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_login_info(self):
667 Get the login info as (username, password)
668 It will look in the netrc file using the _NETRC_MACHINE value
669 If there's no info available, return (None, None)
671 if self._downloader is None:
676 downloader_params = self._downloader.params
678 # Attempt to use provided username and password or .netrc data
679 if downloader_params.get('username') is not None:
680 username = downloader_params['username']
681 password = downloader_params['password']
682 elif downloader_params.get('usenetrc', False):
684 info = netrc.netrc().authenticators(self._NETRC_MACHINE)
689 raise netrc.NetrcParseError('No authenticators for %s' % self._NETRC_MACHINE)
690 except (IOError, netrc.NetrcParseError) as err:
691 self._downloader.report_warning('parsing .netrc: %s' % error_to_compat_str(err))
693 return (username, password)
695 def _get_tfa_info(self, note='two-factor verification code'):
697 Get the two-factor authentication info
698 TODO - asking the user will be required for sms/phone verify
699 currently just uses the command line option
700 If there's no info available, return None
702 if self._downloader is None:
704 downloader_params = self._downloader.params
706 if downloader_params.get('twofactor') is not None:
707 return downloader_params['twofactor']
709 return compat_getpass('Type %s and press [Return]: ' % note)
711 # Helper functions for extracting OpenGraph info
713 def _og_regexes(prop):
714 content_re = r'content=(?:"([^"]+?)"|\'([^\']+?)\'|\s*([^\s"\'=<>`]+?))'
715 property_re = (r'(?:name|property)=(?:\'og:%(prop)s\'|"og:%(prop)s"|\s*og:%(prop)s\b)'
716 % {'prop': re.escape(prop)})
717 template = r'<meta[^>]+?%s[^>]+?%s'
719 template % (property_re, content_re),
720 template % (content_re, property_re),
724 def _meta_regex(prop):
725 return r'''(?isx)<meta
726 (?=[^>]+(?:itemprop|name|property|id|http-equiv)=(["\']?)%s\1)
727 [^>]+?content=(["\'])(?P<content>.*?)\2''' % re.escape(prop)
729 def _og_search_property(self, prop, html, name=None, **kargs):
730 if not isinstance(prop, (list, tuple)):
733 name = 'OpenGraph %s' % prop[0]
736 og_regexes.extend(self._og_regexes(p))
737 escaped = self._search_regex(og_regexes, html, name, flags=re.DOTALL, **kargs)
740 return unescapeHTML(escaped)
742 def _og_search_thumbnail(self, html, **kargs):
743 return self._og_search_property('image', html, 'thumbnail URL', fatal=False, **kargs)
745 def _og_search_description(self, html, **kargs):
746 return self._og_search_property('description', html, fatal=False, **kargs)
748 def _og_search_title(self, html, **kargs):
749 return self._og_search_property('title', html, **kargs)
751 def _og_search_video_url(self, html, name='video url', secure=True, **kargs):
752 regexes = self._og_regexes('video') + self._og_regexes('video:url')
754 regexes = self._og_regexes('video:secure_url') + regexes
755 return self._html_search_regex(regexes, html, name, **kargs)
757 def _og_search_url(self, html, **kargs):
758 return self._og_search_property('url', html, **kargs)
760 def _html_search_meta(self, name, html, display_name=None, fatal=False, **kwargs):
761 if not isinstance(name, (list, tuple)):
763 if display_name is None:
764 display_name = name[0]
765 return self._html_search_regex(
766 [self._meta_regex(n) for n in name],
767 html, display_name, fatal=fatal, group='content', **kwargs)
769 def _dc_search_uploader(self, html):
770 return self._html_search_meta('dc.creator', html, 'uploader')
772 def _rta_search(self, html):
773 # See http://www.rtalabel.org/index.php?content=howtofaq#single
774 if re.search(r'(?ix)<meta\s+name="rating"\s+'
775 r' content="RTA-5042-1996-1400-1577-RTA"',
780 def _media_rating_search(self, html):
781 # See http://www.tjg-designs.com/WP/metadata-code-examples-adding-metadata-to-your-web-pages/
782 rating = self._html_search_meta('rating', html)
794 return RATING_TABLE.get(rating.lower())
796 def _family_friendly_search(self, html):
797 # See http://schema.org/VideoObject
798 family_friendly = self._html_search_meta('isFamilyFriendly', html)
800 if not family_friendly:
809 return RATING_TABLE.get(family_friendly.lower())
811 def _twitter_search_player(self, html):
812 return self._html_search_meta('twitter:player', html,
813 'twitter card player')
815 def _search_json_ld(self, html, video_id, expected_type=None, **kwargs):
816 json_ld = self._search_regex(
817 r'(?s)<script[^>]+type=(["\'])application/ld\+json\1[^>]*>(?P<json_ld>.+?)</script>',
818 html, 'JSON-LD', group='json_ld', **kwargs)
819 default = kwargs.get('default', NO_DEFAULT)
821 return default if default is not NO_DEFAULT else {}
822 # JSON-LD may be malformed and thus `fatal` should be respected.
823 # At the same time `default` may be passed that assumes `fatal=False`
824 # for _search_regex. Let's simulate the same behavior here as well.
825 fatal = kwargs.get('fatal', True) if default == NO_DEFAULT else False
826 return self._json_ld(json_ld, video_id, fatal=fatal, expected_type=expected_type)
828 def _json_ld(self, json_ld, video_id, fatal=True, expected_type=None):
829 if isinstance(json_ld, compat_str):
830 json_ld = self._parse_json(json_ld, video_id, fatal=fatal)
834 if not isinstance(json_ld, (list, tuple, dict)):
836 if isinstance(json_ld, dict):
839 if e.get('@context') == 'http://schema.org':
840 item_type = e.get('@type')
841 if expected_type is not None and expected_type != item_type:
843 if item_type == 'TVEpisode':
845 'episode': unescapeHTML(e.get('name')),
846 'episode_number': int_or_none(e.get('episodeNumber')),
847 'description': unescapeHTML(e.get('description')),
849 part_of_season = e.get('partOfSeason')
850 if isinstance(part_of_season, dict) and part_of_season.get('@type') == 'TVSeason':
851 info['season_number'] = int_or_none(part_of_season.get('seasonNumber'))
852 part_of_series = e.get('partOfSeries') or e.get('partOfTVSeries')
853 if isinstance(part_of_series, dict) and part_of_series.get('@type') == 'TVSeries':
854 info['series'] = unescapeHTML(part_of_series.get('name'))
855 elif item_type == 'Article':
857 'timestamp': parse_iso8601(e.get('datePublished')),
858 'title': unescapeHTML(e.get('headline')),
859 'description': unescapeHTML(e.get('articleBody')),
861 elif item_type == 'VideoObject':
863 'url': e.get('contentUrl'),
864 'title': unescapeHTML(e.get('name')),
865 'description': unescapeHTML(e.get('description')),
866 'thumbnail': e.get('thumbnailUrl'),
867 'duration': parse_duration(e.get('duration')),
868 'timestamp': unified_timestamp(e.get('uploadDate')),
869 'filesize': float_or_none(e.get('contentSize')),
870 'tbr': int_or_none(e.get('bitrate')),
871 'width': int_or_none(e.get('width')),
872 'height': int_or_none(e.get('height')),
875 return dict((k, v) for k, v in info.items() if v is not None)
878 def _hidden_inputs(html):
879 html = re.sub(r'<!--(?:(?!<!--).)*-->', '', html)
881 for input in re.findall(r'(?i)<input([^>]+)>', html):
882 if not re.search(r'type=(["\'])(?:hidden|submit)\1', input):
884 name = re.search(r'(?:name|id)=(["\'])(?P<value>.+?)\1', input)
887 value = re.search(r'value=(["\'])(?P<value>.*?)\1', input)
890 hidden_inputs[name.group('value')] = value.group('value')
893 def _form_hidden_inputs(self, form_id, html):
894 form = self._search_regex(
895 r'(?is)<form[^>]+?id=(["\'])%s\1[^>]*>(?P<form>.+?)</form>' % form_id,
896 html, '%s form' % form_id, group='form')
897 return self._hidden_inputs(form)
899 def _sort_formats(self, formats, field_preference=None):
901 raise ExtractorError('No video formats found')
904 # Automatically determine tbr when missing based on abr and vbr (improves
905 # formats sorting in some cases)
906 if 'tbr' not in f and f.get('abr') is not None and f.get('vbr') is not None:
907 f['tbr'] = f['abr'] + f['vbr']
910 # TODO remove the following workaround
911 from ..utils import determine_ext
912 if not f.get('ext') and 'url' in f:
913 f['ext'] = determine_ext(f['url'])
915 if isinstance(field_preference, (list, tuple)):
918 if f.get(field) is not None
919 else ('' if field == 'format_id' else -1)
920 for field in field_preference)
922 preference = f.get('preference')
923 if preference is None:
925 if f.get('ext') in ['f4f', 'f4m']: # Not yet supported
928 protocol = f.get('protocol') or determine_protocol(f)
929 proto_preference = 0 if protocol in ['http', 'https'] else (-0.5 if protocol == 'rtsp' else -0.1)
931 if f.get('vcodec') == 'none': # audio only
933 if self._downloader.params.get('prefer_free_formats'):
934 ORDER = ['aac', 'mp3', 'm4a', 'webm', 'ogg', 'opus']
936 ORDER = ['webm', 'opus', 'ogg', 'mp3', 'aac', 'm4a']
939 audio_ext_preference = ORDER.index(f['ext'])
941 audio_ext_preference = -1
943 if f.get('acodec') == 'none': # video only
945 if self._downloader.params.get('prefer_free_formats'):
946 ORDER = ['flv', 'mp4', 'webm']
948 ORDER = ['webm', 'flv', 'mp4']
950 ext_preference = ORDER.index(f['ext'])
953 audio_ext_preference = 0
957 f.get('language_preference') if f.get('language_preference') is not None else -1,
958 f.get('quality') if f.get('quality') is not None else -1,
959 f.get('tbr') if f.get('tbr') is not None else -1,
960 f.get('filesize') if f.get('filesize') is not None else -1,
961 f.get('vbr') if f.get('vbr') is not None else -1,
962 f.get('height') if f.get('height') is not None else -1,
963 f.get('width') if f.get('width') is not None else -1,
966 f.get('abr') if f.get('abr') is not None else -1,
967 audio_ext_preference,
968 f.get('fps') if f.get('fps') is not None else -1,
969 f.get('filesize_approx') if f.get('filesize_approx') is not None else -1,
970 f.get('source_preference') if f.get('source_preference') is not None else -1,
971 f.get('format_id') if f.get('format_id') is not None else '',
973 formats.sort(key=_formats_key)
975 def _check_formats(self, formats, video_id):
978 lambda f: self._is_valid_url(
980 item='%s video format' % f.get('format_id') if f.get('format_id') else 'video'),
984 def _remove_duplicate_formats(formats):
988 if f['url'] not in format_urls:
989 format_urls.add(f['url'])
990 unique_formats.append(f)
991 formats[:] = unique_formats
993 def _is_valid_url(self, url, video_id, item='video'):
994 url = self._proto_relative_url(url, scheme='http:')
995 # For now assume non HTTP(S) URLs always valid
996 if not (url.startswith('http://') or url.startswith('https://')):
999 self._request_webpage(url, video_id, 'Checking %s URL' % item)
1001 except ExtractorError as e:
1002 if isinstance(e.cause, compat_urllib_error.URLError):
1004 '%s: %s URL is invalid, skipping' % (video_id, item))
1008 def http_scheme(self):
1009 """ Either "http:" or "https:", depending on the user's preferences """
1012 if self._downloader.params.get('prefer_insecure', False)
1015 def _proto_relative_url(self, url, scheme=None):
1018 if url.startswith('//'):
1020 scheme = self.http_scheme()
1025 def _sleep(self, timeout, video_id, msg_template=None):
1026 if msg_template is None:
1027 msg_template = '%(video_id)s: Waiting for %(timeout)s seconds'
1028 msg = msg_template % {'video_id': video_id, 'timeout': timeout}
1032 def _extract_f4m_formats(self, manifest_url, video_id, preference=None, f4m_id=None,
1033 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1034 fatal=True, m3u8_id=None):
1035 manifest = self._download_xml(
1036 manifest_url, video_id, 'Downloading f4m manifest',
1037 'Unable to download f4m manifest',
1038 # Some manifests may be malformed, e.g. prosiebensat1 generated manifests
1039 # (see https://github.com/rg3/youtube-dl/issues/6215#issuecomment-121704244)
1040 transform_source=transform_source,
1043 if manifest is False:
1046 return self._parse_f4m_formats(
1047 manifest, manifest_url, video_id, preference=preference, f4m_id=f4m_id,
1048 transform_source=transform_source, fatal=fatal, m3u8_id=m3u8_id)
1050 def _parse_f4m_formats(self, manifest, manifest_url, video_id, preference=None, f4m_id=None,
1051 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1052 fatal=True, m3u8_id=None):
1053 # currently youtube-dl cannot decode the playerVerificationChallenge as Akamai uses Adobe Alchemy
1054 akamai_pv = manifest.find('{http://ns.adobe.com/f4m/1.0}pv-2.0')
1055 if akamai_pv is not None and ';' in akamai_pv.text:
1056 playerVerificationChallenge = akamai_pv.text.split(';')[0]
1057 if playerVerificationChallenge.strip() != '':
1061 manifest_version = '1.0'
1062 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/1.0}media')
1064 manifest_version = '2.0'
1065 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/2.0}media')
1066 # Remove unsupported DRM protected media from final formats
1067 # rendition (see https://github.com/rg3/youtube-dl/issues/8573).
1068 media_nodes = remove_encrypted_media(media_nodes)
1071 base_url = xpath_text(
1072 manifest, ['{http://ns.adobe.com/f4m/1.0}baseURL', '{http://ns.adobe.com/f4m/2.0}baseURL'],
1073 'base URL', default=None)
1075 base_url = base_url.strip()
1077 bootstrap_info = xpath_element(
1078 manifest, ['{http://ns.adobe.com/f4m/1.0}bootstrapInfo', '{http://ns.adobe.com/f4m/2.0}bootstrapInfo'],
1079 'bootstrap info', default=None)
1081 for i, media_el in enumerate(media_nodes):
1082 tbr = int_or_none(media_el.attrib.get('bitrate'))
1083 width = int_or_none(media_el.attrib.get('width'))
1084 height = int_or_none(media_el.attrib.get('height'))
1085 format_id = '-'.join(filter(None, [f4m_id, compat_str(i if tbr is None else tbr)]))
1086 # If <bootstrapInfo> is present, the specified f4m is a
1087 # stream-level manifest, and only set-level manifests may refer to
1088 # external resources. See section 11.4 and section 4 of F4M spec
1089 if bootstrap_info is None:
1091 # @href is introduced in 2.0, see section 11.6 of F4M spec
1092 if manifest_version == '2.0':
1093 media_url = media_el.attrib.get('href')
1094 if media_url is None:
1095 media_url = media_el.attrib.get('url')
1099 media_url if media_url.startswith('http://') or media_url.startswith('https://')
1100 else ((base_url or '/'.join(manifest_url.split('/')[:-1])) + '/' + media_url))
1101 # If media_url is itself a f4m manifest do the recursive extraction
1102 # since bitrates in parent manifest (this one) and media_url manifest
1103 # may differ leading to inability to resolve the format by requested
1104 # bitrate in f4m downloader
1105 ext = determine_ext(manifest_url)
1107 f4m_formats = self._extract_f4m_formats(
1108 manifest_url, video_id, preference=preference, f4m_id=f4m_id,
1109 transform_source=transform_source, fatal=fatal)
1110 # Sometimes stream-level manifest contains single media entry that
1111 # does not contain any quality metadata (e.g. http://matchtv.ru/#live-player).
1112 # At the same time parent's media entry in set-level manifest may
1113 # contain it. We will copy it from parent in such cases.
1114 if len(f4m_formats) == 1:
1117 'tbr': f.get('tbr') or tbr,
1118 'width': f.get('width') or width,
1119 'height': f.get('height') or height,
1120 'format_id': f.get('format_id') if not tbr else format_id,
1122 formats.extend(f4m_formats)
1125 formats.extend(self._extract_m3u8_formats(
1126 manifest_url, video_id, 'mp4', preference=preference,
1127 m3u8_id=m3u8_id, fatal=fatal))
1130 'format_id': format_id,
1131 'url': manifest_url,
1132 'ext': 'flv' if bootstrap_info is not None else None,
1136 'preference': preference,
1140 def _m3u8_meta_format(self, m3u8_url, ext=None, preference=None, m3u8_id=None):
1142 'format_id': '-'.join(filter(None, [m3u8_id, 'meta'])),
1146 'preference': preference - 100 if preference else -100,
1147 'resolution': 'multiple',
1148 'format_note': 'Quality selection URL',
1151 def _extract_m3u8_formats(self, m3u8_url, video_id, ext=None,
1152 entry_protocol='m3u8', preference=None,
1153 m3u8_id=None, note=None, errnote=None,
1154 fatal=True, live=False):
1156 formats = [self._m3u8_meta_format(m3u8_url, ext, preference, m3u8_id)]
1158 format_url = lambda u: (
1160 if re.match(r'^https?://', u)
1161 else compat_urlparse.urljoin(m3u8_url, u))
1163 res = self._download_webpage_handle(
1165 note=note or 'Downloading m3u8 information',
1166 errnote=errnote or 'Failed to download m3u8 information',
1170 m3u8_doc, urlh = res
1171 m3u8_url = urlh.geturl()
1173 # We should try extracting formats only from master playlists [1], i.e.
1174 # playlists that describe available qualities. On the other hand media
1175 # playlists [2] should be returned as is since they contain just the media
1176 # without qualities renditions.
1177 # Fortunately, master playlist can be easily distinguished from media
1178 # playlist based on particular tags availability. As of [1, 2] master
1179 # playlist tags MUST NOT appear in a media playist and vice versa.
1180 # As of [3] #EXT-X-TARGETDURATION tag is REQUIRED for every media playlist
1181 # and MUST NOT appear in master playlist thus we can clearly detect media
1182 # playlist with this criterion.
1183 # 1. https://tools.ietf.org/html/draft-pantos-http-live-streaming-17#section-4.3.4
1184 # 2. https://tools.ietf.org/html/draft-pantos-http-live-streaming-17#section-4.3.3
1185 # 3. https://tools.ietf.org/html/draft-pantos-http-live-streaming-17#section-4.3.3.1
1186 if '#EXT-X-TARGETDURATION' in m3u8_doc: # media playlist, return as is
1189 'format_id': m3u8_id,
1191 'protocol': entry_protocol,
1192 'preference': preference,
1196 for line in m3u8_doc.splitlines():
1197 if line.startswith('#EXT-X-STREAM-INF:'):
1198 last_info = parse_m3u8_attributes(line)
1199 elif line.startswith('#EXT-X-MEDIA:'):
1200 last_media = parse_m3u8_attributes(line)
1201 elif line.startswith('#') or not line.strip():
1204 if last_info is None:
1205 formats.append({'url': format_url(line)})
1207 tbr = int_or_none(last_info.get('BANDWIDTH'), scale=1000)
1210 format_id.append(m3u8_id)
1211 last_media_name = last_media.get('NAME') if last_media and last_media.get('TYPE') not in ('SUBTITLES', 'CLOSED-CAPTIONS') else None
1212 # Despite specification does not mention NAME attribute for
1213 # EXT-X-STREAM-INF it still sometimes may be present
1214 stream_name = last_info.get('NAME') or last_media_name
1215 # Bandwidth of live streams may differ over time thus making
1216 # format_id unpredictable. So it's better to keep provided
1219 format_id.append(stream_name if stream_name else '%d' % (tbr if tbr else len(formats)))
1221 'format_id': '-'.join(format_id),
1222 'url': format_url(line.strip()),
1225 'fps': float_or_none(last_info.get('FRAME-RATE')),
1226 'protocol': entry_protocol,
1227 'preference': preference,
1229 resolution = last_info.get('RESOLUTION')
1231 width_str, height_str = resolution.split('x')
1232 f['width'] = int(width_str)
1233 f['height'] = int(height_str)
1234 # Unified Streaming Platform
1236 r'audio.*?(?:%3D|=)(\d+)(?:-video.*?(?:%3D|=)(\d+))?', f['url'])
1238 abr, vbr = mobj.groups()
1239 abr, vbr = float_or_none(abr, 1000), float_or_none(vbr, 1000)
1244 f.update(parse_codecs(last_info.get('CODECS')))
1245 if last_media is not None:
1246 f['m3u8_media'] = last_media
1253 def _xpath_ns(path, namespace=None):
1257 for c in path.split('/'):
1258 if not c or c == '.':
1261 out.append('{%s}%s' % (namespace, c))
1262 return '/'.join(out)
1264 def _extract_smil_formats(self, smil_url, video_id, fatal=True, f4m_params=None, transform_source=None):
1265 smil = self._download_smil(smil_url, video_id, fatal=fatal, transform_source=transform_source)
1271 namespace = self._parse_smil_namespace(smil)
1273 return self._parse_smil_formats(
1274 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
1276 def _extract_smil_info(self, smil_url, video_id, fatal=True, f4m_params=None):
1277 smil = self._download_smil(smil_url, video_id, fatal=fatal)
1280 return self._parse_smil(smil, smil_url, video_id, f4m_params=f4m_params)
1282 def _download_smil(self, smil_url, video_id, fatal=True, transform_source=None):
1283 return self._download_xml(
1284 smil_url, video_id, 'Downloading SMIL file',
1285 'Unable to download SMIL file', fatal=fatal, transform_source=transform_source)
1287 def _parse_smil(self, smil, smil_url, video_id, f4m_params=None):
1288 namespace = self._parse_smil_namespace(smil)
1290 formats = self._parse_smil_formats(
1291 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
1292 subtitles = self._parse_smil_subtitles(smil, namespace=namespace)
1294 video_id = os.path.splitext(url_basename(smil_url))[0]
1298 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
1299 name = meta.attrib.get('name')
1300 content = meta.attrib.get('content')
1301 if not name or not content:
1303 if not title and name == 'title':
1305 elif not description and name in ('description', 'abstract'):
1306 description = content
1307 elif not upload_date and name == 'date':
1308 upload_date = unified_strdate(content)
1311 'id': image.get('type'),
1312 'url': image.get('src'),
1313 'width': int_or_none(image.get('width')),
1314 'height': int_or_none(image.get('height')),
1315 } for image in smil.findall(self._xpath_ns('.//image', namespace)) if image.get('src')]
1319 'title': title or video_id,
1320 'description': description,
1321 'upload_date': upload_date,
1322 'thumbnails': thumbnails,
1324 'subtitles': subtitles,
1327 def _parse_smil_namespace(self, smil):
1328 return self._search_regex(
1329 r'(?i)^{([^}]+)?}smil$', smil.tag, 'namespace', default=None)
1331 def _parse_smil_formats(self, smil, smil_url, video_id, namespace=None, f4m_params=None, transform_rtmp_url=None):
1333 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
1334 b = meta.get('base') or meta.get('httpBase')
1345 media = smil.findall(self._xpath_ns('.//video', namespace)) + smil.findall(self._xpath_ns('.//audio', namespace))
1346 for medium in media:
1347 src = medium.get('src')
1348 if not src or src in srcs:
1352 bitrate = float_or_none(medium.get('system-bitrate') or medium.get('systemBitrate'), 1000)
1353 filesize = int_or_none(medium.get('size') or medium.get('fileSize'))
1354 width = int_or_none(medium.get('width'))
1355 height = int_or_none(medium.get('height'))
1356 proto = medium.get('proto')
1357 ext = medium.get('ext')
1358 src_ext = determine_ext(src)
1359 streamer = medium.get('streamer') or base
1361 if proto == 'rtmp' or streamer.startswith('rtmp'):
1367 'format_id': 'rtmp-%d' % (rtmp_count if bitrate is None else bitrate),
1369 'filesize': filesize,
1373 if transform_rtmp_url:
1374 streamer, src = transform_rtmp_url(streamer, src)
1375 formats[-1].update({
1381 src_url = src if src.startswith('http') else compat_urlparse.urljoin(base, src)
1382 src_url = src_url.strip()
1384 if proto == 'm3u8' or src_ext == 'm3u8':
1385 m3u8_formats = self._extract_m3u8_formats(
1386 src_url, video_id, ext or 'mp4', m3u8_id='hls', fatal=False)
1387 if len(m3u8_formats) == 1:
1389 m3u8_formats[0].update({
1390 'format_id': 'hls-%d' % (m3u8_count if bitrate is None else bitrate),
1395 formats.extend(m3u8_formats)
1398 if src_ext == 'f4m':
1403 'plugin': 'flowplayer-3.2.0.1',
1405 f4m_url += '&' if '?' in f4m_url else '?'
1406 f4m_url += compat_urllib_parse_urlencode(f4m_params)
1407 formats.extend(self._extract_f4m_formats(f4m_url, video_id, f4m_id='hds', fatal=False))
1410 if src_url.startswith('http') and self._is_valid_url(src, video_id):
1414 'ext': ext or src_ext or 'flv',
1415 'format_id': 'http-%d' % (bitrate or http_count),
1417 'filesize': filesize,
1425 def _parse_smil_subtitles(self, smil, namespace=None, subtitles_lang='en'):
1428 for num, textstream in enumerate(smil.findall(self._xpath_ns('.//textstream', namespace))):
1429 src = textstream.get('src')
1430 if not src or src in urls:
1433 ext = textstream.get('ext') or mimetype2ext(textstream.get('type')) or determine_ext(src)
1434 lang = textstream.get('systemLanguage') or textstream.get('systemLanguageName') or textstream.get('lang') or subtitles_lang
1435 subtitles.setdefault(lang, []).append({
1441 def _extract_xspf_playlist(self, playlist_url, playlist_id, fatal=True):
1442 xspf = self._download_xml(
1443 playlist_url, playlist_id, 'Downloading xpsf playlist',
1444 'Unable to download xspf manifest', fatal=fatal)
1447 return self._parse_xspf(xspf, playlist_id)
1449 def _parse_xspf(self, playlist, playlist_id):
1451 'xspf': 'http://xspf.org/ns/0/',
1452 's1': 'http://static.streamone.nl/player/ns/0',
1456 for track in playlist.findall(xpath_with_ns('./xspf:trackList/xspf:track', NS_MAP)):
1458 track, xpath_with_ns('./xspf:title', NS_MAP), 'title', default=playlist_id)
1459 description = xpath_text(
1460 track, xpath_with_ns('./xspf:annotation', NS_MAP), 'description')
1461 thumbnail = xpath_text(
1462 track, xpath_with_ns('./xspf:image', NS_MAP), 'thumbnail')
1463 duration = float_or_none(
1464 xpath_text(track, xpath_with_ns('./xspf:duration', NS_MAP), 'duration'), 1000)
1467 'url': location.text,
1468 'format_id': location.get(xpath_with_ns('s1:label', NS_MAP)),
1469 'width': int_or_none(location.get(xpath_with_ns('s1:width', NS_MAP))),
1470 'height': int_or_none(location.get(xpath_with_ns('s1:height', NS_MAP))),
1471 } for location in track.findall(xpath_with_ns('./xspf:location', NS_MAP))]
1472 self._sort_formats(formats)
1477 'description': description,
1478 'thumbnail': thumbnail,
1479 'duration': duration,
1484 def _extract_mpd_formats(self, mpd_url, video_id, mpd_id=None, note=None, errnote=None, fatal=True, formats_dict={}):
1485 res = self._download_webpage_handle(
1487 note=note or 'Downloading MPD manifest',
1488 errnote=errnote or 'Failed to download MPD manifest',
1493 mpd_base_url = re.match(r'https?://.+/', urlh.geturl()).group()
1495 return self._parse_mpd_formats(
1496 compat_etree_fromstring(mpd.encode('utf-8')), mpd_id, mpd_base_url, formats_dict=formats_dict)
1498 def _parse_mpd_formats(self, mpd_doc, mpd_id=None, mpd_base_url='', formats_dict={}):
1500 Parse formats from MPD manifest.
1502 1. MPEG-DASH Standard, ISO/IEC 23009-1:2014(E),
1503 http://standards.iso.org/ittf/PubliclyAvailableStandards/c065274_ISO_IEC_23009-1_2014.zip
1504 2. https://en.wikipedia.org/wiki/Dynamic_Adaptive_Streaming_over_HTTP
1506 if mpd_doc.get('type') == 'dynamic':
1509 namespace = self._search_regex(r'(?i)^{([^}]+)?}MPD$', mpd_doc.tag, 'namespace', default=None)
1512 return self._xpath_ns(path, namespace)
1514 def is_drm_protected(element):
1515 return element.find(_add_ns('ContentProtection')) is not None
1517 def extract_multisegment_info(element, ms_parent_info):
1518 ms_info = ms_parent_info.copy()
1519 segment_list = element.find(_add_ns('SegmentList'))
1520 if segment_list is not None:
1521 segment_urls_e = segment_list.findall(_add_ns('SegmentURL'))
1523 ms_info['segment_urls'] = [segment.attrib['media'] for segment in segment_urls_e]
1524 initialization = segment_list.find(_add_ns('Initialization'))
1525 if initialization is not None:
1526 ms_info['initialization_url'] = initialization.attrib['sourceURL']
1528 segment_template = element.find(_add_ns('SegmentTemplate'))
1529 if segment_template is not None:
1530 start_number = segment_template.get('startNumber')
1532 ms_info['start_number'] = int(start_number)
1533 segment_timeline = segment_template.find(_add_ns('SegmentTimeline'))
1534 if segment_timeline is not None:
1535 s_e = segment_timeline.findall(_add_ns('S'))
1537 ms_info['total_number'] = 0
1540 r = int(s.get('r', 0))
1541 ms_info['total_number'] += 1 + r
1542 ms_info['s'].append({
1543 't': int(s.get('t', 0)),
1544 # @d is mandatory (see [1, 5.3.9.6.2, Table 17, page 60])
1545 'd': int(s.attrib['d']),
1549 timescale = segment_template.get('timescale')
1551 ms_info['timescale'] = int(timescale)
1552 segment_duration = segment_template.get('duration')
1553 if segment_duration:
1554 ms_info['segment_duration'] = int(segment_duration)
1555 media_template = segment_template.get('media')
1557 ms_info['media_template'] = media_template
1558 initialization = segment_template.get('initialization')
1560 ms_info['initialization_url'] = initialization
1562 initialization = segment_template.find(_add_ns('Initialization'))
1563 if initialization is not None:
1564 ms_info['initialization_url'] = initialization.attrib['sourceURL']
1567 mpd_duration = parse_duration(mpd_doc.get('mediaPresentationDuration'))
1569 for period in mpd_doc.findall(_add_ns('Period')):
1570 period_duration = parse_duration(period.get('duration')) or mpd_duration
1571 period_ms_info = extract_multisegment_info(period, {
1575 for adaptation_set in period.findall(_add_ns('AdaptationSet')):
1576 if is_drm_protected(adaptation_set):
1578 adaption_set_ms_info = extract_multisegment_info(adaptation_set, period_ms_info)
1579 for representation in adaptation_set.findall(_add_ns('Representation')):
1580 if is_drm_protected(representation):
1582 representation_attrib = adaptation_set.attrib.copy()
1583 representation_attrib.update(representation.attrib)
1584 # According to [1, 5.3.7.2, Table 9, page 41], @mimeType is mandatory
1585 mime_type = representation_attrib['mimeType']
1586 content_type = mime_type.split('/')[0]
1587 if content_type == 'text':
1588 # TODO implement WebVTT downloading
1590 elif content_type == 'video' or content_type == 'audio':
1592 for element in (representation, adaptation_set, period, mpd_doc):
1593 base_url_e = element.find(_add_ns('BaseURL'))
1594 if base_url_e is not None:
1595 base_url = base_url_e.text + base_url
1596 if re.match(r'^https?://', base_url):
1598 if mpd_base_url and not re.match(r'^https?://', base_url):
1599 if not mpd_base_url.endswith('/') and not base_url.startswith('/'):
1601 base_url = mpd_base_url + base_url
1602 representation_id = representation_attrib.get('id')
1603 lang = representation_attrib.get('lang')
1604 url_el = representation.find(_add_ns('BaseURL'))
1605 filesize = int_or_none(url_el.attrib.get('{http://youtube.com/yt/2012/10/10}contentLength') if url_el is not None else None)
1607 'format_id': '%s-%s' % (mpd_id, representation_id) if mpd_id else representation_id,
1609 'ext': mimetype2ext(mime_type),
1610 'width': int_or_none(representation_attrib.get('width')),
1611 'height': int_or_none(representation_attrib.get('height')),
1612 'tbr': int_or_none(representation_attrib.get('bandwidth'), 1000),
1613 'asr': int_or_none(representation_attrib.get('audioSamplingRate')),
1614 'fps': int_or_none(representation_attrib.get('frameRate')),
1615 'vcodec': 'none' if content_type == 'audio' else representation_attrib.get('codecs'),
1616 'acodec': 'none' if content_type == 'video' else representation_attrib.get('codecs'),
1617 'language': lang if lang not in ('mul', 'und', 'zxx', 'mis') else None,
1618 'format_note': 'DASH %s' % content_type,
1619 'filesize': filesize,
1621 representation_ms_info = extract_multisegment_info(representation, adaption_set_ms_info)
1622 if 'segment_urls' not in representation_ms_info and 'media_template' in representation_ms_info:
1623 if 'total_number' not in representation_ms_info and 'segment_duration':
1624 segment_duration = float(representation_ms_info['segment_duration']) / float(representation_ms_info['timescale'])
1625 representation_ms_info['total_number'] = int(math.ceil(float(period_duration) / segment_duration))
1626 media_template = representation_ms_info['media_template']
1627 media_template = media_template.replace('$RepresentationID$', representation_id)
1628 media_template = re.sub(r'\$(Number|Bandwidth|Time)\$', r'%(\1)d', media_template)
1629 media_template = re.sub(r'\$(Number|Bandwidth|Time)%([^$]+)\$', r'%(\1)\2', media_template)
1630 media_template.replace('$$', '$')
1632 # As per [1, 5.3.9.4.4, Table 16, page 55] $Number$ and $Time$
1633 # can't be used at the same time
1634 if '%(Number' in media_template:
1635 representation_ms_info['segment_urls'] = [
1637 'Number': segment_number,
1638 'Bandwidth': representation_attrib.get('bandwidth'),
1640 for segment_number in range(
1641 representation_ms_info['start_number'],
1642 representation_ms_info['total_number'] + representation_ms_info['start_number'])]
1644 representation_ms_info['segment_urls'] = []
1647 def add_segment_url():
1648 representation_ms_info['segment_urls'].append(
1650 'Time': segment_time,
1651 'Bandwidth': representation_attrib.get('bandwidth'),
1655 for num, s in enumerate(representation_ms_info['s']):
1656 segment_time = s.get('t') or segment_time
1658 for r in range(s.get('r', 0)):
1659 segment_time += s['d']
1661 segment_time += s['d']
1662 if 'segment_urls' in representation_ms_info:
1664 'segment_urls': representation_ms_info['segment_urls'],
1665 'protocol': 'http_dash_segments',
1667 if 'initialization_url' in representation_ms_info:
1668 initialization_url = representation_ms_info['initialization_url'].replace('$RepresentationID$', representation_id)
1670 'initialization_url': initialization_url,
1672 if not f.get('url'):
1673 f['url'] = initialization_url
1675 existing_format = next(
1676 fo for fo in formats
1677 if fo['format_id'] == representation_id)
1678 except StopIteration:
1679 full_info = formats_dict.get(representation_id, {}).copy()
1681 formats.append(full_info)
1683 existing_format.update(f)
1685 self.report_warning('Unknown MIME type %s in DASH manifest' % mime_type)
1688 def _parse_html5_media_entries(self, base_url, webpage):
1689 def absolute_url(video_url):
1690 return compat_urlparse.urljoin(base_url, video_url)
1692 def parse_content_type(content_type):
1693 if not content_type:
1695 ctr = re.search(r'(?P<mimetype>[^/]+/[^;]+)(?:;\s*codecs="?(?P<codecs>[^"]+))?', content_type)
1697 mimetype, codecs = ctr.groups()
1698 f = parse_codecs(codecs)
1699 f['ext'] = mimetype2ext(mimetype)
1704 for media_tag, media_type, media_content in re.findall(r'(?s)(<(?P<tag>video|audio)[^>]*>)(.*?)</(?P=tag)>', webpage):
1709 media_attributes = extract_attributes(media_tag)
1710 src = media_attributes.get('src')
1712 media_info['formats'].append({
1713 'url': absolute_url(src),
1714 'vcodec': 'none' if media_type == 'audio' else None,
1716 media_info['thumbnail'] = media_attributes.get('poster')
1718 for source_tag in re.findall(r'<source[^>]+>', media_content):
1719 source_attributes = extract_attributes(source_tag)
1720 src = source_attributes.get('src')
1723 f = parse_content_type(source_attributes.get('type'))
1725 'url': absolute_url(src),
1726 'vcodec': 'none' if media_type == 'audio' else None,
1728 media_info['formats'].append(f)
1729 for track_tag in re.findall(r'<track[^>]+>', media_content):
1730 track_attributes = extract_attributes(track_tag)
1731 kind = track_attributes.get('kind')
1732 if not kind or kind == 'subtitles':
1733 src = track_attributes.get('src')
1736 lang = track_attributes.get('srclang') or track_attributes.get('lang') or track_attributes.get('label')
1737 media_info['subtitles'].setdefault(lang, []).append({
1738 'url': absolute_url(src),
1740 if media_info['formats']:
1741 entries.append(media_info)
1744 def _live_title(self, name):
1745 """ Generate the title for a live video """
1746 now = datetime.datetime.now()
1747 now_str = now.strftime('%Y-%m-%d %H:%M')
1748 return name + ' ' + now_str
1750 def _int(self, v, name, fatal=False, **kwargs):
1751 res = int_or_none(v, **kwargs)
1752 if 'get_attr' in kwargs:
1753 print(getattr(v, kwargs['get_attr']))
1755 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
1757 raise ExtractorError(msg)
1759 self._downloader.report_warning(msg)
1762 def _float(self, v, name, fatal=False, **kwargs):
1763 res = float_or_none(v, **kwargs)
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 _set_cookie(self, domain, name, value, expire_time=None):
1773 cookie = compat_cookiejar.Cookie(
1774 0, name, value, None, None, domain, None,
1775 None, '/', True, False, expire_time, '', None, None, None)
1776 self._downloader.cookiejar.set_cookie(cookie)
1778 def _get_cookies(self, url):
1779 """ Return a compat_cookies.SimpleCookie with the cookies for the url """
1780 req = sanitized_Request(url)
1781 self._downloader.cookiejar.add_cookie_header(req)
1782 return compat_cookies.SimpleCookie(req.get_header('Cookie'))
1784 def get_testcases(self, include_onlymatching=False):
1785 t = getattr(self, '_TEST', None)
1787 assert not hasattr(self, '_TESTS'), \
1788 '%s has _TEST and _TESTS' % type(self).__name__
1791 tests = getattr(self, '_TESTS', [])
1793 if not include_onlymatching and t.get('only_matching', False):
1795 t['name'] = type(self).__name__[:-len('IE')]
1798 def is_suitable(self, age_limit):
1799 """ Test whether the extractor is generally suitable for the given
1800 age limit (i.e. pornographic sites are not, all others usually are) """
1802 any_restricted = False
1803 for tc in self.get_testcases(include_onlymatching=False):
1804 if tc.get('playlist', []):
1805 tc = tc['playlist'][0]
1806 is_restricted = age_restricted(
1807 tc.get('info_dict', {}).get('age_limit'), age_limit)
1808 if not is_restricted:
1810 any_restricted = any_restricted or is_restricted
1811 return not any_restricted
1813 def extract_subtitles(self, *args, **kwargs):
1814 if (self._downloader.params.get('writesubtitles', False) or
1815 self._downloader.params.get('listsubtitles')):
1816 return self._get_subtitles(*args, **kwargs)
1819 def _get_subtitles(self, *args, **kwargs):
1820 raise NotImplementedError('This method must be implemented by subclasses')
1823 def _merge_subtitle_items(subtitle_list1, subtitle_list2):
1824 """ Merge subtitle items for one language. Items with duplicated URLs
1825 will be dropped. """
1826 list1_urls = set([item['url'] for item in subtitle_list1])
1827 ret = list(subtitle_list1)
1828 ret.extend([item for item in subtitle_list2 if item['url'] not in list1_urls])
1832 def _merge_subtitles(cls, subtitle_dict1, subtitle_dict2):
1833 """ Merge two subtitle dictionaries, language by language. """
1834 ret = dict(subtitle_dict1)
1835 for lang in subtitle_dict2:
1836 ret[lang] = cls._merge_subtitle_items(subtitle_dict1.get(lang, []), subtitle_dict2[lang])
1839 def extract_automatic_captions(self, *args, **kwargs):
1840 if (self._downloader.params.get('writeautomaticsub', False) or
1841 self._downloader.params.get('listsubtitles')):
1842 return self._get_automatic_captions(*args, **kwargs)
1845 def _get_automatic_captions(self, *args, **kwargs):
1846 raise NotImplementedError('This method must be implemented by subclasses')
1848 def mark_watched(self, *args, **kwargs):
1849 if (self._downloader.params.get('mark_watched', False) and
1850 (self._get_login_info()[0] is not None or
1851 self._downloader.params.get('cookiefile') is not None)):
1852 self._mark_watched(*args, **kwargs)
1854 def _mark_watched(self, *args, **kwargs):
1855 raise NotImplementedError('This method must be implemented by subclasses')
1857 def geo_verification_headers(self):
1859 geo_verification_proxy = self._downloader.params.get('geo_verification_proxy')
1860 if geo_verification_proxy:
1861 headers['Ytdl-request-proxy'] = geo_verification_proxy
1865 class SearchInfoExtractor(InfoExtractor):
1867 Base class for paged search queries extractors.
1868 They accept URLs in the format _SEARCH_KEY(|all|[0-9]):{query}
1869 Instances should define _SEARCH_KEY and _MAX_RESULTS.
1873 def _make_valid_url(cls):
1874 return r'%s(?P<prefix>|[1-9][0-9]*|all):(?P<query>[\s\S]+)' % cls._SEARCH_KEY
1877 def suitable(cls, url):
1878 return re.match(cls._make_valid_url(), url) is not None
1880 def _real_extract(self, query):
1881 mobj = re.match(self._make_valid_url(), query)
1883 raise ExtractorError('Invalid search query "%s"' % query)
1885 prefix = mobj.group('prefix')
1886 query = mobj.group('query')
1888 return self._get_n_results(query, 1)
1889 elif prefix == 'all':
1890 return self._get_n_results(query, self._MAX_RESULTS)
1894 raise ExtractorError('invalid download number %s for query "%s"' % (n, query))
1895 elif n > self._MAX_RESULTS:
1896 self._downloader.report_warning('%s returns max %i results (you requested %i)' % (self._SEARCH_KEY, self._MAX_RESULTS, n))
1897 n = self._MAX_RESULTS
1898 return self._get_n_results(query, n)
1900 def _get_n_results(self, query, n):
1901 """Get a specified number of results for a query"""
1902 raise NotImplementedError('This method must be implemented by subclasses')
1905 def SEARCH_KEY(self):
1906 return self._SEARCH_KEY