference = -1
1099 if f.get('acodec') == 'none': # video only
1101 if self._downloader.params.get('prefer_free_formats'):
1102 ORDER = ['flv', 'mp4', 'webm']
1104 ORDER = ['webm', 'flv', 'mp4']
1106 ext_preference = ORDER.index(f['ext'])
1109 audio_ext_preference = 0
1113 f.get('language_preference') if f.get('language_preference') is not None else -1,
1114 f.get('quality') if f.get('quality') is not None else -1,
1115 f.get('tbr') if f.get('tbr') is not None else -1,
1116 f.get('filesize') if f.get('filesize') is not None else -1,
1117 f.get('vbr') if f.get('vbr') is not None else -1,
1118 f.get('height') if f.get('height') is not None else -1,
1119 f.get('width') if f.get('width') is not None else -1,
1122 f.get('abr') if f.get('abr') is not None else -1,
1123 audio_ext_preference,
1124 f.get('fps') if f.get('fps') is not None else -1,
1125 f.get('filesize_approx') if f.get('filesize_approx') is not None else -1,
1126 f.get('source_preference') if f.get('source_preference') is not None else -1,
1127 f.get('format_id') if f.get('format_id') is not None else '',
1129 formats.sort(key=_formats_key)
1131 def _check_formats(self, formats, video_id):
1133 formats[:] = filter(
1134 lambda f: self._is_valid_url(
1136 item='%s video format' % f.get('format_id') if f.get('format_id') else 'video'),
1140 def _remove_duplicate_formats(formats):
1144 if f['url'] not in format_urls:
1145 format_urls.add(f['url'])
1146 unique_formats.append(f)
1147 formats[:] = unique_formats
1149 def _is_valid_url(self, url, video_id, item='video', headers={}):
1150 url = self._proto_relative_url(url, scheme='http:')
1151 # For now assume non HTTP(S) URLs always valid
1152 if not (url.startswith('http://') or url.startswith('https://')):
1155 self._request_webpage(url, video_id, 'Checking %s URL' % item, headers=headers)
1157 except ExtractorError as e:
1158 if isinstance(e.cause, compat_urllib_error.URLError):
1160 '%s: %s URL is invalid, skipping' % (video_id, item))
1164 def http_scheme(self):
1165 """ Either "http:" or "https:", depending on the user's preferences """
1168 if self._downloader.params.get('prefer_insecure', False)
1171 def _proto_relative_url(self, url, scheme=None):
1174 if url.startswith('//'):
1176 scheme = self.http_scheme()
1181 def _sleep(self, timeout, video_id, msg_template=None):
1182 if msg_template is None:
1183 msg_template = '%(video_id)s: Waiting for %(timeout)s seconds'
1184 msg = msg_template % {'video_id': video_id, 'timeout': timeout}
1188 def _extract_f4m_formats(self, manifest_url, video_id, preference=None, f4m_id=None,
1189 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1190 fatal=True, m3u8_id=None):
1191 manifest = self._download_xml(
1192 manifest_url, video_id, 'Downloading f4m manifest',
1193 'Unable to download f4m manifest',
1194 # Some manifests may be malformed, e.g. prosiebensat1 generated manifests
1195 # (see https://github.com/rg3/youtube-dl/issues/6215#issuecomment-121704244)
1196 transform_source=transform_source,
1199 if manifest is False:
1202 return self._parse_f4m_formats(
1203 manifest, manifest_url, video_id, preference=preference, f4m_id=f4m_id,
1204 transform_source=transform_source, fatal=fatal, m3u8_id=m3u8_id)
1206 def _parse_f4m_formats(self, manifest, manifest_url, video_id, preference=None, f4m_id=None,
1207 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1208 fatal=True, m3u8_id=None):
1209 # currently youtube-dl cannot decode the playerVerificationChallenge as Akamai uses Adobe Alchemy
1210 akamai_pv = manifest.find('{http://ns.adobe.com/f4m/1.0}pv-2.0')
1211 if akamai_pv is not None and ';' in akamai_pv.text:
1212 playerVerificationChallenge = akamai_pv.text.split(';')[0]
1213 if playerVerificationChallenge.strip() != '':
1217 manifest_version = '1.0'
1218 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/1.0}media')
1220 manifest_version = '2.0'
1221 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/2.0}media')
1222 # Remove unsupported DRM protected media from final formats
1223 # rendition (see https://github.com/rg3/youtube-dl/issues/8573).
1224 media_nodes = remove_encrypted_media(media_nodes)
1227 base_url = xpath_text(
1228 manifest, ['{http://ns.adobe.com/f4m/1.0}baseURL', '{http://ns.adobe.com/f4m/2.0}baseURL'],
1229 'base URL', default=None)
1231 base_url = base_url.strip()
1233 bootstrap_info = xpath_element(
1234 manifest, ['{http://ns.adobe.com/f4m/1.0}bootstrapInfo', '{http://ns.adobe.com/f4m/2.0}bootstrapInfo'],
1235 'bootstrap info', default=None)
1238 mime_type = xpath_text(
1239 manifest, ['{http://ns.adobe.com/f4m/1.0}mimeType', '{http://ns.adobe.com/f4m/2.0}mimeType'],
1240 'base URL', default=None)
1241 if mime_type and mime_type.startswith('audio/'):
1244 for i, media_el in enumerate(media_nodes):
1245 tbr = int_or_none(media_el.attrib.get('bitrate'))
1246 width = int_or_none(media_el.attrib.get('width'))
1247 height = int_or_none(media_el.attrib.get('height'))
1248 format_id = '-'.join(filter(None, [f4m_id, compat_str(i if tbr is None else tbr)]))
1249 # If <bootstrapInfo> is present, the specified f4m is a
1250 # stream-level manifest, and only set-level manifests may refer to
1251 # external resources. See section 11.4 and section 4 of F4M spec
1252 if bootstrap_info is None:
1254 # @href is introduced in 2.0, see section 11.6 of F4M spec
1255 if manifest_version == '2.0':
1256 media_url = media_el.attrib.get('href')
1257 if media_url is None:
1258 media_url = media_el.attrib.get('url')
1262 media_url if media_url.startswith('http://') or media_url.startswith('https://')
1263 else ((base_url or '/'.join(manifest_url.split('/')[:-1])) + '/' + media_url))
1264 # If media_url is itself a f4m manifest do the recursive extraction
1265 # since bitrates in parent manifest (this one) and media_url manifest
1266 # may differ leading to inability to resolve the format by requested
1267 # bitrate in f4m downloader
1268 ext = determine_ext(manifest_url)
1270 f4m_formats = self._extract_f4m_formats(
1271 manifest_url, video_id, preference=preference, f4m_id=f4m_id,
1272 transform_source=transform_source, fatal=fatal)
1273 # Sometimes stream-level manifest contains single media entry that
1274 # does not contain any quality metadata (e.g. http://matchtv.ru/#live-player).
1275 # At the same time parent's media entry in set-level manifest may
1276 # contain it. We will copy it from parent in such cases.
1277 if len(f4m_formats) == 1:
1280 'tbr': f.get('tbr') or tbr,
1281 'width': f.get('width') or width,
1282 'height': f.get('height') or height,
1283 'format_id': f.get('format_id') if not tbr else format_id,
1286 formats.extend(f4m_formats)
1289 formats.extend(self._extract_m3u8_formats(
1290 manifest_url, video_id, 'mp4', preference=preference,
1291 m3u8_id=m3u8_id, fatal=fatal))
1294 'format_id': format_id,
1295 'url': manifest_url,
1296 'manifest_url': manifest_url,
1297 'ext': 'flv' if bootstrap_info is not None else None,
1302 'preference': preference,
1306 def _m3u8_meta_format(self, m3u8_url, ext=None, preference=None, m3u8_id=None):
1308 'format_id': '-'.join(filter(None, [m3u8_id, 'meta'])),
1312 'preference': preference - 100 if preference else -100,
1313 'resolution': 'multiple',
1314 'format_note': 'Quality selection URL',
1317 def _extract_m3u8_formats(self, m3u8_url, video_id, ext=None,
1318 entry_protocol='m3u8', preference=None,
1319 m3u8_id=None, note=None, errnote=None,
1320 fatal=True, live=False):
1321 res = self._download_webpage_handle(
1323 note=note or 'Downloading m3u8 information',
1324 errnote=errnote or 'Failed to download m3u8 information',
1330 m3u8_doc, urlh = res
1331 m3u8_url = urlh.geturl()
1333 return self._parse_m3u8_formats(
1334 m3u8_doc, m3u8_url, ext=ext, entry_protocol=entry_protocol,
1335 preference=preference, m3u8_id=m3u8_id, live=live)
1337 def _parse_m3u8_formats(self, m3u8_doc, m3u8_url, ext=None,
1338 entry_protocol='m3u8', preference=None,
1339 m3u8_id=None, live=False):
1340 if '#EXT-X-FAXS-CM:' in m3u8_doc: # Adobe Flash Access
1345 format_url = lambda u: (
1347 if re.match(r'^https?://', u)
1348 else compat_urlparse.urljoin(m3u8_url, u))
1351 # 1. https://tools.ietf.org/html/draft-pantos-http-live-streaming-21
1352 # 2. https://github.com/rg3/youtube-dl/issues/12211
1354 # We should try extracting formats only from master playlists [1, 4.3.4],
1355 # i.e. playlists that describe available qualities. On the other hand
1356 # media playlists [1, 4.3.3] should be returned as is since they contain
1357 # just the media without qualities renditions.
1358 # Fortunately, master playlist can be easily distinguished from media
1359 # playlist based on particular tags availability. As of [1, 4.3.3, 4.3.4]
1360 # master playlist tags MUST NOT appear in a media playist and vice versa.
1361 # As of [1, 4.3.3.1] #EXT-X-TARGETDURATION tag is REQUIRED for every
1362 # media playlist and MUST NOT appear in master playlist thus we can
1363 # clearly detect media playlist with this criterion.
1365 if '#EXT-X-TARGETDURATION' in m3u8_doc: # media playlist, return as is
1368 'format_id': m3u8_id,
1370 'protocol': entry_protocol,
1371 'preference': preference,
1375 last_stream_inf = {}
1377 def extract_media(x_media_line):
1378 media = parse_m3u8_attributes(x_media_line)
1379 # As per [1, 4.3.4.1] TYPE, GROUP-ID and NAME are REQUIRED
1380 media_type, group_id, name = media.get('TYPE'), media.get('GROUP-ID'), media.get('NAME')
1381 if not (media_type and group_id and name):
1383 groups.setdefault(group_id, []).append(media)
1384 if media_type not in ('VIDEO', 'AUDIO'):
1386 media_url = media.get('URI')
1389 for v in (group_id, name):
1393 'format_id': '-'.join(format_id),
1394 'url': format_url(media_url),
1395 'manifest_url': m3u8_url,
1396 'language': media.get('LANGUAGE'),
1398 'protocol': entry_protocol,
1399 'preference': preference,
1401 if media_type == 'AUDIO':
1402 f['vcodec'] = 'none'
1405 def build_stream_name():
1406 # Despite specification does not mention NAME attribute for
1407 # EXT-X-STREAM-INF tag it still sometimes may be present (see [1]
1408 # or vidio test in TestInfoExtractor.test_parse_m3u8_formats)
1409 # 1. http://www.vidio.com/watch/165683-dj_ambred-booyah-live-2015
1410 stream_name = last_stream_inf.get('NAME')
1413 # If there is no NAME in EXT-X-STREAM-INF it will be obtained
1414 # from corresponding rendition group
1415 stream_group_id = last_stream_inf.get('VIDEO')
1416 if not stream_group_id:
1418 stream_group = groups.get(stream_group_id)
1419 if not stream_group:
1420 return stream_group_id
1421 rendition = stream_group[0]
1422 return rendition.get('NAME') or stream_group_id
1424 for line in m3u8_doc.splitlines():
1425 if line.startswith('#EXT-X-STREAM-INF:'):
1426 last_stream_inf = parse_m3u8_attributes(line)
1427 elif line.startswith('#EXT-X-MEDIA:'):
1429 elif line.startswith('#') or not line.strip():
1432 tbr = float_or_none(
1433 last_stream_inf.get('AVERAGE-BANDWIDTH') or
1434 last_stream_inf.get('BANDWIDTH'), scale=1000)
1437 format_id.append(m3u8_id)
1438 stream_name = build_stream_name()
1439 # Bandwidth of live streams may differ over time thus making
1440 # format_id unpredictable. So it's better to keep provided
1443 format_id.append(stream_name if stream_name else '%d' % (tbr if tbr else len(formats)))
1444 manifest_url = format_url(line.strip())
1446 'format_id': '-'.join(format_id),
1447 'url': manifest_url,
1448 'manifest_url': m3u8_url,
1451 'fps': float_or_none(last_stream_inf.get('FRAME-RATE')),
1452 'protocol': entry_protocol,
1453 'preference': preference,
1455 resolution = last_stream_inf.get('RESOLUTION')
1457 mobj = re.search(r'(?P<width>\d+)[xX](?P<height>\d+)', resolution)
1459 f['width'] = int(mobj.group('width'))
1460 f['height'] = int(mobj.group('height'))
1461 # Unified Streaming Platform
1463 r'audio.*?(?:%3D|=)(\d+)(?:-video.*?(?:%3D|=)(\d+))?', f['url'])
1465 abr, vbr = mobj.groups()
1466 abr, vbr = float_or_none(abr, 1000), float_or_none(vbr, 1000)
1471 codecs = parse_codecs(last_stream_inf.get('CODECS'))
1473 audio_group_id = last_stream_inf.get('AUDIO')
1474 # As per [1, 4.3.4.1.1] any EXT-X-STREAM-INF tag which
1475 # references a rendition group MUST have a CODECS attribute.
1476 # However, this is not always respected, for example, [2]
1477 # contains EXT-X-STREAM-INF tag which references AUDIO
1478 # rendition group but does not have CODECS and despite
1479 # referencing audio group an audio group, it represents
1480 # a complete (with audio and video) format. So, for such cases
1481 # we will ignore references to rendition groups and treat them
1482 # as complete formats.
1483 if audio_group_id and codecs and f.get('vcodec') != 'none':
1484 audio_group = groups.get(audio_group_id)
1485 if audio_group and audio_group[0].get('URI'):
1486 # TODO: update acodec for audio only formats with
1488 f['acodec'] = 'none'
1490 last_stream_inf = {}
1494 def _xpath_ns(path, namespace=None):
1498 for c in path.split('/'):
1499 if not c or c == '.':
1502 out.append('{%s}%s' % (namespace, c))
1503 return '/'.join(out)
1505 def _extract_smil_formats(self, smil_url, video_id, fatal=True, f4m_params=None, transform_source=None):
1506 smil = self._download_smil(smil_url, video_id, fatal=fatal, transform_source=transform_source)
1512 namespace = self._parse_smil_namespace(smil)
1514 return self._parse_smil_formats(
1515 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
1517 def _extract_smil_info(self, smil_url, video_id, fatal=True, f4m_params=None):
1518 smil = self._download_smil(smil_url, video_id, fatal=fatal)
1521 return self._parse_smil(smil, smil_url, video_id, f4m_params=f4m_params)
1523 def _download_smil(self, smil_url, video_id, fatal=True, transform_source=None):
1524 return self._download_xml(
1525 smil_url, video_id, 'Downloading SMIL file',
1526 'Unable to download SMIL file', fatal=fatal, transform_source=transform_source)
1528 def _parse_smil(self, smil, smil_url, video_id, f4m_params=None):
1529 namespace = self._parse_smil_namespace(smil)
1531 formats = self._parse_smil_formats(
1532 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
1533 subtitles = self._parse_smil_subtitles(smil, namespace=namespace)
1535 video_id = os.path.splitext(url_basename(smil_url))[0]
1539 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
1540 name = meta.attrib.get('name')
1541 content = meta.attrib.get('content')
1542 if not name or not content:
1544 if not title and name == 'title':
1546 elif not description and name in ('description', 'abstract'):
1547 description = content
1548 elif not upload_date and name == 'date':
1549 upload_date = unified_strdate(content)
1552 'id': image.get('type'),
1553 'url': image.get('src'),
1554 'width': int_or_none(image.get('width')),
1555 'height': int_or_none(image.get('height')),
1556 } for image in smil.findall(self._xpath_ns('.//image', namespace)) if image.get('src')]
1560 'title': title or video_id,
1561 'description': description,
1562 'upload_date': upload_date,
1563 'thumbnails': thumbnails,
1565 'subtitles': subtitles,
1568 def _parse_smil_namespace(self, smil):
1569 return self._search_regex(
1570 r'(?i)^{([^}]+)?}smil$', smil.tag, 'namespace', default=None)
1572 def _parse_smil_formats(self, smil, smil_url, video_id, namespace=None, f4m_params=None, transform_rtmp_url=None):
1574 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
1575 b = meta.get('base') or meta.get('httpBase')
1586 media = smil.findall(self._xpath_ns('.//video', namespace)) + smil.findall(self._xpath_ns('.//audio', namespace))
1587 for medium in media:
1588 src = medium.get('src')
1589 if not src or src in srcs:
1593 bitrate = float_or_none(medium.get('system-bitrate') or medium.get('systemBitrate'), 1000)
1594 filesize = int_or_none(medium.get('size') or medium.get('fileSize'))
1595 width = int_or_none(medium.get('width'))
1596 height = int_or_none(medium.get('height'))
1597 proto = medium.get('proto')
1598 ext = medium.get('ext')
1599 src_ext = determine_ext(src)
1600 streamer = medium.get('streamer') or base
1602 if proto == 'rtmp' or streamer.startswith('rtmp'):
1608 'format_id': 'rtmp-%d' % (rtmp_count if bitrate is None else bitrate),
1610 'filesize': filesize,
1614 if transform_rtmp_url:
1615 streamer, src = transform_rtmp_url(streamer, src)
1616 formats[-1].update({
1622 src_url = src if src.startswith('http') else compat_urlparse.urljoin(base, src)
1623 src_url = src_url.strip()
1625 if proto == 'm3u8' or src_ext == 'm3u8':
1626 m3u8_formats = self._extract_m3u8_formats(
1627 src_url, video_id, ext or 'mp4', m3u8_id='hls', fatal=False)
1628 if len(m3u8_formats) == 1:
1630 m3u8_formats[0].update({
1631 'format_id': 'hls-%d' % (m3u8_count if bitrate is None else bitrate),
1636 formats.extend(m3u8_formats)
1639 if src_ext == 'f4m':
1644 'plugin': 'flowplayer-3.2.0.1',
1646 f4m_url += '&' if '?' in f4m_url else '?'
1647 f4m_url += compat_urllib_parse_urlencode(f4m_params)
1648 formats.extend(self._extract_f4m_formats(f4m_url, video_id, f4m_id='hds', fatal=False))
1651 if src_url.startswith('http') and self._is_valid_url(src, video_id):
1655 'ext': ext or src_ext or 'flv',
1656 'format_id': 'http-%d' % (bitrate or http_count),
1658 'filesize': filesize,
1666 def _parse_smil_subtitles(self, smil, namespace=None, subtitles_lang='en'):
1669 for num, textstream in enumerate(smil.findall(self._xpath_ns('.//textstream', namespace))):
1670 src = textstream.get('src')
1671 if not src or src in urls:
1674 ext = textstream.get('ext') or mimetype2ext(textstream.get('type')) or determine_ext(src)
1675 lang = textstream.get('systemLanguage') or textstream.get('systemLanguageName') or textstream.get('lang') or subtitles_lang
1676 subtitles.setdefault(lang, []).append({
1682 def _extract_xspf_playlist(self, playlist_url, playlist_id, fatal=True):
1683 xspf = self._download_xml(
1684 playlist_url, playlist_id, 'Downloading xpsf playlist',
1685 'Unable to download xspf manifest', fatal=fatal)
1688 return self._parse_xspf(xspf, playlist_id)
1690 def _parse_xspf(self, playlist, playlist_id):
1692 'xspf': 'http://xspf.org/ns/0/',
1693 's1': 'http://static.streamone.nl/player/ns/0',
1697 for track in playlist.findall(xpath_with_ns('./xspf:trackList/xspf:track', NS_MAP)):
1699 track, xpath_with_ns('./xspf:title', NS_MAP), 'title', default=playlist_id)
1700 description = xpath_text(
1701 track, xpath_with_ns('./xspf:annotation', NS_MAP), 'description')
1702 thumbnail = xpath_text(
1703 track, xpath_with_ns('./xspf:image', NS_MAP), 'thumbnail')
1704 duration = float_or_none(
1705 xpath_text(track, xpath_with_ns('./xspf:duration', NS_MAP), 'duration'), 1000)
1708 'url': location.text,
1709 'format_id': location.get(xpath_with_ns('s1:label', NS_MAP)),
1710 'width': int_or_none(location.get(xpath_with_ns('s1:width', NS_MAP))),
1711 'height': int_or_none(location.get(xpath_with_ns('s1:height', NS_MAP))),
1712 } for location in track.findall(xpath_with_ns('./xspf:location', NS_MAP))]
1713 self._sort_formats(formats)
1718 'description': description,
1719 'thumbnail': thumbnail,
1720 'duration': duration,
1725 def _extract_mpd_formats(self, mpd_url, video_id, mpd_id=None, note=None, errnote=None, fatal=True, formats_dict={}):
1726 res = self._download_webpage_handle(
1728 note=note or 'Downloading MPD manifest',
1729 errnote=errnote or 'Failed to download MPD manifest',
1734 mpd_base_url = base_url(urlh.geturl())
1736 return self._parse_mpd_formats(
1737 compat_etree_fromstring(mpd.encode('utf-8')), mpd_id, mpd_base_url,
1738 formats_dict=formats_dict, mpd_url=mpd_url)
1740 def _parse_mpd_formats(self, mpd_doc, mpd_id=None, mpd_base_url='', formats_dict={}, mpd_url=None):
1742 Parse formats from MPD manifest.
1744 1. MPEG-DASH Standard, ISO/IEC 23009-1:2014(E),
1745 http://standards.iso.org/ittf/PubliclyAvailableStandards/c065274_ISO_IEC_23009-1_2014.zip
1746 2. https://en.wikipedia.org/wiki/Dynamic_Adaptive_Streaming_over_HTTP
1748 if mpd_doc.get('type') == 'dynamic':
1751 namespace = self._search_regex(r'(?i)^{([^}]+)?}MPD$', mpd_doc.tag, 'namespace', default=None)
1754 return self._xpath_ns(path, namespace)
1756 def is_drm_protected(element):
1757 return element.find(_add_ns('ContentProtection')) is not None
1759 def extract_multisegment_info(element, ms_parent_info):
1760 ms_info = ms_parent_info.copy()
1762 # As per [1, 5.3.9.2.2] SegmentList and SegmentTemplate share some
1763 # common attributes and elements. We will only extract relevant
1765 def extract_common(source):
1766 segment_timeline = source.find(_add_ns('SegmentTimeline'))
1767 if segment_timeline is not None:
1768 s_e = segment_timeline.findall(_add_ns('S'))
1770 ms_info['total_number'] = 0
1773 r = int(s.get('r', 0))
1774 ms_info['total_number'] += 1 + r
1775 ms_info['s'].append({
1776 't': int(s.get('t', 0)),
1777 # @d is mandatory (see [1, 5.3.9.6.2, Table 17, page 60])
1778 'd': int(s.attrib['d']),
1781 start_number = source.get('startNumber')
1783 ms_info['start_number'] = int(start_number)
1784 timescale = source.get('timescale')
1786 ms_info['timescale'] = int(timescale)
1787 segment_duration = source.get('duration')
1788 if segment_duration:
1789 ms_info['segment_duration'] = int(segment_duration)
1791 def extract_Initialization(source):
1792 initialization = source.find(_add_ns('Initialization'))
1793 if initialization is not None:
1794 ms_info['initialization_url'] = initialization.attrib['sourceURL']
1796 segment_list = element.find(_add_ns('SegmentList'))
1797 if segment_list is not None:
1798 extract_common(segment_list)
1799 extract_Initialization(segment_list)
1800 segment_urls_e = segment_list.findall(_add_ns('SegmentURL'))
1802 ms_info['segment_urls'] = [segment.attrib['media'] for segment in segment_urls_e]
1804 segment_template = element.find(_add_ns('SegmentTemplate'))
1805 if segment_template is not None:
1806 extract_common(segment_template)
1807 media = segment_template.get('media')
1809 ms_info['media'] = media
1810 initialization = segment_template.get('initialization')
1812 ms_info['initialization'] = initialization
1814 extract_Initialization(segment_template)
1817 mpd_duration = parse_duration(mpd_doc.get('mediaPresentationDuration'))
1819 for period in mpd_doc.findall(_add_ns('Period')):
1820 period_duration = parse_duration(period.get('duration')) or mpd_duration
1821 period_ms_info = extract_multisegment_info(period, {
1825 for adaptation_set in period.findall(_add_ns('AdaptationSet')):
1826 if is_drm_protected(adaptation_set):
1828 adaption_set_ms_info = extract_multisegment_info(adaptation_set, period_ms_info)
1829 for representation in adaptation_set.findall(_add_ns('Representation')):
1830 if is_drm_protected(representation):
1832 representation_attrib = adaptation_set.attrib.copy()
1833 representation_attrib.update(representation.attrib)
1834 # According to [1, 5.3.7.2, Table 9, page 41], @mimeType is mandatory
1835 mime_type = representation_attrib['mimeType']
1836 content_type = mime_type.split('/')[0]
1837 if content_type == 'text':
1838 # TODO implement WebVTT downloading
1840 elif content_type in ('video', 'audio'):
1842 for element in (representation, adaptation_set, period, mpd_doc):
1843 base_url_e = element.find(_add_ns('BaseURL'))
1844 if base_url_e is not None:
1845 base_url = base_url_e.text + base_url
1846 if re.match(r'^https?://', base_url):
1848 if mpd_base_url and not re.match(r'^https?://', base_url):
1849 if not mpd_base_url.endswith('/') and not base_url.startswith('/'):
1851 base_url = mpd_base_url + base_url
1852 representation_id = representation_attrib.get('id')
1853 lang = representation_attrib.get('lang')
1854 url_el = representation.find(_add_ns('BaseURL'))
1855 filesize = int_or_none(url_el.attrib.get('{http://youtube.com/yt/2012/10/10}contentLength') if url_el is not None else None)
1856 bandwidth = int_or_none(representation_attrib.get('bandwidth'))
1858 'format_id': '%s-%s' % (mpd_id, representation_id) if mpd_id else representation_id,
1860 'manifest_url': mpd_url,
1861 'ext': mimetype2ext(mime_type),
1862 'width': int_or_none(representation_attrib.get('width')),
1863 'height': int_or_none(representation_attrib.get('height')),
1864 'tbr': float_or_none(bandwidth, 1000),
1865 'asr': int_or_none(representation_attrib.get('audioSamplingRate')),
1866 'fps': int_or_none(representation_attrib.get('frameRate')),
1867 'language': lang if lang not in ('mul', 'und', 'zxx', 'mis') else None,
1868 'format_note': 'DASH %s' % content_type,
1869 'filesize': filesize,
1871 f.update(parse_codecs(representation_attrib.get('codecs')))
1872 representation_ms_info = extract_multisegment_info(representation, adaption_set_ms_info)
1874 def prepare_template(template_name, identifiers):
1875 t = representation_ms_info[template_name]
1876 t = t.replace('$RepresentationID$', representation_id)
1877 t = re.sub(r'\$(%s)\$' % '|'.join(identifiers), r'%(\1)d', t)
1878 t = re.sub(r'\$(%s)%%([^$]+)\$' % '|'.join(identifiers), r'%(\1)\2', t)
1879 t.replace('$$', '$')
1882 # @initialization is a regular template like @media one
1883 # so it should be handled just the same way (see
1884 # https://github.com/rg3/youtube-dl/issues/11605)
1885 if 'initialization' in representation_ms_info:
1886 initialization_template = prepare_template(
1888 # As per [1, 5.3.9.4.2, Table 15, page 54] $Number$ and
1889 # $Time$ shall not be included for @initialization thus
1890 # only $Bandwidth$ remains
1892 representation_ms_info['initialization_url'] = initialization_template % {
1893 'Bandwidth': bandwidth,
1896 def location_key(location):
1897 return 'url' if re.match(r'^https?://', location) else 'path'
1899 if 'segment_urls' not in representation_ms_info and 'media' in representation_ms_info:
1901 media_template = prepare_template('media', ('Number', 'Bandwidth', 'Time'))
1902 media_location_key = location_key(media_template)
1904 # As per [1, 5.3.9.4.4, Table 16, page 55] $Number$ and $Time$
1905 # can't be used at the same time
1906 if '%(Number' in media_template and 's' not in representation_ms_info:
1907 segment_duration = None
1908 if 'total_number' not in representation_ms_info and 'segment_duration':
1909 segment_duration = float_or_none(representation_ms_info['segment_duration'], representation_ms_info['timescale'])
1910 representation_ms_info['total_number'] = int(math.ceil(float(period_duration) / segment_duration))
1911 representation_ms_info['fragments'] = [{
1912 media_location_key: media_template % {
1913 'Number': segment_number,
1914 'Bandwidth': bandwidth,
1916 'duration': segment_duration,
1917 } for segment_number in range(
1918 representation_ms_info['start_number'],
1919 representation_ms_info['total_number'] + representation_ms_info['start_number'])]
1921 # $Number*$ or $Time$ in media template with S list available
1922 # Example $Number*$: http://www.svtplay.se/klipp/9023742/stopptid-om-bjorn-borg
1923 # Example $Time$: https://play.arkena.com/embed/avp/v2/player/media/b41dda37-d8e7-4d3f-b1b5-9a9db578bdfe/1/129411
1924 representation_ms_info['fragments'] = []
1927 segment_number = representation_ms_info['start_number']
1929 def add_segment_url():
1930 segment_url = media_template % {
1931 'Time': segment_time,
1932 'Bandwidth': bandwidth,
1933 'Number': segment_number,
1935 representation_ms_info['fragments'].append({
1936 media_location_key: segment_url,
1937 'duration': float_or_none(segment_d, representation_ms_info['timescale']),
1940 for num, s in enumerate(representation_ms_info['s']):
1941 segment_time = s.get('t') or segment_time
1945 for r in range(s.get('r', 0)):
1946 segment_time += segment_d
1949 segment_time += segment_d
1950 elif 'segment_urls' in representation_ms_info and 's' in representation_ms_info:
1952 # Example: https://www.youtube.com/watch?v=iXZV5uAYMJI
1953 # or any YouTube dashsegments video
1956 timescale = representation_ms_info['timescale']
1957 for s in representation_ms_info['s']:
1958 duration = float_or_none(s['d'], timescale)
1959 for r in range(s.get('r', 0) + 1):
1960 segment_uri = representation_ms_info['segment_urls'][segment_index]
1962 location_key(segment_uri): segment_uri,
1963 'duration': duration,
1966 representation_ms_info['fragments'] = fragments
1967 # NB: MPD manifest may contain direct URLs to unfragmented media.
1968 # No fragments key is present in this case.
1969 if 'fragments' in representation_ms_info:
1971 'fragment_base_url': base_url,
1973 'protocol': 'http_dash_segments',
1975 if 'initialization_url' in representation_ms_info:
1976 initialization_url = representation_ms_info['initialization_url']
1977 if not f.get('url'):
1978 f['url'] = initialization_url
1979 f['fragments'].append({location_key(initialization_url): initialization_url})
1980 f['fragments'].extend(representation_ms_info['fragments'])
1982 existing_format = next(
1983 fo for fo in formats
1984 if fo['format_id'] == representation_id)
1985 except StopIteration:
1986 full_info = formats_dict.get(representation_id, {}).copy()
1988 formats.append(full_info)
1990 existing_format.update(f)
1992 self.report_warning('Unknown MIME type %s in DASH manifest' % mime_type)
1995 def _extract_ism_formats(self, ism_url, video_id, ism_id=None, note=None, errnote=None, fatal=True):
1996 res = self._download_webpage_handle(
1998 note=note or 'Downloading ISM manifest',
1999 errnote=errnote or 'Failed to download ISM manifest',
2005 return self._parse_ism_formats(
2006 compat_etree_fromstring(ism.encode('utf-8')), urlh.geturl(), ism_id)
2008 def _parse_ism_formats(self, ism_doc, ism_url, ism_id=None):
2010 Parse formats from ISM manifest.
2012 1. [MS-SSTR]: Smooth Streaming Protocol,
2013 https://msdn.microsoft.com/en-us/library/ff469518.aspx
2015 if ism_doc.get('IsLive') == 'TRUE' or ism_doc.find('Protection') is not None:
2018 duration = int(ism_doc.attrib['Duration'])
2019 timescale = int_or_none(ism_doc.get('TimeScale')) or 10000000
2022 for stream in ism_doc.findall('StreamIndex'):
2023 stream_type = stream.get('Type')
2024 if stream_type not in ('video', 'audio'):
2026 url_pattern = stream.attrib['Url']
2027 stream_timescale = int_or_none(stream.get('TimeScale')) or timescale
2028 stream_name = stream.get('Name')
2029 for track in stream.findall('QualityLevel'):
2030 fourcc = track.get('FourCC')
2031 # TODO: add support for WVC1 and WMAP
2032 if fourcc not in ('H264', 'AVC1', 'AACL'):
2033 self.report_warning('%s is not a supported codec' % fourcc)
2035 tbr = int(track.attrib['Bitrate']) // 1000
2036 # [1] does not mention Width and Height attributes. However,
2037 # they're often present while MaxWidth and MaxHeight are
2038 # missing, so should be used as fallbacks
2039 width = int_or_none(track.get('MaxWidth') or track.get('Width'))
2040 height = int_or_none(track.get('MaxHeight') or track.get('Height'))
2041 sampling_rate = int_or_none(track.get('SamplingRate'))
2043 track_url_pattern = re.sub(r'{[Bb]itrate}', track.attrib['Bitrate'], url_pattern)
2044 track_url_pattern = compat_urlparse.urljoin(ism_url, track_url_pattern)
2050 stream_fragments = stream.findall('c')
2051 for stream_fragment_index, stream_fragment in enumerate(stream_fragments):
2052 fragment_ctx['time'] = int_or_none(stream_fragment.get('t')) or fragment_ctx['time']
2053 fragment_repeat = int_or_none(stream_fragment.get('r')) or 1
2054 fragment_ctx['duration'] = int_or_none(stream_fragment.get('d'))
2055 if not fragment_ctx['duration']:
2057 next_fragment_time = int(stream_fragment[stream_fragment_index + 1].attrib['t'])
2059 next_fragment_time = duration
2060 fragment_ctx['duration'] = (next_fragment_time - fragment_ctx['time']) / fragment_repeat
2061 for _ in range(fragment_repeat):
2063 'url': re.sub(r'{start[ _]time}', compat_str(fragment_ctx['time']), track_url_pattern),
2064 'duration': fragment_ctx['duration'] / stream_timescale,
2066 fragment_ctx['time'] += fragment_ctx['duration']
2070 format_id.append(ism_id)
2072 format_id.append(stream_name)
2073 format_id.append(compat_str(tbr))
2076 'format_id': '-'.join(format_id),
2078 'manifest_url': ism_url,
2079 'ext': 'ismv' if stream_type == 'video' else 'isma',
2083 'asr': sampling_rate,
2084 'vcodec': 'none' if stream_type == 'audio' else fourcc,
2085 'acodec': 'none' if stream_type == 'video' else fourcc,
2087 'fragments': fragments,
2088 '_download_params': {
2089 'duration': duration,
2090 'timescale': stream_timescale,
2091 'width': width or 0,
2092 'height': height or 0,
2094 'codec_private_data': track.get('CodecPrivateData'),
2095 'sampling_rate': sampling_rate,
2096 'channels': int_or_none(track.get('Channels', 2)),
2097 'bits_per_sample': int_or_none(track.get('BitsPerSample', 16)),
2098 'nal_unit_length_field': int_or_none(track.get('NALUnitLengthField', 4)),
2103 def _parse_html5_media_entries(self, base_url, webpage, video_id, m3u8_id=None, m3u8_entry_protocol='m3u8', mpd_id=None, preference=None):
2104 def absolute_url(video_url):
2105 return compat_urlparse.urljoin(base_url, video_url)
2107 def parse_content_type(content_type):
2108 if not content_type:
2110 ctr = re.search(r'(?P<mimetype>[^/]+/[^;]+)(?:;\s*codecs="?(?P<codecs>[^"]+))?', content_type)
2112 mimetype, codecs = ctr.groups()
2113 f = parse_codecs(codecs)
2114 f['ext'] = mimetype2ext(mimetype)
2118 def _media_formats(src, cur_media_type, type_info={}):
2119 full_url = absolute_url(src)
2120 ext = type_info.get('ext') or determine_ext(full_url)
2122 is_plain_url = False
2123 formats = self._extract_m3u8_formats(
2124 full_url, video_id, ext='mp4',
2125 entry_protocol=m3u8_entry_protocol, m3u8_id=m3u8_id,
2126 preference=preference)
2128 is_plain_url = False
2129 formats = self._extract_mpd_formats(
2130 full_url, video_id, mpd_id=mpd_id)
2135 'vcodec': 'none' if cur_media_type == 'audio' else None,
2137 return is_plain_url, formats
2140 # amp-video and amp-audio are very similar to their HTML5 counterparts
2141 # so we wll include them right here (see
2142 # https://www.ampproject.org/docs/reference/components/amp-video)
2143 media_tags = [(media_tag, media_type, '')
2144 for media_tag, media_type
2145 in re.findall(r'(?s)(<(?:amp-)?(video|audio)[^>]*/>)', webpage)]
2146 media_tags.extend(re.findall(
2147 # We only allow video|audio followed by a whitespace or '>'.
2148 # Allowing more characters may end up in significant slow down (see
2149 # https://github.com/rg3/youtube-dl/issues/11979, example URL:
2150 # http://www.porntrex.com/maps/videositemap.xml).
2151 r'(?s)(<(?P<tag>(?:amp-)?(?:video|audio))(?:\s+[^>]*)?>)(.*?)</(?P=tag)>', webpage))
2152 for media_tag, media_type, media_content in media_tags:
2157 media_attributes = extract_attributes(media_tag)
2158 src = media_attributes.get('src')
2160 _, formats = _media_formats(src, media_type)
2161 media_info['formats'].extend(formats)
2162 media_info['thumbnail'] = media_attributes.get('poster')
2164 for source_tag in re.findall(r'<source[^>]+>', media_content):
2165 source_attributes = extract_attributes(source_tag)
2166 src = source_attributes.get('src')
2169 f = parse_content_type(source_attributes.get('type'))
2170 is_plain_url, formats = _media_formats(src, media_type, f)
2172 f.update(formats[0])
2173 media_info['formats'].append(f)
2175 media_info['formats'].extend(formats)
2176 for track_tag in re.findall(r'<track[^>]+>', media_content):
2177 track_attributes = extract_attributes(track_tag)
2178 kind = track_attributes.get('kind')
2179 if not kind or kind in ('subtitles', 'captions'):
2180 src = track_attributes.get('src')
2183 lang = track_attributes.get('srclang') or track_attributes.get('lang') or track_attributes.get('label')
2184 media_info['subtitles'].setdefault(lang, []).append({
2185 'url': absolute_url(src),
2187 if media_info['formats'] or media_info['subtitles']:
2188 entries.append(media_info)
2191 def _extract_akamai_formats(self, manifest_url, video_id, hosts={}):
2193 hdcore_sign = 'hdcore=3.7.0'
2194 f4m_url = re.sub(r'(https?://[^/]+)/i/', r'\1/z/', manifest_url).replace('/master.m3u8', '/manifest.f4m')
2195 hds_host = hosts.get('hds')
2197 f4m_url = re.sub(r'(https?://)[^/]+', r'\1' + hds_host, f4m_url)
2198 if 'hdcore=' not in f4m_url:
2199 f4m_url += ('&' if '?' in f4m_url else '?') + hdcore_sign
2200 f4m_formats = self._extract_f4m_formats(
2201 f4m_url, video_id, f4m_id='hds', fatal=False)
2202 for entry in f4m_formats:
2203 entry.update({'extra_param_to_segment_url': hdcore_sign})
2204 formats.extend(f4m_formats)
2205 m3u8_url = re.sub(r'(https?://[^/]+)/z/', r'\1/i/', manifest_url).replace('/manifest.f4m', '/master.m3u8')
2206 hls_host = hosts.get('hls')
2208 m3u8_url = re.sub(r'(https?://)[^/]+', r'\1' + hls_host, m3u8_url)
2209 formats.extend(self._extract_m3u8_formats(
2210 m3u8_url, video_id, 'mp4', 'm3u8_native',
2211 m3u8_id='hls', fatal=False))
2214 def _extract_wowza_formats(self, url, video_id, m3u8_entry_protocol='m3u8_native', skip_protocols=[]):
2215 url = re.sub(r'/(?:manifest|playlist|jwplayer)\.(?:m3u8|f4m|mpd|smil)', '', url)
2216 url_base = self._search_regex(
2217 r'(?:(?:https?|rtmp|rtsp):)?(//[^?]+)', url, 'format url')
2218 http_base_url = '%s:%s' % ('http', url_base)
2220 if 'm3u8' not in skip_protocols:
2221 formats.extend(self._extract_m3u8_formats(
2222 http_base_url + '/playlist.m3u8', video_id, 'mp4',
2223 m3u8_entry_protocol, m3u8_id='hls', fatal=False))
2224 if 'f4m' not in skip_protocols:
2225 formats.extend(self._extract_f4m_formats(
2226 http_base_url + '/manifest.f4m',
2227 video_id, f4m_id='hds', fatal=False))
2228 if 'dash' not in skip_protocols:
2229 formats.extend(self._extract_mpd_formats(
2230 http_base_url + '/manifest.mpd',
2231 video_id, mpd_id='dash', fatal=False))
2232 if re.search(r'(?:/smil:|\.smil)', url_base):
2233 if 'smil' not in skip_protocols:
2234 rtmp_formats = self._extract_smil_formats(
2235 http_base_url + '/jwplayer.smil',
2236 video_id, fatal=False)
2237 for rtmp_format in rtmp_formats:
2238 rtsp_format = rtmp_format.copy()
2239 rtsp_format['url'] = '%s/%s' % (rtmp_format['url'], rtmp_format['play_path'])
2240 del rtsp_format['play_path']
2241 del rtsp_format['ext']
2242 rtsp_format.update({
2243 'url': rtsp_format['url'].replace('rtmp://', 'rtsp://'),
2244 'format_id': rtmp_format['format_id'].replace('rtmp', 'rtsp'),
2247 formats.extend([rtmp_format, rtsp_format])
2249 for protocol in ('rtmp', 'rtsp'):
2250 if protocol not in skip_protocols:
2252 'url': '%s:%s' % (protocol, url_base),
2253 'format_id': protocol,
2254 'protocol': protocol,
2258 def _find_jwplayer_data(self, webpage, video_id=None, transform_source=js_to_json):
2260 r'(?s)jwplayer\((?P<quote>[\'"])[^\'" ]+(?P=quote)\)(?!</script>).*?\.setup\s*\((?P<options>[^)]+)\)',
2264 jwplayer_data = self._parse_json(mobj.group('options'),
2266 transform_source=transform_source)
2267 except ExtractorError:
2270 if isinstance(jwplayer_data, dict):
2271 return jwplayer_data
2273 def _extract_jwplayer_data(self, webpage, video_id, *args, **kwargs):
2274 jwplayer_data = self._find_jwplayer_data(
2275 webpage, video_id, transform_source=js_to_json)
2276 return self._parse_jwplayer_data(
2277 jwplayer_data, video_id, *args, **kwargs)
2279 def _parse_jwplayer_data(self, jwplayer_data, video_id=None, require_title=True,
2280 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
2281 # JWPlayer backward compatibility: flattened playlists
2282 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/api/config.js#L81-L96
2283 if 'playlist' not in jwplayer_data:
2284 jwplayer_data = {'playlist': [jwplayer_data]}
2288 # JWPlayer backward compatibility: single playlist item
2289 # https://github.com/jwplayer/jwplayer/blob/v7.7.0/src/js/playlist/playlist.js#L10
2290 if not isinstance(jwplayer_data['playlist'], list):
2291 jwplayer_data['playlist'] = [jwplayer_data['playlist']]
2293 for video_data in jwplayer_data['playlist']:
2294 # JWPlayer backward compatibility: flattened sources
2295 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/playlist/item.js#L29-L35
2296 if 'sources' not in video_data:
2297 video_data['sources'] = [video_data]
2299 this_video_id = video_id or video_data['mediaid']
2301 formats = self._parse_jwplayer_formats(
2302 video_data['sources'], video_id=this_video_id, m3u8_id=m3u8_id,
2303 mpd_id=mpd_id, rtmp_params=rtmp_params, base_url=base_url)
2304 self._sort_formats(formats)
2307 tracks = video_data.get('tracks')
2308 if tracks and isinstance(tracks, list):
2309 for track in tracks:
2310 if not isinstance(track, dict):
2312 if track.get('kind') != 'captions':
2314 track_url = urljoin(base_url, track.get('file'))
2317 subtitles.setdefault(track.get('label') or 'en', []).append({
2318 'url': self._proto_relative_url(track_url)
2322 'id': this_video_id,
2323 'title': video_data['title'] if require_title else video_data.get('title'),
2324 'description': video_data.get('description'),
2325 'thumbnail': self._proto_relative_url(video_data.get('image')),
2326 'timestamp': int_or_none(video_data.get('pubdate')),
2327 'duration': float_or_none(jwplayer_data.get('duration') or video_data.get('duration')),
2328 'subtitles': subtitles,
2331 if len(entries) == 1:
2334 return self.playlist_result(entries)
2336 def _parse_jwplayer_formats(self, jwplayer_sources_data, video_id=None,
2337 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
2340 for source in jwplayer_sources_data:
2341 if not isinstance(source, dict):
2343 source_url = self._proto_relative_url(source.get('file'))
2347 source_url = compat_urlparse.urljoin(base_url, source_url)
2348 if source_url in urls:
2350 urls.append(source_url)
2351 source_type = source.get('type') or ''
2352 ext = mimetype2ext(source_type) or determine_ext(source_url)
2353 if source_type == 'hls' or ext == 'm3u8':
2354 formats.extend(self._extract_m3u8_formats(
2355 source_url, video_id, 'mp4', entry_protocol='m3u8_native',
2356 m3u8_id=m3u8_id, fatal=False))
2358 formats.extend(self._extract_mpd_formats(
2359 source_url, video_id, mpd_id=mpd_id, fatal=False))
2361 formats.extend(self._extract_smil_formats(
2362 source_url, video_id, fatal=False))
2363 # https://github.com/jwplayer/jwplayer/blob/master/src/js/providers/default.js#L67
2364 elif source_type.startswith('audio') or ext in (
2365 'oga', 'aac', 'mp3', 'mpeg', 'vorbis'):
2372 height = int_or_none(source.get('height'))
2374 # Often no height is provided but there is a label in
2375 # format like "1080p", "720p SD", or 1080.
2376 height = int_or_none(self._search_regex(
2377 r'^(\d{3,4})[pP]?(?:\b|$)', compat_str(source.get('label') or ''),
2378 'height', default=None))
2381 'width': int_or_none(source.get('width')),
2383 'tbr': int_or_none(source.get('bitrate')),
2386 if source_url.startswith('rtmp'):
2387 a_format['ext'] = 'flv'
2388 # See com/longtailvideo/jwplayer/media/RTMPMediaProvider.as
2389 # of jwplayer.flash.swf
2390 rtmp_url_parts = re.split(
2391 r'((?:mp4|mp3|flv):)', source_url, 1)
2392 if len(rtmp_url_parts) == 3:
2393 rtmp_url, prefix, play_path = rtmp_url_parts
2396 'play_path': prefix + play_path,
2399 a_format.update(rtmp_params)
2400 formats.append(a_format)
2403 def _live_title(self, name):
2404 """ Generate the title for a live video """
2405 now = datetime.datetime.now()
2406 now_str = now.strftime('%Y-%m-%d %H:%M')
2407 return name + ' ' + now_str
2409 def _int(self, v, name, fatal=False, **kwargs):
2410 res = int_or_none(v, **kwargs)
2411 if 'get_attr' in kwargs:
2412 print(getattr(v, kwargs['get_attr']))
2414 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
2416 raise ExtractorError(msg)
2418 self._downloader.report_warning(msg)
2421 def _float(self, v, name, fatal=False, **kwargs):
2422 res = float_or_none(v, **kwargs)
2424 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
2426 raise ExtractorError(msg)
2428 self._downloader.report_warning(msg)
2431 def _set_cookie(self, domain, name, value, expire_time=None):
2432 cookie = compat_cookiejar.Cookie(
2433 0, name, value, None, None, domain, None,
2434 None, '/', True, False, expire_time, '', None, None, None)
2435 self._downloader.cookiejar.set_cookie(cookie)
2437 def _get_cookies(self, url):
2438 """ Return a compat_cookies.SimpleCookie with the cookies for the url """
2439 req = sanitized_Request(url)
2440 self._downloader.cookiejar.add_cookie_header(req)
2441 return compat_cookies.SimpleCookie(req.get_header('Cookie'))
2443 def get_testcases(self, include_onlymatching=False):
2444 t = getattr(self, '_TEST', None)
2446 assert not hasattr(self, '_TESTS'), \
2447 '%s has _TEST and _TESTS' % type(self).__name__
2450 tests = getattr(self, '_TESTS', [])
2452 if not include_onlymatching and t.get('only_matching', False):
2454 t['name'] = type(self).__name__[:-len('IE')]
2457 def is_suitable(self, age_limit):
2458 """ Test whether the extractor is generally suitable for the given
2459 age limit (i.e. pornographic sites are not, all others usually are) """
2461 any_restricted = False
2462 for tc in self.get_testcases(include_onlymatching=False):
2463 if tc.get('playlist', []):
2464 tc = tc['playlist'][0]
2465 is_restricted = age_restricted(
2466 tc.get('info_dict', {}).get('age_limit'), age_limit)
2467 if not is_restricted:
2469 any_restricted = any_restricted or is_restricted
2470 return not any_restricted
2472 def extract_subtitles(self, *args, **kwargs):
2473 if (self._downloader.params.get('writesubtitles', False) or
2474 self._downloader.params.get('listsubtitles')):
2475 return self._get_subtitles(*args, **kwargs)
2478 def _get_subtitles(self, *args, **kwargs):
2479 raise NotImplementedError('This method must be implemented by subclasses')
2482 def _merge_subtitle_items(subtitle_list1, subtitle_list2):
2483 """ Merge subtitle items for one language. Items with duplicated URLs
2484 will be dropped. """
2485 list1_urls = set([item['url'] for item in subtitle_list1])
2486 ret = list(subtitle_list1)
2487 ret.extend([item for item in subtitle_list2 if item['url'] not in list1_urls])
2491 def _merge_subtitles(cls, subtitle_dict1, subtitle_dict2):
2492 """ Merge two subtitle dictionaries, language by language. """
2493 ret = dict(subtitle_dict1)
2494 for lang in subtitle_dict2:
2495 ret[lang] = cls._merge_subtitle_items(subtitle_dict1.get(lang, []), subtitle_dict2[lang])
2498 def extract_automatic_captions(self, *args, **kwargs):
2499 if (self._downloader.params.get('writeautomaticsub', False) or
2500 self._downloader.params.get('listsubtitles')):
2501 return self._get_automatic_captions(*args, **kwargs)
2504 def _get_automatic_captions(self, *args, **kwargs):
2505 raise NotImplementedError('This method must be implemented by subclasses')
2507 def mark_watched(self, *args, **kwargs):
2508 if (self._downloader.params.get('mark_watched', False) and
2509 (self._get_login_info()[0] is not None or
2510 self._downloader.params.get('cookiefile') is not None)):
2511 self._mark_watched(*args, **kwargs)
2513 def _mark_watched(self, *args, **kwargs):
2514 raise NotImplementedError('This method must be implemented by subclasses')
2516 def geo_verification_headers(self):
2518 geo_verification_proxy = self._downloader.params.get('geo_verification_proxy')
2519 if geo_verification_proxy:
2520 headers['Ytdl-request-proxy'] = geo_verification_proxy
2523 def _generic_id(self, url):
2524 return compat_urllib_parse_unquote(os.path.splitext(url.rstrip('/').split('/')[-1])[0])
2526 def _generic_title(self, url):
2527 return compat_urllib_parse_unquote(os.path.splitext(url_basename(url))[0])
2530 class SearchInfoExtractor(InfoExtractor):
2532 Base class for paged search queries extractors.
2533 They accept URLs in the format _SEARCH_KEY(|all|[0-9]):{query}
2534 Instances should define _SEARCH_KEY and _MAX_RESULTS.
2538 def _make_valid_url(cls):
2539 return r'%s(?P<prefix>|[1-9][0-9]*|all):(?P<query>[\s\S]+)' % cls._SEARCH_KEY
2542 def suitable(cls, url):
2543 return re.match(cls._make_valid_url(), url) is not None
2545 def _real_extract(self, query):
2546 mobj = re.match(self._make_valid_url(), query)
2548 raise ExtractorError('Invalid search query "%s"' % query)
2550 prefix = mobj.group('prefix')
2551 query = mobj.group('query')
2553 return self._get_n_results(query, 1)
2554 elif prefix == 'all':
2555 return self._get_n_results(query, self._MAX_RESULTS)
2559 raise ExtractorError('invalid download number %s for query "%s"' % (n, query))
2560 elif n > self._MAX_RESULTS:
2561 self._downloader.report_warning('%s returns max %i results (you requested %i)' % (self._SEARCH_KEY, self._MAX_RESULTS, n))
2562 n = self._MAX_RESULTS
2563 return self._get_n_results(query, n)
2565 def _get_n_results(self, query, n):
2566 """Get a specified number of results for a query"""
2567 raise NotImplementedError('This method must be implemented by subclasses')
2570 def SEARCH_KEY(self):
2571 return self._SEARCH_KEY