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
56 parse_m3u8_attributes,
60 class InfoExtractor(object):
61 """Information Extractor class.
63 Information extractors are the classes that, given a URL, extract
64 information about the video (or videos) the URL refers to. This
65 information includes the real video URL, the video title, author and
66 others. The information is stored in a dictionary which is then
67 passed to the YoutubeDL. The YoutubeDL processes this
68 information possibly downloading the video to the file system, among
69 other possible outcomes.
71 The type field determines the type of the result.
72 By far the most common value (and the default if _type is missing) is
73 "video", which indicates a single video.
75 For a video, the dictionaries must include the following fields:
78 title: Video title, unescaped.
80 Additionally, it must contain either a formats entry or a url one:
82 formats: A list of dictionaries for each format available, ordered
83 from worst to best quality.
86 * url Mandatory. The URL of the video file
87 * ext Will be calculated from URL if missing
88 * format A human-readable description of the format
89 ("mp4 container with h264/opus").
90 Calculated from the format_id, width, height.
91 and format_note fields if missing.
92 * format_id A short description of the format
93 ("mp4_h264_opus" or "19").
94 Technically optional, but strongly recommended.
95 * format_note Additional info about the format
96 ("3D" or "DASH video")
97 * width Width of the video, if known
98 * height Height of the video, if known
99 * resolution Textual description of width and height
100 * tbr Average bitrate of audio and video in KBit/s
101 * abr Average audio bitrate in KBit/s
102 * acodec Name of the audio codec in use
103 * asr Audio sampling rate in Hertz
104 * vbr Average video bitrate in KBit/s
106 * vcodec Name of the video codec in use
107 * container Name of the container format
108 * filesize The number of bytes, if known in advance
109 * filesize_approx An estimate for the number of bytes
110 * player_url SWF Player URL (used for rtmpdump).
111 * protocol The protocol that will be used for the actual
112 download, lower-case.
113 "http", "https", "rtsp", "rtmp", "rtmpe",
114 "m3u8", "m3u8_native" or "http_dash_segments".
115 * preference Order number of this format. If this field is
116 present and not None, the formats get sorted
117 by this field, regardless of all other values.
118 -1 for default (order by other properties),
119 -2 or smaller for less than default.
120 < -1000 to hide the format (if there is
121 another one which is strictly better)
122 * language Language code, e.g. "de" or "en-US".
123 * language_preference Is this in the language mentioned in
125 10 if it's what the URL is about,
126 -1 for default (don't know),
127 -10 otherwise, other values reserved for now.
128 * quality Order number of the video quality of this
129 format, irrespective of the file format.
130 -1 for default (order by other properties),
131 -2 or smaller for less than default.
132 * source_preference Order number for this video source
133 (quality takes higher priority)
134 -1 for default (order by other properties),
135 -2 or smaller for less than default.
136 * http_headers A dictionary of additional HTTP headers
137 to add to the request.
138 * stretched_ratio If given and not 1, indicates that the
139 video's pixels are not square.
140 width : height ratio as float.
141 * no_resume The server does not support resuming the
142 (HTTP or RTMP) download. Boolean.
144 url: Final video URL.
145 ext: Video filename extension.
146 format: The video format, defaults to ext (used for --get-format)
147 player_url: SWF Player URL (used for rtmpdump).
149 The following fields are optional:
151 alt_title: A secondary title of the video.
152 display_id An alternative identifier for the video, not necessarily
153 unique, but available before title. Typically, id is
154 something like "4234987", title "Dancing naked mole rats",
155 and display_id "dancing-naked-mole-rats"
156 thumbnails: A list of dictionaries, with the following entries:
157 * "id" (optional, string) - Thumbnail format ID
159 * "preference" (optional, int) - quality of the image
160 * "width" (optional, int)
161 * "height" (optional, int)
162 * "resolution" (optional, string "{width}x{height"},
164 * "filesize" (optional, int)
165 thumbnail: Full URL to a video thumbnail image.
166 description: Full video description.
167 uploader: Full name of the video uploader.
168 license: License name the video is licensed under.
169 creator: The creator of the video.
170 release_date: The date (YYYYMMDD) when the video was released.
171 timestamp: UNIX timestamp of the moment the video became available.
172 upload_date: Video upload date (YYYYMMDD).
173 If not explicitly set, calculated from timestamp.
174 uploader_id: Nickname or id of the video uploader.
175 uploader_url: Full URL to a personal webpage of the video uploader.
176 location: Physical location where the video was filmed.
177 subtitles: The available subtitles as a dictionary in the format
178 {language: subformats}. "subformats" is a list sorted from
179 lower to higher preference, each element is a dictionary
180 with the "ext" entry and one of:
181 * "data": The subtitles file contents
182 * "url": A URL pointing to the subtitles file
183 "ext" will be calculated from URL if missing
184 automatic_captions: Like 'subtitles', used by the YoutubeIE for
185 automatically generated captions
186 duration: Length of the video in seconds, as an integer or float.
187 view_count: How many users have watched the video on the platform.
188 like_count: Number of positive ratings of the video
189 dislike_count: Number of negative ratings of the video
190 repost_count: Number of reposts of the video
191 average_rating: Average rating give by users, the scale used depends on the webpage
192 comment_count: Number of comments on the video
193 comments: A list of comments, each with one or more of the following
194 properties (all but one of text or html optional):
195 * "author" - human-readable name of the comment author
196 * "author_id" - user ID of the comment author
198 * "html" - Comment as HTML
199 * "text" - Plain text of the comment
200 * "timestamp" - UNIX timestamp of comment
201 * "parent" - ID of the comment this one is replying to.
202 Set to "root" to indicate that this is a
203 comment to the original video.
204 age_limit: Age restriction for the video, as an integer (years)
205 webpage_url: The URL to the video webpage, if given to youtube-dl it
206 should allow to get the same result again. (It will be set
207 by YoutubeDL if it's missing)
208 categories: A list of categories that the video falls in, for example
210 tags: A list of tags assigned to the video, e.g. ["sweden", "pop music"]
211 is_live: True, False, or None (=unknown). Whether this video is a
212 live stream that goes on instead of a fixed-length video.
213 start_time: Time in seconds where the reproduction should start, as
214 specified in the URL.
215 end_time: Time in seconds where the reproduction should end, as
216 specified in the URL.
218 The following fields should only be used when the video belongs to some logical
221 chapter: Name or title of the chapter the video belongs to.
222 chapter_number: Number of the chapter the video belongs to, as an integer.
223 chapter_id: Id of the chapter the video belongs to, as a unicode string.
225 The following fields should only be used when the video is an episode of some
228 series: Title of the series or programme the video episode belongs to.
229 season: Title of the season the video episode belongs to.
230 season_number: Number of the season the video episode belongs to, as an integer.
231 season_id: Id of the season the video episode belongs to, as a unicode string.
232 episode: Title of the video episode. Unlike mandatory video title field,
233 this field should denote the exact title of the video episode
234 without any kind of decoration.
235 episode_number: Number of the video episode within a season, as an integer.
236 episode_id: Id of the video episode, as a unicode string.
238 The following fields should only be used when the media is a track or a part of
241 track: Title of the track.
242 track_number: Number of the track within an album or a disc, as an integer.
243 track_id: Id of the track (useful in case of custom indexing, e.g. 6.iii),
245 artist: Artist(s) of the track.
246 genre: Genre(s) of the track.
247 album: Title of the album the track belongs to.
248 album_type: Type of the album (e.g. "Demo", "Full-length", "Split", "Compilation", etc).
249 album_artist: List of all artists appeared on the album (e.g.
250 "Ash Borer / Fell Voices" or "Various Artists", useful for splits
252 disc_number: Number of the disc or other physical medium the track belongs to,
254 release_year: Year (YYYY) when the album was released.
256 Unless mentioned otherwise, the fields should be Unicode strings.
258 Unless mentioned otherwise, None is equivalent to absence of information.
261 _type "playlist" indicates multiple videos.
262 There must be a key "entries", which is a list, an iterable, or a PagedList
263 object, each element of which is a valid dictionary by this specification.
265 Additionally, playlists can have "title", "description" and "id" attributes
266 with the same semantics as videos (see above).
269 _type "multi_video" indicates that there are multiple videos that
270 form a single show, for examples multiple acts of an opera or TV episode.
271 It must have an entries key like a playlist and contain all the keys
272 required for a video at the same time.
275 _type "url" indicates that the video must be extracted from another
276 location, possibly by a different extractor. Its only required key is:
277 "url" - the next URL to extract.
278 The key "ie_key" can be set to the class name (minus the trailing "IE",
279 e.g. "Youtube") if the extractor class is known in advance.
280 Additionally, the dictionary may have any properties of the resolved entity
281 known in advance, for example "title" if the title of the referred video is
285 _type "url_transparent" entities have the same specification as "url", but
286 indicate that the given additional information is more precise than the one
287 associated with the resolved URL.
288 This is useful when a site employs a video service that hosts the video and
289 its technical metadata, but that video service does not embed a useful
290 title, description etc.
293 Subclasses of this one should re-define the _real_initialize() and
294 _real_extract() methods and define a _VALID_URL regexp.
295 Probably, they should also be added to the list of extractors.
297 Finally, the _WORKING attribute should be set to False for broken IEs
298 in order to warn the users and skip the tests.
305 def __init__(self, downloader=None):
306 """Constructor. Receives an optional downloader."""
308 self.set_downloader(downloader)
311 def suitable(cls, url):
312 """Receives a URL and returns True if suitable for this IE."""
314 # This does not use has/getattr intentionally - we want to know whether
315 # we have cached the regexp for *this* class, whereas getattr would also
316 # match the superclass
317 if '_VALID_URL_RE' not in cls.__dict__:
318 cls._VALID_URL_RE = re.compile(cls._VALID_URL)
319 return cls._VALID_URL_RE.match(url) is not None
322 def _match_id(cls, url):
323 if '_VALID_URL_RE' not in cls.__dict__:
324 cls._VALID_URL_RE = re.compile(cls._VALID_URL)
325 m = cls._VALID_URL_RE.match(url)
331 """Getter method for _WORKING."""
334 def initialize(self):
335 """Initializes an instance (authentication, etc)."""
337 self._real_initialize()
340 def extract(self, url):
341 """Extracts URL information and returns it in list of dicts."""
344 return self._real_extract(url)
345 except ExtractorError:
347 except compat_http_client.IncompleteRead as e:
348 raise ExtractorError('A network error has occurred.', cause=e, expected=True)
349 except (KeyError, StopIteration) as e:
350 raise ExtractorError('An extractor error has occurred.', cause=e)
352 def set_downloader(self, downloader):
353 """Sets the downloader for this IE."""
354 self._downloader = downloader
356 def _real_initialize(self):
357 """Real initialization process. Redefine in subclasses."""
360 def _real_extract(self, url):
361 """Real extraction process. Redefine in subclasses."""
366 """A string for getting the InfoExtractor with get_info_extractor"""
367 return compat_str(cls.__name__[:-2])
371 return compat_str(type(self).__name__[:-2])
373 def _request_webpage(self, url_or_request, video_id, note=None, errnote=None, fatal=True, data=None, headers={}, query={}):
374 """ Returns the response handle """
376 self.report_download_webpage(video_id)
377 elif note is not False:
379 self.to_screen('%s' % (note,))
381 self.to_screen('%s: %s' % (video_id, note))
382 if isinstance(url_or_request, compat_urllib_request.Request):
383 url_or_request = update_Request(
384 url_or_request, data=data, headers=headers, query=query)
387 url_or_request = update_url_query(url_or_request, query)
388 if data is not None or headers:
389 url_or_request = sanitized_Request(url_or_request, data, headers)
391 return self._downloader.urlopen(url_or_request)
392 except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err:
396 errnote = 'Unable to download webpage'
398 errmsg = '%s: %s' % (errnote, error_to_compat_str(err))
400 raise ExtractorError(errmsg, sys.exc_info()[2], cause=err)
402 self._downloader.report_warning(errmsg)
405 def _download_webpage_handle(self, url_or_request, video_id, note=None, errnote=None, fatal=True, encoding=None, data=None, headers={}, query={}):
406 """ Returns a tuple (page content as string, URL handle) """
407 # Strip hashes from the URL (#1038)
408 if isinstance(url_or_request, (compat_str, str)):
409 url_or_request = url_or_request.partition('#')[0]
411 urlh = self._request_webpage(url_or_request, video_id, note, errnote, fatal, data=data, headers=headers, query=query)
415 content = self._webpage_read_content(urlh, url_or_request, video_id, note, errnote, fatal, encoding=encoding)
416 return (content, urlh)
419 def _guess_encoding_from_content(content_type, webpage_bytes):
420 m = re.match(r'[a-zA-Z0-9_.-]+/[a-zA-Z0-9_.-]+\s*;\s*charset=(.+)', content_type)
422 encoding = m.group(1)
424 m = re.search(br'<meta[^>]+charset=[\'"]?([^\'")]+)[ /\'">]',
425 webpage_bytes[:1024])
427 encoding = m.group(1).decode('ascii')
428 elif webpage_bytes.startswith(b'\xff\xfe'):
435 def _webpage_read_content(self, urlh, url_or_request, video_id, note=None, errnote=None, fatal=True, prefix=None, encoding=None):
436 content_type = urlh.headers.get('Content-Type', '')
437 webpage_bytes = urlh.read()
438 if prefix is not None:
439 webpage_bytes = prefix + webpage_bytes
441 encoding = self._guess_encoding_from_content(content_type, webpage_bytes)
442 if self._downloader.params.get('dump_intermediate_pages', False):
444 url = url_or_request.get_full_url()
445 except AttributeError:
447 self.to_screen('Dumping request to ' + url)
448 dump = base64.b64encode(webpage_bytes).decode('ascii')
449 self._downloader.to_screen(dump)
450 if self._downloader.params.get('write_pages', False):
452 url = url_or_request.get_full_url()
453 except AttributeError:
455 basen = '%s_%s' % (video_id, url)
457 h = '___' + hashlib.md5(basen.encode('utf-8')).hexdigest()
458 basen = basen[:240 - len(h)] + h
459 raw_filename = basen + '.dump'
460 filename = sanitize_filename(raw_filename, restricted=True)
461 self.to_screen('Saving request to ' + filename)
462 # Working around MAX_PATH limitation on Windows (see
463 # http://msdn.microsoft.com/en-us/library/windows/desktop/aa365247(v=vs.85).aspx)
464 if compat_os_name == 'nt':
465 absfilepath = os.path.abspath(filename)
466 if len(absfilepath) > 259:
467 filename = '\\\\?\\' + absfilepath
468 with open(filename, 'wb') as outf:
469 outf.write(webpage_bytes)
472 content = webpage_bytes.decode(encoding, 'replace')
474 content = webpage_bytes.decode('utf-8', 'replace')
476 if ('<title>Access to this site is blocked</title>' in content and
477 'Websense' in content[:512]):
478 msg = 'Access to this webpage has been blocked by Websense filtering software in your network.'
479 blocked_iframe = self._html_search_regex(
480 r'<iframe src="([^"]+)"', content,
481 'Websense information URL', default=None)
483 msg += ' Visit %s for more details' % blocked_iframe
484 raise ExtractorError(msg, expected=True)
485 if '<title>The URL you requested has been blocked</title>' in content[:512]:
487 'Access to this webpage has been blocked by Indian censorship. '
488 'Use a VPN or proxy server (with --proxy) to route around it.')
489 block_msg = self._html_search_regex(
490 r'</h1><p>(.*?)</p>',
491 content, 'block message', default=None)
493 msg += ' (Message: "%s")' % block_msg.replace('\n', ' ')
494 raise ExtractorError(msg, expected=True)
498 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={}):
499 """ Returns the data of the page as a string """
502 while success is False:
504 res = self._download_webpage_handle(url_or_request, video_id, note, errnote, fatal, encoding=encoding, data=data, headers=headers, query=query)
506 except compat_http_client.IncompleteRead as e:
508 if try_count >= tries:
510 self._sleep(timeout, video_id)
517 def _download_xml(self, url_or_request, video_id,
518 note='Downloading XML', errnote='Unable to download XML',
519 transform_source=None, fatal=True, encoding=None, data=None, headers={}, query={}):
520 """Return the xml as an xml.etree.ElementTree.Element"""
521 xml_string = self._download_webpage(
522 url_or_request, video_id, note, errnote, fatal=fatal, encoding=encoding, data=data, headers=headers, query=query)
523 if xml_string is False:
526 xml_string = transform_source(xml_string)
527 return compat_etree_fromstring(xml_string.encode('utf-8'))
529 def _download_json(self, url_or_request, video_id,
530 note='Downloading JSON metadata',
531 errnote='Unable to download JSON metadata',
532 transform_source=None,
533 fatal=True, encoding=None, data=None, headers={}, query={}):
534 json_string = self._download_webpage(
535 url_or_request, video_id, note, errnote, fatal=fatal,
536 encoding=encoding, data=data, headers=headers, query=query)
537 if (not fatal) and json_string is False:
539 return self._parse_json(
540 json_string, video_id, transform_source=transform_source, fatal=fatal)
542 def _parse_json(self, json_string, video_id, transform_source=None, fatal=True):
544 json_string = transform_source(json_string)
546 return json.loads(json_string)
547 except ValueError as ve:
548 errmsg = '%s: Failed to parse JSON ' % video_id
550 raise ExtractorError(errmsg, cause=ve)
552 self.report_warning(errmsg + str(ve))
554 def report_warning(self, msg, video_id=None):
555 idstr = '' if video_id is None else '%s: ' % video_id
556 self._downloader.report_warning(
557 '[%s] %s%s' % (self.IE_NAME, idstr, msg))
559 def to_screen(self, msg):
560 """Print msg to screen, prefixing it with '[ie_name]'"""
561 self._downloader.to_screen('[%s] %s' % (self.IE_NAME, msg))
563 def report_extraction(self, id_or_name):
564 """Report information extraction."""
565 self.to_screen('%s: Extracting information' % id_or_name)
567 def report_download_webpage(self, video_id):
568 """Report webpage download."""
569 self.to_screen('%s: Downloading webpage' % video_id)
571 def report_age_confirmation(self):
572 """Report attempt to confirm age."""
573 self.to_screen('Confirming age')
575 def report_login(self):
576 """Report attempt to log in."""
577 self.to_screen('Logging in')
580 def raise_login_required(msg='This video is only available for registered users'):
581 raise ExtractorError(
582 '%s. Use --username and --password or --netrc to provide account credentials.' % msg,
586 def raise_geo_restricted(msg='This video is not available from your location due to geo restriction'):
587 raise ExtractorError(
588 '%s. You might want to use --proxy to workaround.' % msg,
591 # Methods for following #608
593 def url_result(url, ie=None, video_id=None, video_title=None):
594 """Returns a URL that points to a page that should be processed"""
595 # TODO: ie should be the class used for getting the info
596 video_info = {'_type': 'url',
599 if video_id is not None:
600 video_info['id'] = video_id
601 if video_title is not None:
602 video_info['title'] = video_title
606 def playlist_result(entries, playlist_id=None, playlist_title=None, playlist_description=None):
607 """Returns a playlist"""
608 video_info = {'_type': 'playlist',
611 video_info['id'] = playlist_id
613 video_info['title'] = playlist_title
614 if playlist_description:
615 video_info['description'] = playlist_description
618 def _search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
620 Perform a regex search on the given string, using a single or a list of
621 patterns returning the first matching group.
622 In case of failure return a default value or raise a WARNING or a
623 RegexNotFoundError, depending on fatal, specifying the field name.
625 if isinstance(pattern, (str, compat_str, compiled_regex_type)):
626 mobj = re.search(pattern, string, flags)
629 mobj = re.search(p, string, flags)
633 if not self._downloader.params.get('no_color') and compat_os_name != 'nt' and sys.stderr.isatty():
634 _name = '\033[0;34m%s\033[0m' % name
640 # return the first matching group
641 return next(g for g in mobj.groups() if g is not None)
643 return mobj.group(group)
644 elif default is not NO_DEFAULT:
647 raise RegexNotFoundError('Unable to extract %s' % _name)
649 self._downloader.report_warning('unable to extract %s' % _name + bug_reports_message())
652 def _html_search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
654 Like _search_regex, but strips HTML tags and unescapes entities.
656 res = self._search_regex(pattern, string, name, default, fatal, flags, group)
658 return clean_html(res).strip()
662 def _get_login_info(self):
664 Get the login info as (username, password)
665 It will look in the netrc file using the _NETRC_MACHINE value
666 If there's no info available, return (None, None)
668 if self._downloader is None:
673 downloader_params = self._downloader.params
675 # Attempt to use provided username and password or .netrc data
676 if downloader_params.get('username') is not None:
677 username = downloader_params['username']
678 password = downloader_params['password']
679 elif downloader_params.get('usenetrc', False):
681 info = netrc.netrc().authenticators(self._NETRC_MACHINE)
686 raise netrc.NetrcParseError('No authenticators for %s' % self._NETRC_MACHINE)
687 except (IOError, netrc.NetrcParseError) as err:
688 self._downloader.report_warning('parsing .netrc: %s' % error_to_compat_str(err))
690 return (username, password)
692 def _get_tfa_info(self, note='two-factor verification code'):
694 Get the two-factor authentication info
695 TODO - asking the user will be required for sms/phone verify
696 currently just uses the command line option
697 If there's no info available, return None
699 if self._downloader is None:
701 downloader_params = self._downloader.params
703 if downloader_params.get('twofactor') is not None:
704 return downloader_params['twofactor']
706 return compat_getpass('Type %s and press [Return]: ' % note)
708 # Helper functions for extracting OpenGraph info
710 def _og_regexes(prop):
711 content_re = r'content=(?:"([^"]+?)"|\'([^\']+?)\'|\s*([^\s"\'=<>`]+?))'
712 property_re = (r'(?:name|property)=(?:\'og:%(prop)s\'|"og:%(prop)s"|\s*og:%(prop)s\b)'
713 % {'prop': re.escape(prop)})
714 template = r'<meta[^>]+?%s[^>]+?%s'
716 template % (property_re, content_re),
717 template % (content_re, property_re),
721 def _meta_regex(prop):
722 return r'''(?isx)<meta
723 (?=[^>]+(?:itemprop|name|property|id|http-equiv)=(["\']?)%s\1)
724 [^>]+?content=(["\'])(?P<content>.*?)\2''' % re.escape(prop)
726 def _og_search_property(self, prop, html, name=None, **kargs):
728 name = 'OpenGraph %s' % prop
729 escaped = self._search_regex(self._og_regexes(prop), html, name, flags=re.DOTALL, **kargs)
732 return unescapeHTML(escaped)
734 def _og_search_thumbnail(self, html, **kargs):
735 return self._og_search_property('image', html, 'thumbnail URL', fatal=False, **kargs)
737 def _og_search_description(self, html, **kargs):
738 return self._og_search_property('description', html, fatal=False, **kargs)
740 def _og_search_title(self, html, **kargs):
741 return self._og_search_property('title', html, **kargs)
743 def _og_search_video_url(self, html, name='video url', secure=True, **kargs):
744 regexes = self._og_regexes('video') + self._og_regexes('video:url')
746 regexes = self._og_regexes('video:secure_url') + regexes
747 return self._html_search_regex(regexes, html, name, **kargs)
749 def _og_search_url(self, html, **kargs):
750 return self._og_search_property('url', html, **kargs)
752 def _html_search_meta(self, name, html, display_name=None, fatal=False, **kwargs):
753 if not isinstance(name, (list, tuple)):
755 if display_name is None:
756 display_name = name[0]
757 return self._html_search_regex(
758 [self._meta_regex(n) for n in name],
759 html, display_name, fatal=fatal, group='content', **kwargs)
761 def _dc_search_uploader(self, html):
762 return self._html_search_meta('dc.creator', html, 'uploader')
764 def _rta_search(self, html):
765 # See http://www.rtalabel.org/index.php?content=howtofaq#single
766 if re.search(r'(?ix)<meta\s+name="rating"\s+'
767 r' content="RTA-5042-1996-1400-1577-RTA"',
772 def _media_rating_search(self, html):
773 # See http://www.tjg-designs.com/WP/metadata-code-examples-adding-metadata-to-your-web-pages/
774 rating = self._html_search_meta('rating', html)
786 return RATING_TABLE.get(rating.lower())
788 def _family_friendly_search(self, html):
789 # See http://schema.org/VideoObject
790 family_friendly = self._html_search_meta('isFamilyFriendly', html)
792 if not family_friendly:
801 return RATING_TABLE.get(family_friendly.lower())
803 def _twitter_search_player(self, html):
804 return self._html_search_meta('twitter:player', html,
805 'twitter card player')
807 def _search_json_ld(self, html, video_id, **kwargs):
808 json_ld = self._search_regex(
809 r'(?s)<script[^>]+type=(["\'])application/ld\+json\1[^>]*>(?P<json_ld>.+?)</script>',
810 html, 'JSON-LD', group='json_ld', **kwargs)
813 return self._json_ld(json_ld, video_id, fatal=kwargs.get('fatal', True))
815 def _json_ld(self, json_ld, video_id, fatal=True):
816 if isinstance(json_ld, compat_str):
817 json_ld = self._parse_json(json_ld, video_id, fatal=fatal)
821 if json_ld.get('@context') == 'http://schema.org':
822 item_type = json_ld.get('@type')
823 if item_type == 'TVEpisode':
825 'episode': unescapeHTML(json_ld.get('name')),
826 'episode_number': int_or_none(json_ld.get('episodeNumber')),
827 'description': unescapeHTML(json_ld.get('description')),
829 part_of_season = json_ld.get('partOfSeason')
830 if isinstance(part_of_season, dict) and part_of_season.get('@type') == 'TVSeason':
831 info['season_number'] = int_or_none(part_of_season.get('seasonNumber'))
832 part_of_series = json_ld.get('partOfSeries')
833 if isinstance(part_of_series, dict) and part_of_series.get('@type') == 'TVSeries':
834 info['series'] = unescapeHTML(part_of_series.get('name'))
835 elif item_type == 'Article':
837 'timestamp': parse_iso8601(json_ld.get('datePublished')),
838 'title': unescapeHTML(json_ld.get('headline')),
839 'description': unescapeHTML(json_ld.get('articleBody')),
841 elif item_type == 'VideoObject':
843 'title': unescapeHTML(json_ld.get('name')),
844 'description': unescapeHTML(json_ld.get('description')),
845 'upload_date': unified_strdate(json_ld.get('upload_date')),
846 'url': unescapeHTML(json_ld.get('contentUrl')),
848 return dict((k, v) for k, v in info.items() if v is not None)
851 def _hidden_inputs(html):
852 html = re.sub(r'<!--(?:(?!<!--).)*-->', '', html)
854 for input in re.findall(r'(?i)<input([^>]+)>', html):
855 if not re.search(r'type=(["\'])(?:hidden|submit)\1', input):
857 name = re.search(r'(?:name|id)=(["\'])(?P<value>.+?)\1', input)
860 value = re.search(r'value=(["\'])(?P<value>.*?)\1', input)
863 hidden_inputs[name.group('value')] = value.group('value')
866 def _form_hidden_inputs(self, form_id, html):
867 form = self._search_regex(
868 r'(?is)<form[^>]+?id=(["\'])%s\1[^>]*>(?P<form>.+?)</form>' % form_id,
869 html, '%s form' % form_id, group='form')
870 return self._hidden_inputs(form)
872 def _sort_formats(self, formats, field_preference=None):
874 raise ExtractorError('No video formats found')
877 # Automatically determine tbr when missing based on abr and vbr (improves
878 # formats sorting in some cases)
879 if 'tbr' not in f and f.get('abr') is not None and f.get('vbr') is not None:
880 f['tbr'] = f['abr'] + f['vbr']
883 # TODO remove the following workaround
884 from ..utils import determine_ext
885 if not f.get('ext') and 'url' in f:
886 f['ext'] = determine_ext(f['url'])
888 if isinstance(field_preference, (list, tuple)):
891 if f.get(field) is not None
892 else ('' if field == 'format_id' else -1)
893 for field in field_preference)
895 preference = f.get('preference')
896 if preference is None:
898 if f.get('ext') in ['f4f', 'f4m']: # Not yet supported
901 proto_preference = 0 if determine_protocol(f) in ['http', 'https'] else -0.1
903 if f.get('vcodec') == 'none': # audio only
905 if self._downloader.params.get('prefer_free_formats'):
906 ORDER = ['aac', 'mp3', 'm4a', 'webm', 'ogg', 'opus']
908 ORDER = ['webm', 'opus', 'ogg', 'mp3', 'aac', 'm4a']
911 audio_ext_preference = ORDER.index(f['ext'])
913 audio_ext_preference = -1
915 if f.get('acodec') == 'none': # video only
917 if self._downloader.params.get('prefer_free_formats'):
918 ORDER = ['flv', 'mp4', 'webm']
920 ORDER = ['webm', 'flv', 'mp4']
922 ext_preference = ORDER.index(f['ext'])
925 audio_ext_preference = 0
929 f.get('language_preference') if f.get('language_preference') is not None else -1,
930 f.get('quality') if f.get('quality') is not None else -1,
931 f.get('tbr') if f.get('tbr') is not None else -1,
932 f.get('filesize') if f.get('filesize') is not None else -1,
933 f.get('vbr') if f.get('vbr') is not None else -1,
934 f.get('height') if f.get('height') is not None else -1,
935 f.get('width') if f.get('width') is not None else -1,
938 f.get('abr') if f.get('abr') is not None else -1,
939 audio_ext_preference,
940 f.get('fps') if f.get('fps') is not None else -1,
941 f.get('filesize_approx') if f.get('filesize_approx') is not None else -1,
942 f.get('source_preference') if f.get('source_preference') is not None else -1,
943 f.get('format_id') if f.get('format_id') is not None else '',
945 formats.sort(key=_formats_key)
947 def _check_formats(self, formats, video_id):
950 lambda f: self._is_valid_url(
952 item='%s video format' % f.get('format_id') if f.get('format_id') else 'video'),
956 def _remove_duplicate_formats(formats):
960 if f['url'] not in format_urls:
961 format_urls.add(f['url'])
962 unique_formats.append(f)
963 formats[:] = unique_formats
965 def _is_valid_url(self, url, video_id, item='video'):
966 url = self._proto_relative_url(url, scheme='http:')
967 # For now assume non HTTP(S) URLs always valid
968 if not (url.startswith('http://') or url.startswith('https://')):
971 self._request_webpage(url, video_id, 'Checking %s URL' % item)
973 except ExtractorError as e:
974 if isinstance(e.cause, compat_urllib_error.URLError):
976 '%s: %s URL is invalid, skipping' % (video_id, item))
980 def http_scheme(self):
981 """ Either "http:" or "https:", depending on the user's preferences """
984 if self._downloader.params.get('prefer_insecure', False)
987 def _proto_relative_url(self, url, scheme=None):
990 if url.startswith('//'):
992 scheme = self.http_scheme()
997 def _sleep(self, timeout, video_id, msg_template=None):
998 if msg_template is None:
999 msg_template = '%(video_id)s: Waiting for %(timeout)s seconds'
1000 msg = msg_template % {'video_id': video_id, 'timeout': timeout}
1004 def _extract_f4m_formats(self, manifest_url, video_id, preference=None, f4m_id=None,
1005 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1006 fatal=True, m3u8_id=None):
1007 manifest = self._download_xml(
1008 manifest_url, video_id, 'Downloading f4m manifest',
1009 'Unable to download f4m manifest',
1010 # Some manifests may be malformed, e.g. prosiebensat1 generated manifests
1011 # (see https://github.com/rg3/youtube-dl/issues/6215#issuecomment-121704244)
1012 transform_source=transform_source,
1015 if manifest is False:
1018 return self._parse_f4m_formats(
1019 manifest, manifest_url, video_id, preference=preference, f4m_id=f4m_id,
1020 transform_source=transform_source, fatal=fatal, m3u8_id=m3u8_id)
1022 def _parse_f4m_formats(self, manifest, manifest_url, video_id, preference=None, f4m_id=None,
1023 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1024 fatal=True, m3u8_id=None):
1025 # currently youtube-dl cannot decode the playerVerificationChallenge as Akamai uses Adobe Alchemy
1026 akamai_pv = manifest.find('{http://ns.adobe.com/f4m/1.0}pv-2.0')
1027 if akamai_pv is not None and ';' in akamai_pv.text:
1028 playerVerificationChallenge = akamai_pv.text.split(';')[0]
1029 if playerVerificationChallenge.strip() != '':
1033 manifest_version = '1.0'
1034 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/1.0}media')
1036 manifest_version = '2.0'
1037 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/2.0}media')
1038 # Remove unsupported DRM protected media from final formats
1039 # rendition (see https://github.com/rg3/youtube-dl/issues/8573).
1040 media_nodes = remove_encrypted_media(media_nodes)
1043 base_url = xpath_text(
1044 manifest, ['{http://ns.adobe.com/f4m/1.0}baseURL', '{http://ns.adobe.com/f4m/2.0}baseURL'],
1045 'base URL', default=None)
1047 base_url = base_url.strip()
1049 bootstrap_info = xpath_element(
1050 manifest, ['{http://ns.adobe.com/f4m/1.0}bootstrapInfo', '{http://ns.adobe.com/f4m/2.0}bootstrapInfo'],
1051 'bootstrap info', default=None)
1053 for i, media_el in enumerate(media_nodes):
1054 tbr = int_or_none(media_el.attrib.get('bitrate'))
1055 width = int_or_none(media_el.attrib.get('width'))
1056 height = int_or_none(media_el.attrib.get('height'))
1057 format_id = '-'.join(filter(None, [f4m_id, compat_str(i if tbr is None else tbr)]))
1058 # If <bootstrapInfo> is present, the specified f4m is a
1059 # stream-level manifest, and only set-level manifests may refer to
1060 # external resources. See section 11.4 and section 4 of F4M spec
1061 if bootstrap_info is None:
1063 # @href is introduced in 2.0, see section 11.6 of F4M spec
1064 if manifest_version == '2.0':
1065 media_url = media_el.attrib.get('href')
1066 if media_url is None:
1067 media_url = media_el.attrib.get('url')
1071 media_url if media_url.startswith('http://') or media_url.startswith('https://')
1072 else ((base_url or '/'.join(manifest_url.split('/')[:-1])) + '/' + media_url))
1073 # If media_url is itself a f4m manifest do the recursive extraction
1074 # since bitrates in parent manifest (this one) and media_url manifest
1075 # may differ leading to inability to resolve the format by requested
1076 # bitrate in f4m downloader
1077 ext = determine_ext(manifest_url)
1079 f4m_formats = self._extract_f4m_formats(
1080 manifest_url, video_id, preference=preference, f4m_id=f4m_id,
1081 transform_source=transform_source, fatal=fatal)
1082 # Sometimes stream-level manifest contains single media entry that
1083 # does not contain any quality metadata (e.g. http://matchtv.ru/#live-player).
1084 # At the same time parent's media entry in set-level manifest may
1085 # contain it. We will copy it from parent in such cases.
1086 if len(f4m_formats) == 1:
1089 'tbr': f.get('tbr') or tbr,
1090 'width': f.get('width') or width,
1091 'height': f.get('height') or height,
1092 'format_id': f.get('format_id') if not tbr else format_id,
1094 formats.extend(f4m_formats)
1097 formats.extend(self._extract_m3u8_formats(
1098 manifest_url, video_id, 'mp4', preference=preference,
1099 m3u8_id=m3u8_id, fatal=fatal))
1102 'format_id': format_id,
1103 'url': manifest_url,
1104 'ext': 'flv' if bootstrap_info is not None else None,
1108 'preference': preference,
1112 def _m3u8_meta_format(self, m3u8_url, ext=None, preference=None, m3u8_id=None):
1114 'format_id': '-'.join(filter(None, [m3u8_id, 'meta'])),
1118 'preference': preference - 1 if preference else -1,
1119 'resolution': 'multiple',
1120 'format_note': 'Quality selection URL',
1123 def _extract_m3u8_formats(self, m3u8_url, video_id, ext=None,
1124 entry_protocol='m3u8', preference=None,
1125 m3u8_id=None, note=None, errnote=None,
1126 fatal=True, live=False):
1128 formats = [self._m3u8_meta_format(m3u8_url, ext, preference, m3u8_id)]
1130 format_url = lambda u: (
1132 if re.match(r'^https?://', u)
1133 else compat_urlparse.urljoin(m3u8_url, u))
1135 res = self._download_webpage_handle(
1137 note=note or 'Downloading m3u8 information',
1138 errnote=errnote or 'Failed to download m3u8 information',
1142 m3u8_doc, urlh = res
1143 m3u8_url = urlh.geturl()
1145 # We should try extracting formats only from master playlists [1], i.e.
1146 # playlists that describe available qualities. On the other hand media
1147 # playlists [2] should be returned as is since they contain just the media
1148 # without qualities renditions.
1149 # Fortunately, master playlist can be easily distinguished from media
1150 # playlist based on particular tags availability. As of [1, 2] master
1151 # playlist tags MUST NOT appear in a media playist and vice versa.
1152 # As of [3] #EXT-X-TARGETDURATION tag is REQUIRED for every media playlist
1153 # and MUST NOT appear in master playlist thus we can clearly detect media
1154 # playlist with this criterion.
1155 # 1. https://tools.ietf.org/html/draft-pantos-http-live-streaming-17#section-4.3.4
1156 # 2. https://tools.ietf.org/html/draft-pantos-http-live-streaming-17#section-4.3.3
1157 # 3. https://tools.ietf.org/html/draft-pantos-http-live-streaming-17#section-4.3.3.1
1158 if '#EXT-X-TARGETDURATION' in m3u8_doc: # media playlist, return as is
1161 'format_id': m3u8_id,
1163 'protocol': entry_protocol,
1164 'preference': preference,
1168 for line in m3u8_doc.splitlines():
1169 if line.startswith('#EXT-X-STREAM-INF:'):
1170 last_info = parse_m3u8_attributes(line)
1171 elif line.startswith('#EXT-X-MEDIA:'):
1172 last_media = parse_m3u8_attributes(line)
1173 elif line.startswith('#') or not line.strip():
1176 if last_info is None:
1177 formats.append({'url': format_url(line)})
1179 tbr = int_or_none(last_info.get('BANDWIDTH'), scale=1000)
1182 format_id.append(m3u8_id)
1183 last_media_name = last_media.get('NAME') if last_media and last_media.get('TYPE') not in ('SUBTITLES', 'CLOSED-CAPTIONS') else None
1184 # Despite specification does not mention NAME attribute for
1185 # EXT-X-STREAM-INF it still sometimes may be present
1186 stream_name = last_info.get('NAME') or last_media_name
1187 # Bandwidth of live streams may differ over time thus making
1188 # format_id unpredictable. So it's better to keep provided
1191 format_id.append(stream_name if stream_name else '%d' % (tbr if tbr else len(formats)))
1193 'format_id': '-'.join(format_id),
1194 'url': format_url(line.strip()),
1197 'protocol': entry_protocol,
1198 'preference': preference,
1200 resolution = last_info.get('RESOLUTION')
1202 width_str, height_str = resolution.split('x')
1203 f['width'] = int(width_str)
1204 f['height'] = int(height_str)
1205 codecs = last_info.get('CODECS')
1207 vcodec, acodec = [None] * 2
1208 va_codecs = codecs.split(',')
1209 if len(va_codecs) == 1:
1210 # Audio only entries usually come with single codec and
1211 # no resolution. For more robustness we also check it to
1213 if not resolution and va_codecs[0].startswith('mp4a'):
1214 vcodec, acodec = 'none', va_codecs[0]
1216 vcodec = va_codecs[0]
1218 vcodec, acodec = va_codecs[:2]
1223 if last_media is not None:
1224 f['m3u8_media'] = last_media
1231 def _xpath_ns(path, namespace=None):
1235 for c in path.split('/'):
1236 if not c or c == '.':
1239 out.append('{%s}%s' % (namespace, c))
1240 return '/'.join(out)
1242 def _extract_smil_formats(self, smil_url, video_id, fatal=True, f4m_params=None, transform_source=None):
1243 smil = self._download_smil(smil_url, video_id, fatal=fatal, transform_source=transform_source)
1249 namespace = self._parse_smil_namespace(smil)
1251 return self._parse_smil_formats(
1252 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
1254 def _extract_smil_info(self, smil_url, video_id, fatal=True, f4m_params=None):
1255 smil = self._download_smil(smil_url, video_id, fatal=fatal)
1258 return self._parse_smil(smil, smil_url, video_id, f4m_params=f4m_params)
1260 def _download_smil(self, smil_url, video_id, fatal=True, transform_source=None):
1261 return self._download_xml(
1262 smil_url, video_id, 'Downloading SMIL file',
1263 'Unable to download SMIL file', fatal=fatal, transform_source=transform_source)
1265 def _parse_smil(self, smil, smil_url, video_id, f4m_params=None):
1266 namespace = self._parse_smil_namespace(smil)
1268 formats = self._parse_smil_formats(
1269 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
1270 subtitles = self._parse_smil_subtitles(smil, namespace=namespace)
1272 video_id = os.path.splitext(url_basename(smil_url))[0]
1276 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
1277 name = meta.attrib.get('name')
1278 content = meta.attrib.get('content')
1279 if not name or not content:
1281 if not title and name == 'title':
1283 elif not description and name in ('description', 'abstract'):
1284 description = content
1285 elif not upload_date and name == 'date':
1286 upload_date = unified_strdate(content)
1289 'id': image.get('type'),
1290 'url': image.get('src'),
1291 'width': int_or_none(image.get('width')),
1292 'height': int_or_none(image.get('height')),
1293 } for image in smil.findall(self._xpath_ns('.//image', namespace)) if image.get('src')]
1297 'title': title or video_id,
1298 'description': description,
1299 'upload_date': upload_date,
1300 'thumbnails': thumbnails,
1302 'subtitles': subtitles,
1305 def _parse_smil_namespace(self, smil):
1306 return self._search_regex(
1307 r'(?i)^{([^}]+)?}smil$', smil.tag, 'namespace', default=None)
1309 def _parse_smil_formats(self, smil, smil_url, video_id, namespace=None, f4m_params=None, transform_rtmp_url=None):
1311 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
1312 b = meta.get('base') or meta.get('httpBase')
1323 media = smil.findall(self._xpath_ns('.//video', namespace)) + smil.findall(self._xpath_ns('.//audio', namespace))
1324 for medium in media:
1325 src = medium.get('src')
1326 if not src or src in srcs:
1330 bitrate = float_or_none(medium.get('system-bitrate') or medium.get('systemBitrate'), 1000)
1331 filesize = int_or_none(medium.get('size') or medium.get('fileSize'))
1332 width = int_or_none(medium.get('width'))
1333 height = int_or_none(medium.get('height'))
1334 proto = medium.get('proto')
1335 ext = medium.get('ext')
1336 src_ext = determine_ext(src)
1337 streamer = medium.get('streamer') or base
1339 if proto == 'rtmp' or streamer.startswith('rtmp'):
1345 'format_id': 'rtmp-%d' % (rtmp_count if bitrate is None else bitrate),
1347 'filesize': filesize,
1351 if transform_rtmp_url:
1352 streamer, src = transform_rtmp_url(streamer, src)
1353 formats[-1].update({
1359 src_url = src if src.startswith('http') else compat_urlparse.urljoin(base, src)
1360 src_url = src_url.strip()
1362 if proto == 'm3u8' or src_ext == 'm3u8':
1363 m3u8_formats = self._extract_m3u8_formats(
1364 src_url, video_id, ext or 'mp4', m3u8_id='hls', fatal=False)
1365 if len(m3u8_formats) == 1:
1367 m3u8_formats[0].update({
1368 'format_id': 'hls-%d' % (m3u8_count if bitrate is None else bitrate),
1373 formats.extend(m3u8_formats)
1376 if src_ext == 'f4m':
1381 'plugin': 'flowplayer-3.2.0.1',
1383 f4m_url += '&' if '?' in f4m_url else '?'
1384 f4m_url += compat_urllib_parse_urlencode(f4m_params)
1385 formats.extend(self._extract_f4m_formats(f4m_url, video_id, f4m_id='hds', fatal=False))
1388 if src_url.startswith('http') and self._is_valid_url(src, video_id):
1392 'ext': ext or src_ext or 'flv',
1393 'format_id': 'http-%d' % (bitrate or http_count),
1395 'filesize': filesize,
1403 def _parse_smil_subtitles(self, smil, namespace=None, subtitles_lang='en'):
1406 for num, textstream in enumerate(smil.findall(self._xpath_ns('.//textstream', namespace))):
1407 src = textstream.get('src')
1408 if not src or src in urls:
1411 ext = textstream.get('ext') or mimetype2ext(textstream.get('type')) or determine_ext(src)
1412 lang = textstream.get('systemLanguage') or textstream.get('systemLanguageName') or textstream.get('lang') or subtitles_lang
1413 subtitles.setdefault(lang, []).append({
1419 def _extract_xspf_playlist(self, playlist_url, playlist_id, fatal=True):
1420 xspf = self._download_xml(
1421 playlist_url, playlist_id, 'Downloading xpsf playlist',
1422 'Unable to download xspf manifest', fatal=fatal)
1425 return self._parse_xspf(xspf, playlist_id)
1427 def _parse_xspf(self, playlist, playlist_id):
1429 'xspf': 'http://xspf.org/ns/0/',
1430 's1': 'http://static.streamone.nl/player/ns/0',
1434 for track in playlist.findall(xpath_with_ns('./xspf:trackList/xspf:track', NS_MAP)):
1436 track, xpath_with_ns('./xspf:title', NS_MAP), 'title', default=playlist_id)
1437 description = xpath_text(
1438 track, xpath_with_ns('./xspf:annotation', NS_MAP), 'description')
1439 thumbnail = xpath_text(
1440 track, xpath_with_ns('./xspf:image', NS_MAP), 'thumbnail')
1441 duration = float_or_none(
1442 xpath_text(track, xpath_with_ns('./xspf:duration', NS_MAP), 'duration'), 1000)
1445 'url': location.text,
1446 'format_id': location.get(xpath_with_ns('s1:label', NS_MAP)),
1447 'width': int_or_none(location.get(xpath_with_ns('s1:width', NS_MAP))),
1448 'height': int_or_none(location.get(xpath_with_ns('s1:height', NS_MAP))),
1449 } for location in track.findall(xpath_with_ns('./xspf:location', NS_MAP))]
1450 self._sort_formats(formats)
1455 'description': description,
1456 'thumbnail': thumbnail,
1457 'duration': duration,
1462 def _extract_mpd_formats(self, mpd_url, video_id, mpd_id=None, note=None, errnote=None, fatal=True, formats_dict={}):
1463 res = self._download_webpage_handle(
1465 note=note or 'Downloading MPD manifest',
1466 errnote=errnote or 'Failed to download MPD manifest',
1471 mpd_base_url = re.match(r'https?://.+/', urlh.geturl()).group()
1473 return self._parse_mpd_formats(
1474 compat_etree_fromstring(mpd.encode('utf-8')), mpd_id, mpd_base_url, formats_dict=formats_dict)
1476 def _parse_mpd_formats(self, mpd_doc, mpd_id=None, mpd_base_url='', formats_dict={}):
1477 if mpd_doc.get('type') == 'dynamic':
1480 namespace = self._search_regex(r'(?i)^{([^}]+)?}MPD$', mpd_doc.tag, 'namespace', default=None)
1483 return self._xpath_ns(path, namespace)
1485 def is_drm_protected(element):
1486 return element.find(_add_ns('ContentProtection')) is not None
1488 def extract_multisegment_info(element, ms_parent_info):
1489 ms_info = ms_parent_info.copy()
1490 segment_list = element.find(_add_ns('SegmentList'))
1491 if segment_list is not None:
1492 segment_urls_e = segment_list.findall(_add_ns('SegmentURL'))
1494 ms_info['segment_urls'] = [segment.attrib['media'] for segment in segment_urls_e]
1495 initialization = segment_list.find(_add_ns('Initialization'))
1496 if initialization is not None:
1497 ms_info['initialization_url'] = initialization.attrib['sourceURL']
1499 segment_template = element.find(_add_ns('SegmentTemplate'))
1500 if segment_template is not None:
1501 start_number = segment_template.get('startNumber')
1503 ms_info['start_number'] = int(start_number)
1504 segment_timeline = segment_template.find(_add_ns('SegmentTimeline'))
1505 if segment_timeline is not None:
1506 s_e = segment_timeline.findall(_add_ns('S'))
1508 ms_info['total_number'] = 0
1510 ms_info['total_number'] += 1 + int(s.get('r', '0'))
1512 timescale = segment_template.get('timescale')
1514 ms_info['timescale'] = int(timescale)
1515 segment_duration = segment_template.get('duration')
1516 if segment_duration:
1517 ms_info['segment_duration'] = int(segment_duration)
1518 media_template = segment_template.get('media')
1520 ms_info['media_template'] = media_template
1521 initialization = segment_template.get('initialization')
1523 ms_info['initialization_url'] = initialization
1525 initialization = segment_template.find(_add_ns('Initialization'))
1526 if initialization is not None:
1527 ms_info['initialization_url'] = initialization.attrib['sourceURL']
1530 mpd_duration = parse_duration(mpd_doc.get('mediaPresentationDuration'))
1532 for period in mpd_doc.findall(_add_ns('Period')):
1533 period_duration = parse_duration(period.get('duration')) or mpd_duration
1534 period_ms_info = extract_multisegment_info(period, {
1538 for adaptation_set in period.findall(_add_ns('AdaptationSet')):
1539 if is_drm_protected(adaptation_set):
1541 adaption_set_ms_info = extract_multisegment_info(adaptation_set, period_ms_info)
1542 for representation in adaptation_set.findall(_add_ns('Representation')):
1543 if is_drm_protected(representation):
1545 representation_attrib = adaptation_set.attrib.copy()
1546 representation_attrib.update(representation.attrib)
1547 # According to page 41 of ISO/IEC 29001-1:2014, @mimeType is mandatory
1548 mime_type = representation_attrib['mimeType']
1549 content_type = mime_type.split('/')[0]
1550 if content_type == 'text':
1551 # TODO implement WebVTT downloading
1553 elif content_type == 'video' or content_type == 'audio':
1555 for element in (representation, adaptation_set, period, mpd_doc):
1556 base_url_e = element.find(_add_ns('BaseURL'))
1557 if base_url_e is not None:
1558 base_url = base_url_e.text + base_url
1559 if re.match(r'^https?://', base_url):
1561 if mpd_base_url and not re.match(r'^https?://', base_url):
1562 if not mpd_base_url.endswith('/') and not base_url.startswith('/'):
1564 base_url = mpd_base_url + base_url
1565 representation_id = representation_attrib.get('id')
1566 lang = representation_attrib.get('lang')
1567 url_el = representation.find(_add_ns('BaseURL'))
1568 filesize = int_or_none(url_el.attrib.get('{http://youtube.com/yt/2012/10/10}contentLength') if url_el is not None else None)
1570 'format_id': '%s-%s' % (mpd_id, representation_id) if mpd_id else representation_id,
1572 'ext': mimetype2ext(mime_type),
1573 'width': int_or_none(representation_attrib.get('width')),
1574 'height': int_or_none(representation_attrib.get('height')),
1575 'tbr': int_or_none(representation_attrib.get('bandwidth'), 1000),
1576 'asr': int_or_none(representation_attrib.get('audioSamplingRate')),
1577 'fps': int_or_none(representation_attrib.get('frameRate')),
1578 'vcodec': 'none' if content_type == 'audio' else representation_attrib.get('codecs'),
1579 'acodec': 'none' if content_type == 'video' else representation_attrib.get('codecs'),
1580 'language': lang if lang not in ('mul', 'und', 'zxx', 'mis') else None,
1581 'format_note': 'DASH %s' % content_type,
1582 'filesize': filesize,
1584 representation_ms_info = extract_multisegment_info(representation, adaption_set_ms_info)
1585 if 'segment_urls' not in representation_ms_info and 'media_template' in representation_ms_info:
1586 if 'total_number' not in representation_ms_info and 'segment_duration':
1587 segment_duration = float(representation_ms_info['segment_duration']) / float(representation_ms_info['timescale'])
1588 representation_ms_info['total_number'] = int(math.ceil(float(period_duration) / segment_duration))
1589 media_template = representation_ms_info['media_template']
1590 media_template = media_template.replace('$RepresentationID$', representation_id)
1591 media_template = re.sub(r'\$(Number|Bandwidth)\$', r'%(\1)d', media_template)
1592 media_template = re.sub(r'\$(Number|Bandwidth)%([^$]+)\$', r'%(\1)\2', media_template)
1593 media_template.replace('$$', '$')
1594 representation_ms_info['segment_urls'] = [
1596 'Number': segment_number,
1597 'Bandwidth': representation_attrib.get('bandwidth')}
1598 for segment_number in range(
1599 representation_ms_info['start_number'],
1600 representation_ms_info['total_number'] + representation_ms_info['start_number'])]
1601 if 'segment_urls' in representation_ms_info:
1603 'segment_urls': representation_ms_info['segment_urls'],
1604 'protocol': 'http_dash_segments',
1606 if 'initialization_url' in representation_ms_info:
1607 initialization_url = representation_ms_info['initialization_url'].replace('$RepresentationID$', representation_id)
1609 'initialization_url': initialization_url,
1611 if not f.get('url'):
1612 f['url'] = initialization_url
1614 existing_format = next(
1615 fo for fo in formats
1616 if fo['format_id'] == representation_id)
1617 except StopIteration:
1618 full_info = formats_dict.get(representation_id, {}).copy()
1620 formats.append(full_info)
1622 existing_format.update(f)
1624 self.report_warning('Unknown MIME type %s in DASH manifest' % mime_type)
1627 def _live_title(self, name):
1628 """ Generate the title for a live video """
1629 now = datetime.datetime.now()
1630 now_str = now.strftime('%Y-%m-%d %H:%M')
1631 return name + ' ' + now_str
1633 def _int(self, v, name, fatal=False, **kwargs):
1634 res = int_or_none(v, **kwargs)
1635 if 'get_attr' in kwargs:
1636 print(getattr(v, kwargs['get_attr']))
1638 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
1640 raise ExtractorError(msg)
1642 self._downloader.report_warning(msg)
1645 def _float(self, v, name, fatal=False, **kwargs):
1646 res = float_or_none(v, **kwargs)
1648 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
1650 raise ExtractorError(msg)
1652 self._downloader.report_warning(msg)
1655 def _set_cookie(self, domain, name, value, expire_time=None):
1656 cookie = compat_cookiejar.Cookie(
1657 0, name, value, None, None, domain, None,
1658 None, '/', True, False, expire_time, '', None, None, None)
1659 self._downloader.cookiejar.set_cookie(cookie)
1661 def _get_cookies(self, url):
1662 """ Return a compat_cookies.SimpleCookie with the cookies for the url """
1663 req = sanitized_Request(url)
1664 self._downloader.cookiejar.add_cookie_header(req)
1665 return compat_cookies.SimpleCookie(req.get_header('Cookie'))
1667 def get_testcases(self, include_onlymatching=False):
1668 t = getattr(self, '_TEST', None)
1670 assert not hasattr(self, '_TESTS'), \
1671 '%s has _TEST and _TESTS' % type(self).__name__
1674 tests = getattr(self, '_TESTS', [])
1676 if not include_onlymatching and t.get('only_matching', False):
1678 t['name'] = type(self).__name__[:-len('IE')]
1681 def is_suitable(self, age_limit):
1682 """ Test whether the extractor is generally suitable for the given
1683 age limit (i.e. pornographic sites are not, all others usually are) """
1685 any_restricted = False
1686 for tc in self.get_testcases(include_onlymatching=False):
1687 if 'playlist' in tc:
1688 tc = tc['playlist'][0]
1689 is_restricted = age_restricted(
1690 tc.get('info_dict', {}).get('age_limit'), age_limit)
1691 if not is_restricted:
1693 any_restricted = any_restricted or is_restricted
1694 return not any_restricted
1696 def extract_subtitles(self, *args, **kwargs):
1697 if (self._downloader.params.get('writesubtitles', False) or
1698 self._downloader.params.get('listsubtitles')):
1699 return self._get_subtitles(*args, **kwargs)
1702 def _get_subtitles(self, *args, **kwargs):
1703 raise NotImplementedError('This method must be implemented by subclasses')
1706 def _merge_subtitle_items(subtitle_list1, subtitle_list2):
1707 """ Merge subtitle items for one language. Items with duplicated URLs
1708 will be dropped. """
1709 list1_urls = set([item['url'] for item in subtitle_list1])
1710 ret = list(subtitle_list1)
1711 ret.extend([item for item in subtitle_list2 if item['url'] not in list1_urls])
1715 def _merge_subtitles(cls, subtitle_dict1, subtitle_dict2):
1716 """ Merge two subtitle dictionaries, language by language. """
1717 ret = dict(subtitle_dict1)
1718 for lang in subtitle_dict2:
1719 ret[lang] = cls._merge_subtitle_items(subtitle_dict1.get(lang, []), subtitle_dict2[lang])
1722 def extract_automatic_captions(self, *args, **kwargs):
1723 if (self._downloader.params.get('writeautomaticsub', False) or
1724 self._downloader.params.get('listsubtitles')):
1725 return self._get_automatic_captions(*args, **kwargs)
1728 def _get_automatic_captions(self, *args, **kwargs):
1729 raise NotImplementedError('This method must be implemented by subclasses')
1731 def mark_watched(self, *args, **kwargs):
1732 if (self._downloader.params.get('mark_watched', False) and
1733 (self._get_login_info()[0] is not None or
1734 self._downloader.params.get('cookiefile') is not None)):
1735 self._mark_watched(*args, **kwargs)
1737 def _mark_watched(self, *args, **kwargs):
1738 raise NotImplementedError('This method must be implemented by subclasses')
1740 def geo_verification_headers(self):
1742 geo_verification_proxy = self._downloader.params.get('geo_verification_proxy')
1743 if geo_verification_proxy:
1744 headers['Ytdl-request-proxy'] = geo_verification_proxy
1748 class SearchInfoExtractor(InfoExtractor):
1750 Base class for paged search queries extractors.
1751 They accept URLs in the format _SEARCH_KEY(|all|[0-9]):{query}
1752 Instances should define _SEARCH_KEY and _MAX_RESULTS.
1756 def _make_valid_url(cls):
1757 return r'%s(?P<prefix>|[1-9][0-9]*|all):(?P<query>[\s\S]+)' % cls._SEARCH_KEY
1760 def suitable(cls, url):
1761 return re.match(cls._make_valid_url(), url) is not None
1763 def _real_extract(self, query):
1764 mobj = re.match(self._make_valid_url(), query)
1766 raise ExtractorError('Invalid search query "%s"' % query)
1768 prefix = mobj.group('prefix')
1769 query = mobj.group('query')
1771 return self._get_n_results(query, 1)
1772 elif prefix == 'all':
1773 return self._get_n_results(query, self._MAX_RESULTS)
1777 raise ExtractorError('invalid download number %s for query "%s"' % (n, query))
1778 elif n > self._MAX_RESULTS:
1779 self._downloader.report_warning('%s returns max %i results (you requested %i)' % (self._SEARCH_KEY, self._MAX_RESULTS, n))
1780 n = self._MAX_RESULTS
1781 return self._get_n_results(query, n)
1783 def _get_n_results(self, query, n):
1784 """Get a specified number of results for a query"""
1785 raise NotImplementedError('This method must be implemented by subclasses')
1788 def SEARCH_KEY(self):
1789 return self._SEARCH_KEY