1 from __future__ import unicode_literals
15 from ..compat import (
18 compat_etree_fromstring,
24 compat_urllib_parse_urlencode,
25 compat_urllib_request,
28 from ..downloader.f4m import remove_encrypted_media
57 parse_m3u8_attributes,
63 class InfoExtractor(object):
64 """Information Extractor class.
66 Information extractors are the classes that, given a URL, extract
67 information about the video (or videos) the URL refers to. This
68 information includes the real video URL, the video title, author and
69 others. The information is stored in a dictionary which is then
70 passed to the YoutubeDL. The YoutubeDL processes this
71 information possibly downloading the video to the file system, among
72 other possible outcomes.
74 The type field determines the type of the result.
75 By far the most common value (and the default if _type is missing) is
76 "video", which indicates a single video.
78 For a video, the dictionaries must include the following fields:
81 title: Video title, unescaped.
83 Additionally, it must contain either a formats entry or a url one:
85 formats: A list of dictionaries for each format available, ordered
86 from worst to best quality.
89 * url Mandatory. The URL of the video file
90 * ext Will be calculated from URL if missing
91 * format A human-readable description of the format
92 ("mp4 container with h264/opus").
93 Calculated from the format_id, width, height.
94 and format_note fields if missing.
95 * format_id A short description of the format
96 ("mp4_h264_opus" or "19").
97 Technically optional, but strongly recommended.
98 * format_note Additional info about the format
99 ("3D" or "DASH video")
100 * width Width of the video, if known
101 * height Height of the video, if known
102 * resolution Textual description of width and height
103 * tbr Average bitrate of audio and video in KBit/s
104 * abr Average audio bitrate in KBit/s
105 * acodec Name of the audio codec in use
106 * asr Audio sampling rate in Hertz
107 * vbr Average video bitrate in KBit/s
109 * vcodec Name of the video codec in use
110 * container Name of the container format
111 * filesize The number of bytes, if known in advance
112 * filesize_approx An estimate for the number of bytes
113 * player_url SWF Player URL (used for rtmpdump).
114 * protocol The protocol that will be used for the actual
115 download, lower-case.
116 "http", "https", "rtsp", "rtmp", "rtmpe",
117 "m3u8", "m3u8_native" or "http_dash_segments".
118 * preference Order number of this format. If this field is
119 present and not None, the formats get sorted
120 by this field, regardless of all other values.
121 -1 for default (order by other properties),
122 -2 or smaller for less than default.
123 < -1000 to hide the format (if there is
124 another one which is strictly better)
125 * language Language code, e.g. "de" or "en-US".
126 * language_preference Is this in the language mentioned in
128 10 if it's what the URL is about,
129 -1 for default (don't know),
130 -10 otherwise, other values reserved for now.
131 * quality Order number of the video quality of this
132 format, irrespective of the file format.
133 -1 for default (order by other properties),
134 -2 or smaller for less than default.
135 * source_preference Order number for this video source
136 (quality takes higher priority)
137 -1 for default (order by other properties),
138 -2 or smaller for less than default.
139 * http_headers A dictionary of additional HTTP headers
140 to add to the request.
141 * stretched_ratio If given and not 1, indicates that the
142 video's pixels are not square.
143 width : height ratio as float.
144 * no_resume The server does not support resuming the
145 (HTTP or RTMP) download. Boolean.
147 url: Final video URL.
148 ext: Video filename extension.
149 format: The video format, defaults to ext (used for --get-format)
150 player_url: SWF Player URL (used for rtmpdump).
152 The following fields are optional:
154 alt_title: A secondary title of the video.
155 display_id An alternative identifier for the video, not necessarily
156 unique, but available before title. Typically, id is
157 something like "4234987", title "Dancing naked mole rats",
158 and display_id "dancing-naked-mole-rats"
159 thumbnails: A list of dictionaries, with the following entries:
160 * "id" (optional, string) - Thumbnail format ID
162 * "preference" (optional, int) - quality of the image
163 * "width" (optional, int)
164 * "height" (optional, int)
165 * "resolution" (optional, string "{width}x{height"},
167 * "filesize" (optional, int)
168 thumbnail: Full URL to a video thumbnail image.
169 description: Full video description.
170 uploader: Full name of the video uploader.
171 license: License name the video is licensed under.
172 creator: The creator of the video.
173 release_date: The date (YYYYMMDD) when the video was released.
174 timestamp: UNIX timestamp of the moment the video became available.
175 upload_date: Video upload date (YYYYMMDD).
176 If not explicitly set, calculated from timestamp.
177 uploader_id: Nickname or id of the video uploader.
178 uploader_url: Full URL to a personal webpage of the video uploader.
179 location: Physical location where the video was filmed.
180 subtitles: The available subtitles as a dictionary in the format
181 {language: subformats}. "subformats" is a list sorted from
182 lower to higher preference, each element is a dictionary
183 with the "ext" entry and one of:
184 * "data": The subtitles file contents
185 * "url": A URL pointing to the subtitles file
186 "ext" will be calculated from URL if missing
187 automatic_captions: Like 'subtitles', used by the YoutubeIE for
188 automatically generated captions
189 duration: Length of the video in seconds, as an integer or float.
190 view_count: How many users have watched the video on the platform.
191 like_count: Number of positive ratings of the video
192 dislike_count: Number of negative ratings of the video
193 repost_count: Number of reposts of the video
194 average_rating: Average rating give by users, the scale used depends on the webpage
195 comment_count: Number of comments on the video
196 comments: A list of comments, each with one or more of the following
197 properties (all but one of text or html optional):
198 * "author" - human-readable name of the comment author
199 * "author_id" - user ID of the comment author
201 * "html" - Comment as HTML
202 * "text" - Plain text of the comment
203 * "timestamp" - UNIX timestamp of comment
204 * "parent" - ID of the comment this one is replying to.
205 Set to "root" to indicate that this is a
206 comment to the original video.
207 age_limit: Age restriction for the video, as an integer (years)
208 webpage_url: The URL to the video webpage, if given to youtube-dl it
209 should allow to get the same result again. (It will be set
210 by YoutubeDL if it's missing)
211 categories: A list of categories that the video falls in, for example
213 tags: A list of tags assigned to the video, e.g. ["sweden", "pop music"]
214 is_live: True, False, or None (=unknown). Whether this video is a
215 live stream that goes on instead of a fixed-length video.
216 start_time: Time in seconds where the reproduction should start, as
217 specified in the URL.
218 end_time: Time in seconds where the reproduction should end, as
219 specified in the URL.
221 The following fields should only be used when the video belongs to some logical
224 chapter: Name or title of the chapter the video belongs to.
225 chapter_number: Number of the chapter the video belongs to, as an integer.
226 chapter_id: Id of the chapter the video belongs to, as a unicode string.
228 The following fields should only be used when the video is an episode of some
231 series: Title of the series or programme the video episode belongs to.
232 season: Title of the season the video episode belongs to.
233 season_number: Number of the season the video episode belongs to, as an integer.
234 season_id: Id of the season the video episode belongs to, as a unicode string.
235 episode: Title of the video episode. Unlike mandatory video title field,
236 this field should denote the exact title of the video episode
237 without any kind of decoration.
238 episode_number: Number of the video episode within a season, as an integer.
239 episode_id: Id of the video episode, as a unicode string.
241 The following fields should only be used when the media is a track or a part of
244 track: Title of the track.
245 track_number: Number of the track within an album or a disc, as an integer.
246 track_id: Id of the track (useful in case of custom indexing, e.g. 6.iii),
248 artist: Artist(s) of the track.
249 genre: Genre(s) of the track.
250 album: Title of the album the track belongs to.
251 album_type: Type of the album (e.g. "Demo", "Full-length", "Split", "Compilation", etc).
252 album_artist: List of all artists appeared on the album (e.g.
253 "Ash Borer / Fell Voices" or "Various Artists", useful for splits
255 disc_number: Number of the disc or other physical medium the track belongs to,
257 release_year: Year (YYYY) when the album was released.
259 Unless mentioned otherwise, the fields should be Unicode strings.
261 Unless mentioned otherwise, None is equivalent to absence of information.
264 _type "playlist" indicates multiple videos.
265 There must be a key "entries", which is a list, an iterable, or a PagedList
266 object, each element of which is a valid dictionary by this specification.
268 Additionally, playlists can have "title", "description" and "id" attributes
269 with the same semantics as videos (see above).
272 _type "multi_video" indicates that there are multiple videos that
273 form a single show, for examples multiple acts of an opera or TV episode.
274 It must have an entries key like a playlist and contain all the keys
275 required for a video at the same time.
278 _type "url" indicates that the video must be extracted from another
279 location, possibly by a different extractor. Its only required key is:
280 "url" - the next URL to extract.
281 The key "ie_key" can be set to the class name (minus the trailing "IE",
282 e.g. "Youtube") if the extractor class is known in advance.
283 Additionally, the dictionary may have any properties of the resolved entity
284 known in advance, for example "title" if the title of the referred video is
288 _type "url_transparent" entities have the same specification as "url", but
289 indicate that the given additional information is more precise than the one
290 associated with the resolved URL.
291 This is useful when a site employs a video service that hosts the video and
292 its technical metadata, but that video service does not embed a useful
293 title, description etc.
296 Subclasses of this one should re-define the _real_initialize() and
297 _real_extract() methods and define a _VALID_URL regexp.
298 Probably, they should also be added to the list of extractors.
300 Finally, the _WORKING attribute should be set to False for broken IEs
301 in order to warn the users and skip the tests.
308 def __init__(self, downloader=None):
309 """Constructor. Receives an optional downloader."""
311 self.set_downloader(downloader)
314 def suitable(cls, url):
315 """Receives a URL and returns True if suitable for this IE."""
317 # This does not use has/getattr intentionally - we want to know whether
318 # we have cached the regexp for *this* class, whereas getattr would also
319 # match the superclass
320 if '_VALID_URL_RE' not in cls.__dict__:
321 cls._VALID_URL_RE = re.compile(cls._VALID_URL)
322 return cls._VALID_URL_RE.match(url) is not None
325 def _match_id(cls, url):
326 if '_VALID_URL_RE' not in cls.__dict__:
327 cls._VALID_URL_RE = re.compile(cls._VALID_URL)
328 m = cls._VALID_URL_RE.match(url)
334 """Getter method for _WORKING."""
337 def initialize(self):
338 """Initializes an instance (authentication, etc)."""
340 self._real_initialize()
343 def extract(self, url):
344 """Extracts URL information and returns it in list of dicts."""
347 return self._real_extract(url)
348 except ExtractorError:
350 except compat_http_client.IncompleteRead as e:
351 raise ExtractorError('A network error has occurred.', cause=e, expected=True)
352 except (KeyError, StopIteration) as e:
353 raise ExtractorError('An extractor error has occurred.', cause=e)
355 def set_downloader(self, downloader):
356 """Sets the downloader for this IE."""
357 self._downloader = downloader
359 def _real_initialize(self):
360 """Real initialization process. Redefine in subclasses."""
363 def _real_extract(self, url):
364 """Real extraction process. Redefine in subclasses."""
369 """A string for getting the InfoExtractor with get_info_extractor"""
370 return compat_str(cls.__name__[:-2])
374 return compat_str(type(self).__name__[:-2])
376 def _request_webpage(self, url_or_request, video_id, note=None, errnote=None, fatal=True, data=None, headers={}, query={}):
377 """ Returns the response handle """
379 self.report_download_webpage(video_id)
380 elif note is not False:
382 self.to_screen('%s' % (note,))
384 self.to_screen('%s: %s' % (video_id, note))
385 if isinstance(url_or_request, compat_urllib_request.Request):
386 url_or_request = update_Request(
387 url_or_request, data=data, headers=headers, query=query)
390 url_or_request = update_url_query(url_or_request, query)
391 if data is not None or headers:
392 url_or_request = sanitized_Request(url_or_request, data, headers)
394 return self._downloader.urlopen(url_or_request)
395 except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err:
399 errnote = 'Unable to download webpage'
401 errmsg = '%s: %s' % (errnote, error_to_compat_str(err))
403 raise ExtractorError(errmsg, sys.exc_info()[2], cause=err)
405 self._downloader.report_warning(errmsg)
408 def _download_webpage_handle(self, url_or_request, video_id, note=None, errnote=None, fatal=True, encoding=None, data=None, headers={}, query={}):
409 """ Returns a tuple (page content as string, URL handle) """
410 # Strip hashes from the URL (#1038)
411 if isinstance(url_or_request, (compat_str, str)):
412 url_or_request = url_or_request.partition('#')[0]
414 urlh = self._request_webpage(url_or_request, video_id, note, errnote, fatal, data=data, headers=headers, query=query)
418 content = self._webpage_read_content(urlh, url_or_request, video_id, note, errnote, fatal, encoding=encoding)
419 return (content, urlh)
422 def _guess_encoding_from_content(content_type, webpage_bytes):
423 m = re.match(r'[a-zA-Z0-9_.-]+/[a-zA-Z0-9_.-]+\s*;\s*charset=(.+)', content_type)
425 encoding = m.group(1)
427 m = re.search(br'<meta[^>]+charset=[\'"]?([^\'")]+)[ /\'">]',
428 webpage_bytes[:1024])
430 encoding = m.group(1).decode('ascii')
431 elif webpage_bytes.startswith(b'\xff\xfe'):
438 def _webpage_read_content(self, urlh, url_or_request, video_id, note=None, errnote=None, fatal=True, prefix=None, encoding=None):
439 content_type = urlh.headers.get('Content-Type', '')
440 webpage_bytes = urlh.read()
441 if prefix is not None:
442 webpage_bytes = prefix + webpage_bytes
444 encoding = self._guess_encoding_from_content(content_type, webpage_bytes)
445 if self._downloader.params.get('dump_intermediate_pages', False):
447 url = url_or_request.get_full_url()
448 except AttributeError:
450 self.to_screen('Dumping request to ' + url)
451 dump = base64.b64encode(webpage_bytes).decode('ascii')
452 self._downloader.to_screen(dump)
453 if self._downloader.params.get('write_pages', False):
455 url = url_or_request.get_full_url()
456 except AttributeError:
458 basen = '%s_%s' % (video_id, url)
460 h = '___' + hashlib.md5(basen.encode('utf-8')).hexdigest()
461 basen = basen[:240 - len(h)] + h
462 raw_filename = basen + '.dump'
463 filename = sanitize_filename(raw_filename, restricted=True)
464 self.to_screen('Saving request to ' + filename)
465 # Working around MAX_PATH limitation on Windows (see
466 # http://msdn.microsoft.com/en-us/library/windows/desktop/aa365247(v=vs.85).aspx)
467 if compat_os_name == 'nt':
468 absfilepath = os.path.abspath(filename)
469 if len(absfilepath) > 259:
470 filename = '\\\\?\\' + absfilepath
471 with open(filename, 'wb') as outf:
472 outf.write(webpage_bytes)
475 content = webpage_bytes.decode(encoding, 'replace')
477 content = webpage_bytes.decode('utf-8', 'replace')
479 if ('<title>Access to this site is blocked</title>' in content and
480 'Websense' in content[:512]):
481 msg = 'Access to this webpage has been blocked by Websense filtering software in your network.'
482 blocked_iframe = self._html_search_regex(
483 r'<iframe src="([^"]+)"', content,
484 'Websense information URL', default=None)
486 msg += ' Visit %s for more details' % blocked_iframe
487 raise ExtractorError(msg, expected=True)
488 if '<title>The URL you requested has been blocked</title>' in content[:512]:
490 'Access to this webpage has been blocked by Indian censorship. '
491 'Use a VPN or proxy server (with --proxy) to route around it.')
492 block_msg = self._html_search_regex(
493 r'</h1><p>(.*?)</p>',
494 content, 'block message', default=None)
496 msg += ' (Message: "%s")' % block_msg.replace('\n', ' ')
497 raise ExtractorError(msg, expected=True)
501 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={}):
502 """ Returns the data of the page as a string """
505 while success is False:
507 res = self._download_webpage_handle(url_or_request, video_id, note, errnote, fatal, encoding=encoding, data=data, headers=headers, query=query)
509 except compat_http_client.IncompleteRead as e:
511 if try_count >= tries:
513 self._sleep(timeout, video_id)
520 def _download_xml(self, url_or_request, video_id,
521 note='Downloading XML', errnote='Unable to download XML',
522 transform_source=None, fatal=True, encoding=None, data=None, headers={}, query={}):
523 """Return the xml as an xml.etree.ElementTree.Element"""
524 xml_string = self._download_webpage(
525 url_or_request, video_id, note, errnote, fatal=fatal, encoding=encoding, data=data, headers=headers, query=query)
526 if xml_string is False:
529 xml_string = transform_source(xml_string)
530 return compat_etree_fromstring(xml_string.encode('utf-8'))
532 def _download_json(self, url_or_request, video_id,
533 note='Downloading JSON metadata',
534 errnote='Unable to download JSON metadata',
535 transform_source=None,
536 fatal=True, encoding=None, data=None, headers={}, query={}):
537 json_string = self._download_webpage(
538 url_or_request, video_id, note, errnote, fatal=fatal,
539 encoding=encoding, data=data, headers=headers, query=query)
540 if (not fatal) and json_string is False:
542 return self._parse_json(
543 json_string, video_id, transform_source=transform_source, fatal=fatal)
545 def _parse_json(self, json_string, video_id, transform_source=None, fatal=True):
547 json_string = transform_source(json_string)
549 return json.loads(json_string)
550 except ValueError as ve:
551 errmsg = '%s: Failed to parse JSON ' % video_id
553 raise ExtractorError(errmsg, cause=ve)
555 self.report_warning(errmsg + str(ve))
557 def report_warning(self, msg, video_id=None):
558 idstr = '' if video_id is None else '%s: ' % video_id
559 self._downloader.report_warning(
560 '[%s] %s%s' % (self.IE_NAME, idstr, msg))
562 def to_screen(self, msg):
563 """Print msg to screen, prefixing it with '[ie_name]'"""
564 self._downloader.to_screen('[%s] %s' % (self.IE_NAME, msg))
566 def report_extraction(self, id_or_name):
567 """Report information extraction."""
568 self.to_screen('%s: Extracting information' % id_or_name)
570 def report_download_webpage(self, video_id):
571 """Report webpage download."""
572 self.to_screen('%s: Downloading webpage' % video_id)
574 def report_age_confirmation(self):
575 """Report attempt to confirm age."""
576 self.to_screen('Confirming age')
578 def report_login(self):
579 """Report attempt to log in."""
580 self.to_screen('Logging in')
583 def raise_login_required(msg='This video is only available for registered users'):
584 raise ExtractorError(
585 '%s. Use --username and --password or --netrc to provide account credentials.' % msg,
589 def raise_geo_restricted(msg='This video is not available from your location due to geo restriction'):
590 raise ExtractorError(
591 '%s. You might want to use --proxy to workaround.' % msg,
594 # Methods for following #608
596 def url_result(url, ie=None, video_id=None, video_title=None):
597 """Returns a URL that points to a page that should be processed"""
598 # TODO: ie should be the class used for getting the info
599 video_info = {'_type': 'url',
602 if video_id is not None:
603 video_info['id'] = video_id
604 if video_title is not None:
605 video_info['title'] = video_title
609 def playlist_result(entries, playlist_id=None, playlist_title=None, playlist_description=None):
610 """Returns a playlist"""
611 video_info = {'_type': 'playlist',
614 video_info['id'] = playlist_id
616 video_info['title'] = playlist_title
617 if playlist_description:
618 video_info['description'] = playlist_description
621 def _search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
623 Perform a regex search on the given string, using a single or a list of
624 patterns returning the first matching group.
625 In case of failure return a default value or raise a WARNING or a
626 RegexNotFoundError, depending on fatal, specifying the field name.
628 if isinstance(pattern, (str, compat_str, compiled_regex_type)):
629 mobj = re.search(pattern, string, flags)
632 mobj = re.search(p, string, flags)
636 if not self._downloader.params.get('no_color') and compat_os_name != 'nt' and sys.stderr.isatty():
637 _name = '\033[0;34m%s\033[0m' % name
643 # return the first matching group
644 return next(g for g in mobj.groups() if g is not None)
646 return mobj.group(group)
647 elif default is not NO_DEFAULT:
650 raise RegexNotFoundError('Unable to extract %s' % _name)
652 self._downloader.report_warning('unable to extract %s' % _name + bug_reports_message())
655 def _html_search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
657 Like _search_regex, but strips HTML tags and unescapes entities.
659 res = self._search_regex(pattern, string, name, default, fatal, flags, group)
661 return clean_html(res).strip()
665 def _get_netrc_login_info(self, netrc_machine=None):
668 netrc_machine = netrc_machine or self._NETRC_MACHINE
670 if self._downloader.params.get('usenetrc', False):
672 info = netrc.netrc().authenticators(netrc_machine)
677 raise netrc.NetrcParseError('No authenticators for %s' % netrc_machine)
678 except (IOError, netrc.NetrcParseError) as err:
679 self._downloader.report_warning('parsing .netrc: %s' % error_to_compat_str(err))
681 return (username, password)
683 def _get_login_info(self):
685 Get the login info as (username, password)
686 It will look in the netrc file using the _NETRC_MACHINE value
687 If there's no info available, return (None, None)
689 if self._downloader is None:
694 downloader_params = self._downloader.params
696 # Attempt to use provided username and password or .netrc data
697 if downloader_params.get('username') is not None:
698 username = downloader_params['username']
699 password = downloader_params['password']
701 username, password = self._get_netrc_login_info()
703 return (username, password)
705 def _get_tfa_info(self, note='two-factor verification code'):
707 Get the two-factor authentication info
708 TODO - asking the user will be required for sms/phone verify
709 currently just uses the command line option
710 If there's no info available, return None
712 if self._downloader is None:
714 downloader_params = self._downloader.params
716 if downloader_params.get('twofactor') is not None:
717 return downloader_params['twofactor']
719 return compat_getpass('Type %s and press [Return]: ' % note)
721 # Helper functions for extracting OpenGraph info
723 def _og_regexes(prop):
724 content_re = r'content=(?:"([^"]+?)"|\'([^\']+?)\'|\s*([^\s"\'=<>`]+?))'
725 property_re = (r'(?:name|property)=(?:\'og:%(prop)s\'|"og:%(prop)s"|\s*og:%(prop)s\b)'
726 % {'prop': re.escape(prop)})
727 template = r'<meta[^>]+?%s[^>]+?%s'
729 template % (property_re, content_re),
730 template % (content_re, property_re),
734 def _meta_regex(prop):
735 return r'''(?isx)<meta
736 (?=[^>]+(?:itemprop|name|property|id|http-equiv)=(["\']?)%s\1)
737 [^>]+?content=(["\'])(?P<content>.*?)\2''' % re.escape(prop)
739 def _og_search_property(self, prop, html, name=None, **kargs):
740 if not isinstance(prop, (list, tuple)):
743 name = 'OpenGraph %s' % prop[0]
746 og_regexes.extend(self._og_regexes(p))
747 escaped = self._search_regex(og_regexes, html, name, flags=re.DOTALL, **kargs)
750 return unescapeHTML(escaped)
752 def _og_search_thumbnail(self, html, **kargs):
753 return self._og_search_property('image', html, 'thumbnail URL', fatal=False, **kargs)
755 def _og_search_description(self, html, **kargs):
756 return self._og_search_property('description', html, fatal=False, **kargs)
758 def _og_search_title(self, html, **kargs):
759 return self._og_search_property('title', html, **kargs)
761 def _og_search_video_url(self, html, name='video url', secure=True, **kargs):
762 regexes = self._og_regexes('video') + self._og_regexes('video:url')
764 regexes = self._og_regexes('video:secure_url') + regexes
765 return self._html_search_regex(regexes, html, name, **kargs)
767 def _og_search_url(self, html, **kargs):
768 return self._og_search_property('url', html, **kargs)
770 def _html_search_meta(self, name, html, display_name=None, fatal=False, **kwargs):
771 if not isinstance(name, (list, tuple)):
773 if display_name is None:
774 display_name = name[0]
775 return self._html_search_regex(
776 [self._meta_regex(n) for n in name],
777 html, display_name, fatal=fatal, group='content', **kwargs)
779 def _dc_search_uploader(self, html):
780 return self._html_search_meta('dc.creator', html, 'uploader')
782 def _rta_search(self, html):
783 # See http://www.rtalabel.org/index.php?content=howtofaq#single
784 if re.search(r'(?ix)<meta\s+name="rating"\s+'
785 r' content="RTA-5042-1996-1400-1577-RTA"',
790 def _media_rating_search(self, html):
791 # See http://www.tjg-designs.com/WP/metadata-code-examples-adding-metadata-to-your-web-pages/
792 rating = self._html_search_meta('rating', html)
804 return RATING_TABLE.get(rating.lower())
806 def _family_friendly_search(self, html):
807 # See http://schema.org/VideoObject
808 family_friendly = self._html_search_meta('isFamilyFriendly', html)
810 if not family_friendly:
819 return RATING_TABLE.get(family_friendly.lower())
821 def _twitter_search_player(self, html):
822 return self._html_search_meta('twitter:player', html,
823 'twitter card player')
825 def _search_json_ld(self, html, video_id, expected_type=None, **kwargs):
826 json_ld = self._search_regex(
827 r'(?s)<script[^>]+type=(["\'])application/ld\+json\1[^>]*>(?P<json_ld>.+?)</script>',
828 html, 'JSON-LD', group='json_ld', **kwargs)
829 default = kwargs.get('default', NO_DEFAULT)
831 return default if default is not NO_DEFAULT else {}
832 # JSON-LD may be malformed and thus `fatal` should be respected.
833 # At the same time `default` may be passed that assumes `fatal=False`
834 # for _search_regex. Let's simulate the same behavior here as well.
835 fatal = kwargs.get('fatal', True) if default == NO_DEFAULT else False
836 return self._json_ld(json_ld, video_id, fatal=fatal, expected_type=expected_type)
838 def _json_ld(self, json_ld, video_id, fatal=True, expected_type=None):
839 if isinstance(json_ld, compat_str):
840 json_ld = self._parse_json(json_ld, video_id, fatal=fatal)
844 if not isinstance(json_ld, (list, tuple, dict)):
846 if isinstance(json_ld, dict):
849 if e.get('@context') == 'http://schema.org':
850 item_type = e.get('@type')
851 if expected_type is not None and expected_type != item_type:
853 if item_type == 'TVEpisode':
855 'episode': unescapeHTML(e.get('name')),
856 'episode_number': int_or_none(e.get('episodeNumber')),
857 'description': unescapeHTML(e.get('description')),
859 part_of_season = e.get('partOfSeason')
860 if isinstance(part_of_season, dict) and part_of_season.get('@type') == 'TVSeason':
861 info['season_number'] = int_or_none(part_of_season.get('seasonNumber'))
862 part_of_series = e.get('partOfSeries') or e.get('partOfTVSeries')
863 if isinstance(part_of_series, dict) and part_of_series.get('@type') == 'TVSeries':
864 info['series'] = unescapeHTML(part_of_series.get('name'))
865 elif item_type == 'Article':
867 'timestamp': parse_iso8601(e.get('datePublished')),
868 'title': unescapeHTML(e.get('headline')),
869 'description': unescapeHTML(e.get('articleBody')),
871 elif item_type == 'VideoObject':
873 'url': e.get('contentUrl'),
874 'title': unescapeHTML(e.get('name')),
875 'description': unescapeHTML(e.get('description')),
876 'thumbnail': e.get('thumbnailUrl'),
877 'duration': parse_duration(e.get('duration')),
878 'timestamp': unified_timestamp(e.get('uploadDate')),
879 'filesize': float_or_none(e.get('contentSize')),
880 'tbr': int_or_none(e.get('bitrate')),
881 'width': int_or_none(e.get('width')),
882 'height': int_or_none(e.get('height')),
885 return dict((k, v) for k, v in info.items() if v is not None)
888 def _hidden_inputs(html):
889 html = re.sub(r'<!--(?:(?!<!--).)*-->', '', html)
891 for input in re.findall(r'(?i)<input([^>]+)>', html):
892 if not re.search(r'type=(["\'])(?:hidden|submit)\1', input):
894 name = re.search(r'(?:name|id)=(["\'])(?P<value>.+?)\1', input)
897 value = re.search(r'value=(["\'])(?P<value>.*?)\1', input)
900 hidden_inputs[name.group('value')] = value.group('value')
903 def _form_hidden_inputs(self, form_id, html):
904 form = self._search_regex(
905 r'(?is)<form[^>]+?id=(["\'])%s\1[^>]*>(?P<form>.+?)</form>' % form_id,
906 html, '%s form' % form_id, group='form')
907 return self._hidden_inputs(form)
909 def _sort_formats(self, formats, field_preference=None):
911 raise ExtractorError('No video formats found')
914 # Automatically determine tbr when missing based on abr and vbr (improves
915 # formats sorting in some cases)
916 if 'tbr' not in f and f.get('abr') is not None and f.get('vbr') is not None:
917 f['tbr'] = f['abr'] + f['vbr']
920 # TODO remove the following workaround
921 from ..utils import determine_ext
922 if not f.get('ext') and 'url' in f:
923 f['ext'] = determine_ext(f['url'])
925 if isinstance(field_preference, (list, tuple)):
928 if f.get(field) is not None
929 else ('' if field == 'format_id' else -1)
930 for field in field_preference)
932 preference = f.get('preference')
933 if preference is None:
935 if f.get('ext') in ['f4f', 'f4m']: # Not yet supported
938 protocol = f.get('protocol') or determine_protocol(f)
939 proto_preference = 0 if protocol in ['http', 'https'] else (-0.5 if protocol == 'rtsp' else -0.1)
941 if f.get('vcodec') == 'none': # audio only
943 if self._downloader.params.get('prefer_free_formats'):
944 ORDER = ['aac', 'mp3', 'm4a', 'webm', 'ogg', 'opus']
946 ORDER = ['webm', 'opus', 'ogg', 'mp3', 'aac', 'm4a']
949 audio_ext_preference = ORDER.index(f['ext'])
951 audio_ext_preference = -1
953 if f.get('acodec') == 'none': # video only
955 if self._downloader.params.get('prefer_free_formats'):
956 ORDER = ['flv', 'mp4', 'webm']
958 ORDER = ['webm', 'flv', 'mp4']
960 ext_preference = ORDER.index(f['ext'])
963 audio_ext_preference = 0
967 f.get('language_preference') if f.get('language_preference') is not None else -1,
968 f.get('quality') if f.get('quality') is not None else -1,
969 f.get('tbr') if f.get('tbr') is not None else -1,
970 f.get('filesize') if f.get('filesize') is not None else -1,
971 f.get('vbr') if f.get('vbr') is not None else -1,
972 f.get('height') if f.get('height') is not None else -1,
973 f.get('width') if f.get('width') is not None else -1,
976 f.get('abr') if f.get('abr') is not None else -1,
977 audio_ext_preference,
978 f.get('fps') if f.get('fps') is not None else -1,
979 f.get('filesize_approx') if f.get('filesize_approx') is not None else -1,
980 f.get('source_preference') if f.get('source_preference') is not None else -1,
981 f.get('format_id') if f.get('format_id') is not None else '',
983 formats.sort(key=_formats_key)
985 def _check_formats(self, formats, video_id):
988 lambda f: self._is_valid_url(
990 item='%s video format' % f.get('format_id') if f.get('format_id') else 'video'),
994 def _remove_duplicate_formats(formats):
998 if f['url'] not in format_urls:
999 format_urls.add(f['url'])
1000 unique_formats.append(f)
1001 formats[:] = unique_formats
1003 def _is_valid_url(self, url, video_id, item='video'):
1004 url = self._proto_relative_url(url, scheme='http:')
1005 # For now assume non HTTP(S) URLs always valid
1006 if not (url.startswith('http://') or url.startswith('https://')):
1009 self._request_webpage(url, video_id, 'Checking %s URL' % item)
1011 except ExtractorError as e:
1012 if isinstance(e.cause, compat_urllib_error.URLError):
1014 '%s: %s URL is invalid, skipping' % (video_id, item))
1018 def http_scheme(self):
1019 """ Either "http:" or "https:", depending on the user's preferences """
1022 if self._downloader.params.get('prefer_insecure', False)
1025 def _proto_relative_url(self, url, scheme=None):
1028 if url.startswith('//'):
1030 scheme = self.http_scheme()
1035 def _sleep(self, timeout, video_id, msg_template=None):
1036 if msg_template is None:
1037 msg_template = '%(video_id)s: Waiting for %(timeout)s seconds'
1038 msg = msg_template % {'video_id': video_id, 'timeout': timeout}
1042 def _extract_f4m_formats(self, manifest_url, video_id, preference=None, f4m_id=None,
1043 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1044 fatal=True, m3u8_id=None):
1045 manifest = self._download_xml(
1046 manifest_url, video_id, 'Downloading f4m manifest',
1047 'Unable to download f4m manifest',
1048 # Some manifests may be malformed, e.g. prosiebensat1 generated manifests
1049 # (see https://github.com/rg3/youtube-dl/issues/6215#issuecomment-121704244)
1050 transform_source=transform_source,
1053 if manifest is False:
1056 return self._parse_f4m_formats(
1057 manifest, manifest_url, video_id, preference=preference, f4m_id=f4m_id,
1058 transform_source=transform_source, fatal=fatal, m3u8_id=m3u8_id)
1060 def _parse_f4m_formats(self, manifest, manifest_url, video_id, preference=None, f4m_id=None,
1061 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1062 fatal=True, m3u8_id=None):
1063 # currently youtube-dl cannot decode the playerVerificationChallenge as Akamai uses Adobe Alchemy
1064 akamai_pv = manifest.find('{http://ns.adobe.com/f4m/1.0}pv-2.0')
1065 if akamai_pv is not None and ';' in akamai_pv.text:
1066 playerVerificationChallenge = akamai_pv.text.split(';')[0]
1067 if playerVerificationChallenge.strip() != '':
1071 manifest_version = '1.0'
1072 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/1.0}media')
1074 manifest_version = '2.0'
1075 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/2.0}media')
1076 # Remove unsupported DRM protected media from final formats
1077 # rendition (see https://github.com/rg3/youtube-dl/issues/8573).
1078 media_nodes = remove_encrypted_media(media_nodes)
1081 base_url = xpath_text(
1082 manifest, ['{http://ns.adobe.com/f4m/1.0}baseURL', '{http://ns.adobe.com/f4m/2.0}baseURL'],
1083 'base URL', default=None)
1085 base_url = base_url.strip()
1087 bootstrap_info = xpath_element(
1088 manifest, ['{http://ns.adobe.com/f4m/1.0}bootstrapInfo', '{http://ns.adobe.com/f4m/2.0}bootstrapInfo'],
1089 'bootstrap info', default=None)
1091 for i, media_el in enumerate(media_nodes):
1092 tbr = int_or_none(media_el.attrib.get('bitrate'))
1093 width = int_or_none(media_el.attrib.get('width'))
1094 height = int_or_none(media_el.attrib.get('height'))
1095 format_id = '-'.join(filter(None, [f4m_id, compat_str(i if tbr is None else tbr)]))
1096 # If <bootstrapInfo> is present, the specified f4m is a
1097 # stream-level manifest, and only set-level manifests may refer to
1098 # external resources. See section 11.4 and section 4 of F4M spec
1099 if bootstrap_info is None:
1101 # @href is introduced in 2.0, see section 11.6 of F4M spec
1102 if manifest_version == '2.0':
1103 media_url = media_el.attrib.get('href')
1104 if media_url is None:
1105 media_url = media_el.attrib.get('url')
1109 media_url if media_url.startswith('http://') or media_url.startswith('https://')
1110 else ((base_url or '/'.join(manifest_url.split('/')[:-1])) + '/' + media_url))
1111 # If media_url is itself a f4m manifest do the recursive extraction
1112 # since bitrates in parent manifest (this one) and media_url manifest
1113 # may differ leading to inability to resolve the format by requested
1114 # bitrate in f4m downloader
1115 ext = determine_ext(manifest_url)
1117 f4m_formats = self._extract_f4m_formats(
1118 manifest_url, video_id, preference=preference, f4m_id=f4m_id,
1119 transform_source=transform_source, fatal=fatal)
1120 # Sometimes stream-level manifest contains single media entry that
1121 # does not contain any quality metadata (e.g. http://matchtv.ru/#live-player).
1122 # At the same time parent's media entry in set-level manifest may
1123 # contain it. We will copy it from parent in such cases.
1124 if len(f4m_formats) == 1:
1127 'tbr': f.get('tbr') or tbr,
1128 'width': f.get('width') or width,
1129 'height': f.get('height') or height,
1130 'format_id': f.get('format_id') if not tbr else format_id,
1132 formats.extend(f4m_formats)
1135 formats.extend(self._extract_m3u8_formats(
1136 manifest_url, video_id, 'mp4', preference=preference,
1137 m3u8_id=m3u8_id, fatal=fatal))
1140 'format_id': format_id,
1141 'url': manifest_url,
1142 'ext': 'flv' if bootstrap_info is not None else None,
1146 'preference': preference,
1150 def _m3u8_meta_format(self, m3u8_url, ext=None, preference=None, m3u8_id=None):
1152 'format_id': '-'.join(filter(None, [m3u8_id, 'meta'])),
1156 'preference': preference - 100 if preference else -100,
1157 'resolution': 'multiple',
1158 'format_note': 'Quality selection URL',
1161 def _extract_m3u8_formats(self, m3u8_url, video_id, ext=None,
1162 entry_protocol='m3u8', preference=None,
1163 m3u8_id=None, note=None, errnote=None,
1164 fatal=True, live=False):
1166 formats = [self._m3u8_meta_format(m3u8_url, ext, preference, m3u8_id)]
1168 format_url = lambda u: (
1170 if re.match(r'^https?://', u)
1171 else compat_urlparse.urljoin(m3u8_url, u))
1173 res = self._download_webpage_handle(
1175 note=note or 'Downloading m3u8 information',
1176 errnote=errnote or 'Failed to download m3u8 information',
1180 m3u8_doc, urlh = res
1181 m3u8_url = urlh.geturl()
1183 # We should try extracting formats only from master playlists [1], i.e.
1184 # playlists that describe available qualities. On the other hand media
1185 # playlists [2] should be returned as is since they contain just the media
1186 # without qualities renditions.
1187 # Fortunately, master playlist can be easily distinguished from media
1188 # playlist based on particular tags availability. As of [1, 2] master
1189 # playlist tags MUST NOT appear in a media playist and vice versa.
1190 # As of [3] #EXT-X-TARGETDURATION tag is REQUIRED for every media playlist
1191 # and MUST NOT appear in master playlist thus we can clearly detect media
1192 # playlist with this criterion.
1193 # 1. https://tools.ietf.org/html/draft-pantos-http-live-streaming-17#section-4.3.4
1194 # 2. https://tools.ietf.org/html/draft-pantos-http-live-streaming-17#section-4.3.3
1195 # 3. https://tools.ietf.org/html/draft-pantos-http-live-streaming-17#section-4.3.3.1
1196 if '#EXT-X-TARGETDURATION' in m3u8_doc: # media playlist, return as is
1199 'format_id': m3u8_id,
1201 'protocol': entry_protocol,
1202 'preference': preference,
1206 for line in m3u8_doc.splitlines():
1207 if line.startswith('#EXT-X-STREAM-INF:'):
1208 last_info = parse_m3u8_attributes(line)
1209 elif line.startswith('#EXT-X-MEDIA:'):
1210 last_media = parse_m3u8_attributes(line)
1211 elif line.startswith('#') or not line.strip():
1214 if last_info is None:
1215 formats.append({'url': format_url(line)})
1217 tbr = int_or_none(last_info.get('BANDWIDTH'), scale=1000)
1220 format_id.append(m3u8_id)
1221 last_media_name = last_media.get('NAME') if last_media and last_media.get('TYPE') not in ('SUBTITLES', 'CLOSED-CAPTIONS') else None
1222 # Despite specification does not mention NAME attribute for
1223 # EXT-X-STREAM-INF it still sometimes may be present
1224 stream_name = last_info.get('NAME') or last_media_name
1225 # Bandwidth of live streams may differ over time thus making
1226 # format_id unpredictable. So it's better to keep provided
1229 format_id.append(stream_name if stream_name else '%d' % (tbr if tbr else len(formats)))
1231 'format_id': '-'.join(format_id),
1232 'url': format_url(line.strip()),
1235 'fps': float_or_none(last_info.get('FRAME-RATE')),
1236 'protocol': entry_protocol,
1237 'preference': preference,
1239 resolution = last_info.get('RESOLUTION')
1241 width_str, height_str = resolution.split('x')
1242 f['width'] = int(width_str)
1243 f['height'] = int(height_str)
1244 # Unified Streaming Platform
1246 r'audio.*?(?:%3D|=)(\d+)(?:-video.*?(?:%3D|=)(\d+))?', f['url'])
1248 abr, vbr = mobj.groups()
1249 abr, vbr = float_or_none(abr, 1000), float_or_none(vbr, 1000)
1254 f.update(parse_codecs(last_info.get('CODECS')))
1255 if last_media is not None:
1256 f['m3u8_media'] = last_media
1263 def _xpath_ns(path, namespace=None):
1267 for c in path.split('/'):
1268 if not c or c == '.':
1271 out.append('{%s}%s' % (namespace, c))
1272 return '/'.join(out)
1274 def _extract_smil_formats(self, smil_url, video_id, fatal=True, f4m_params=None, transform_source=None):
1275 smil = self._download_smil(smil_url, video_id, fatal=fatal, transform_source=transform_source)
1281 namespace = self._parse_smil_namespace(smil)
1283 return self._parse_smil_formats(
1284 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
1286 def _extract_smil_info(self, smil_url, video_id, fatal=True, f4m_params=None):
1287 smil = self._download_smil(smil_url, video_id, fatal=fatal)
1290 return self._parse_smil(smil, smil_url, video_id, f4m_params=f4m_params)
1292 def _download_smil(self, smil_url, video_id, fatal=True, transform_source=None):
1293 return self._download_xml(
1294 smil_url, video_id, 'Downloading SMIL file',
1295 'Unable to download SMIL file', fatal=fatal, transform_source=transform_source)
1297 def _parse_smil(self, smil, smil_url, video_id, f4m_params=None):
1298 namespace = self._parse_smil_namespace(smil)
1300 formats = self._parse_smil_formats(
1301 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
1302 subtitles = self._parse_smil_subtitles(smil, namespace=namespace)
1304 video_id = os.path.splitext(url_basename(smil_url))[0]
1308 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
1309 name = meta.attrib.get('name')
1310 content = meta.attrib.get('content')
1311 if not name or not content:
1313 if not title and name == 'title':
1315 elif not description and name in ('description', 'abstract'):
1316 description = content
1317 elif not upload_date and name == 'date':
1318 upload_date = unified_strdate(content)
1321 'id': image.get('type'),
1322 'url': image.get('src'),
1323 'width': int_or_none(image.get('width')),
1324 'height': int_or_none(image.get('height')),
1325 } for image in smil.findall(self._xpath_ns('.//image', namespace)) if image.get('src')]
1329 'title': title or video_id,
1330 'description': description,
1331 'upload_date': upload_date,
1332 'thumbnails': thumbnails,
1334 'subtitles': subtitles,
1337 def _parse_smil_namespace(self, smil):
1338 return self._search_regex(
1339 r'(?i)^{([^}]+)?}smil$', smil.tag, 'namespace', default=None)
1341 def _parse_smil_formats(self, smil, smil_url, video_id, namespace=None, f4m_params=None, transform_rtmp_url=None):
1343 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
1344 b = meta.get('base') or meta.get('httpBase')
1355 media = smil.findall(self._xpath_ns('.//video', namespace)) + smil.findall(self._xpath_ns('.//audio', namespace))
1356 for medium in media:
1357 src = medium.get('src')
1358 if not src or src in srcs:
1362 bitrate = float_or_none(medium.get('system-bitrate') or medium.get('systemBitrate'), 1000)
1363 filesize = int_or_none(medium.get('size') or medium.get('fileSize'))
1364 width = int_or_none(medium.get('width'))
1365 height = int_or_none(medium.get('height'))
1366 proto = medium.get('proto')
1367 ext = medium.get('ext')
1368 src_ext = determine_ext(src)
1369 streamer = medium.get('streamer') or base
1371 if proto == 'rtmp' or streamer.startswith('rtmp'):
1377 'format_id': 'rtmp-%d' % (rtmp_count if bitrate is None else bitrate),
1379 'filesize': filesize,
1383 if transform_rtmp_url:
1384 streamer, src = transform_rtmp_url(streamer, src)
1385 formats[-1].update({
1391 src_url = src if src.startswith('http') else compat_urlparse.urljoin(base, src)
1392 src_url = src_url.strip()
1394 if proto == 'm3u8' or src_ext == 'm3u8':
1395 m3u8_formats = self._extract_m3u8_formats(
1396 src_url, video_id, ext or 'mp4', m3u8_id='hls', fatal=False)
1397 if len(m3u8_formats) == 1:
1399 m3u8_formats[0].update({
1400 'format_id': 'hls-%d' % (m3u8_count if bitrate is None else bitrate),
1405 formats.extend(m3u8_formats)
1408 if src_ext == 'f4m':
1413 'plugin': 'flowplayer-3.2.0.1',
1415 f4m_url += '&' if '?' in f4m_url else '?'
1416 f4m_url += compat_urllib_parse_urlencode(f4m_params)
1417 formats.extend(self._extract_f4m_formats(f4m_url, video_id, f4m_id='hds', fatal=False))
1420 if src_url.startswith('http') and self._is_valid_url(src, video_id):
1424 'ext': ext or src_ext or 'flv',
1425 'format_id': 'http-%d' % (bitrate or http_count),
1427 'filesize': filesize,
1435 def _parse_smil_subtitles(self, smil, namespace=None, subtitles_lang='en'):
1438 for num, textstream in enumerate(smil.findall(self._xpath_ns('.//textstream', namespace))):
1439 src = textstream.get('src')
1440 if not src or src in urls:
1443 ext = textstream.get('ext') or mimetype2ext(textstream.get('type')) or determine_ext(src)
1444 lang = textstream.get('systemLanguage') or textstream.get('systemLanguageName') or textstream.get('lang') or subtitles_lang
1445 subtitles.setdefault(lang, []).append({
1451 def _extract_xspf_playlist(self, playlist_url, playlist_id, fatal=True):
1452 xspf = self._download_xml(
1453 playlist_url, playlist_id, 'Downloading xpsf playlist',
1454 'Unable to download xspf manifest', fatal=fatal)
1457 return self._parse_xspf(xspf, playlist_id)
1459 def _parse_xspf(self, playlist, playlist_id):
1461 'xspf': 'http://xspf.org/ns/0/',
1462 's1': 'http://static.streamone.nl/player/ns/0',
1466 for track in playlist.findall(xpath_with_ns('./xspf:trackList/xspf:track', NS_MAP)):
1468 track, xpath_with_ns('./xspf:title', NS_MAP), 'title', default=playlist_id)
1469 description = xpath_text(
1470 track, xpath_with_ns('./xspf:annotation', NS_MAP), 'description')
1471 thumbnail = xpath_text(
1472 track, xpath_with_ns('./xspf:image', NS_MAP), 'thumbnail')
1473 duration = float_or_none(
1474 xpath_text(track, xpath_with_ns('./xspf:duration', NS_MAP), 'duration'), 1000)
1477 'url': location.text,
1478 'format_id': location.get(xpath_with_ns('s1:label', NS_MAP)),
1479 'width': int_or_none(location.get(xpath_with_ns('s1:width', NS_MAP))),
1480 'height': int_or_none(location.get(xpath_with_ns('s1:height', NS_MAP))),
1481 } for location in track.findall(xpath_with_ns('./xspf:location', NS_MAP))]
1482 self._sort_formats(formats)
1487 'description': description,
1488 'thumbnail': thumbnail,
1489 'duration': duration,
1494 def _extract_mpd_formats(self, mpd_url, video_id, mpd_id=None, note=None, errnote=None, fatal=True, formats_dict={}):
1495 res = self._download_webpage_handle(
1497 note=note or 'Downloading MPD manifest',
1498 errnote=errnote or 'Failed to download MPD manifest',
1503 mpd_base_url = re.match(r'https?://.+/', urlh.geturl()).group()
1505 return self._parse_mpd_formats(
1506 compat_etree_fromstring(mpd.encode('utf-8')), mpd_id, mpd_base_url, formats_dict=formats_dict)
1508 def _parse_mpd_formats(self, mpd_doc, mpd_id=None, mpd_base_url='', formats_dict={}):
1510 Parse formats from MPD manifest.
1512 1. MPEG-DASH Standard, ISO/IEC 23009-1:2014(E),
1513 http://standards.iso.org/ittf/PubliclyAvailableStandards/c065274_ISO_IEC_23009-1_2014.zip
1514 2. https://en.wikipedia.org/wiki/Dynamic_Adaptive_Streaming_over_HTTP
1516 if mpd_doc.get('type') == 'dynamic':
1519 namespace = self._search_regex(r'(?i)^{([^}]+)?}MPD$', mpd_doc.tag, 'namespace', default=None)
1522 return self._xpath_ns(path, namespace)
1524 def is_drm_protected(element):
1525 return element.find(_add_ns('ContentProtection')) is not None
1527 def extract_multisegment_info(element, ms_parent_info):
1528 ms_info = ms_parent_info.copy()
1529 segment_list = element.find(_add_ns('SegmentList'))
1530 if segment_list is not None:
1531 segment_urls_e = segment_list.findall(_add_ns('SegmentURL'))
1533 ms_info['segment_urls'] = [segment.attrib['media'] for segment in segment_urls_e]
1534 initialization = segment_list.find(_add_ns('Initialization'))
1535 if initialization is not None:
1536 ms_info['initialization_url'] = initialization.attrib['sourceURL']
1538 segment_template = element.find(_add_ns('SegmentTemplate'))
1539 if segment_template is not None:
1540 start_number = segment_template.get('startNumber')
1542 ms_info['start_number'] = int(start_number)
1543 segment_timeline = segment_template.find(_add_ns('SegmentTimeline'))
1544 if segment_timeline is not None:
1545 s_e = segment_timeline.findall(_add_ns('S'))
1547 ms_info['total_number'] = 0
1550 r = int(s.get('r', 0))
1551 ms_info['total_number'] += 1 + r
1552 ms_info['s'].append({
1553 't': int(s.get('t', 0)),
1554 # @d is mandatory (see [1, 5.3.9.6.2, Table 17, page 60])
1555 'd': int(s.attrib['d']),
1559 timescale = segment_template.get('timescale')
1561 ms_info['timescale'] = int(timescale)
1562 segment_duration = segment_template.get('duration')
1563 if segment_duration:
1564 ms_info['segment_duration'] = int(segment_duration)
1565 media_template = segment_template.get('media')
1567 ms_info['media_template'] = media_template
1568 initialization = segment_template.get('initialization')
1570 ms_info['initialization_url'] = initialization
1572 initialization = segment_template.find(_add_ns('Initialization'))
1573 if initialization is not None:
1574 ms_info['initialization_url'] = initialization.attrib['sourceURL']
1577 mpd_duration = parse_duration(mpd_doc.get('mediaPresentationDuration'))
1579 for period in mpd_doc.findall(_add_ns('Period')):
1580 period_duration = parse_duration(period.get('duration')) or mpd_duration
1581 period_ms_info = extract_multisegment_info(period, {
1585 for adaptation_set in period.findall(_add_ns('AdaptationSet')):
1586 if is_drm_protected(adaptation_set):
1588 adaption_set_ms_info = extract_multisegment_info(adaptation_set, period_ms_info)
1589 for representation in adaptation_set.findall(_add_ns('Representation')):
1590 if is_drm_protected(representation):
1592 representation_attrib = adaptation_set.attrib.copy()
1593 representation_attrib.update(representation.attrib)
1594 # According to [1, 5.3.7.2, Table 9, page 41], @mimeType is mandatory
1595 mime_type = representation_attrib['mimeType']
1596 content_type = mime_type.split('/')[0]
1597 if content_type == 'text':
1598 # TODO implement WebVTT downloading
1600 elif content_type == 'video' or content_type == 'audio':
1602 for element in (representation, adaptation_set, period, mpd_doc):
1603 base_url_e = element.find(_add_ns('BaseURL'))
1604 if base_url_e is not None:
1605 base_url = base_url_e.text + base_url
1606 if re.match(r'^https?://', base_url):
1608 if mpd_base_url and not re.match(r'^https?://', base_url):
1609 if not mpd_base_url.endswith('/') and not base_url.startswith('/'):
1611 base_url = mpd_base_url + base_url
1612 representation_id = representation_attrib.get('id')
1613 lang = representation_attrib.get('lang')
1614 url_el = representation.find(_add_ns('BaseURL'))
1615 filesize = int_or_none(url_el.attrib.get('{http://youtube.com/yt/2012/10/10}contentLength') if url_el is not None else None)
1617 'format_id': '%s-%s' % (mpd_id, representation_id) if mpd_id else representation_id,
1619 'ext': mimetype2ext(mime_type),
1620 'width': int_or_none(representation_attrib.get('width')),
1621 'height': int_or_none(representation_attrib.get('height')),
1622 'tbr': int_or_none(representation_attrib.get('bandwidth'), 1000),
1623 'asr': int_or_none(representation_attrib.get('audioSamplingRate')),
1624 'fps': int_or_none(representation_attrib.get('frameRate')),
1625 'vcodec': 'none' if content_type == 'audio' else representation_attrib.get('codecs'),
1626 'acodec': 'none' if content_type == 'video' else representation_attrib.get('codecs'),
1627 'language': lang if lang not in ('mul', 'und', 'zxx', 'mis') else None,
1628 'format_note': 'DASH %s' % content_type,
1629 'filesize': filesize,
1631 representation_ms_info = extract_multisegment_info(representation, adaption_set_ms_info)
1632 if 'segment_urls' not in representation_ms_info and 'media_template' in representation_ms_info:
1633 if 'total_number' not in representation_ms_info and 'segment_duration':
1634 segment_duration = float(representation_ms_info['segment_duration']) / float(representation_ms_info['timescale'])
1635 representation_ms_info['total_number'] = int(math.ceil(float(period_duration) / segment_duration))
1636 media_template = representation_ms_info['media_template']
1637 media_template = media_template.replace('$RepresentationID$', representation_id)
1638 media_template = re.sub(r'\$(Number|Bandwidth|Time)\$', r'%(\1)d', media_template)
1639 media_template = re.sub(r'\$(Number|Bandwidth|Time)%([^$]+)\$', r'%(\1)\2', media_template)
1640 media_template.replace('$$', '$')
1642 # As per [1, 5.3.9.4.4, Table 16, page 55] $Number$ and $Time$
1643 # can't be used at the same time
1644 if '%(Number' in media_template:
1645 representation_ms_info['segment_urls'] = [
1647 'Number': segment_number,
1648 'Bandwidth': representation_attrib.get('bandwidth'),
1650 for segment_number in range(
1651 representation_ms_info['start_number'],
1652 representation_ms_info['total_number'] + representation_ms_info['start_number'])]
1654 representation_ms_info['segment_urls'] = []
1657 def add_segment_url():
1658 representation_ms_info['segment_urls'].append(
1660 'Time': segment_time,
1661 'Bandwidth': representation_attrib.get('bandwidth'),
1665 for num, s in enumerate(representation_ms_info['s']):
1666 segment_time = s.get('t') or segment_time
1668 for r in range(s.get('r', 0)):
1669 segment_time += s['d']
1671 segment_time += s['d']
1672 if 'segment_urls' in representation_ms_info:
1674 'segment_urls': representation_ms_info['segment_urls'],
1675 'protocol': 'http_dash_segments',
1677 if 'initialization_url' in representation_ms_info:
1678 initialization_url = representation_ms_info['initialization_url'].replace('$RepresentationID$', representation_id)
1680 'initialization_url': initialization_url,
1682 if not f.get('url'):
1683 f['url'] = initialization_url
1685 existing_format = next(
1686 fo for fo in formats
1687 if fo['format_id'] == representation_id)
1688 except StopIteration:
1689 full_info = formats_dict.get(representation_id, {}).copy()
1691 formats.append(full_info)
1693 existing_format.update(f)
1695 self.report_warning('Unknown MIME type %s in DASH manifest' % mime_type)
1698 def _parse_html5_media_entries(self, base_url, webpage):
1699 def absolute_url(video_url):
1700 return compat_urlparse.urljoin(base_url, video_url)
1702 def parse_content_type(content_type):
1703 if not content_type:
1705 ctr = re.search(r'(?P<mimetype>[^/]+/[^;]+)(?:;\s*codecs="?(?P<codecs>[^"]+))?', content_type)
1707 mimetype, codecs = ctr.groups()
1708 f = parse_codecs(codecs)
1709 f['ext'] = mimetype2ext(mimetype)
1714 for media_tag, media_type, media_content in re.findall(r'(?s)(<(?P<tag>video|audio)[^>]*>)(.*?)</(?P=tag)>', webpage):
1719 media_attributes = extract_attributes(media_tag)
1720 src = media_attributes.get('src')
1722 media_info['formats'].append({
1723 'url': absolute_url(src),
1724 'vcodec': 'none' if media_type == 'audio' else None,
1726 media_info['thumbnail'] = media_attributes.get('poster')
1728 for source_tag in re.findall(r'<source[^>]+>', media_content):
1729 source_attributes = extract_attributes(source_tag)
1730 src = source_attributes.get('src')
1733 f = parse_content_type(source_attributes.get('type'))
1735 'url': absolute_url(src),
1736 'vcodec': 'none' if media_type == 'audio' else None,
1738 media_info['formats'].append(f)
1739 for track_tag in re.findall(r'<track[^>]+>', media_content):
1740 track_attributes = extract_attributes(track_tag)
1741 kind = track_attributes.get('kind')
1742 if not kind or kind == 'subtitles':
1743 src = track_attributes.get('src')
1746 lang = track_attributes.get('srclang') or track_attributes.get('lang') or track_attributes.get('label')
1747 media_info['subtitles'].setdefault(lang, []).append({
1748 'url': absolute_url(src),
1750 if media_info['formats']:
1751 entries.append(media_info)
1754 def _live_title(self, name):
1755 """ Generate the title for a live video """
1756 now = datetime.datetime.now()
1757 now_str = now.strftime('%Y-%m-%d %H:%M')
1758 return name + ' ' + now_str
1760 def _int(self, v, name, fatal=False, **kwargs):
1761 res = int_or_none(v, **kwargs)
1762 if 'get_attr' in kwargs:
1763 print(getattr(v, kwargs['get_attr']))
1765 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
1767 raise ExtractorError(msg)
1769 self._downloader.report_warning(msg)
1772 def _float(self, v, name, fatal=False, **kwargs):
1773 res = float_or_none(v, **kwargs)
1775 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
1777 raise ExtractorError(msg)
1779 self._downloader.report_warning(msg)
1782 def _set_cookie(self, domain, name, value, expire_time=None):
1783 cookie = compat_cookiejar.Cookie(
1784 0, name, value, None, None, domain, None,
1785 None, '/', True, False, expire_time, '', None, None, None)
1786 self._downloader.cookiejar.set_cookie(cookie)
1788 def _get_cookies(self, url):
1789 """ Return a compat_cookies.SimpleCookie with the cookies for the url """
1790 req = sanitized_Request(url)
1791 self._downloader.cookiejar.add_cookie_header(req)
1792 return compat_cookies.SimpleCookie(req.get_header('Cookie'))
1794 def get_testcases(self, include_onlymatching=False):
1795 t = getattr(self, '_TEST', None)
1797 assert not hasattr(self, '_TESTS'), \
1798 '%s has _TEST and _TESTS' % type(self).__name__
1801 tests = getattr(self, '_TESTS', [])
1803 if not include_onlymatching and t.get('only_matching', False):
1805 t['name'] = type(self).__name__[:-len('IE')]
1808 def is_suitable(self, age_limit):
1809 """ Test whether the extractor is generally suitable for the given
1810 age limit (i.e. pornographic sites are not, all others usually are) """
1812 any_restricted = False
1813 for tc in self.get_testcases(include_onlymatching=False):
1814 if tc.get('playlist', []):
1815 tc = tc['playlist'][0]
1816 is_restricted = age_restricted(
1817 tc.get('info_dict', {}).get('age_limit'), age_limit)
1818 if not is_restricted:
1820 any_restricted = any_restricted or is_restricted
1821 return not any_restricted
1823 def extract_subtitles(self, *args, **kwargs):
1824 if (self._downloader.params.get('writesubtitles', False) or
1825 self._downloader.params.get('listsubtitles')):
1826 return self._get_subtitles(*args, **kwargs)
1829 def _get_subtitles(self, *args, **kwargs):
1830 raise NotImplementedError('This method must be implemented by subclasses')
1833 def _merge_subtitle_items(subtitle_list1, subtitle_list2):
1834 """ Merge subtitle items for one language. Items with duplicated URLs
1835 will be dropped. """
1836 list1_urls = set([item['url'] for item in subtitle_list1])
1837 ret = list(subtitle_list1)
1838 ret.extend([item for item in subtitle_list2 if item['url'] not in list1_urls])
1842 def _merge_subtitles(cls, subtitle_dict1, subtitle_dict2):
1843 """ Merge two subtitle dictionaries, language by language. """
1844 ret = dict(subtitle_dict1)
1845 for lang in subtitle_dict2:
1846 ret[lang] = cls._merge_subtitle_items(subtitle_dict1.get(lang, []), subtitle_dict2[lang])
1849 def extract_automatic_captions(self, *args, **kwargs):
1850 if (self._downloader.params.get('writeautomaticsub', False) or
1851 self._downloader.params.get('listsubtitles')):
1852 return self._get_automatic_captions(*args, **kwargs)
1855 def _get_automatic_captions(self, *args, **kwargs):
1856 raise NotImplementedError('This method must be implemented by subclasses')
1858 def mark_watched(self, *args, **kwargs):
1859 if (self._downloader.params.get('mark_watched', False) and
1860 (self._get_login_info()[0] is not None or
1861 self._downloader.params.get('cookiefile') is not None)):
1862 self._mark_watched(*args, **kwargs)
1864 def _mark_watched(self, *args, **kwargs):
1865 raise NotImplementedError('This method must be implemented by subclasses')
1867 def geo_verification_headers(self):
1869 geo_verification_proxy = self._downloader.params.get('geo_verification_proxy')
1870 if geo_verification_proxy:
1871 headers['Ytdl-request-proxy'] = geo_verification_proxy
1875 class SearchInfoExtractor(InfoExtractor):
1877 Base class for paged search queries extractors.
1878 They accept URLs in the format _SEARCH_KEY(|all|[0-9]):{query}
1879 Instances should define _SEARCH_KEY and _MAX_RESULTS.
1883 def _make_valid_url(cls):
1884 return r'%s(?P<prefix>|[1-9][0-9]*|all):(?P<query>[\s\S]+)' % cls._SEARCH_KEY
1887 def suitable(cls, url):
1888 return re.match(cls._make_valid_url(), url) is not None
1890 def _real_extract(self, query):
1891 mobj = re.match(self._make_valid_url(), query)
1893 raise ExtractorError('Invalid search query "%s"' % query)
1895 prefix = mobj.group('prefix')
1896 query = mobj.group('query')
1898 return self._get_n_results(query, 1)
1899 elif prefix == 'all':
1900 return self._get_n_results(query, self._MAX_RESULTS)
1904 raise ExtractorError('invalid download number %s for query "%s"' % (n, query))
1905 elif n > self._MAX_RESULTS:
1906 self._downloader.report_warning('%s returns max %i results (you requested %i)' % (self._SEARCH_KEY, self._MAX_RESULTS, n))
1907 n = self._MAX_RESULTS
1908 return self._get_n_results(query, n)
1910 def _get_n_results(self, query, n):
1911 """Get a specified number of results for a query"""
1912 raise NotImplementedError('This method must be implemented by subclasses')
1915 def SEARCH_KEY(self):
1916 return self._SEARCH_KEY