1 from __future__ import unicode_literals
15 from ..compat import (
18 compat_etree_fromstring,
24 compat_urllib_parse_urlencode,
25 compat_urllib_request,
28 from ..downloader.f4m import remove_encrypted_media
57 parse_m3u8_attributes,
63 class InfoExtractor(object):
64 """Information Extractor class.
66 Information extractors are the classes that, given a URL, extract
67 information about the video (or videos) the URL refers to. This
68 information includes the real video URL, the video title, author and
69 others. The information is stored in a dictionary which is then
70 passed to the YoutubeDL. The YoutubeDL processes this
71 information possibly downloading the video to the file system, among
72 other possible outcomes.
74 The type field determines the type of the result.
75 By far the most common value (and the default if _type is missing) is
76 "video", which indicates a single video.
78 For a video, the dictionaries must include the following fields:
81 title: Video title, unescaped.
83 Additionally, it must contain either a formats entry or a url one:
85 formats: A list of dictionaries for each format available, ordered
86 from worst to best quality.
89 * url Mandatory. The URL of the video file
91 The URL of the manifest file in case of
92 fragmented media (DASH, hls, hds)
93 * ext Will be calculated from URL if missing
94 * format A human-readable description of the format
95 ("mp4 container with h264/opus").
96 Calculated from the format_id, width, height.
97 and format_note fields if missing.
98 * format_id A short description of the format
99 ("mp4_h264_opus" or "19").
100 Technically optional, but strongly recommended.
101 * format_note Additional info about the format
102 ("3D" or "DASH video")
103 * width Width of the video, if known
104 * height Height of the video, if known
105 * resolution Textual description of width and height
106 * tbr Average bitrate of audio and video in KBit/s
107 * abr Average audio bitrate in KBit/s
108 * acodec Name of the audio codec in use
109 * asr Audio sampling rate in Hertz
110 * vbr Average video bitrate in KBit/s
112 * vcodec Name of the video codec in use
113 * container Name of the container format
114 * filesize The number of bytes, if known in advance
115 * filesize_approx An estimate for the number of bytes
116 * player_url SWF Player URL (used for rtmpdump).
117 * protocol The protocol that will be used for the actual
118 download, lower-case.
119 "http", "https", "rtsp", "rtmp", "rtmpe",
120 "m3u8", "m3u8_native" or "http_dash_segments".
121 * fragments A list of fragments of the fragmented media,
122 with the following entries:
123 * "url" (mandatory) - fragment's URL
124 * "duration" (optional, int or float)
125 * "filesize" (optional, int)
126 * preference Order number of this format. If this field is
127 present and not None, the formats get sorted
128 by this field, regardless of all other values.
129 -1 for default (order by other properties),
130 -2 or smaller for less than default.
131 < -1000 to hide the format (if there is
132 another one which is strictly better)
133 * language Language code, e.g. "de" or "en-US".
134 * language_preference Is this in the language mentioned in
136 10 if it's what the URL is about,
137 -1 for default (don't know),
138 -10 otherwise, other values reserved for now.
139 * quality Order number of the video quality of this
140 format, irrespective of the file format.
141 -1 for default (order by other properties),
142 -2 or smaller for less than default.
143 * source_preference Order number for this video source
144 (quality takes higher priority)
145 -1 for default (order by other properties),
146 -2 or smaller for less than default.
147 * http_headers A dictionary of additional HTTP headers
148 to add to the request.
149 * stretched_ratio If given and not 1, indicates that the
150 video's pixels are not square.
151 width : height ratio as float.
152 * no_resume The server does not support resuming the
153 (HTTP or RTMP) download. Boolean.
155 url: Final video URL.
156 ext: Video filename extension.
157 format: The video format, defaults to ext (used for --get-format)
158 player_url: SWF Player URL (used for rtmpdump).
160 The following fields are optional:
162 alt_title: A secondary title of the video.
163 display_id An alternative identifier for the video, not necessarily
164 unique, but available before title. Typically, id is
165 something like "4234987", title "Dancing naked mole rats",
166 and display_id "dancing-naked-mole-rats"
167 thumbnails: A list of dictionaries, with the following entries:
168 * "id" (optional, string) - Thumbnail format ID
170 * "preference" (optional, int) - quality of the image
171 * "width" (optional, int)
172 * "height" (optional, int)
173 * "resolution" (optional, string "{width}x{height"},
175 * "filesize" (optional, int)
176 thumbnail: Full URL to a video thumbnail image.
177 description: Full video description.
178 uploader: Full name of the video uploader.
179 license: License name the video is licensed under.
180 creator: The creator of the video.
181 release_date: The date (YYYYMMDD) when the video was released.
182 timestamp: UNIX timestamp of the moment the video became available.
183 upload_date: Video upload date (YYYYMMDD).
184 If not explicitly set, calculated from timestamp.
185 uploader_id: Nickname or id of the video uploader.
186 uploader_url: Full URL to a personal webpage of the video uploader.
187 location: Physical location where the video was filmed.
188 subtitles: The available subtitles as a dictionary in the format
189 {language: subformats}. "subformats" is a list sorted from
190 lower to higher preference, each element is a dictionary
191 with the "ext" entry and one of:
192 * "data": The subtitles file contents
193 * "url": A URL pointing to the subtitles file
194 "ext" will be calculated from URL if missing
195 automatic_captions: Like 'subtitles', used by the YoutubeIE for
196 automatically generated captions
197 duration: Length of the video in seconds, as an integer or float.
198 view_count: How many users have watched the video on the platform.
199 like_count: Number of positive ratings of the video
200 dislike_count: Number of negative ratings of the video
201 repost_count: Number of reposts of the video
202 average_rating: Average rating give by users, the scale used depends on the webpage
203 comment_count: Number of comments on the video
204 comments: A list of comments, each with one or more of the following
205 properties (all but one of text or html optional):
206 * "author" - human-readable name of the comment author
207 * "author_id" - user ID of the comment author
209 * "html" - Comment as HTML
210 * "text" - Plain text of the comment
211 * "timestamp" - UNIX timestamp of comment
212 * "parent" - ID of the comment this one is replying to.
213 Set to "root" to indicate that this is a
214 comment to the original video.
215 age_limit: Age restriction for the video, as an integer (years)
216 webpage_url: The URL to the video webpage, if given to youtube-dl it
217 should allow to get the same result again. (It will be set
218 by YoutubeDL if it's missing)
219 categories: A list of categories that the video falls in, for example
221 tags: A list of tags assigned to the video, e.g. ["sweden", "pop music"]
222 is_live: True, False, or None (=unknown). Whether this video is a
223 live stream that goes on instead of a fixed-length video.
224 start_time: Time in seconds where the reproduction should start, as
225 specified in the URL.
226 end_time: Time in seconds where the reproduction should end, as
227 specified in the URL.
229 The following fields should only be used when the video belongs to some logical
232 chapter: Name or title of the chapter the video belongs to.
233 chapter_number: Number of the chapter the video belongs to, as an integer.
234 chapter_id: Id of the chapter the video belongs to, as a unicode string.
236 The following fields should only be used when the video is an episode of some
239 series: Title of the series or programme the video episode belongs to.
240 season: Title of the season the video episode belongs to.
241 season_number: Number of the season the video episode belongs to, as an integer.
242 season_id: Id of the season the video episode belongs to, as a unicode string.
243 episode: Title of the video episode. Unlike mandatory video title field,
244 this field should denote the exact title of the video episode
245 without any kind of decoration.
246 episode_number: Number of the video episode within a season, as an integer.
247 episode_id: Id of the video episode, as a unicode string.
249 The following fields should only be used when the media is a track or a part of
252 track: Title of the track.
253 track_number: Number of the track within an album or a disc, as an integer.
254 track_id: Id of the track (useful in case of custom indexing, e.g. 6.iii),
256 artist: Artist(s) of the track.
257 genre: Genre(s) of the track.
258 album: Title of the album the track belongs to.
259 album_type: Type of the album (e.g. "Demo", "Full-length", "Split", "Compilation", etc).
260 album_artist: List of all artists appeared on the album (e.g.
261 "Ash Borer / Fell Voices" or "Various Artists", useful for splits
263 disc_number: Number of the disc or other physical medium the track belongs to,
265 release_year: Year (YYYY) when the album was released.
267 Unless mentioned otherwise, the fields should be Unicode strings.
269 Unless mentioned otherwise, None is equivalent to absence of information.
272 _type "playlist" indicates multiple videos.
273 There must be a key "entries", which is a list, an iterable, or a PagedList
274 object, each element of which is a valid dictionary by this specification.
276 Additionally, playlists can have "title", "description" and "id" attributes
277 with the same semantics as videos (see above).
280 _type "multi_video" indicates that there are multiple videos that
281 form a single show, for examples multiple acts of an opera or TV episode.
282 It must have an entries key like a playlist and contain all the keys
283 required for a video at the same time.
286 _type "url" indicates that the video must be extracted from another
287 location, possibly by a different extractor. Its only required key is:
288 "url" - the next URL to extract.
289 The key "ie_key" can be set to the class name (minus the trailing "IE",
290 e.g. "Youtube") if the extractor class is known in advance.
291 Additionally, the dictionary may have any properties of the resolved entity
292 known in advance, for example "title" if the title of the referred video is
296 _type "url_transparent" entities have the same specification as "url", but
297 indicate that the given additional information is more precise than the one
298 associated with the resolved URL.
299 This is useful when a site employs a video service that hosts the video and
300 its technical metadata, but that video service does not embed a useful
301 title, description etc.
304 Subclasses of this one should re-define the _real_initialize() and
305 _real_extract() methods and define a _VALID_URL regexp.
306 Probably, they should also be added to the list of extractors.
308 Finally, the _WORKING attribute should be set to False for broken IEs
309 in order to warn the users and skip the tests.
316 def __init__(self, downloader=None):
317 """Constructor. Receives an optional downloader."""
319 self.set_downloader(downloader)
322 def suitable(cls, url):
323 """Receives a URL and returns True if suitable for this IE."""
325 # This does not use has/getattr intentionally - we want to know whether
326 # we have cached the regexp for *this* class, whereas getattr would also
327 # match the superclass
328 if '_VALID_URL_RE' not in cls.__dict__:
329 cls._VALID_URL_RE = re.compile(cls._VALID_URL)
330 return cls._VALID_URL_RE.match(url) is not None
333 def _match_id(cls, url):
334 if '_VALID_URL_RE' not in cls.__dict__:
335 cls._VALID_URL_RE = re.compile(cls._VALID_URL)
336 m = cls._VALID_URL_RE.match(url)
342 """Getter method for _WORKING."""
345 def initialize(self):
346 """Initializes an instance (authentication, etc)."""
348 self._real_initialize()
351 def extract(self, url):
352 """Extracts URL information and returns it in list of dicts."""
355 return self._real_extract(url)
356 except ExtractorError:
358 except compat_http_client.IncompleteRead as e:
359 raise ExtractorError('A network error has occurred.', cause=e, expected=True)
360 except (KeyError, StopIteration) as e:
361 raise ExtractorError('An extractor error has occurred.', cause=e)
363 def set_downloader(self, downloader):
364 """Sets the downloader for this IE."""
365 self._downloader = downloader
367 def _real_initialize(self):
368 """Real initialization process. Redefine in subclasses."""
371 def _real_extract(self, url):
372 """Real extraction process. Redefine in subclasses."""
377 """A string for getting the InfoExtractor with get_info_extractor"""
378 return compat_str(cls.__name__[:-2])
382 return compat_str(type(self).__name__[:-2])
384 def _request_webpage(self, url_or_request, video_id, note=None, errnote=None, fatal=True, data=None, headers={}, query={}):
385 """ Returns the response handle """
387 self.report_download_webpage(video_id)
388 elif note is not False:
390 self.to_screen('%s' % (note,))
392 self.to_screen('%s: %s' % (video_id, note))
393 if isinstance(url_or_request, compat_urllib_request.Request):
394 url_or_request = update_Request(
395 url_or_request, data=data, headers=headers, query=query)
398 url_or_request = update_url_query(url_or_request, query)
399 if data is not None or headers:
400 url_or_request = sanitized_Request(url_or_request, data, headers)
402 return self._downloader.urlopen(url_or_request)
403 except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err:
407 errnote = 'Unable to download webpage'
409 errmsg = '%s: %s' % (errnote, error_to_compat_str(err))
411 raise ExtractorError(errmsg, sys.exc_info()[2], cause=err)
413 self._downloader.report_warning(errmsg)
416 def _download_webpage_handle(self, url_or_request, video_id, note=None, errnote=None, fatal=True, encoding=None, data=None, headers={}, query={}):
417 """ Returns a tuple (page content as string, URL handle) """
418 # Strip hashes from the URL (#1038)
419 if isinstance(url_or_request, (compat_str, str)):
420 url_or_request = url_or_request.partition('#')[0]
422 urlh = self._request_webpage(url_or_request, video_id, note, errnote, fatal, data=data, headers=headers, query=query)
426 content = self._webpage_read_content(urlh, url_or_request, video_id, note, errnote, fatal, encoding=encoding)
427 return (content, urlh)
430 def _guess_encoding_from_content(content_type, webpage_bytes):
431 m = re.match(r'[a-zA-Z0-9_.-]+/[a-zA-Z0-9_.-]+\s*;\s*charset=(.+)', content_type)
433 encoding = m.group(1)
435 m = re.search(br'<meta[^>]+charset=[\'"]?([^\'")]+)[ /\'">]',
436 webpage_bytes[:1024])
438 encoding = m.group(1).decode('ascii')
439 elif webpage_bytes.startswith(b'\xff\xfe'):
446 def _webpage_read_content(self, urlh, url_or_request, video_id, note=None, errnote=None, fatal=True, prefix=None, encoding=None):
447 content_type = urlh.headers.get('Content-Type', '')
448 webpage_bytes = urlh.read()
449 if prefix is not None:
450 webpage_bytes = prefix + webpage_bytes
452 encoding = self._guess_encoding_from_content(content_type, webpage_bytes)
453 if self._downloader.params.get('dump_intermediate_pages', False):
455 url = url_or_request.get_full_url()
456 except AttributeError:
458 self.to_screen('Dumping request to ' + url)
459 dump = base64.b64encode(webpage_bytes).decode('ascii')
460 self._downloader.to_screen(dump)
461 if self._downloader.params.get('write_pages', False):
463 url = url_or_request.get_full_url()
464 except AttributeError:
466 basen = '%s_%s' % (video_id, url)
468 h = '___' + hashlib.md5(basen.encode('utf-8')).hexdigest()
469 basen = basen[:240 - len(h)] + h
470 raw_filename = basen + '.dump'
471 filename = sanitize_filename(raw_filename, restricted=True)
472 self.to_screen('Saving request to ' + filename)
473 # Working around MAX_PATH limitation on Windows (see
474 # http://msdn.microsoft.com/en-us/library/windows/desktop/aa365247(v=vs.85).aspx)
475 if compat_os_name == 'nt':
476 absfilepath = os.path.abspath(filename)
477 if len(absfilepath) > 259:
478 filename = '\\\\?\\' + absfilepath
479 with open(filename, 'wb') as outf:
480 outf.write(webpage_bytes)
483 content = webpage_bytes.decode(encoding, 'replace')
485 content = webpage_bytes.decode('utf-8', 'replace')
487 if ('<title>Access to this site is blocked</title>' in content and
488 'Websense' in content[:512]):
489 msg = 'Access to this webpage has been blocked by Websense filtering software in your network.'
490 blocked_iframe = self._html_search_regex(
491 r'<iframe src="([^"]+)"', content,
492 'Websense information URL', default=None)
494 msg += ' Visit %s for more details' % blocked_iframe
495 raise ExtractorError(msg, expected=True)
496 if '<title>The URL you requested has been blocked</title>' in content[:512]:
498 'Access to this webpage has been blocked by Indian censorship. '
499 'Use a VPN or proxy server (with --proxy) to route around it.')
500 block_msg = self._html_search_regex(
501 r'</h1><p>(.*?)</p>',
502 content, 'block message', default=None)
504 msg += ' (Message: "%s")' % block_msg.replace('\n', ' ')
505 raise ExtractorError(msg, expected=True)
509 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={}):
510 """ Returns the data of the page as a string """
513 while success is False:
515 res = self._download_webpage_handle(url_or_request, video_id, note, errnote, fatal, encoding=encoding, data=data, headers=headers, query=query)
517 except compat_http_client.IncompleteRead as e:
519 if try_count >= tries:
521 self._sleep(timeout, video_id)
528 def _download_xml(self, url_or_request, video_id,
529 note='Downloading XML', errnote='Unable to download XML',
530 transform_source=None, fatal=True, encoding=None, data=None, headers={}, query={}):
531 """Return the xml as an xml.etree.ElementTree.Element"""
532 xml_string = self._download_webpage(
533 url_or_request, video_id, note, errnote, fatal=fatal, encoding=encoding, data=data, headers=headers, query=query)
534 if xml_string is False:
537 xml_string = transform_source(xml_string)
538 return compat_etree_fromstring(xml_string.encode('utf-8'))
540 def _download_json(self, url_or_request, video_id,
541 note='Downloading JSON metadata',
542 errnote='Unable to download JSON metadata',
543 transform_source=None,
544 fatal=True, encoding=None, data=None, headers={}, query={}):
545 json_string = self._download_webpage(
546 url_or_request, video_id, note, errnote, fatal=fatal,
547 encoding=encoding, data=data, headers=headers, query=query)
548 if (not fatal) and json_string is False:
550 return self._parse_json(
551 json_string, video_id, transform_source=transform_source, fatal=fatal)
553 def _parse_json(self, json_string, video_id, transform_source=None, fatal=True):
555 json_string = transform_source(json_string)
557 return json.loads(json_string)
558 except ValueError as ve:
559 errmsg = '%s: Failed to parse JSON ' % video_id
561 raise ExtractorError(errmsg, cause=ve)
563 self.report_warning(errmsg + str(ve))
565 def report_warning(self, msg, video_id=None):
566 idstr = '' if video_id is None else '%s: ' % video_id
567 self._downloader.report_warning(
568 '[%s] %s%s' % (self.IE_NAME, idstr, msg))
570 def to_screen(self, msg):
571 """Print msg to screen, prefixing it with '[ie_name]'"""
572 self._downloader.to_screen('[%s] %s' % (self.IE_NAME, msg))
574 def report_extraction(self, id_or_name):
575 """Report information extraction."""
576 self.to_screen('%s: Extracting information' % id_or_name)
578 def report_download_webpage(self, video_id):
579 """Report webpage download."""
580 self.to_screen('%s: Downloading webpage' % video_id)
582 def report_age_confirmation(self):
583 """Report attempt to confirm age."""
584 self.to_screen('Confirming age')
586 def report_login(self):
587 """Report attempt to log in."""
588 self.to_screen('Logging in')
591 def raise_login_required(msg='This video is only available for registered users'):
592 raise ExtractorError(
593 '%s. Use --username and --password or --netrc to provide account credentials.' % msg,
597 def raise_geo_restricted(msg='This video is not available from your location due to geo restriction'):
598 raise ExtractorError(
599 '%s. You might want to use --proxy to workaround.' % msg,
602 # Methods for following #608
604 def url_result(url, ie=None, video_id=None, video_title=None):
605 """Returns a URL that points to a page that should be processed"""
606 # TODO: ie should be the class used for getting the info
607 video_info = {'_type': 'url',
610 if video_id is not None:
611 video_info['id'] = video_id
612 if video_title is not None:
613 video_info['title'] = video_title
617 def playlist_result(entries, playlist_id=None, playlist_title=None, playlist_description=None):
618 """Returns a playlist"""
619 video_info = {'_type': 'playlist',
622 video_info['id'] = playlist_id
624 video_info['title'] = playlist_title
625 if playlist_description:
626 video_info['description'] = playlist_description
629 def _search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
631 Perform a regex search on the given string, using a single or a list of
632 patterns returning the first matching group.
633 In case of failure return a default value or raise a WARNING or a
634 RegexNotFoundError, depending on fatal, specifying the field name.
636 if isinstance(pattern, (str, compat_str, compiled_regex_type)):
637 mobj = re.search(pattern, string, flags)
640 mobj = re.search(p, string, flags)
644 if not self._downloader.params.get('no_color') and compat_os_name != 'nt' and sys.stderr.isatty():
645 _name = '\033[0;34m%s\033[0m' % name
651 # return the first matching group
652 return next(g for g in mobj.groups() if g is not None)
654 return mobj.group(group)
655 elif default is not NO_DEFAULT:
658 raise RegexNotFoundError('Unable to extract %s' % _name)
660 self._downloader.report_warning('unable to extract %s' % _name + bug_reports_message())
663 def _html_search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
665 Like _search_regex, but strips HTML tags and unescapes entities.
667 res = self._search_regex(pattern, string, name, default, fatal, flags, group)
669 return clean_html(res).strip()
673 def _get_netrc_login_info(self, netrc_machine=None):
676 netrc_machine = netrc_machine or self._NETRC_MACHINE
678 if self._downloader.params.get('usenetrc', False):
680 info = netrc.netrc().authenticators(netrc_machine)
685 raise netrc.NetrcParseError(
686 'No authenticators for %s' % netrc_machine)
687 except (IOError, netrc.NetrcParseError) as err:
688 self._downloader.report_warning(
689 'parsing .netrc: %s' % error_to_compat_str(err))
691 return username, password
693 def _get_login_info(self, username_option='username', password_option='password', netrc_machine=None):
695 Get the login info as (username, password)
696 First look for the manually specified credentials using username_option
697 and password_option as keys in params dictionary. If no such credentials
698 available look in the netrc file using the netrc_machine or _NETRC_MACHINE
700 If there's no info available, return (None, None)
702 if self._downloader is None:
705 downloader_params = self._downloader.params
707 # Attempt to use provided username and password or .netrc data
708 if downloader_params.get(username_option) is not None:
709 username = downloader_params[username_option]
710 password = downloader_params[password_option]
712 username, password = self._get_netrc_login_info(netrc_machine)
714 return username, password
716 def _get_tfa_info(self, note='two-factor verification code'):
718 Get the two-factor authentication info
719 TODO - asking the user will be required for sms/phone verify
720 currently just uses the command line option
721 If there's no info available, return None
723 if self._downloader is None:
725 downloader_params = self._downloader.params
727 if downloader_params.get('twofactor') is not None:
728 return downloader_params['twofactor']
730 return compat_getpass('Type %s and press [Return]: ' % note)
732 # Helper functions for extracting OpenGraph info
734 def _og_regexes(prop):
735 content_re = r'content=(?:"([^"]+?)"|\'([^\']+?)\'|\s*([^\s"\'=<>`]+?))'
736 property_re = (r'(?:name|property)=(?:\'og:%(prop)s\'|"og:%(prop)s"|\s*og:%(prop)s\b)'
737 % {'prop': re.escape(prop)})
738 template = r'<meta[^>]+?%s[^>]+?%s'
740 template % (property_re, content_re),
741 template % (content_re, property_re),
745 def _meta_regex(prop):
746 return r'''(?isx)<meta
747 (?=[^>]+(?:itemprop|name|property|id|http-equiv)=(["\']?)%s\1)
748 [^>]+?content=(["\'])(?P<content>.*?)\2''' % re.escape(prop)
750 def _og_search_property(self, prop, html, name=None, **kargs):
751 if not isinstance(prop, (list, tuple)):
754 name = 'OpenGraph %s' % prop[0]
757 og_regexes.extend(self._og_regexes(p))
758 escaped = self._search_regex(og_regexes, html, name, flags=re.DOTALL, **kargs)
761 return unescapeHTML(escaped)
763 def _og_search_thumbnail(self, html, **kargs):
764 return self._og_search_property('image', html, 'thumbnail URL', fatal=False, **kargs)
766 def _og_search_description(self, html, **kargs):
767 return self._og_search_property('description', html, fatal=False, **kargs)
769 def _og_search_title(self, html, **kargs):
770 return self._og_search_property('title', html, **kargs)
772 def _og_search_video_url(self, html, name='video url', secure=True, **kargs):
773 regexes = self._og_regexes('video') + self._og_regexes('video:url')
775 regexes = self._og_regexes('video:secure_url') + regexes
776 return self._html_search_regex(regexes, html, name, **kargs)
778 def _og_search_url(self, html, **kargs):
779 return self._og_search_property('url', html, **kargs)
781 def _html_search_meta(self, name, html, display_name=None, fatal=False, **kwargs):
782 if not isinstance(name, (list, tuple)):
784 if display_name is None:
785 display_name = name[0]
786 return self._html_search_regex(
787 [self._meta_regex(n) for n in name],
788 html, display_name, fatal=fatal, group='content', **kwargs)
790 def _dc_search_uploader(self, html):
791 return self._html_search_meta('dc.creator', html, 'uploader')
793 def _rta_search(self, html):
794 # See http://www.rtalabel.org/index.php?content=howtofaq#single
795 if re.search(r'(?ix)<meta\s+name="rating"\s+'
796 r' content="RTA-5042-1996-1400-1577-RTA"',
801 def _media_rating_search(self, html):
802 # See http://www.tjg-designs.com/WP/metadata-code-examples-adding-metadata-to-your-web-pages/
803 rating = self._html_search_meta('rating', html)
815 return RATING_TABLE.get(rating.lower())
817 def _family_friendly_search(self, html):
818 # See http://schema.org/VideoObject
819 family_friendly = self._html_search_meta('isFamilyFriendly', html)
821 if not family_friendly:
830 return RATING_TABLE.get(family_friendly.lower())
832 def _twitter_search_player(self, html):
833 return self._html_search_meta('twitter:player', html,
834 'twitter card player')
836 def _search_json_ld(self, html, video_id, expected_type=None, **kwargs):
837 json_ld = self._search_regex(
838 r'(?s)<script[^>]+type=(["\'])application/ld\+json\1[^>]*>(?P<json_ld>.+?)</script>',
839 html, 'JSON-LD', group='json_ld', **kwargs)
840 default = kwargs.get('default', NO_DEFAULT)
842 return default if default is not NO_DEFAULT else {}
843 # JSON-LD may be malformed and thus `fatal` should be respected.
844 # At the same time `default` may be passed that assumes `fatal=False`
845 # for _search_regex. Let's simulate the same behavior here as well.
846 fatal = kwargs.get('fatal', True) if default == NO_DEFAULT else False
847 return self._json_ld(json_ld, video_id, fatal=fatal, expected_type=expected_type)
849 def _json_ld(self, json_ld, video_id, fatal=True, expected_type=None):
850 if isinstance(json_ld, compat_str):
851 json_ld = self._parse_json(json_ld, video_id, fatal=fatal)
855 if not isinstance(json_ld, (list, tuple, dict)):
857 if isinstance(json_ld, dict):
860 if e.get('@context') == 'http://schema.org':
861 item_type = e.get('@type')
862 if expected_type is not None and expected_type != item_type:
864 if item_type == 'TVEpisode':
866 'episode': unescapeHTML(e.get('name')),
867 'episode_number': int_or_none(e.get('episodeNumber')),
868 'description': unescapeHTML(e.get('description')),
870 part_of_season = e.get('partOfSeason')
871 if isinstance(part_of_season, dict) and part_of_season.get('@type') == 'TVSeason':
872 info['season_number'] = int_or_none(part_of_season.get('seasonNumber'))
873 part_of_series = e.get('partOfSeries') or e.get('partOfTVSeries')
874 if isinstance(part_of_series, dict) and part_of_series.get('@type') == 'TVSeries':
875 info['series'] = unescapeHTML(part_of_series.get('name'))
876 elif item_type == 'Article':
878 'timestamp': parse_iso8601(e.get('datePublished')),
879 'title': unescapeHTML(e.get('headline')),
880 'description': unescapeHTML(e.get('articleBody')),
882 elif item_type == 'VideoObject':
884 'url': e.get('contentUrl'),
885 'title': unescapeHTML(e.get('name')),
886 'description': unescapeHTML(e.get('description')),
887 'thumbnail': e.get('thumbnailUrl'),
888 'duration': parse_duration(e.get('duration')),
889 'timestamp': unified_timestamp(e.get('uploadDate')),
890 'filesize': float_or_none(e.get('contentSize')),
891 'tbr': int_or_none(e.get('bitrate')),
892 'width': int_or_none(e.get('width')),
893 'height': int_or_none(e.get('height')),
896 return dict((k, v) for k, v in info.items() if v is not None)
899 def _hidden_inputs(html):
900 html = re.sub(r'<!--(?:(?!<!--).)*-->', '', html)
902 for input in re.findall(r'(?i)(<input[^>]+>)', html):
903 attrs = extract_attributes(input)
906 if attrs.get('type') not in ('hidden', 'submit'):
908 name = attrs.get('name') or attrs.get('id')
909 value = attrs.get('value')
910 if name and value is not None:
911 hidden_inputs[name] = value
914 def _form_hidden_inputs(self, form_id, html):
915 form = self._search_regex(
916 r'(?is)<form[^>]+?id=(["\'])%s\1[^>]*>(?P<form>.+?)</form>' % form_id,
917 html, '%s form' % form_id, group='form')
918 return self._hidden_inputs(form)
920 def _sort_formats(self, formats, field_preference=None):
922 raise ExtractorError('No video formats found')
925 # Automatically determine tbr when missing based on abr and vbr (improves
926 # formats sorting in some cases)
927 if 'tbr' not in f and f.get('abr') is not None and f.get('vbr') is not None:
928 f['tbr'] = f['abr'] + f['vbr']
931 # TODO remove the following workaround
932 from ..utils import determine_ext
933 if not f.get('ext') and 'url' in f:
934 f['ext'] = determine_ext(f['url'])
936 if isinstance(field_preference, (list, tuple)):
939 if f.get(field) is not None
940 else ('' if field == 'format_id' else -1)
941 for field in field_preference)
943 preference = f.get('preference')
944 if preference is None:
946 if f.get('ext') in ['f4f', 'f4m']: # Not yet supported
949 protocol = f.get('protocol') or determine_protocol(f)
950 proto_preference = 0 if protocol in ['http', 'https'] else (-0.5 if protocol == 'rtsp' else -0.1)
952 if f.get('vcodec') == 'none': # audio only
954 if self._downloader.params.get('prefer_free_formats'):
955 ORDER = ['aac', 'mp3', 'm4a', 'webm', 'ogg', 'opus']
957 ORDER = ['webm', 'opus', 'ogg', 'mp3', 'aac', 'm4a']
960 audio_ext_preference = ORDER.index(f['ext'])
962 audio_ext_preference = -1
964 if f.get('acodec') == 'none': # video only
966 if self._downloader.params.get('prefer_free_formats'):
967 ORDER = ['flv', 'mp4', 'webm']
969 ORDER = ['webm', 'flv', 'mp4']
971 ext_preference = ORDER.index(f['ext'])
974 audio_ext_preference = 0
978 f.get('language_preference') if f.get('language_preference') is not None else -1,
979 f.get('quality') if f.get('quality') is not None else -1,
980 f.get('tbr') if f.get('tbr') is not None else -1,
981 f.get('filesize') if f.get('filesize') is not None else -1,
982 f.get('vbr') if f.get('vbr') is not None else -1,
983 f.get('height') if f.get('height') is not None else -1,
984 f.get('width') if f.get('width') is not None else -1,
987 f.get('abr') if f.get('abr') is not None else -1,
988 audio_ext_preference,
989 f.get('fps') if f.get('fps') is not None else -1,
990 f.get('filesize_approx') if f.get('filesize_approx') is not None else -1,
991 f.get('source_preference') if f.get('source_preference') is not None else -1,
992 f.get('format_id') if f.get('format_id') is not None else '',
994 formats.sort(key=_formats_key)
996 def _check_formats(self, formats, video_id):
999 lambda f: self._is_valid_url(
1001 item='%s video format' % f.get('format_id') if f.get('format_id') else 'video'),
1005 def _remove_duplicate_formats(formats):
1009 if f['url'] not in format_urls:
1010 format_urls.add(f['url'])
1011 unique_formats.append(f)
1012 formats[:] = unique_formats
1014 def _is_valid_url(self, url, video_id, item='video'):
1015 url = self._proto_relative_url(url, scheme='http:')
1016 # For now assume non HTTP(S) URLs always valid
1017 if not (url.startswith('http://') or url.startswith('https://')):
1020 self._request_webpage(url, video_id, 'Checking %s URL' % item)
1022 except ExtractorError as e:
1023 if isinstance(e.cause, compat_urllib_error.URLError):
1025 '%s: %s URL is invalid, skipping' % (video_id, item))
1029 def http_scheme(self):
1030 """ Either "http:" or "https:", depending on the user's preferences """
1033 if self._downloader.params.get('prefer_insecure', False)
1036 def _proto_relative_url(self, url, scheme=None):
1039 if url.startswith('//'):
1041 scheme = self.http_scheme()
1046 def _sleep(self, timeout, video_id, msg_template=None):
1047 if msg_template is None:
1048 msg_template = '%(video_id)s: Waiting for %(timeout)s seconds'
1049 msg = msg_template % {'video_id': video_id, 'timeout': timeout}
1053 def _extract_f4m_formats(self, manifest_url, video_id, preference=None, f4m_id=None,
1054 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1055 fatal=True, m3u8_id=None):
1056 manifest = self._download_xml(
1057 manifest_url, video_id, 'Downloading f4m manifest',
1058 'Unable to download f4m manifest',
1059 # Some manifests may be malformed, e.g. prosiebensat1 generated manifests
1060 # (see https://github.com/rg3/youtube-dl/issues/6215#issuecomment-121704244)
1061 transform_source=transform_source,
1064 if manifest is False:
1067 return self._parse_f4m_formats(
1068 manifest, manifest_url, video_id, preference=preference, f4m_id=f4m_id,
1069 transform_source=transform_source, fatal=fatal, m3u8_id=m3u8_id)
1071 def _parse_f4m_formats(self, manifest, manifest_url, video_id, preference=None, f4m_id=None,
1072 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1073 fatal=True, m3u8_id=None):
1074 # currently youtube-dl cannot decode the playerVerificationChallenge as Akamai uses Adobe Alchemy
1075 akamai_pv = manifest.find('{http://ns.adobe.com/f4m/1.0}pv-2.0')
1076 if akamai_pv is not None and ';' in akamai_pv.text:
1077 playerVerificationChallenge = akamai_pv.text.split(';')[0]
1078 if playerVerificationChallenge.strip() != '':
1082 manifest_version = '1.0'
1083 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/1.0}media')
1085 manifest_version = '2.0'
1086 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/2.0}media')
1087 # Remove unsupported DRM protected media from final formats
1088 # rendition (see https://github.com/rg3/youtube-dl/issues/8573).
1089 media_nodes = remove_encrypted_media(media_nodes)
1092 base_url = xpath_text(
1093 manifest, ['{http://ns.adobe.com/f4m/1.0}baseURL', '{http://ns.adobe.com/f4m/2.0}baseURL'],
1094 'base URL', default=None)
1096 base_url = base_url.strip()
1098 bootstrap_info = xpath_element(
1099 manifest, ['{http://ns.adobe.com/f4m/1.0}bootstrapInfo', '{http://ns.adobe.com/f4m/2.0}bootstrapInfo'],
1100 'bootstrap info', default=None)
1102 for i, media_el in enumerate(media_nodes):
1103 tbr = int_or_none(media_el.attrib.get('bitrate'))
1104 width = int_or_none(media_el.attrib.get('width'))
1105 height = int_or_none(media_el.attrib.get('height'))
1106 format_id = '-'.join(filter(None, [f4m_id, compat_str(i if tbr is None else tbr)]))
1107 # If <bootstrapInfo> is present, the specified f4m is a
1108 # stream-level manifest, and only set-level manifests may refer to
1109 # external resources. See section 11.4 and section 4 of F4M spec
1110 if bootstrap_info is None:
1112 # @href is introduced in 2.0, see section 11.6 of F4M spec
1113 if manifest_version == '2.0':
1114 media_url = media_el.attrib.get('href')
1115 if media_url is None:
1116 media_url = media_el.attrib.get('url')
1120 media_url if media_url.startswith('http://') or media_url.startswith('https://')
1121 else ((base_url or '/'.join(manifest_url.split('/')[:-1])) + '/' + media_url))
1122 # If media_url is itself a f4m manifest do the recursive extraction
1123 # since bitrates in parent manifest (this one) and media_url manifest
1124 # may differ leading to inability to resolve the format by requested
1125 # bitrate in f4m downloader
1126 ext = determine_ext(manifest_url)
1128 f4m_formats = self._extract_f4m_formats(
1129 manifest_url, video_id, preference=preference, f4m_id=f4m_id,
1130 transform_source=transform_source, fatal=fatal)
1131 # Sometimes stream-level manifest contains single media entry that
1132 # does not contain any quality metadata (e.g. http://matchtv.ru/#live-player).
1133 # At the same time parent's media entry in set-level manifest may
1134 # contain it. We will copy it from parent in such cases.
1135 if len(f4m_formats) == 1:
1138 'tbr': f.get('tbr') or tbr,
1139 'width': f.get('width') or width,
1140 'height': f.get('height') or height,
1141 'format_id': f.get('format_id') if not tbr else format_id,
1143 formats.extend(f4m_formats)
1146 formats.extend(self._extract_m3u8_formats(
1147 manifest_url, video_id, 'mp4', preference=preference,
1148 m3u8_id=m3u8_id, fatal=fatal))
1151 'format_id': format_id,
1152 'url': manifest_url,
1153 'manifest_url': manifest_url,
1154 'ext': 'flv' if bootstrap_info is not None else None,
1158 'preference': preference,
1162 def _m3u8_meta_format(self, m3u8_url, ext=None, preference=None, m3u8_id=None):
1164 'format_id': '-'.join(filter(None, [m3u8_id, 'meta'])),
1168 'preference': preference - 100 if preference else -100,
1169 'resolution': 'multiple',
1170 'format_note': 'Quality selection URL',
1173 def _extract_m3u8_formats(self, m3u8_url, video_id, ext=None,
1174 entry_protocol='m3u8', preference=None,
1175 m3u8_id=None, note=None, errnote=None,
1176 fatal=True, live=False):
1178 res = self._download_webpage_handle(
1180 note=note or 'Downloading m3u8 information',
1181 errnote=errnote or 'Failed to download m3u8 information',
1185 m3u8_doc, urlh = res
1186 m3u8_url = urlh.geturl()
1188 formats = [self._m3u8_meta_format(m3u8_url, ext, preference, m3u8_id)]
1190 format_url = lambda u: (
1192 if re.match(r'^https?://', u)
1193 else compat_urlparse.urljoin(m3u8_url, u))
1195 # We should try extracting formats only from master playlists [1], i.e.
1196 # playlists that describe available qualities. On the other hand media
1197 # playlists [2] should be returned as is since they contain just the media
1198 # without qualities renditions.
1199 # Fortunately, master playlist can be easily distinguished from media
1200 # playlist based on particular tags availability. As of [1, 2] master
1201 # playlist tags MUST NOT appear in a media playist and vice versa.
1202 # As of [3] #EXT-X-TARGETDURATION tag is REQUIRED for every media playlist
1203 # and MUST NOT appear in master playlist thus we can clearly detect media
1204 # playlist with this criterion.
1205 # 1. https://tools.ietf.org/html/draft-pantos-http-live-streaming-17#section-4.3.4
1206 # 2. https://tools.ietf.org/html/draft-pantos-http-live-streaming-17#section-4.3.3
1207 # 3. https://tools.ietf.org/html/draft-pantos-http-live-streaming-17#section-4.3.3.1
1208 if '#EXT-X-TARGETDURATION' in m3u8_doc: # media playlist, return as is
1211 'format_id': m3u8_id,
1213 'protocol': entry_protocol,
1214 'preference': preference,
1218 for line in m3u8_doc.splitlines():
1219 if line.startswith('#EXT-X-STREAM-INF:'):
1220 last_info = parse_m3u8_attributes(line)
1221 elif line.startswith('#EXT-X-MEDIA:'):
1222 media = parse_m3u8_attributes(line)
1223 media_type = media.get('TYPE')
1224 if media_type in ('VIDEO', 'AUDIO'):
1225 media_url = media.get('URI')
1228 for v in (media.get('GROUP-ID'), media.get('NAME')):
1232 'format_id': '-'.join(format_id),
1233 'url': format_url(media_url),
1234 'language': media.get('LANGUAGE'),
1235 'vcodec': 'none' if media_type == 'AUDIO' else None,
1237 'protocol': entry_protocol,
1238 'preference': preference,
1241 # When there is no URI in EXT-X-MEDIA let this tag's
1242 # data be used by regular URI lines below
1244 elif line.startswith('#') or not line.strip():
1247 tbr = int_or_none(last_info.get('AVERAGE-BANDWIDTH') or last_info.get('BANDWIDTH'), scale=1000)
1250 format_id.append(m3u8_id)
1251 # Despite specification does not mention NAME attribute for
1252 # EXT-X-STREAM-INF it still sometimes may be present
1253 stream_name = last_info.get('NAME') or last_media.get('NAME')
1254 # Bandwidth of live streams may differ over time thus making
1255 # format_id unpredictable. So it's better to keep provided
1258 format_id.append(stream_name if stream_name else '%d' % (tbr if tbr else len(formats)))
1259 manifest_url = format_url(line.strip())
1261 'format_id': '-'.join(format_id),
1262 'url': manifest_url,
1263 'manifest_url': manifest_url,
1266 'fps': float_or_none(last_info.get('FRAME-RATE')),
1267 'protocol': entry_protocol,
1268 'preference': preference,
1270 resolution = last_info.get('RESOLUTION')
1272 width_str, height_str = resolution.split('x')
1273 f['width'] = int(width_str)
1274 f['height'] = int(height_str)
1275 # Unified Streaming Platform
1277 r'audio.*?(?:%3D|=)(\d+)(?:-video.*?(?:%3D|=)(\d+))?', f['url'])
1279 abr, vbr = mobj.groups()
1280 abr, vbr = float_or_none(abr, 1000), float_or_none(vbr, 1000)
1285 f.update(parse_codecs(last_info.get('CODECS')))
1292 def _xpath_ns(path, namespace=None):
1296 for c in path.split('/'):
1297 if not c or c == '.':
1300 out.append('{%s}%s' % (namespace, c))
1301 return '/'.join(out)
1303 def _extract_smil_formats(self, smil_url, video_id, fatal=True, f4m_params=None, transform_source=None):
1304 smil = self._download_smil(smil_url, video_id, fatal=fatal, transform_source=transform_source)
1310 namespace = self._parse_smil_namespace(smil)
1312 return self._parse_smil_formats(
1313 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
1315 def _extract_smil_info(self, smil_url, video_id, fatal=True, f4m_params=None):
1316 smil = self._download_smil(smil_url, video_id, fatal=fatal)
1319 return self._parse_smil(smil, smil_url, video_id, f4m_params=f4m_params)
1321 def _download_smil(self, smil_url, video_id, fatal=True, transform_source=None):
1322 return self._download_xml(
1323 smil_url, video_id, 'Downloading SMIL file',
1324 'Unable to download SMIL file', fatal=fatal, transform_source=transform_source)
1326 def _parse_smil(self, smil, smil_url, video_id, f4m_params=None):
1327 namespace = self._parse_smil_namespace(smil)
1329 formats = self._parse_smil_formats(
1330 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
1331 subtitles = self._parse_smil_subtitles(smil, namespace=namespace)
1333 video_id = os.path.splitext(url_basename(smil_url))[0]
1337 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
1338 name = meta.attrib.get('name')
1339 content = meta.attrib.get('content')
1340 if not name or not content:
1342 if not title and name == 'title':
1344 elif not description and name in ('description', 'abstract'):
1345 description = content
1346 elif not upload_date and name == 'date':
1347 upload_date = unified_strdate(content)
1350 'id': image.get('type'),
1351 'url': image.get('src'),
1352 'width': int_or_none(image.get('width')),
1353 'height': int_or_none(image.get('height')),
1354 } for image in smil.findall(self._xpath_ns('.//image', namespace)) if image.get('src')]
1358 'title': title or video_id,
1359 'description': description,
1360 'upload_date': upload_date,
1361 'thumbnails': thumbnails,
1363 'subtitles': subtitles,
1366 def _parse_smil_namespace(self, smil):
1367 return self._search_regex(
1368 r'(?i)^{([^}]+)?}smil$', smil.tag, 'namespace', default=None)
1370 def _parse_smil_formats(self, smil, smil_url, video_id, namespace=None, f4m_params=None, transform_rtmp_url=None):
1372 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
1373 b = meta.get('base') or meta.get('httpBase')
1384 media = smil.findall(self._xpath_ns('.//video', namespace)) + smil.findall(self._xpath_ns('.//audio', namespace))
1385 for medium in media:
1386 src = medium.get('src')
1387 if not src or src in srcs:
1391 bitrate = float_or_none(medium.get('system-bitrate') or medium.get('systemBitrate'), 1000)
1392 filesize = int_or_none(medium.get('size') or medium.get('fileSize'))
1393 width = int_or_none(medium.get('width'))
1394 height = int_or_none(medium.get('height'))
1395 proto = medium.get('proto')
1396 ext = medium.get('ext')
1397 src_ext = determine_ext(src)
1398 streamer = medium.get('streamer') or base
1400 if proto == 'rtmp' or streamer.startswith('rtmp'):
1406 'format_id': 'rtmp-%d' % (rtmp_count if bitrate is None else bitrate),
1408 'filesize': filesize,
1412 if transform_rtmp_url:
1413 streamer, src = transform_rtmp_url(streamer, src)
1414 formats[-1].update({
1420 src_url = src if src.startswith('http') else compat_urlparse.urljoin(base, src)
1421 src_url = src_url.strip()
1423 if proto == 'm3u8' or src_ext == 'm3u8':
1424 m3u8_formats = self._extract_m3u8_formats(
1425 src_url, video_id, ext or 'mp4', m3u8_id='hls', fatal=False)
1426 if len(m3u8_formats) == 1:
1428 m3u8_formats[0].update({
1429 'format_id': 'hls-%d' % (m3u8_count if bitrate is None else bitrate),
1434 formats.extend(m3u8_formats)
1437 if src_ext == 'f4m':
1442 'plugin': 'flowplayer-3.2.0.1',
1444 f4m_url += '&' if '?' in f4m_url else '?'
1445 f4m_url += compat_urllib_parse_urlencode(f4m_params)
1446 formats.extend(self._extract_f4m_formats(f4m_url, video_id, f4m_id='hds', fatal=False))
1449 if src_url.startswith('http') and self._is_valid_url(src, video_id):
1453 'ext': ext or src_ext or 'flv',
1454 'format_id': 'http-%d' % (bitrate or http_count),
1456 'filesize': filesize,
1464 def _parse_smil_subtitles(self, smil, namespace=None, subtitles_lang='en'):
1467 for num, textstream in enumerate(smil.findall(self._xpath_ns('.//textstream', namespace))):
1468 src = textstream.get('src')
1469 if not src or src in urls:
1472 ext = textstream.get('ext') or mimetype2ext(textstream.get('type')) or determine_ext(src)
1473 lang = textstream.get('systemLanguage') or textstream.get('systemLanguageName') or textstream.get('lang') or subtitles_lang
1474 subtitles.setdefault(lang, []).append({
1480 def _extract_xspf_playlist(self, playlist_url, playlist_id, fatal=True):
1481 xspf = self._download_xml(
1482 playlist_url, playlist_id, 'Downloading xpsf playlist',
1483 'Unable to download xspf manifest', fatal=fatal)
1486 return self._parse_xspf(xspf, playlist_id)
1488 def _parse_xspf(self, playlist, playlist_id):
1490 'xspf': 'http://xspf.org/ns/0/',
1491 's1': 'http://static.streamone.nl/player/ns/0',
1495 for track in playlist.findall(xpath_with_ns('./xspf:trackList/xspf:track', NS_MAP)):
1497 track, xpath_with_ns('./xspf:title', NS_MAP), 'title', default=playlist_id)
1498 description = xpath_text(
1499 track, xpath_with_ns('./xspf:annotation', NS_MAP), 'description')
1500 thumbnail = xpath_text(
1501 track, xpath_with_ns('./xspf:image', NS_MAP), 'thumbnail')
1502 duration = float_or_none(
1503 xpath_text(track, xpath_with_ns('./xspf:duration', NS_MAP), 'duration'), 1000)
1506 'url': location.text,
1507 'format_id': location.get(xpath_with_ns('s1:label', NS_MAP)),
1508 'width': int_or_none(location.get(xpath_with_ns('s1:width', NS_MAP))),
1509 'height': int_or_none(location.get(xpath_with_ns('s1:height', NS_MAP))),
1510 } for location in track.findall(xpath_with_ns('./xspf:location', NS_MAP))]
1511 self._sort_formats(formats)
1516 'description': description,
1517 'thumbnail': thumbnail,
1518 'duration': duration,
1523 def _extract_mpd_formats(self, mpd_url, video_id, mpd_id=None, note=None, errnote=None, fatal=True, formats_dict={}):
1524 res = self._download_webpage_handle(
1526 note=note or 'Downloading MPD manifest',
1527 errnote=errnote or 'Failed to download MPD manifest',
1532 mpd_base_url = re.match(r'https?://.+/', urlh.geturl()).group()
1534 return self._parse_mpd_formats(
1535 compat_etree_fromstring(mpd.encode('utf-8')), mpd_id, mpd_base_url,
1536 formats_dict=formats_dict, mpd_url=mpd_url)
1538 def _parse_mpd_formats(self, mpd_doc, mpd_id=None, mpd_base_url='', formats_dict={}, mpd_url=None):
1540 Parse formats from MPD manifest.
1542 1. MPEG-DASH Standard, ISO/IEC 23009-1:2014(E),
1543 http://standards.iso.org/ittf/PubliclyAvailableStandards/c065274_ISO_IEC_23009-1_2014.zip
1544 2. https://en.wikipedia.org/wiki/Dynamic_Adaptive_Streaming_over_HTTP
1546 if mpd_doc.get('type') == 'dynamic':
1549 namespace = self._search_regex(r'(?i)^{([^}]+)?}MPD$', mpd_doc.tag, 'namespace', default=None)
1552 return self._xpath_ns(path, namespace)
1554 def is_drm_protected(element):
1555 return element.find(_add_ns('ContentProtection')) is not None
1557 def extract_multisegment_info(element, ms_parent_info):
1558 ms_info = ms_parent_info.copy()
1560 # As per [1, 5.3.9.2.2] SegmentList and SegmentTemplate share some
1561 # common attributes and elements. We will only extract relevant
1563 def extract_common(source):
1564 segment_timeline = source.find(_add_ns('SegmentTimeline'))
1565 if segment_timeline is not None:
1566 s_e = segment_timeline.findall(_add_ns('S'))
1568 ms_info['total_number'] = 0
1571 r = int(s.get('r', 0))
1572 ms_info['total_number'] += 1 + r
1573 ms_info['s'].append({
1574 't': int(s.get('t', 0)),
1575 # @d is mandatory (see [1, 5.3.9.6.2, Table 17, page 60])
1576 'd': int(s.attrib['d']),
1579 start_number = source.get('startNumber')
1581 ms_info['start_number'] = int(start_number)
1582 timescale = source.get('timescale')
1584 ms_info['timescale'] = int(timescale)
1585 segment_duration = source.get('duration')
1586 if segment_duration:
1587 ms_info['segment_duration'] = int(segment_duration)
1589 def extract_Initialization(source):
1590 initialization = source.find(_add_ns('Initialization'))
1591 if initialization is not None:
1592 ms_info['initialization_url'] = initialization.attrib['sourceURL']
1594 segment_list = element.find(_add_ns('SegmentList'))
1595 if segment_list is not None:
1596 extract_common(segment_list)
1597 extract_Initialization(segment_list)
1598 segment_urls_e = segment_list.findall(_add_ns('SegmentURL'))
1600 ms_info['segment_urls'] = [segment.attrib['media'] for segment in segment_urls_e]
1602 segment_template = element.find(_add_ns('SegmentTemplate'))
1603 if segment_template is not None:
1604 extract_common(segment_template)
1605 media_template = segment_template.get('media')
1607 ms_info['media_template'] = media_template
1608 initialization = segment_template.get('initialization')
1610 ms_info['initialization_url'] = initialization
1612 extract_Initialization(segment_template)
1615 def combine_url(base_url, target_url):
1616 if re.match(r'^https?://', target_url):
1618 return '%s%s%s' % (base_url, '' if base_url.endswith('/') else '/', target_url)
1620 mpd_duration = parse_duration(mpd_doc.get('mediaPresentationDuration'))
1622 for period in mpd_doc.findall(_add_ns('Period')):
1623 period_duration = parse_duration(period.get('duration')) or mpd_duration
1624 period_ms_info = extract_multisegment_info(period, {
1628 for adaptation_set in period.findall(_add_ns('AdaptationSet')):
1629 if is_drm_protected(adaptation_set):
1631 adaption_set_ms_info = extract_multisegment_info(adaptation_set, period_ms_info)
1632 for representation in adaptation_set.findall(_add_ns('Representation')):
1633 if is_drm_protected(representation):
1635 representation_attrib = adaptation_set.attrib.copy()
1636 representation_attrib.update(representation.attrib)
1637 # According to [1, 5.3.7.2, Table 9, page 41], @mimeType is mandatory
1638 mime_type = representation_attrib['mimeType']
1639 content_type = mime_type.split('/')[0]
1640 if content_type == 'text':
1641 # TODO implement WebVTT downloading
1643 elif content_type == 'video' or content_type == 'audio':
1645 for element in (representation, adaptation_set, period, mpd_doc):
1646 base_url_e = element.find(_add_ns('BaseURL'))
1647 if base_url_e is not None:
1648 base_url = base_url_e.text + base_url
1649 if re.match(r'^https?://', base_url):
1651 if mpd_base_url and not re.match(r'^https?://', base_url):
1652 if not mpd_base_url.endswith('/') and not base_url.startswith('/'):
1654 base_url = mpd_base_url + base_url
1655 representation_id = representation_attrib.get('id')
1656 lang = representation_attrib.get('lang')
1657 url_el = representation.find(_add_ns('BaseURL'))
1658 filesize = int_or_none(url_el.attrib.get('{http://youtube.com/yt/2012/10/10}contentLength') if url_el is not None else None)
1660 'format_id': '%s-%s' % (mpd_id, representation_id) if mpd_id else representation_id,
1662 'manifest_url': mpd_url,
1663 'ext': mimetype2ext(mime_type),
1664 'width': int_or_none(representation_attrib.get('width')),
1665 'height': int_or_none(representation_attrib.get('height')),
1666 'tbr': int_or_none(representation_attrib.get('bandwidth'), 1000),
1667 'asr': int_or_none(representation_attrib.get('audioSamplingRate')),
1668 'fps': int_or_none(representation_attrib.get('frameRate')),
1669 'vcodec': 'none' if content_type == 'audio' else representation_attrib.get('codecs'),
1670 'acodec': 'none' if content_type == 'video' else representation_attrib.get('codecs'),
1671 'language': lang if lang not in ('mul', 'und', 'zxx', 'mis') else None,
1672 'format_note': 'DASH %s' % content_type,
1673 'filesize': filesize,
1675 representation_ms_info = extract_multisegment_info(representation, adaption_set_ms_info)
1676 if 'segment_urls' not in representation_ms_info and 'media_template' in representation_ms_info:
1678 media_template = representation_ms_info['media_template']
1679 media_template = media_template.replace('$RepresentationID$', representation_id)
1680 media_template = re.sub(r'\$(Number|Bandwidth|Time)\$', r'%(\1)d', media_template)
1681 media_template = re.sub(r'\$(Number|Bandwidth|Time)%([^$]+)\$', r'%(\1)\2', media_template)
1682 media_template.replace('$$', '$')
1684 # As per [1, 5.3.9.4.4, Table 16, page 55] $Number$ and $Time$
1685 # can't be used at the same time
1686 if '%(Number' in media_template and 's' not in representation_ms_info:
1687 segment_duration = None
1688 if 'total_number' not in representation_ms_info and 'segment_duration':
1689 segment_duration = float_or_none(representation_ms_info['segment_duration'], representation_ms_info['timescale'])
1690 representation_ms_info['total_number'] = int(math.ceil(float(period_duration) / segment_duration))
1691 representation_ms_info['fragments'] = [{
1692 'url': media_template % {
1693 'Number': segment_number,
1694 'Bandwidth': representation_attrib.get('bandwidth'),
1696 'duration': segment_duration,
1697 } for segment_number in range(
1698 representation_ms_info['start_number'],
1699 representation_ms_info['total_number'] + representation_ms_info['start_number'])]
1701 # $Number*$ or $Time$ in media template with S list available
1702 # Example $Number*$: http://www.svtplay.se/klipp/9023742/stopptid-om-bjorn-borg
1703 # Example $Time$: https://play.arkena.com/embed/avp/v2/player/media/b41dda37-d8e7-4d3f-b1b5-9a9db578bdfe/1/129411
1704 representation_ms_info['fragments'] = []
1707 segment_number = representation_ms_info['start_number']
1709 def add_segment_url():
1710 segment_url = media_template % {
1711 'Time': segment_time,
1712 'Bandwidth': representation_attrib.get('bandwidth'),
1713 'Number': segment_number,
1715 representation_ms_info['fragments'].append({
1717 'duration': float_or_none(segment_d, representation_ms_info['timescale']),
1720 for num, s in enumerate(representation_ms_info['s']):
1721 segment_time = s.get('t') or segment_time
1725 for r in range(s.get('r', 0)):
1726 segment_time += segment_d
1729 segment_time += segment_d
1730 elif 'segment_urls' in representation_ms_info and 's' in representation_ms_info:
1732 # Example: https://www.youtube.com/watch?v=iXZV5uAYMJI
1733 # or any YouTube dashsegments video
1736 for segment_url in representation_ms_info['segment_urls']:
1737 s = representation_ms_info['s'][s_num]
1738 for r in range(s.get('r', 0) + 1):
1741 'duration': float_or_none(s['d'], representation_ms_info['timescale']),
1743 representation_ms_info['fragments'] = fragments
1744 # NB: MPD manifest may contain direct URLs to unfragmented media.
1745 # No fragments key is present in this case.
1746 if 'fragments' in representation_ms_info:
1749 'protocol': 'http_dash_segments',
1751 if 'initialization_url' in representation_ms_info:
1752 initialization_url = representation_ms_info['initialization_url'].replace('$RepresentationID$', representation_id)
1753 if not f.get('url'):
1754 f['url'] = initialization_url
1755 f['fragments'].append({'url': initialization_url})
1756 f['fragments'].extend(representation_ms_info['fragments'])
1757 for fragment in f['fragments']:
1758 fragment['url'] = combine_url(base_url, fragment['url'])
1760 existing_format = next(
1761 fo for fo in formats
1762 if fo['format_id'] == representation_id)
1763 except StopIteration:
1764 full_info = formats_dict.get(representation_id, {}).copy()
1766 formats.append(full_info)
1768 existing_format.update(f)
1770 self.report_warning('Unknown MIME type %s in DASH manifest' % mime_type)
1773 def _parse_html5_media_entries(self, base_url, webpage, video_id, m3u8_id=None, m3u8_entry_protocol='m3u8'):
1774 def absolute_url(video_url):
1775 return compat_urlparse.urljoin(base_url, video_url)
1777 def parse_content_type(content_type):
1778 if not content_type:
1780 ctr = re.search(r'(?P<mimetype>[^/]+/[^;]+)(?:;\s*codecs="?(?P<codecs>[^"]+))?', content_type)
1782 mimetype, codecs = ctr.groups()
1783 f = parse_codecs(codecs)
1784 f['ext'] = mimetype2ext(mimetype)
1788 def _media_formats(src, cur_media_type):
1789 full_url = absolute_url(src)
1790 if determine_ext(full_url) == 'm3u8':
1791 is_plain_url = False
1792 formats = self._extract_m3u8_formats(
1793 full_url, video_id, ext='mp4',
1794 entry_protocol=m3u8_entry_protocol, m3u8_id=m3u8_id)
1799 'vcodec': 'none' if cur_media_type == 'audio' else None,
1801 return is_plain_url, formats
1804 for media_tag, media_type, media_content in re.findall(r'(?s)(<(?P<tag>video|audio)[^>]*>)(.*?)</(?P=tag)>', webpage):
1809 media_attributes = extract_attributes(media_tag)
1810 src = media_attributes.get('src')
1812 _, formats = _media_formats(src, media_type)
1813 media_info['formats'].extend(formats)
1814 media_info['thumbnail'] = media_attributes.get('poster')
1816 for source_tag in re.findall(r'<source[^>]+>', media_content):
1817 source_attributes = extract_attributes(source_tag)
1818 src = source_attributes.get('src')
1821 is_plain_url, formats = _media_formats(src, media_type)
1823 f = parse_content_type(source_attributes.get('type'))
1824 f.update(formats[0])
1825 media_info['formats'].append(f)
1827 media_info['formats'].extend(formats)
1828 for track_tag in re.findall(r'<track[^>]+>', media_content):
1829 track_attributes = extract_attributes(track_tag)
1830 kind = track_attributes.get('kind')
1831 if not kind or kind in ('subtitles', 'captions'):
1832 src = track_attributes.get('src')
1835 lang = track_attributes.get('srclang') or track_attributes.get('lang') or track_attributes.get('label')
1836 media_info['subtitles'].setdefault(lang, []).append({
1837 'url': absolute_url(src),
1839 if media_info['formats'] or media_info['subtitles']:
1840 entries.append(media_info)
1843 def _extract_akamai_formats(self, manifest_url, video_id):
1845 hdcore_sign = 'hdcore=3.7.0'
1846 f4m_url = re.sub(r'(https?://.+?)/i/', r'\1/z/', manifest_url).replace('/master.m3u8', '/manifest.f4m')
1847 if 'hdcore=' not in f4m_url:
1848 f4m_url += ('&' if '?' in f4m_url else '?') + hdcore_sign
1849 f4m_formats = self._extract_f4m_formats(
1850 f4m_url, video_id, f4m_id='hds', fatal=False)
1851 for entry in f4m_formats:
1852 entry.update({'extra_param_to_segment_url': hdcore_sign})
1853 formats.extend(f4m_formats)
1854 m3u8_url = re.sub(r'(https?://.+?)/z/', r'\1/i/', manifest_url).replace('/manifest.f4m', '/master.m3u8')
1855 formats.extend(self._extract_m3u8_formats(
1856 m3u8_url, video_id, 'mp4', 'm3u8_native',
1857 m3u8_id='hls', fatal=False))
1860 def _extract_wowza_formats(self, url, video_id, m3u8_entry_protocol='m3u8_native', skip_protocols=[]):
1861 url = re.sub(r'/(?:manifest|playlist|jwplayer)\.(?:m3u8|f4m|mpd|smil)', '', url)
1862 url_base = self._search_regex(r'(?:https?|rtmp|rtsp)(://[^?]+)', url, 'format url')
1863 http_base_url = 'http' + url_base
1865 if 'm3u8' not in skip_protocols:
1866 formats.extend(self._extract_m3u8_formats(
1867 http_base_url + '/playlist.m3u8', video_id, 'mp4',
1868 m3u8_entry_protocol, m3u8_id='hls', fatal=False))
1869 if 'f4m' not in skip_protocols:
1870 formats.extend(self._extract_f4m_formats(
1871 http_base_url + '/manifest.f4m',
1872 video_id, f4m_id='hds', fatal=False))
1873 if re.search(r'(?:/smil:|\.smil)', url_base):
1874 if 'dash' not in skip_protocols:
1875 formats.extend(self._extract_mpd_formats(
1876 http_base_url + '/manifest.mpd',
1877 video_id, mpd_id='dash', fatal=False))
1878 if 'smil' not in skip_protocols:
1879 rtmp_formats = self._extract_smil_formats(
1880 http_base_url + '/jwplayer.smil',
1881 video_id, fatal=False)
1882 for rtmp_format in rtmp_formats:
1883 rtsp_format = rtmp_format.copy()
1884 rtsp_format['url'] = '%s/%s' % (rtmp_format['url'], rtmp_format['play_path'])
1885 del rtsp_format['play_path']
1886 del rtsp_format['ext']
1887 rtsp_format.update({
1888 'url': rtsp_format['url'].replace('rtmp://', 'rtsp://'),
1889 'format_id': rtmp_format['format_id'].replace('rtmp', 'rtsp'),
1892 formats.extend([rtmp_format, rtsp_format])
1894 for protocol in ('rtmp', 'rtsp'):
1895 if protocol not in skip_protocols:
1897 'url': protocol + url_base,
1898 'format_id': protocol,
1899 'protocol': protocol,
1903 def _live_title(self, name):
1904 """ Generate the title for a live video """
1905 now = datetime.datetime.now()
1906 now_str = now.strftime('%Y-%m-%d %H:%M')
1907 return name + ' ' + now_str
1909 def _int(self, v, name, fatal=False, **kwargs):
1910 res = int_or_none(v, **kwargs)
1911 if 'get_attr' in kwargs:
1912 print(getattr(v, kwargs['get_attr']))
1914 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
1916 raise ExtractorError(msg)
1918 self._downloader.report_warning(msg)
1921 def _float(self, v, name, fatal=False, **kwargs):
1922 res = float_or_none(v, **kwargs)
1924 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
1926 raise ExtractorError(msg)
1928 self._downloader.report_warning(msg)
1931 def _set_cookie(self, domain, name, value, expire_time=None):
1932 cookie = compat_cookiejar.Cookie(
1933 0, name, value, None, None, domain, None,
1934 None, '/', True, False, expire_time, '', None, None, None)
1935 self._downloader.cookiejar.set_cookie(cookie)
1937 def _get_cookies(self, url):
1938 """ Return a compat_cookies.SimpleCookie with the cookies for the url """
1939 req = sanitized_Request(url)
1940 self._downloader.cookiejar.add_cookie_header(req)
1941 return compat_cookies.SimpleCookie(req.get_header('Cookie'))
1943 def get_testcases(self, include_onlymatching=False):
1944 t = getattr(self, '_TEST', None)
1946 assert not hasattr(self, '_TESTS'), \
1947 '%s has _TEST and _TESTS' % type(self).__name__
1950 tests = getattr(self, '_TESTS', [])
1952 if not include_onlymatching and t.get('only_matching', False):
1954 t['name'] = type(self).__name__[:-len('IE')]
1957 def is_suitable(self, age_limit):
1958 """ Test whether the extractor is generally suitable for the given
1959 age limit (i.e. pornographic sites are not, all others usually are) """
1961 any_restricted = False
1962 for tc in self.get_testcases(include_onlymatching=False):
1963 if tc.get('playlist', []):
1964 tc = tc['playlist'][0]
1965 is_restricted = age_restricted(
1966 tc.get('info_dict', {}).get('age_limit'), age_limit)
1967 if not is_restricted:
1969 any_restricted = any_restricted or is_restricted
1970 return not any_restricted
1972 def extract_subtitles(self, *args, **kwargs):
1973 if (self._downloader.params.get('writesubtitles', False) or
1974 self._downloader.params.get('listsubtitles')):
1975 return self._get_subtitles(*args, **kwargs)
1978 def _get_subtitles(self, *args, **kwargs):
1979 raise NotImplementedError('This method must be implemented by subclasses')
1982 def _merge_subtitle_items(subtitle_list1, subtitle_list2):
1983 """ Merge subtitle items for one language. Items with duplicated URLs
1984 will be dropped. """
1985 list1_urls = set([item['url'] for item in subtitle_list1])
1986 ret = list(subtitle_list1)
1987 ret.extend([item for item in subtitle_list2 if item['url'] not in list1_urls])
1991 def _merge_subtitles(cls, subtitle_dict1, subtitle_dict2):
1992 """ Merge two subtitle dictionaries, language by language. """
1993 ret = dict(subtitle_dict1)
1994 for lang in subtitle_dict2:
1995 ret[lang] = cls._merge_subtitle_items(subtitle_dict1.get(lang, []), subtitle_dict2[lang])
1998 def extract_automatic_captions(self, *args, **kwargs):
1999 if (self._downloader.params.get('writeautomaticsub', False) or
2000 self._downloader.params.get('listsubtitles')):
2001 return self._get_automatic_captions(*args, **kwargs)
2004 def _get_automatic_captions(self, *args, **kwargs):
2005 raise NotImplementedError('This method must be implemented by subclasses')
2007 def mark_watched(self, *args, **kwargs):
2008 if (self._downloader.params.get('mark_watched', False) and
2009 (self._get_login_info()[0] is not None or
2010 self._downloader.params.get('cookiefile') is not None)):
2011 self._mark_watched(*args, **kwargs)
2013 def _mark_watched(self, *args, **kwargs):
2014 raise NotImplementedError('This method must be implemented by subclasses')
2016 def geo_verification_headers(self):
2018 geo_verification_proxy = self._downloader.params.get('geo_verification_proxy')
2019 if geo_verification_proxy:
2020 headers['Ytdl-request-proxy'] = geo_verification_proxy
2024 class SearchInfoExtractor(InfoExtractor):
2026 Base class for paged search queries extractors.
2027 They accept URLs in the format _SEARCH_KEY(|all|[0-9]):{query}
2028 Instances should define _SEARCH_KEY and _MAX_RESULTS.
2032 def _make_valid_url(cls):
2033 return r'%s(?P<prefix>|[1-9][0-9]*|all):(?P<query>[\s\S]+)' % cls._SEARCH_KEY
2036 def suitable(cls, url):
2037 return re.match(cls._make_valid_url(), url) is not None
2039 def _real_extract(self, query):
2040 mobj = re.match(self._make_valid_url(), query)
2042 raise ExtractorError('Invalid search query "%s"' % query)
2044 prefix = mobj.group('prefix')
2045 query = mobj.group('query')
2047 return self._get_n_results(query, 1)
2048 elif prefix == 'all':
2049 return self._get_n_results(query, self._MAX_RESULTS)
2053 raise ExtractorError('invalid download number %s for query "%s"' % (n, query))
2054 elif n > self._MAX_RESULTS:
2055 self._downloader.report_warning('%s returns max %i results (you requested %i)' % (self._SEARCH_KEY, self._MAX_RESULTS, n))
2056 n = self._MAX_RESULTS
2057 return self._get_n_results(query, n)
2059 def _get_n_results(self, query, n):
2060 """Get a specified number of results for a query"""
2061 raise NotImplementedError('This method must be implemented by subclasses')
2064 def SEARCH_KEY(self):
2065 return self._SEARCH_KEY