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,
61 class InfoExtractor(object):
62 """Information Extractor class.
64 Information extractors are the classes that, given a URL, extract
65 information about the video (or videos) the URL refers to. This
66 information includes the real video URL, the video title, author and
67 others. The information is stored in a dictionary which is then
68 passed to the YoutubeDL. The YoutubeDL processes this
69 information possibly downloading the video to the file system, among
70 other possible outcomes.
72 The type field determines the type of the result.
73 By far the most common value (and the default if _type is missing) is
74 "video", which indicates a single video.
76 For a video, the dictionaries must include the following fields:
79 title: Video title, unescaped.
81 Additionally, it must contain either a formats entry or a url one:
83 formats: A list of dictionaries for each format available, ordered
84 from worst to best quality.
87 * url Mandatory. The URL of the video file
88 * ext Will be calculated from URL if missing
89 * format A human-readable description of the format
90 ("mp4 container with h264/opus").
91 Calculated from the format_id, width, height.
92 and format_note fields if missing.
93 * format_id A short description of the format
94 ("mp4_h264_opus" or "19").
95 Technically optional, but strongly recommended.
96 * format_note Additional info about the format
97 ("3D" or "DASH video")
98 * width Width of the video, if known
99 * height Height of the video, if known
100 * resolution Textual description of width and height
101 * tbr Average bitrate of audio and video in KBit/s
102 * abr Average audio bitrate in KBit/s
103 * acodec Name of the audio codec in use
104 * asr Audio sampling rate in Hertz
105 * vbr Average video bitrate in KBit/s
107 * vcodec Name of the video codec in use
108 * container Name of the container format
109 * filesize The number of bytes, if known in advance
110 * filesize_approx An estimate for the number of bytes
111 * player_url SWF Player URL (used for rtmpdump).
112 * protocol The protocol that will be used for the actual
113 download, lower-case.
114 "http", "https", "rtsp", "rtmp", "rtmpe",
115 "m3u8", "m3u8_native" or "http_dash_segments".
116 * preference Order number of this format. If this field is
117 present and not None, the formats get sorted
118 by this field, regardless of all other values.
119 -1 for default (order by other properties),
120 -2 or smaller for less than default.
121 < -1000 to hide the format (if there is
122 another one which is strictly better)
123 * language Language code, e.g. "de" or "en-US".
124 * language_preference Is this in the language mentioned in
126 10 if it's what the URL is about,
127 -1 for default (don't know),
128 -10 otherwise, other values reserved for now.
129 * quality Order number of the video quality of this
130 format, irrespective of the file format.
131 -1 for default (order by other properties),
132 -2 or smaller for less than default.
133 * source_preference Order number for this video source
134 (quality takes higher priority)
135 -1 for default (order by other properties),
136 -2 or smaller for less than default.
137 * http_headers A dictionary of additional HTTP headers
138 to add to the request.
139 * stretched_ratio If given and not 1, indicates that the
140 video's pixels are not square.
141 width : height ratio as float.
142 * no_resume The server does not support resuming the
143 (HTTP or RTMP) download. Boolean.
145 url: Final video URL.
146 ext: Video filename extension.
147 format: The video format, defaults to ext (used for --get-format)
148 player_url: SWF Player URL (used for rtmpdump).
150 The following fields are optional:
152 alt_title: A secondary title of the video.
153 display_id An alternative identifier for the video, not necessarily
154 unique, but available before title. Typically, id is
155 something like "4234987", title "Dancing naked mole rats",
156 and display_id "dancing-naked-mole-rats"
157 thumbnails: A list of dictionaries, with the following entries:
158 * "id" (optional, string) - Thumbnail format ID
160 * "preference" (optional, int) - quality of the image
161 * "width" (optional, int)
162 * "height" (optional, int)
163 * "resolution" (optional, string "{width}x{height"},
165 * "filesize" (optional, int)
166 thumbnail: Full URL to a video thumbnail image.
167 description: Full video description.
168 uploader: Full name of the video uploader.
169 license: License name the video is licensed under.
170 creator: The creator of the video.
171 release_date: The date (YYYYMMDD) when the video was released.
172 timestamp: UNIX timestamp of the moment the video became available.
173 upload_date: Video upload date (YYYYMMDD).
174 If not explicitly set, calculated from timestamp.
175 uploader_id: Nickname or id of the video uploader.
176 uploader_url: Full URL to a personal webpage of the video uploader.
177 location: Physical location where the video was filmed.
178 subtitles: The available subtitles as a dictionary in the format
179 {language: subformats}. "subformats" is a list sorted from
180 lower to higher preference, each element is a dictionary
181 with the "ext" entry and one of:
182 * "data": The subtitles file contents
183 * "url": A URL pointing to the subtitles file
184 "ext" will be calculated from URL if missing
185 automatic_captions: Like 'subtitles', used by the YoutubeIE for
186 automatically generated captions
187 duration: Length of the video in seconds, as an integer or float.
188 view_count: How many users have watched the video on the platform.
189 like_count: Number of positive ratings of the video
190 dislike_count: Number of negative ratings of the video
191 repost_count: Number of reposts of the video
192 average_rating: Average rating give by users, the scale used depends on the webpage
193 comment_count: Number of comments on the video
194 comments: A list of comments, each with one or more of the following
195 properties (all but one of text or html optional):
196 * "author" - human-readable name of the comment author
197 * "author_id" - user ID of the comment author
199 * "html" - Comment as HTML
200 * "text" - Plain text of the comment
201 * "timestamp" - UNIX timestamp of comment
202 * "parent" - ID of the comment this one is replying to.
203 Set to "root" to indicate that this is a
204 comment to the original video.
205 age_limit: Age restriction for the video, as an integer (years)
206 webpage_url: The URL to the video webpage, if given to youtube-dl it
207 should allow to get the same result again. (It will be set
208 by YoutubeDL if it's missing)
209 categories: A list of categories that the video falls in, for example
211 tags: A list of tags assigned to the video, e.g. ["sweden", "pop music"]
212 is_live: True, False, or None (=unknown). Whether this video is a
213 live stream that goes on instead of a fixed-length video.
214 start_time: Time in seconds where the reproduction should start, as
215 specified in the URL.
216 end_time: Time in seconds where the reproduction should end, as
217 specified in the URL.
219 The following fields should only be used when the video belongs to some logical
222 chapter: Name or title of the chapter the video belongs to.
223 chapter_number: Number of the chapter the video belongs to, as an integer.
224 chapter_id: Id of the chapter the video belongs to, as a unicode string.
226 The following fields should only be used when the video is an episode of some
229 series: Title of the series or programme the video episode belongs to.
230 season: Title of the season the video episode belongs to.
231 season_number: Number of the season the video episode belongs to, as an integer.
232 season_id: Id of the season the video episode belongs to, as a unicode string.
233 episode: Title of the video episode. Unlike mandatory video title field,
234 this field should denote the exact title of the video episode
235 without any kind of decoration.
236 episode_number: Number of the video episode within a season, as an integer.
237 episode_id: Id of the video episode, as a unicode string.
239 The following fields should only be used when the media is a track or a part of
242 track: Title of the track.
243 track_number: Number of the track within an album or a disc, as an integer.
244 track_id: Id of the track (useful in case of custom indexing, e.g. 6.iii),
246 artist: Artist(s) of the track.
247 genre: Genre(s) of the track.
248 album: Title of the album the track belongs to.
249 album_type: Type of the album (e.g. "Demo", "Full-length", "Split", "Compilation", etc).
250 album_artist: List of all artists appeared on the album (e.g.
251 "Ash Borer / Fell Voices" or "Various Artists", useful for splits
253 disc_number: Number of the disc or other physical medium the track belongs to,
255 release_year: Year (YYYY) when the album was released.
257 Unless mentioned otherwise, the fields should be Unicode strings.
259 Unless mentioned otherwise, None is equivalent to absence of information.
262 _type "playlist" indicates multiple videos.
263 There must be a key "entries", which is a list, an iterable, or a PagedList
264 object, each element of which is a valid dictionary by this specification.
266 Additionally, playlists can have "title", "description" and "id" attributes
267 with the same semantics as videos (see above).
270 _type "multi_video" indicates that there are multiple videos that
271 form a single show, for examples multiple acts of an opera or TV episode.
272 It must have an entries key like a playlist and contain all the keys
273 required for a video at the same time.
276 _type "url" indicates that the video must be extracted from another
277 location, possibly by a different extractor. Its only required key is:
278 "url" - the next URL to extract.
279 The key "ie_key" can be set to the class name (minus the trailing "IE",
280 e.g. "Youtube") if the extractor class is known in advance.
281 Additionally, the dictionary may have any properties of the resolved entity
282 known in advance, for example "title" if the title of the referred video is
286 _type "url_transparent" entities have the same specification as "url", but
287 indicate that the given additional information is more precise than the one
288 associated with the resolved URL.
289 This is useful when a site employs a video service that hosts the video and
290 its technical metadata, but that video service does not embed a useful
291 title, description etc.
294 Subclasses of this one should re-define the _real_initialize() and
295 _real_extract() methods and define a _VALID_URL regexp.
296 Probably, they should also be added to the list of extractors.
298 Finally, the _WORKING attribute should be set to False for broken IEs
299 in order to warn the users and skip the tests.
306 def __init__(self, downloader=None):
307 """Constructor. Receives an optional downloader."""
309 self.set_downloader(downloader)
312 def suitable(cls, url):
313 """Receives a URL and returns True if suitable for this IE."""
315 # This does not use has/getattr intentionally - we want to know whether
316 # we have cached the regexp for *this* class, whereas getattr would also
317 # match the superclass
318 if '_VALID_URL_RE' not in cls.__dict__:
319 cls._VALID_URL_RE = re.compile(cls._VALID_URL)
320 return cls._VALID_URL_RE.match(url) is not None
323 def _match_id(cls, url):
324 if '_VALID_URL_RE' not in cls.__dict__:
325 cls._VALID_URL_RE = re.compile(cls._VALID_URL)
326 m = cls._VALID_URL_RE.match(url)
332 """Getter method for _WORKING."""
335 def initialize(self):
336 """Initializes an instance (authentication, etc)."""
338 self._real_initialize()
341 def extract(self, url):
342 """Extracts URL information and returns it in list of dicts."""
345 return self._real_extract(url)
346 except ExtractorError:
348 except compat_http_client.IncompleteRead as e:
349 raise ExtractorError('A network error has occurred.', cause=e, expected=True)
350 except (KeyError, StopIteration) as e:
351 raise ExtractorError('An extractor error has occurred.', cause=e)
353 def set_downloader(self, downloader):
354 """Sets the downloader for this IE."""
355 self._downloader = downloader
357 def _real_initialize(self):
358 """Real initialization process. Redefine in subclasses."""
361 def _real_extract(self, url):
362 """Real extraction process. Redefine in subclasses."""
367 """A string for getting the InfoExtractor with get_info_extractor"""
368 return compat_str(cls.__name__[:-2])
372 return compat_str(type(self).__name__[:-2])
374 def _request_webpage(self, url_or_request, video_id, note=None, errnote=None, fatal=True, data=None, headers={}, query={}):
375 """ Returns the response handle """
377 self.report_download_webpage(video_id)
378 elif note is not False:
380 self.to_screen('%s' % (note,))
382 self.to_screen('%s: %s' % (video_id, note))
383 if isinstance(url_or_request, compat_urllib_request.Request):
384 url_or_request = update_Request(
385 url_or_request, data=data, headers=headers, query=query)
388 url_or_request = update_url_query(url_or_request, query)
389 if data is not None or headers:
390 url_or_request = sanitized_Request(url_or_request, data, headers)
392 return self._downloader.urlopen(url_or_request)
393 except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err:
397 errnote = 'Unable to download webpage'
399 errmsg = '%s: %s' % (errnote, error_to_compat_str(err))
401 raise ExtractorError(errmsg, sys.exc_info()[2], cause=err)
403 self._downloader.report_warning(errmsg)
406 def _download_webpage_handle(self, url_or_request, video_id, note=None, errnote=None, fatal=True, encoding=None, data=None, headers={}, query={}):
407 """ Returns a tuple (page content as string, URL handle) """
408 # Strip hashes from the URL (#1038)
409 if isinstance(url_or_request, (compat_str, str)):
410 url_or_request = url_or_request.partition('#')[0]
412 urlh = self._request_webpage(url_or_request, video_id, note, errnote, fatal, data=data, headers=headers, query=query)
416 content = self._webpage_read_content(urlh, url_or_request, video_id, note, errnote, fatal, encoding=encoding)
417 return (content, urlh)
420 def _guess_encoding_from_content(content_type, webpage_bytes):
421 m = re.match(r'[a-zA-Z0-9_.-]+/[a-zA-Z0-9_.-]+\s*;\s*charset=(.+)', content_type)
423 encoding = m.group(1)
425 m = re.search(br'<meta[^>]+charset=[\'"]?([^\'")]+)[ /\'">]',
426 webpage_bytes[:1024])
428 encoding = m.group(1).decode('ascii')
429 elif webpage_bytes.startswith(b'\xff\xfe'):
436 def _webpage_read_content(self, urlh, url_or_request, video_id, note=None, errnote=None, fatal=True, prefix=None, encoding=None):
437 content_type = urlh.headers.get('Content-Type', '')
438 webpage_bytes = urlh.read()
439 if prefix is not None:
440 webpage_bytes = prefix + webpage_bytes
442 encoding = self._guess_encoding_from_content(content_type, webpage_bytes)
443 if self._downloader.params.get('dump_intermediate_pages', False):
445 url = url_or_request.get_full_url()
446 except AttributeError:
448 self.to_screen('Dumping request to ' + url)
449 dump = base64.b64encode(webpage_bytes).decode('ascii')
450 self._downloader.to_screen(dump)
451 if self._downloader.params.get('write_pages', False):
453 url = url_or_request.get_full_url()
454 except AttributeError:
456 basen = '%s_%s' % (video_id, url)
458 h = '___' + hashlib.md5(basen.encode('utf-8')).hexdigest()
459 basen = basen[:240 - len(h)] + h
460 raw_filename = basen + '.dump'
461 filename = sanitize_filename(raw_filename, restricted=True)
462 self.to_screen('Saving request to ' + filename)
463 # Working around MAX_PATH limitation on Windows (see
464 # http://msdn.microsoft.com/en-us/library/windows/desktop/aa365247(v=vs.85).aspx)
465 if compat_os_name == 'nt':
466 absfilepath = os.path.abspath(filename)
467 if len(absfilepath) > 259:
468 filename = '\\\\?\\' + absfilepath
469 with open(filename, 'wb') as outf:
470 outf.write(webpage_bytes)
473 content = webpage_bytes.decode(encoding, 'replace')
475 content = webpage_bytes.decode('utf-8', 'replace')
477 if ('<title>Access to this site is blocked</title>' in content and
478 'Websense' in content[:512]):
479 msg = 'Access to this webpage has been blocked by Websense filtering software in your network.'
480 blocked_iframe = self._html_search_regex(
481 r'<iframe src="([^"]+)"', content,
482 'Websense information URL', default=None)
484 msg += ' Visit %s for more details' % blocked_iframe
485 raise ExtractorError(msg, expected=True)
486 if '<title>The URL you requested has been blocked</title>' in content[:512]:
488 'Access to this webpage has been blocked by Indian censorship. '
489 'Use a VPN or proxy server (with --proxy) to route around it.')
490 block_msg = self._html_search_regex(
491 r'</h1><p>(.*?)</p>',
492 content, 'block message', default=None)
494 msg += ' (Message: "%s")' % block_msg.replace('\n', ' ')
495 raise ExtractorError(msg, expected=True)
499 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={}):
500 """ Returns the data of the page as a string """
503 while success is False:
505 res = self._download_webpage_handle(url_or_request, video_id, note, errnote, fatal, encoding=encoding, data=data, headers=headers, query=query)
507 except compat_http_client.IncompleteRead as e:
509 if try_count >= tries:
511 self._sleep(timeout, video_id)
518 def _download_xml(self, url_or_request, video_id,
519 note='Downloading XML', errnote='Unable to download XML',
520 transform_source=None, fatal=True, encoding=None, data=None, headers={}, query={}):
521 """Return the xml as an xml.etree.ElementTree.Element"""
522 xml_string = self._download_webpage(
523 url_or_request, video_id, note, errnote, fatal=fatal, encoding=encoding, data=data, headers=headers, query=query)
524 if xml_string is False:
527 xml_string = transform_source(xml_string)
528 return compat_etree_fromstring(xml_string.encode('utf-8'))
530 def _download_json(self, url_or_request, video_id,
531 note='Downloading JSON metadata',
532 errnote='Unable to download JSON metadata',
533 transform_source=None,
534 fatal=True, encoding=None, data=None, headers={}, query={}):
535 json_string = self._download_webpage(
536 url_or_request, video_id, note, errnote, fatal=fatal,
537 encoding=encoding, data=data, headers=headers, query=query)
538 if (not fatal) and json_string is False:
540 return self._parse_json(
541 json_string, video_id, transform_source=transform_source, fatal=fatal)
543 def _parse_json(self, json_string, video_id, transform_source=None, fatal=True):
545 json_string = transform_source(json_string)
547 return json.loads(json_string)
548 except ValueError as ve:
549 errmsg = '%s: Failed to parse JSON ' % video_id
551 raise ExtractorError(errmsg, cause=ve)
553 self.report_warning(errmsg + str(ve))
555 def report_warning(self, msg, video_id=None):
556 idstr = '' if video_id is None else '%s: ' % video_id
557 self._downloader.report_warning(
558 '[%s] %s%s' % (self.IE_NAME, idstr, msg))
560 def to_screen(self, msg):
561 """Print msg to screen, prefixing it with '[ie_name]'"""
562 self._downloader.to_screen('[%s] %s' % (self.IE_NAME, msg))
564 def report_extraction(self, id_or_name):
565 """Report information extraction."""
566 self.to_screen('%s: Extracting information' % id_or_name)
568 def report_download_webpage(self, video_id):
569 """Report webpage download."""
570 self.to_screen('%s: Downloading webpage' % video_id)
572 def report_age_confirmation(self):
573 """Report attempt to confirm age."""
574 self.to_screen('Confirming age')
576 def report_login(self):
577 """Report attempt to log in."""
578 self.to_screen('Logging in')
581 def raise_login_required(msg='This video is only available for registered users'):
582 raise ExtractorError(
583 '%s. Use --username and --password or --netrc to provide account credentials.' % msg,
587 def raise_geo_restricted(msg='This video is not available from your location due to geo restriction'):
588 raise ExtractorError(
589 '%s. You might want to use --proxy to workaround.' % msg,
592 # Methods for following #608
594 def url_result(url, ie=None, video_id=None, video_title=None):
595 """Returns a URL that points to a page that should be processed"""
596 # TODO: ie should be the class used for getting the info
597 video_info = {'_type': 'url',
600 if video_id is not None:
601 video_info['id'] = video_id
602 if video_title is not None:
603 video_info['title'] = video_title
607 def playlist_result(entries, playlist_id=None, playlist_title=None, playlist_description=None):
608 """Returns a playlist"""
609 video_info = {'_type': 'playlist',
612 video_info['id'] = playlist_id
614 video_info['title'] = playlist_title
615 if playlist_description:
616 video_info['description'] = playlist_description
619 def _search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
621 Perform a regex search on the given string, using a single or a list of
622 patterns returning the first matching group.
623 In case of failure return a default value or raise a WARNING or a
624 RegexNotFoundError, depending on fatal, specifying the field name.
626 if isinstance(pattern, (str, compat_str, compiled_regex_type)):
627 mobj = re.search(pattern, string, flags)
630 mobj = re.search(p, string, flags)
634 if not self._downloader.params.get('no_color') and compat_os_name != 'nt' and sys.stderr.isatty():
635 _name = '\033[0;34m%s\033[0m' % name
641 # return the first matching group
642 return next(g for g in mobj.groups() if g is not None)
644 return mobj.group(group)
645 elif default is not NO_DEFAULT:
648 raise RegexNotFoundError('Unable to extract %s' % _name)
650 self._downloader.report_warning('unable to extract %s' % _name + bug_reports_message())
653 def _html_search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
655 Like _search_regex, but strips HTML tags and unescapes entities.
657 res = self._search_regex(pattern, string, name, default, fatal, flags, group)
659 return clean_html(res).strip()
663 def _get_login_info(self):
665 Get the login info as (username, password)
666 It will look in the netrc file using the _NETRC_MACHINE value
667 If there's no info available, return (None, None)
669 if self._downloader is None:
674 downloader_params = self._downloader.params
676 # Attempt to use provided username and password or .netrc data
677 if downloader_params.get('username') is not None:
678 username = downloader_params['username']
679 password = downloader_params['password']
680 elif downloader_params.get('usenetrc', False):
682 info = netrc.netrc().authenticators(self._NETRC_MACHINE)
687 raise netrc.NetrcParseError('No authenticators for %s' % self._NETRC_MACHINE)
688 except (IOError, netrc.NetrcParseError) as err:
689 self._downloader.report_warning('parsing .netrc: %s' % error_to_compat_str(err))
691 return (username, password)
693 def _get_tfa_info(self, note='two-factor verification code'):
695 Get the two-factor authentication info
696 TODO - asking the user will be required for sms/phone verify
697 currently just uses the command line option
698 If there's no info available, return None
700 if self._downloader is None:
702 downloader_params = self._downloader.params
704 if downloader_params.get('twofactor') is not None:
705 return downloader_params['twofactor']
707 return compat_getpass('Type %s and press [Return]: ' % note)
709 # Helper functions for extracting OpenGraph info
711 def _og_regexes(prop):
712 content_re = r'content=(?:"([^"]+?)"|\'([^\']+?)\'|\s*([^\s"\'=<>`]+?))'
713 property_re = (r'(?:name|property)=(?:\'og:%(prop)s\'|"og:%(prop)s"|\s*og:%(prop)s\b)'
714 % {'prop': re.escape(prop)})
715 template = r'<meta[^>]+?%s[^>]+?%s'
717 template % (property_re, content_re),
718 template % (content_re, property_re),
722 def _meta_regex(prop):
723 return r'''(?isx)<meta
724 (?=[^>]+(?:itemprop|name|property|id|http-equiv)=(["\']?)%s\1)
725 [^>]+?content=(["\'])(?P<content>.*?)\2''' % re.escape(prop)
727 def _og_search_property(self, prop, html, name=None, **kargs):
729 name = 'OpenGraph %s' % prop
730 escaped = self._search_regex(self._og_regexes(prop), html, name, flags=re.DOTALL, **kargs)
733 return unescapeHTML(escaped)
735 def _og_search_thumbnail(self, html, **kargs):
736 return self._og_search_property('image', html, 'thumbnail URL', fatal=False, **kargs)
738 def _og_search_description(self, html, **kargs):
739 return self._og_search_property('description', html, fatal=False, **kargs)
741 def _og_search_title(self, html, **kargs):
742 return self._og_search_property('title', html, **kargs)
744 def _og_search_video_url(self, html, name='video url', secure=True, **kargs):
745 regexes = self._og_regexes('video') + self._og_regexes('video:url')
747 regexes = self._og_regexes('video:secure_url') + regexes
748 return self._html_search_regex(regexes, html, name, **kargs)
750 def _og_search_url(self, html, **kargs):
751 return self._og_search_property('url', html, **kargs)
753 def _html_search_meta(self, name, html, display_name=None, fatal=False, **kwargs):
754 if not isinstance(name, (list, tuple)):
756 if display_name is None:
757 display_name = name[0]
758 return self._html_search_regex(
759 [self._meta_regex(n) for n in name],
760 html, display_name, fatal=fatal, group='content', **kwargs)
762 def _dc_search_uploader(self, html):
763 return self._html_search_meta('dc.creator', html, 'uploader')
765 def _rta_search(self, html):
766 # See http://www.rtalabel.org/index.php?content=howtofaq#single
767 if re.search(r'(?ix)<meta\s+name="rating"\s+'
768 r' content="RTA-5042-1996-1400-1577-RTA"',
773 def _media_rating_search(self, html):
774 # See http://www.tjg-designs.com/WP/metadata-code-examples-adding-metadata-to-your-web-pages/
775 rating = self._html_search_meta('rating', html)
787 return RATING_TABLE.get(rating.lower())
789 def _family_friendly_search(self, html):
790 # See http://schema.org/VideoObject
791 family_friendly = self._html_search_meta('isFamilyFriendly', html)
793 if not family_friendly:
802 return RATING_TABLE.get(family_friendly.lower())
804 def _twitter_search_player(self, html):
805 return self._html_search_meta('twitter:player', html,
806 'twitter card player')
808 def _search_json_ld(self, html, video_id, expected_type=None, **kwargs):
809 json_ld = self._search_regex(
810 r'(?s)<script[^>]+type=(["\'])application/ld\+json\1[^>]*>(?P<json_ld>.+?)</script>',
811 html, 'JSON-LD', group='json_ld', **kwargs)
814 return self._json_ld(
815 json_ld, video_id, fatal=kwargs.get('fatal', True),
816 expected_type=expected_type)
818 def _json_ld(self, json_ld, video_id, fatal=True, expected_type=None):
819 if isinstance(json_ld, compat_str):
820 json_ld = self._parse_json(json_ld, video_id, fatal=fatal)
824 if json_ld.get('@context') == 'http://schema.org':
825 item_type = json_ld.get('@type')
826 if expected_type is not None and expected_type != item_type:
828 if item_type == 'TVEpisode':
830 'episode': unescapeHTML(json_ld.get('name')),
831 'episode_number': int_or_none(json_ld.get('episodeNumber')),
832 'description': unescapeHTML(json_ld.get('description')),
834 part_of_season = json_ld.get('partOfSeason')
835 if isinstance(part_of_season, dict) and part_of_season.get('@type') == 'TVSeason':
836 info['season_number'] = int_or_none(part_of_season.get('seasonNumber'))
837 part_of_series = json_ld.get('partOfSeries')
838 if isinstance(part_of_series, dict) and part_of_series.get('@type') == 'TVSeries':
839 info['series'] = unescapeHTML(part_of_series.get('name'))
840 elif item_type == 'Article':
842 'timestamp': parse_iso8601(json_ld.get('datePublished')),
843 'title': unescapeHTML(json_ld.get('headline')),
844 'description': unescapeHTML(json_ld.get('articleBody')),
846 elif item_type == 'VideoObject':
848 'url': json_ld.get('contentUrl'),
849 'title': unescapeHTML(json_ld.get('name')),
850 'description': unescapeHTML(json_ld.get('description')),
851 'thumbnail': json_ld.get('thumbnailUrl'),
852 'duration': parse_duration(json_ld.get('duration')),
853 'timestamp': unified_timestamp(json_ld.get('uploadDate')),
854 'filesize': float_or_none(json_ld.get('contentSize')),
855 'tbr': int_or_none(json_ld.get('bitrate')),
856 'width': int_or_none(json_ld.get('width')),
857 'height': int_or_none(json_ld.get('height')),
859 return dict((k, v) for k, v in info.items() if v is not None)
862 def _hidden_inputs(html):
863 html = re.sub(r'<!--(?:(?!<!--).)*-->', '', html)
865 for input in re.findall(r'(?i)<input([^>]+)>', html):
866 if not re.search(r'type=(["\'])(?:hidden|submit)\1', input):
868 name = re.search(r'(?:name|id)=(["\'])(?P<value>.+?)\1', input)
871 value = re.search(r'value=(["\'])(?P<value>.*?)\1', input)
874 hidden_inputs[name.group('value')] = value.group('value')
877 def _form_hidden_inputs(self, form_id, html):
878 form = self._search_regex(
879 r'(?is)<form[^>]+?id=(["\'])%s\1[^>]*>(?P<form>.+?)</form>' % form_id,
880 html, '%s form' % form_id, group='form')
881 return self._hidden_inputs(form)
883 def _sort_formats(self, formats, field_preference=None):
885 raise ExtractorError('No video formats found')
888 # Automatically determine tbr when missing based on abr and vbr (improves
889 # formats sorting in some cases)
890 if 'tbr' not in f and f.get('abr') is not None and f.get('vbr') is not None:
891 f['tbr'] = f['abr'] + f['vbr']
894 # TODO remove the following workaround
895 from ..utils import determine_ext
896 if not f.get('ext') and 'url' in f:
897 f['ext'] = determine_ext(f['url'])
899 if isinstance(field_preference, (list, tuple)):
902 if f.get(field) is not None
903 else ('' if field == 'format_id' else -1)
904 for field in field_preference)
906 preference = f.get('preference')
907 if preference is None:
909 if f.get('ext') in ['f4f', 'f4m']: # Not yet supported
912 proto_preference = 0 if determine_protocol(f) in ['http', 'https'] else -0.1
914 if f.get('vcodec') == 'none': # audio only
916 if self._downloader.params.get('prefer_free_formats'):
917 ORDER = ['aac', 'mp3', 'm4a', 'webm', 'ogg', 'opus']
919 ORDER = ['webm', 'opus', 'ogg', 'mp3', 'aac', 'm4a']
922 audio_ext_preference = ORDER.index(f['ext'])
924 audio_ext_preference = -1
926 if f.get('acodec') == 'none': # video only
928 if self._downloader.params.get('prefer_free_formats'):
929 ORDER = ['flv', 'mp4', 'webm']
931 ORDER = ['webm', 'flv', 'mp4']
933 ext_preference = ORDER.index(f['ext'])
936 audio_ext_preference = 0
940 f.get('language_preference') if f.get('language_preference') is not None else -1,
941 f.get('quality') if f.get('quality') is not None else -1,
942 f.get('tbr') if f.get('tbr') is not None else -1,
943 f.get('filesize') if f.get('filesize') is not None else -1,
944 f.get('vbr') if f.get('vbr') is not None else -1,
945 f.get('height') if f.get('height') is not None else -1,
946 f.get('width') if f.get('width') is not None else -1,
949 f.get('abr') if f.get('abr') is not None else -1,
950 audio_ext_preference,
951 f.get('fps') if f.get('fps') is not None else -1,
952 f.get('filesize_approx') if f.get('filesize_approx') is not None else -1,
953 f.get('source_preference') if f.get('source_preference') is not None else -1,
954 f.get('format_id') if f.get('format_id') is not None else '',
956 formats.sort(key=_formats_key)
958 def _check_formats(self, formats, video_id):
961 lambda f: self._is_valid_url(
963 item='%s video format' % f.get('format_id') if f.get('format_id') else 'video'),
967 def _remove_duplicate_formats(formats):
971 if f['url'] not in format_urls:
972 format_urls.add(f['url'])
973 unique_formats.append(f)
974 formats[:] = unique_formats
976 def _is_valid_url(self, url, video_id, item='video'):
977 url = self._proto_relative_url(url, scheme='http:')
978 # For now assume non HTTP(S) URLs always valid
979 if not (url.startswith('http://') or url.startswith('https://')):
982 self._request_webpage(url, video_id, 'Checking %s URL' % item)
984 except ExtractorError as e:
985 if isinstance(e.cause, compat_urllib_error.URLError):
987 '%s: %s URL is invalid, skipping' % (video_id, item))
991 def http_scheme(self):
992 """ Either "http:" or "https:", depending on the user's preferences """
995 if self._downloader.params.get('prefer_insecure', False)
998 def _proto_relative_url(self, url, scheme=None):
1001 if url.startswith('//'):
1003 scheme = self.http_scheme()
1008 def _sleep(self, timeout, video_id, msg_template=None):
1009 if msg_template is None:
1010 msg_template = '%(video_id)s: Waiting for %(timeout)s seconds'
1011 msg = msg_template % {'video_id': video_id, 'timeout': timeout}
1015 def _extract_f4m_formats(self, manifest_url, video_id, preference=None, f4m_id=None,
1016 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1017 fatal=True, m3u8_id=None):
1018 manifest = self._download_xml(
1019 manifest_url, video_id, 'Downloading f4m manifest',
1020 'Unable to download f4m manifest',
1021 # Some manifests may be malformed, e.g. prosiebensat1 generated manifests
1022 # (see https://github.com/rg3/youtube-dl/issues/6215#issuecomment-121704244)
1023 transform_source=transform_source,
1026 if manifest is False:
1029 return self._parse_f4m_formats(
1030 manifest, manifest_url, video_id, preference=preference, f4m_id=f4m_id,
1031 transform_source=transform_source, fatal=fatal, m3u8_id=m3u8_id)
1033 def _parse_f4m_formats(self, manifest, manifest_url, video_id, preference=None, f4m_id=None,
1034 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1035 fatal=True, m3u8_id=None):
1036 # currently youtube-dl cannot decode the playerVerificationChallenge as Akamai uses Adobe Alchemy
1037 akamai_pv = manifest.find('{http://ns.adobe.com/f4m/1.0}pv-2.0')
1038 if akamai_pv is not None and ';' in akamai_pv.text:
1039 playerVerificationChallenge = akamai_pv.text.split(';')[0]
1040 if playerVerificationChallenge.strip() != '':
1044 manifest_version = '1.0'
1045 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/1.0}media')
1047 manifest_version = '2.0'
1048 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/2.0}media')
1049 # Remove unsupported DRM protected media from final formats
1050 # rendition (see https://github.com/rg3/youtube-dl/issues/8573).
1051 media_nodes = remove_encrypted_media(media_nodes)
1054 base_url = xpath_text(
1055 manifest, ['{http://ns.adobe.com/f4m/1.0}baseURL', '{http://ns.adobe.com/f4m/2.0}baseURL'],
1056 'base URL', default=None)
1058 base_url = base_url.strip()
1060 bootstrap_info = xpath_element(
1061 manifest, ['{http://ns.adobe.com/f4m/1.0}bootstrapInfo', '{http://ns.adobe.com/f4m/2.0}bootstrapInfo'],
1062 'bootstrap info', default=None)
1064 for i, media_el in enumerate(media_nodes):
1065 tbr = int_or_none(media_el.attrib.get('bitrate'))
1066 width = int_or_none(media_el.attrib.get('width'))
1067 height = int_or_none(media_el.attrib.get('height'))
1068 format_id = '-'.join(filter(None, [f4m_id, compat_str(i if tbr is None else tbr)]))
1069 # If <bootstrapInfo> is present, the specified f4m is a
1070 # stream-level manifest, and only set-level manifests may refer to
1071 # external resources. See section 11.4 and section 4 of F4M spec
1072 if bootstrap_info is None:
1074 # @href is introduced in 2.0, see section 11.6 of F4M spec
1075 if manifest_version == '2.0':
1076 media_url = media_el.attrib.get('href')
1077 if media_url is None:
1078 media_url = media_el.attrib.get('url')
1082 media_url if media_url.startswith('http://') or media_url.startswith('https://')
1083 else ((base_url or '/'.join(manifest_url.split('/')[:-1])) + '/' + media_url))
1084 # If media_url is itself a f4m manifest do the recursive extraction
1085 # since bitrates in parent manifest (this one) and media_url manifest
1086 # may differ leading to inability to resolve the format by requested
1087 # bitrate in f4m downloader
1088 ext = determine_ext(manifest_url)
1090 f4m_formats = self._extract_f4m_formats(
1091 manifest_url, video_id, preference=preference, f4m_id=f4m_id,
1092 transform_source=transform_source, fatal=fatal)
1093 # Sometimes stream-level manifest contains single media entry that
1094 # does not contain any quality metadata (e.g. http://matchtv.ru/#live-player).
1095 # At the same time parent's media entry in set-level manifest may
1096 # contain it. We will copy it from parent in such cases.
1097 if len(f4m_formats) == 1:
1100 'tbr': f.get('tbr') or tbr,
1101 'width': f.get('width') or width,
1102 'height': f.get('height') or height,
1103 'format_id': f.get('format_id') if not tbr else format_id,
1105 formats.extend(f4m_formats)
1108 formats.extend(self._extract_m3u8_formats(
1109 manifest_url, video_id, 'mp4', preference=preference,
1110 m3u8_id=m3u8_id, fatal=fatal))
1113 'format_id': format_id,
1114 'url': manifest_url,
1115 'ext': 'flv' if bootstrap_info is not None else None,
1119 'preference': preference,
1123 def _m3u8_meta_format(self, m3u8_url, ext=None, preference=None, m3u8_id=None):
1125 'format_id': '-'.join(filter(None, [m3u8_id, 'meta'])),
1129 'preference': preference - 1 if preference else -1,
1130 'resolution': 'multiple',
1131 'format_note': 'Quality selection URL',
1134 def _extract_m3u8_formats(self, m3u8_url, video_id, ext=None,
1135 entry_protocol='m3u8', preference=None,
1136 m3u8_id=None, note=None, errnote=None,
1137 fatal=True, live=False):
1139 formats = [self._m3u8_meta_format(m3u8_url, ext, preference, m3u8_id)]
1141 format_url = lambda u: (
1143 if re.match(r'^https?://', u)
1144 else compat_urlparse.urljoin(m3u8_url, u))
1146 res = self._download_webpage_handle(
1148 note=note or 'Downloading m3u8 information',
1149 errnote=errnote or 'Failed to download m3u8 information',
1153 m3u8_doc, urlh = res
1154 m3u8_url = urlh.geturl()
1156 # We should try extracting formats only from master playlists [1], i.e.
1157 # playlists that describe available qualities. On the other hand media
1158 # playlists [2] should be returned as is since they contain just the media
1159 # without qualities renditions.
1160 # Fortunately, master playlist can be easily distinguished from media
1161 # playlist based on particular tags availability. As of [1, 2] master
1162 # playlist tags MUST NOT appear in a media playist and vice versa.
1163 # As of [3] #EXT-X-TARGETDURATION tag is REQUIRED for every media playlist
1164 # and MUST NOT appear in master playlist thus we can clearly detect media
1165 # playlist with this criterion.
1166 # 1. https://tools.ietf.org/html/draft-pantos-http-live-streaming-17#section-4.3.4
1167 # 2. https://tools.ietf.org/html/draft-pantos-http-live-streaming-17#section-4.3.3
1168 # 3. https://tools.ietf.org/html/draft-pantos-http-live-streaming-17#section-4.3.3.1
1169 if '#EXT-X-TARGETDURATION' in m3u8_doc: # media playlist, return as is
1172 'format_id': m3u8_id,
1174 'protocol': entry_protocol,
1175 'preference': preference,
1179 for line in m3u8_doc.splitlines():
1180 if line.startswith('#EXT-X-STREAM-INF:'):
1181 last_info = parse_m3u8_attributes(line)
1182 elif line.startswith('#EXT-X-MEDIA:'):
1183 last_media = parse_m3u8_attributes(line)
1184 elif line.startswith('#') or not line.strip():
1187 if last_info is None:
1188 formats.append({'url': format_url(line)})
1190 tbr = int_or_none(last_info.get('BANDWIDTH'), scale=1000)
1193 format_id.append(m3u8_id)
1194 last_media_name = last_media.get('NAME') if last_media and last_media.get('TYPE') not in ('SUBTITLES', 'CLOSED-CAPTIONS') else None
1195 # Despite specification does not mention NAME attribute for
1196 # EXT-X-STREAM-INF it still sometimes may be present
1197 stream_name = last_info.get('NAME') or last_media_name
1198 # Bandwidth of live streams may differ over time thus making
1199 # format_id unpredictable. So it's better to keep provided
1202 format_id.append(stream_name if stream_name else '%d' % (tbr if tbr else len(formats)))
1204 'format_id': '-'.join(format_id),
1205 'url': format_url(line.strip()),
1208 'protocol': entry_protocol,
1209 'preference': preference,
1211 resolution = last_info.get('RESOLUTION')
1213 width_str, height_str = resolution.split('x')
1214 f['width'] = int(width_str)
1215 f['height'] = int(height_str)
1216 codecs = last_info.get('CODECS')
1218 vcodec, acodec = [None] * 2
1219 va_codecs = codecs.split(',')
1220 if len(va_codecs) == 1:
1221 # Audio only entries usually come with single codec and
1222 # no resolution. For more robustness we also check it to
1224 if not resolution and va_codecs[0].startswith('mp4a'):
1225 vcodec, acodec = 'none', va_codecs[0]
1227 vcodec = va_codecs[0]
1229 vcodec, acodec = va_codecs[:2]
1234 if last_media is not None:
1235 f['m3u8_media'] = last_media
1242 def _xpath_ns(path, namespace=None):
1246 for c in path.split('/'):
1247 if not c or c == '.':
1250 out.append('{%s}%s' % (namespace, c))
1251 return '/'.join(out)
1253 def _extract_smil_formats(self, smil_url, video_id, fatal=True, f4m_params=None, transform_source=None):
1254 smil = self._download_smil(smil_url, video_id, fatal=fatal, transform_source=transform_source)
1260 namespace = self._parse_smil_namespace(smil)
1262 return self._parse_smil_formats(
1263 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
1265 def _extract_smil_info(self, smil_url, video_id, fatal=True, f4m_params=None):
1266 smil = self._download_smil(smil_url, video_id, fatal=fatal)
1269 return self._parse_smil(smil, smil_url, video_id, f4m_params=f4m_params)
1271 def _download_smil(self, smil_url, video_id, fatal=True, transform_source=None):
1272 return self._download_xml(
1273 smil_url, video_id, 'Downloading SMIL file',
1274 'Unable to download SMIL file', fatal=fatal, transform_source=transform_source)
1276 def _parse_smil(self, smil, smil_url, video_id, f4m_params=None):
1277 namespace = self._parse_smil_namespace(smil)
1279 formats = self._parse_smil_formats(
1280 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
1281 subtitles = self._parse_smil_subtitles(smil, namespace=namespace)
1283 video_id = os.path.splitext(url_basename(smil_url))[0]
1287 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
1288 name = meta.attrib.get('name')
1289 content = meta.attrib.get('content')
1290 if not name or not content:
1292 if not title and name == 'title':
1294 elif not description and name in ('description', 'abstract'):
1295 description = content
1296 elif not upload_date and name == 'date':
1297 upload_date = unified_strdate(content)
1300 'id': image.get('type'),
1301 'url': image.get('src'),
1302 'width': int_or_none(image.get('width')),
1303 'height': int_or_none(image.get('height')),
1304 } for image in smil.findall(self._xpath_ns('.//image', namespace)) if image.get('src')]
1308 'title': title or video_id,
1309 'description': description,
1310 'upload_date': upload_date,
1311 'thumbnails': thumbnails,
1313 'subtitles': subtitles,
1316 def _parse_smil_namespace(self, smil):
1317 return self._search_regex(
1318 r'(?i)^{([^}]+)?}smil$', smil.tag, 'namespace', default=None)
1320 def _parse_smil_formats(self, smil, smil_url, video_id, namespace=None, f4m_params=None, transform_rtmp_url=None):
1322 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
1323 b = meta.get('base') or meta.get('httpBase')
1334 media = smil.findall(self._xpath_ns('.//video', namespace)) + smil.findall(self._xpath_ns('.//audio', namespace))
1335 for medium in media:
1336 src = medium.get('src')
1337 if not src or src in srcs:
1341 bitrate = float_or_none(medium.get('system-bitrate') or medium.get('systemBitrate'), 1000)
1342 filesize = int_or_none(medium.get('size') or medium.get('fileSize'))
1343 width = int_or_none(medium.get('width'))
1344 height = int_or_none(medium.get('height'))
1345 proto = medium.get('proto')
1346 ext = medium.get('ext')
1347 src_ext = determine_ext(src)
1348 streamer = medium.get('streamer') or base
1350 if proto == 'rtmp' or streamer.startswith('rtmp'):
1356 'format_id': 'rtmp-%d' % (rtmp_count if bitrate is None else bitrate),
1358 'filesize': filesize,
1362 if transform_rtmp_url:
1363 streamer, src = transform_rtmp_url(streamer, src)
1364 formats[-1].update({
1370 src_url = src if src.startswith('http') else compat_urlparse.urljoin(base, src)
1371 src_url = src_url.strip()
1373 if proto == 'm3u8' or src_ext == 'm3u8':
1374 m3u8_formats = self._extract_m3u8_formats(
1375 src_url, video_id, ext or 'mp4', m3u8_id='hls', fatal=False)
1376 if len(m3u8_formats) == 1:
1378 m3u8_formats[0].update({
1379 'format_id': 'hls-%d' % (m3u8_count if bitrate is None else bitrate),
1384 formats.extend(m3u8_formats)
1387 if src_ext == 'f4m':
1392 'plugin': 'flowplayer-3.2.0.1',
1394 f4m_url += '&' if '?' in f4m_url else '?'
1395 f4m_url += compat_urllib_parse_urlencode(f4m_params)
1396 formats.extend(self._extract_f4m_formats(f4m_url, video_id, f4m_id='hds', fatal=False))
1399 if src_url.startswith('http') and self._is_valid_url(src, video_id):
1403 'ext': ext or src_ext or 'flv',
1404 'format_id': 'http-%d' % (bitrate or http_count),
1406 'filesize': filesize,
1414 def _parse_smil_subtitles(self, smil, namespace=None, subtitles_lang='en'):
1417 for num, textstream in enumerate(smil.findall(self._xpath_ns('.//textstream', namespace))):
1418 src = textstream.get('src')
1419 if not src or src in urls:
1422 ext = textstream.get('ext') or mimetype2ext(textstream.get('type')) or determine_ext(src)
1423 lang = textstream.get('systemLanguage') or textstream.get('systemLanguageName') or textstream.get('lang') or subtitles_lang
1424 subtitles.setdefault(lang, []).append({
1430 def _extract_xspf_playlist(self, playlist_url, playlist_id, fatal=True):
1431 xspf = self._download_xml(
1432 playlist_url, playlist_id, 'Downloading xpsf playlist',
1433 'Unable to download xspf manifest', fatal=fatal)
1436 return self._parse_xspf(xspf, playlist_id)
1438 def _parse_xspf(self, playlist, playlist_id):
1440 'xspf': 'http://xspf.org/ns/0/',
1441 's1': 'http://static.streamone.nl/player/ns/0',
1445 for track in playlist.findall(xpath_with_ns('./xspf:trackList/xspf:track', NS_MAP)):
1447 track, xpath_with_ns('./xspf:title', NS_MAP), 'title', default=playlist_id)
1448 description = xpath_text(
1449 track, xpath_with_ns('./xspf:annotation', NS_MAP), 'description')
1450 thumbnail = xpath_text(
1451 track, xpath_with_ns('./xspf:image', NS_MAP), 'thumbnail')
1452 duration = float_or_none(
1453 xpath_text(track, xpath_with_ns('./xspf:duration', NS_MAP), 'duration'), 1000)
1456 'url': location.text,
1457 'format_id': location.get(xpath_with_ns('s1:label', NS_MAP)),
1458 'width': int_or_none(location.get(xpath_with_ns('s1:width', NS_MAP))),
1459 'height': int_or_none(location.get(xpath_with_ns('s1:height', NS_MAP))),
1460 } for location in track.findall(xpath_with_ns('./xspf:location', NS_MAP))]
1461 self._sort_formats(formats)
1466 'description': description,
1467 'thumbnail': thumbnail,
1468 'duration': duration,
1473 def _extract_mpd_formats(self, mpd_url, video_id, mpd_id=None, note=None, errnote=None, fatal=True, formats_dict={}):
1474 res = self._download_webpage_handle(
1476 note=note or 'Downloading MPD manifest',
1477 errnote=errnote or 'Failed to download MPD manifest',
1482 mpd_base_url = re.match(r'https?://.+/', urlh.geturl()).group()
1484 return self._parse_mpd_formats(
1485 compat_etree_fromstring(mpd.encode('utf-8')), mpd_id, mpd_base_url, formats_dict=formats_dict)
1487 def _parse_mpd_formats(self, mpd_doc, mpd_id=None, mpd_base_url='', formats_dict={}):
1488 if mpd_doc.get('type') == 'dynamic':
1491 namespace = self._search_regex(r'(?i)^{([^}]+)?}MPD$', mpd_doc.tag, 'namespace', default=None)
1494 return self._xpath_ns(path, namespace)
1496 def is_drm_protected(element):
1497 return element.find(_add_ns('ContentProtection')) is not None
1499 def extract_multisegment_info(element, ms_parent_info):
1500 ms_info = ms_parent_info.copy()
1501 segment_list = element.find(_add_ns('SegmentList'))
1502 if segment_list is not None:
1503 segment_urls_e = segment_list.findall(_add_ns('SegmentURL'))
1505 ms_info['segment_urls'] = [segment.attrib['media'] for segment in segment_urls_e]
1506 initialization = segment_list.find(_add_ns('Initialization'))
1507 if initialization is not None:
1508 ms_info['initialization_url'] = initialization.attrib['sourceURL']
1510 segment_template = element.find(_add_ns('SegmentTemplate'))
1511 if segment_template is not None:
1512 start_number = segment_template.get('startNumber')
1514 ms_info['start_number'] = int(start_number)
1515 segment_timeline = segment_template.find(_add_ns('SegmentTimeline'))
1516 if segment_timeline is not None:
1517 s_e = segment_timeline.findall(_add_ns('S'))
1519 ms_info['total_number'] = 0
1521 ms_info['total_number'] += 1 + int(s.get('r', '0'))
1523 timescale = segment_template.get('timescale')
1525 ms_info['timescale'] = int(timescale)
1526 segment_duration = segment_template.get('duration')
1527 if segment_duration:
1528 ms_info['segment_duration'] = int(segment_duration)
1529 media_template = segment_template.get('media')
1531 ms_info['media_template'] = media_template
1532 initialization = segment_template.get('initialization')
1534 ms_info['initialization_url'] = initialization
1536 initialization = segment_template.find(_add_ns('Initialization'))
1537 if initialization is not None:
1538 ms_info['initialization_url'] = initialization.attrib['sourceURL']
1541 mpd_duration = parse_duration(mpd_doc.get('mediaPresentationDuration'))
1543 for period in mpd_doc.findall(_add_ns('Period')):
1544 period_duration = parse_duration(period.get('duration')) or mpd_duration
1545 period_ms_info = extract_multisegment_info(period, {
1549 for adaptation_set in period.findall(_add_ns('AdaptationSet')):
1550 if is_drm_protected(adaptation_set):
1552 adaption_set_ms_info = extract_multisegment_info(adaptation_set, period_ms_info)
1553 for representation in adaptation_set.findall(_add_ns('Representation')):
1554 if is_drm_protected(representation):
1556 representation_attrib = adaptation_set.attrib.copy()
1557 representation_attrib.update(representation.attrib)
1558 # According to page 41 of ISO/IEC 29001-1:2014, @mimeType is mandatory
1559 mime_type = representation_attrib['mimeType']
1560 content_type = mime_type.split('/')[0]
1561 if content_type == 'text':
1562 # TODO implement WebVTT downloading
1564 elif content_type == 'video' or content_type == 'audio':
1566 for element in (representation, adaptation_set, period, mpd_doc):
1567 base_url_e = element.find(_add_ns('BaseURL'))
1568 if base_url_e is not None:
1569 base_url = base_url_e.text + base_url
1570 if re.match(r'^https?://', base_url):
1572 if mpd_base_url and not re.match(r'^https?://', base_url):
1573 if not mpd_base_url.endswith('/') and not base_url.startswith('/'):
1575 base_url = mpd_base_url + base_url
1576 representation_id = representation_attrib.get('id')
1577 lang = representation_attrib.get('lang')
1578 url_el = representation.find(_add_ns('BaseURL'))
1579 filesize = int_or_none(url_el.attrib.get('{http://youtube.com/yt/2012/10/10}contentLength') if url_el is not None else None)
1581 'format_id': '%s-%s' % (mpd_id, representation_id) if mpd_id else representation_id,
1583 'ext': mimetype2ext(mime_type),
1584 'width': int_or_none(representation_attrib.get('width')),
1585 'height': int_or_none(representation_attrib.get('height')),
1586 'tbr': int_or_none(representation_attrib.get('bandwidth'), 1000),
1587 'asr': int_or_none(representation_attrib.get('audioSamplingRate')),
1588 'fps': int_or_none(representation_attrib.get('frameRate')),
1589 'vcodec': 'none' if content_type == 'audio' else representation_attrib.get('codecs'),
1590 'acodec': 'none' if content_type == 'video' else representation_attrib.get('codecs'),
1591 'language': lang if lang not in ('mul', 'und', 'zxx', 'mis') else None,
1592 'format_note': 'DASH %s' % content_type,
1593 'filesize': filesize,
1595 representation_ms_info = extract_multisegment_info(representation, adaption_set_ms_info)
1596 if 'segment_urls' not in representation_ms_info and 'media_template' in representation_ms_info:
1597 if 'total_number' not in representation_ms_info and 'segment_duration':
1598 segment_duration = float(representation_ms_info['segment_duration']) / float(representation_ms_info['timescale'])
1599 representation_ms_info['total_number'] = int(math.ceil(float(period_duration) / segment_duration))
1600 media_template = representation_ms_info['media_template']
1601 media_template = media_template.replace('$RepresentationID$', representation_id)
1602 media_template = re.sub(r'\$(Number|Bandwidth)\$', r'%(\1)d', media_template)
1603 media_template = re.sub(r'\$(Number|Bandwidth)%([^$]+)\$', r'%(\1)\2', media_template)
1604 media_template.replace('$$', '$')
1605 representation_ms_info['segment_urls'] = [
1607 'Number': segment_number,
1608 'Bandwidth': representation_attrib.get('bandwidth')}
1609 for segment_number in range(
1610 representation_ms_info['start_number'],
1611 representation_ms_info['total_number'] + representation_ms_info['start_number'])]
1612 if 'segment_urls' in representation_ms_info:
1614 'segment_urls': representation_ms_info['segment_urls'],
1615 'protocol': 'http_dash_segments',
1617 if 'initialization_url' in representation_ms_info:
1618 initialization_url = representation_ms_info['initialization_url'].replace('$RepresentationID$', representation_id)
1620 'initialization_url': initialization_url,
1622 if not f.get('url'):
1623 f['url'] = initialization_url
1625 existing_format = next(
1626 fo for fo in formats
1627 if fo['format_id'] == representation_id)
1628 except StopIteration:
1629 full_info = formats_dict.get(representation_id, {}).copy()
1631 formats.append(full_info)
1633 existing_format.update(f)
1635 self.report_warning('Unknown MIME type %s in DASH manifest' % mime_type)
1638 def _live_title(self, name):
1639 """ Generate the title for a live video """
1640 now = datetime.datetime.now()
1641 now_str = now.strftime('%Y-%m-%d %H:%M')
1642 return name + ' ' + now_str
1644 def _int(self, v, name, fatal=False, **kwargs):
1645 res = int_or_none(v, **kwargs)
1646 if 'get_attr' in kwargs:
1647 print(getattr(v, kwargs['get_attr']))
1649 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
1651 raise ExtractorError(msg)
1653 self._downloader.report_warning(msg)
1656 def _float(self, v, name, fatal=False, **kwargs):
1657 res = float_or_none(v, **kwargs)
1659 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
1661 raise ExtractorError(msg)
1663 self._downloader.report_warning(msg)
1666 def _set_cookie(self, domain, name, value, expire_time=None):
1667 cookie = compat_cookiejar.Cookie(
1668 0, name, value, None, None, domain, None,
1669 None, '/', True, False, expire_time, '', None, None, None)
1670 self._downloader.cookiejar.set_cookie(cookie)
1672 def _get_cookies(self, url):
1673 """ Return a compat_cookies.SimpleCookie with the cookies for the url """
1674 req = sanitized_Request(url)
1675 self._downloader.cookiejar.add_cookie_header(req)
1676 return compat_cookies.SimpleCookie(req.get_header('Cookie'))
1678 def get_testcases(self, include_onlymatching=False):
1679 t = getattr(self, '_TEST', None)
1681 assert not hasattr(self, '_TESTS'), \
1682 '%s has _TEST and _TESTS' % type(self).__name__
1685 tests = getattr(self, '_TESTS', [])
1687 if not include_onlymatching and t.get('only_matching', False):
1689 t['name'] = type(self).__name__[:-len('IE')]
1692 def is_suitable(self, age_limit):
1693 """ Test whether the extractor is generally suitable for the given
1694 age limit (i.e. pornographic sites are not, all others usually are) """
1696 any_restricted = False
1697 for tc in self.get_testcases(include_onlymatching=False):
1698 if 'playlist' in tc:
1699 tc = tc['playlist'][0]
1700 is_restricted = age_restricted(
1701 tc.get('info_dict', {}).get('age_limit'), age_limit)
1702 if not is_restricted:
1704 any_restricted = any_restricted or is_restricted
1705 return not any_restricted
1707 def extract_subtitles(self, *args, **kwargs):
1708 if (self._downloader.params.get('writesubtitles', False) or
1709 self._downloader.params.get('listsubtitles')):
1710 return self._get_subtitles(*args, **kwargs)
1713 def _get_subtitles(self, *args, **kwargs):
1714 raise NotImplementedError('This method must be implemented by subclasses')
1717 def _merge_subtitle_items(subtitle_list1, subtitle_list2):
1718 """ Merge subtitle items for one language. Items with duplicated URLs
1719 will be dropped. """
1720 list1_urls = set([item['url'] for item in subtitle_list1])
1721 ret = list(subtitle_list1)
1722 ret.extend([item for item in subtitle_list2 if item['url'] not in list1_urls])
1726 def _merge_subtitles(cls, subtitle_dict1, subtitle_dict2):
1727 """ Merge two subtitle dictionaries, language by language. """
1728 ret = dict(subtitle_dict1)
1729 for lang in subtitle_dict2:
1730 ret[lang] = cls._merge_subtitle_items(subtitle_dict1.get(lang, []), subtitle_dict2[lang])
1733 def extract_automatic_captions(self, *args, **kwargs):
1734 if (self._downloader.params.get('writeautomaticsub', False) or
1735 self._downloader.params.get('listsubtitles')):
1736 return self._get_automatic_captions(*args, **kwargs)
1739 def _get_automatic_captions(self, *args, **kwargs):
1740 raise NotImplementedError('This method must be implemented by subclasses')
1742 def mark_watched(self, *args, **kwargs):
1743 if (self._downloader.params.get('mark_watched', False) and
1744 (self._get_login_info()[0] is not None or
1745 self._downloader.params.get('cookiefile') is not None)):
1746 self._mark_watched(*args, **kwargs)
1748 def _mark_watched(self, *args, **kwargs):
1749 raise NotImplementedError('This method must be implemented by subclasses')
1751 def geo_verification_headers(self):
1753 geo_verification_proxy = self._downloader.params.get('geo_verification_proxy')
1754 if geo_verification_proxy:
1755 headers['Ytdl-request-proxy'] = geo_verification_proxy
1759 class SearchInfoExtractor(InfoExtractor):
1761 Base class for paged search queries extractors.
1762 They accept URLs in the format _SEARCH_KEY(|all|[0-9]):{query}
1763 Instances should define _SEARCH_KEY and _MAX_RESULTS.
1767 def _make_valid_url(cls):
1768 return r'%s(?P<prefix>|[1-9][0-9]*|all):(?P<query>[\s\S]+)' % cls._SEARCH_KEY
1771 def suitable(cls, url):
1772 return re.match(cls._make_valid_url(), url) is not None
1774 def _real_extract(self, query):
1775 mobj = re.match(self._make_valid_url(), query)
1777 raise ExtractorError('Invalid search query "%s"' % query)
1779 prefix = mobj.group('prefix')
1780 query = mobj.group('query')
1782 return self._get_n_results(query, 1)
1783 elif prefix == 'all':
1784 return self._get_n_results(query, self._MAX_RESULTS)
1788 raise ExtractorError('invalid download number %s for query "%s"' % (n, query))
1789 elif n > self._MAX_RESULTS:
1790 self._downloader.report_warning('%s returns max %i results (you requested %i)' % (self._SEARCH_KEY, self._MAX_RESULTS, n))
1791 n = self._MAX_RESULTS
1792 return self._get_n_results(query, n)
1794 def _get_n_results(self, query, n):
1795 """Get a specified number of results for a query"""
1796 raise NotImplementedError('This method must be implemented by subclasses')
1799 def SEARCH_KEY(self):
1800 return self._SEARCH_KEY