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 or URL of
90 the manifest file in case of fragmented media
92 * ext Will be calculated from URL if missing
93 * format A human-readable description of the format
94 ("mp4 container with h264/opus").
95 Calculated from the format_id, width, height.
96 and format_note fields if missing.
97 * format_id A short description of the format
98 ("mp4_h264_opus" or "19").
99 Technically optional, but strongly recommended.
100 * format_note Additional info about the format
101 ("3D" or "DASH video")
102 * width Width of the video, if known
103 * height Height of the video, if known
104 * resolution Textual description of width and height
105 * tbr Average bitrate of audio and video in KBit/s
106 * abr Average audio bitrate in KBit/s
107 * acodec Name of the audio codec in use
108 * asr Audio sampling rate in Hertz
109 * vbr Average video bitrate in KBit/s
111 * vcodec Name of the video codec in use
112 * container Name of the container format
113 * filesize The number of bytes, if known in advance
114 * filesize_approx An estimate for the number of bytes
115 * player_url SWF Player URL (used for rtmpdump).
116 * protocol The protocol that will be used for the actual
117 download, lower-case.
118 "http", "https", "rtsp", "rtmp", "rtmpe",
119 "m3u8", "m3u8_native" or "http_dash_segments".
120 * fragments A list of fragments of the fragmented media,
121 with the following entries:
122 * "url" (mandatory) - fragment's URL
123 * "duration" (optional, int or float)
124 * "filesize" (optional, int)
125 * preference Order number of this format. If this field is
126 present and not None, the formats get sorted
127 by this field, regardless of all other values.
128 -1 for default (order by other properties),
129 -2 or smaller for less than default.
130 < -1000 to hide the format (if there is
131 another one which is strictly better)
132 * language Language code, e.g. "de" or "en-US".
133 * language_preference Is this in the language mentioned in
135 10 if it's what the URL is about,
136 -1 for default (don't know),
137 -10 otherwise, other values reserved for now.
138 * quality Order number of the video quality of this
139 format, irrespective of the file format.
140 -1 for default (order by other properties),
141 -2 or smaller for less than default.
142 * source_preference Order number for this video source
143 (quality takes higher priority)
144 -1 for default (order by other properties),
145 -2 or smaller for less than default.
146 * http_headers A dictionary of additional HTTP headers
147 to add to the request.
148 * stretched_ratio If given and not 1, indicates that the
149 video's pixels are not square.
150 width : height ratio as float.
151 * no_resume The server does not support resuming the
152 (HTTP or RTMP) download. Boolean.
154 url: Final video URL.
155 ext: Video filename extension.
156 format: The video format, defaults to ext (used for --get-format)
157 player_url: SWF Player URL (used for rtmpdump).
159 The following fields are optional:
161 alt_title: A secondary title of the video.
162 display_id An alternative identifier for the video, not necessarily
163 unique, but available before title. Typically, id is
164 something like "4234987", title "Dancing naked mole rats",
165 and display_id "dancing-naked-mole-rats"
166 thumbnails: A list of dictionaries, with the following entries:
167 * "id" (optional, string) - Thumbnail format ID
169 * "preference" (optional, int) - quality of the image
170 * "width" (optional, int)
171 * "height" (optional, int)
172 * "resolution" (optional, string "{width}x{height"},
174 * "filesize" (optional, int)
175 thumbnail: Full URL to a video thumbnail image.
176 description: Full video description.
177 uploader: Full name of the video uploader.
178 license: License name the video is licensed under.
179 creator: The creator of the video.
180 release_date: The date (YYYYMMDD) when the video was released.
181 timestamp: UNIX timestamp of the moment the video became available.
182 upload_date: Video upload date (YYYYMMDD).
183 If not explicitly set, calculated from timestamp.
184 uploader_id: Nickname or id of the video uploader.
185 uploader_url: Full URL to a personal webpage of the video uploader.
186 location: Physical location where the video was filmed.
187 subtitles: The available subtitles as a dictionary in the format
188 {language: subformats}. "subformats" is a list sorted from
189 lower to higher preference, each element is a dictionary
190 with the "ext" entry and one of:
191 * "data": The subtitles file contents
192 * "url": A URL pointing to the subtitles file
193 "ext" will be calculated from URL if missing
194 automatic_captions: Like 'subtitles', used by the YoutubeIE for
195 automatically generated captions
196 duration: Length of the video in seconds, as an integer or float.
197 view_count: How many users have watched the video on the platform.
198 like_count: Number of positive ratings of the video
199 dislike_count: Number of negative ratings of the video
200 repost_count: Number of reposts of the video
201 average_rating: Average rating give by users, the scale used depends on the webpage
202 comment_count: Number of comments on the video
203 comments: A list of comments, each with one or more of the following
204 properties (all but one of text or html optional):
205 * "author" - human-readable name of the comment author
206 * "author_id" - user ID of the comment author
208 * "html" - Comment as HTML
209 * "text" - Plain text of the comment
210 * "timestamp" - UNIX timestamp of comment
211 * "parent" - ID of the comment this one is replying to.
212 Set to "root" to indicate that this is a
213 comment to the original video.
214 age_limit: Age restriction for the video, as an integer (years)
215 webpage_url: The URL to the video webpage, if given to youtube-dl it
216 should allow to get the same result again. (It will be set
217 by YoutubeDL if it's missing)
218 categories: A list of categories that the video falls in, for example
220 tags: A list of tags assigned to the video, e.g. ["sweden", "pop music"]
221 is_live: True, False, or None (=unknown). Whether this video is a
222 live stream that goes on instead of a fixed-length video.
223 start_time: Time in seconds where the reproduction should start, as
224 specified in the URL.
225 end_time: Time in seconds where the reproduction should end, as
226 specified in the URL.
228 The following fields should only be used when the video belongs to some logical
231 chapter: Name or title of the chapter the video belongs to.
232 chapter_number: Number of the chapter the video belongs to, as an integer.
233 chapter_id: Id of the chapter the video belongs to, as a unicode string.
235 The following fields should only be used when the video is an episode of some
238 series: Title of the series or programme the video episode belongs to.
239 season: Title of the season the video episode belongs to.
240 season_number: Number of the season the video episode belongs to, as an integer.
241 season_id: Id of the season the video episode belongs to, as a unicode string.
242 episode: Title of the video episode. Unlike mandatory video title field,
243 this field should denote the exact title of the video episode
244 without any kind of decoration.
245 episode_number: Number of the video episode within a season, as an integer.
246 episode_id: Id of the video episode, as a unicode string.
248 The following fields should only be used when the media is a track or a part of
251 track: Title of the track.
252 track_number: Number of the track within an album or a disc, as an integer.
253 track_id: Id of the track (useful in case of custom indexing, e.g. 6.iii),
255 artist: Artist(s) of the track.
256 genre: Genre(s) of the track.
257 album: Title of the album the track belongs to.
258 album_type: Type of the album (e.g. "Demo", "Full-length", "Split", "Compilation", etc).
259 album_artist: List of all artists appeared on the album (e.g.
260 "Ash Borer / Fell Voices" or "Various Artists", useful for splits
262 disc_number: Number of the disc or other physical medium the track belongs to,
264 release_year: Year (YYYY) when the album was released.
266 Unless mentioned otherwise, the fields should be Unicode strings.
268 Unless mentioned otherwise, None is equivalent to absence of information.
271 _type "playlist" indicates multiple videos.
272 There must be a key "entries", which is a list, an iterable, or a PagedList
273 object, each element of which is a valid dictionary by this specification.
275 Additionally, playlists can have "title", "description" and "id" attributes
276 with the same semantics as videos (see above).
279 _type "multi_video" indicates that there are multiple videos that
280 form a single show, for examples multiple acts of an opera or TV episode.
281 It must have an entries key like a playlist and contain all the keys
282 required for a video at the same time.
285 _type "url" indicates that the video must be extracted from another
286 location, possibly by a different extractor. Its only required key is:
287 "url" - the next URL to extract.
288 The key "ie_key" can be set to the class name (minus the trailing "IE",
289 e.g. "Youtube") if the extractor class is known in advance.
290 Additionally, the dictionary may have any properties of the resolved entity
291 known in advance, for example "title" if the title of the referred video is
295 _type "url_transparent" entities have the same specification as "url", but
296 indicate that the given additional information is more precise than the one
297 associated with the resolved URL.
298 This is useful when a site employs a video service that hosts the video and
299 its technical metadata, but that video service does not embed a useful
300 title, description etc.
303 Subclasses of this one should re-define the _real_initialize() and
304 _real_extract() methods and define a _VALID_URL regexp.
305 Probably, they should also be added to the list of extractors.
307 Finally, the _WORKING attribute should be set to False for broken IEs
308 in order to warn the users and skip the tests.
315 def __init__(self, downloader=None):
316 """Constructor. Receives an optional downloader."""
318 self.set_downloader(downloader)
321 def suitable(cls, url):
322 """Receives a URL and returns True if suitable for this IE."""
324 # This does not use has/getattr intentionally - we want to know whether
325 # we have cached the regexp for *this* class, whereas getattr would also
326 # match the superclass
327 if '_VALID_URL_RE' not in cls.__dict__:
328 cls._VALID_URL_RE = re.compile(cls._VALID_URL)
329 return cls._VALID_URL_RE.match(url) is not None
332 def _match_id(cls, url):
333 if '_VALID_URL_RE' not in cls.__dict__:
334 cls._VALID_URL_RE = re.compile(cls._VALID_URL)
335 m = cls._VALID_URL_RE.match(url)
341 """Getter method for _WORKING."""
344 def initialize(self):
345 """Initializes an instance (authentication, etc)."""
347 self._real_initialize()
350 def extract(self, url):
351 """Extracts URL information and returns it in list of dicts."""
354 return self._real_extract(url)
355 except ExtractorError:
357 except compat_http_client.IncompleteRead as e:
358 raise ExtractorError('A network error has occurred.', cause=e, expected=True)
359 except (KeyError, StopIteration) as e:
360 raise ExtractorError('An extractor error has occurred.', cause=e)
362 def set_downloader(self, downloader):
363 """Sets the downloader for this IE."""
364 self._downloader = downloader
366 def _real_initialize(self):
367 """Real initialization process. Redefine in subclasses."""
370 def _real_extract(self, url):
371 """Real extraction process. Redefine in subclasses."""
376 """A string for getting the InfoExtractor with get_info_extractor"""
377 return compat_str(cls.__name__[:-2])
381 return compat_str(type(self).__name__[:-2])
383 def _request_webpage(self, url_or_request, video_id, note=None, errnote=None, fatal=True, data=None, headers={}, query={}):
384 """ Returns the response handle """
386 self.report_download_webpage(video_id)
387 elif note is not False:
389 self.to_screen('%s' % (note,))
391 self.to_screen('%s: %s' % (video_id, note))
392 if isinstance(url_or_request, compat_urllib_request.Request):
393 url_or_request = update_Request(
394 url_or_request, data=data, headers=headers, query=query)
397 url_or_request = update_url_query(url_or_request, query)
398 if data is not None or headers:
399 url_or_request = sanitized_Request(url_or_request, data, headers)
401 return self._downloader.urlopen(url_or_request)
402 except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err:
406 errnote = 'Unable to download webpage'
408 errmsg = '%s: %s' % (errnote, error_to_compat_str(err))
410 raise ExtractorError(errmsg, sys.exc_info()[2], cause=err)
412 self._downloader.report_warning(errmsg)
415 def _download_webpage_handle(self, url_or_request, video_id, note=None, errnote=None, fatal=True, encoding=None, data=None, headers={}, query={}):
416 """ Returns a tuple (page content as string, URL handle) """
417 # Strip hashes from the URL (#1038)
418 if isinstance(url_or_request, (compat_str, str)):
419 url_or_request = url_or_request.partition('#')[0]
421 urlh = self._request_webpage(url_or_request, video_id, note, errnote, fatal, data=data, headers=headers, query=query)
425 content = self._webpage_read_content(urlh, url_or_request, video_id, note, errnote, fatal, encoding=encoding)
426 return (content, urlh)
429 def _guess_encoding_from_content(content_type, webpage_bytes):
430 m = re.match(r'[a-zA-Z0-9_.-]+/[a-zA-Z0-9_.-]+\s*;\s*charset=(.+)', content_type)
432 encoding = m.group(1)
434 m = re.search(br'<meta[^>]+charset=[\'"]?([^\'")]+)[ /\'">]',
435 webpage_bytes[:1024])
437 encoding = m.group(1).decode('ascii')
438 elif webpage_bytes.startswith(b'\xff\xfe'):
445 def _webpage_read_content(self, urlh, url_or_request, video_id, note=None, errnote=None, fatal=True, prefix=None, encoding=None):
446 content_type = urlh.headers.get('Content-Type', '')
447 webpage_bytes = urlh.read()
448 if prefix is not None:
449 webpage_bytes = prefix + webpage_bytes
451 encoding = self._guess_encoding_from_content(content_type, webpage_bytes)
452 if self._downloader.params.get('dump_intermediate_pages', False):
454 url = url_or_request.get_full_url()
455 except AttributeError:
457 self.to_screen('Dumping request to ' + url)
458 dump = base64.b64encode(webpage_bytes).decode('ascii')
459 self._downloader.to_screen(dump)
460 if self._downloader.params.get('write_pages', False):
462 url = url_or_request.get_full_url()
463 except AttributeError:
465 basen = '%s_%s' % (video_id, url)
467 h = '___' + hashlib.md5(basen.encode('utf-8')).hexdigest()
468 basen = basen[:240 - len(h)] + h
469 raw_filename = basen + '.dump'
470 filename = sanitize_filename(raw_filename, restricted=True)
471 self.to_screen('Saving request to ' + filename)
472 # Working around MAX_PATH limitation on Windows (see
473 # http://msdn.microsoft.com/en-us/library/windows/desktop/aa365247(v=vs.85).aspx)
474 if compat_os_name == 'nt':
475 absfilepath = os.path.abspath(filename)
476 if len(absfilepath) > 259:
477 filename = '\\\\?\\' + absfilepath
478 with open(filename, 'wb') as outf:
479 outf.write(webpage_bytes)
482 content = webpage_bytes.decode(encoding, 'replace')
484 content = webpage_bytes.decode('utf-8', 'replace')
486 if ('<title>Access to this site is blocked</title>' in content and
487 'Websense' in content[:512]):
488 msg = 'Access to this webpage has been blocked by Websense filtering software in your network.'
489 blocked_iframe = self._html_search_regex(
490 r'<iframe src="([^"]+)"', content,
491 'Websense information URL', default=None)
493 msg += ' Visit %s for more details' % blocked_iframe
494 raise ExtractorError(msg, expected=True)
495 if '<title>The URL you requested has been blocked</title>' in content[:512]:
497 'Access to this webpage has been blocked by Indian censorship. '
498 'Use a VPN or proxy server (with --proxy) to route around it.')
499 block_msg = self._html_search_regex(
500 r'</h1><p>(.*?)</p>',
501 content, 'block message', default=None)
503 msg += ' (Message: "%s")' % block_msg.replace('\n', ' ')
504 raise ExtractorError(msg, expected=True)
508 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={}):
509 """ Returns the data of the page as a string """
512 while success is False:
514 res = self._download_webpage_handle(url_or_request, video_id, note, errnote, fatal, encoding=encoding, data=data, headers=headers, query=query)
516 except compat_http_client.IncompleteRead as e:
518 if try_count >= tries:
520 self._sleep(timeout, video_id)
527 def _download_xml(self, url_or_request, video_id,
528 note='Downloading XML', errnote='Unable to download XML',
529 transform_source=None, fatal=True, encoding=None, data=None, headers={}, query={}):
530 """Return the xml as an xml.etree.ElementTree.Element"""
531 xml_string = self._download_webpage(
532 url_or_request, video_id, note, errnote, fatal=fatal, encoding=encoding, data=data, headers=headers, query=query)
533 if xml_string is False:
536 xml_string = transform_source(xml_string)
537 return compat_etree_fromstring(xml_string.encode('utf-8'))
539 def _download_json(self, url_or_request, video_id,
540 note='Downloading JSON metadata',
541 errnote='Unable to download JSON metadata',
542 transform_source=None,
543 fatal=True, encoding=None, data=None, headers={}, query={}):
544 json_string = self._download_webpage(
545 url_or_request, video_id, note, errnote, fatal=fatal,
546 encoding=encoding, data=data, headers=headers, query=query)
547 if (not fatal) and json_string is False:
549 return self._parse_json(
550 json_string, video_id, transform_source=transform_source, fatal=fatal)
552 def _parse_json(self, json_string, video_id, transform_source=None, fatal=True):
554 json_string = transform_source(json_string)
556 return json.loads(json_string)
557 except ValueError as ve:
558 errmsg = '%s: Failed to parse JSON ' % video_id
560 raise ExtractorError(errmsg, cause=ve)
562 self.report_warning(errmsg + str(ve))
564 def report_warning(self, msg, video_id=None):
565 idstr = '' if video_id is None else '%s: ' % video_id
566 self._downloader.report_warning(
567 '[%s] %s%s' % (self.IE_NAME, idstr, msg))
569 def to_screen(self, msg):
570 """Print msg to screen, prefixing it with '[ie_name]'"""
571 self._downloader.to_screen('[%s] %s' % (self.IE_NAME, msg))
573 def report_extraction(self, id_or_name):
574 """Report information extraction."""
575 self.to_screen('%s: Extracting information' % id_or_name)
577 def report_download_webpage(self, video_id):
578 """Report webpage download."""
579 self.to_screen('%s: Downloading webpage' % video_id)
581 def report_age_confirmation(self):
582 """Report attempt to confirm age."""
583 self.to_screen('Confirming age')
585 def report_login(self):
586 """Report attempt to log in."""
587 self.to_screen('Logging in')
590 def raise_login_required(msg='This video is only available for registered users'):
591 raise ExtractorError(
592 '%s. Use --username and --password or --netrc to provide account credentials.' % msg,
596 def raise_geo_restricted(msg='This video is not available from your location due to geo restriction'):
597 raise ExtractorError(
598 '%s. You might want to use --proxy to workaround.' % msg,
601 # Methods for following #608
603 def url_result(url, ie=None, video_id=None, video_title=None):
604 """Returns a URL that points to a page that should be processed"""
605 # TODO: ie should be the class used for getting the info
606 video_info = {'_type': 'url',
609 if video_id is not None:
610 video_info['id'] = video_id
611 if video_title is not None:
612 video_info['title'] = video_title
616 def playlist_result(entries, playlist_id=None, playlist_title=None, playlist_description=None):
617 """Returns a playlist"""
618 video_info = {'_type': 'playlist',
621 video_info['id'] = playlist_id
623 video_info['title'] = playlist_title
624 if playlist_description:
625 video_info['description'] = playlist_description
628 def _search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
630 Perform a regex search on the given string, using a single or a list of
631 patterns returning the first matching group.
632 In case of failure return a default value or raise a WARNING or a
633 RegexNotFoundError, depending on fatal, specifying the field name.
635 if isinstance(pattern, (str, compat_str, compiled_regex_type)):
636 mobj = re.search(pattern, string, flags)
639 mobj = re.search(p, string, flags)
643 if not self._downloader.params.get('no_color') and compat_os_name != 'nt' and sys.stderr.isatty():
644 _name = '\033[0;34m%s\033[0m' % name
650 # return the first matching group
651 return next(g for g in mobj.groups() if g is not None)
653 return mobj.group(group)
654 elif default is not NO_DEFAULT:
657 raise RegexNotFoundError('Unable to extract %s' % _name)
659 self._downloader.report_warning('unable to extract %s' % _name + bug_reports_message())
662 def _html_search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
664 Like _search_regex, but strips HTML tags and unescapes entities.
666 res = self._search_regex(pattern, string, name, default, fatal, flags, group)
668 return clean_html(res).strip()
672 def _get_netrc_login_info(self, netrc_machine=None):
675 netrc_machine = netrc_machine or self._NETRC_MACHINE
677 if self._downloader.params.get('usenetrc', False):
679 info = netrc.netrc().authenticators(netrc_machine)
684 raise netrc.NetrcParseError(
685 'No authenticators for %s' % netrc_machine)
686 except (IOError, netrc.NetrcParseError) as err:
687 self._downloader.report_warning(
688 'parsing .netrc: %s' % error_to_compat_str(err))
690 return username, password
692 def _get_login_info(self, username_option='username', password_option='password', netrc_machine=None):
694 Get the login info as (username, password)
695 First look for the manually specified credentials using username_option
696 and password_option as keys in params dictionary. If no such credentials
697 available look in the netrc file using the netrc_machine or _NETRC_MACHINE
699 If there's no info available, return (None, None)
701 if self._downloader is None:
704 downloader_params = self._downloader.params
706 # Attempt to use provided username and password or .netrc data
707 if downloader_params.get(username_option) is not None:
708 username = downloader_params[username_option]
709 password = downloader_params[password_option]
711 username, password = self._get_netrc_login_info(netrc_machine)
713 return username, password
715 def _get_tfa_info(self, note='two-factor verification code'):
717 Get the two-factor authentication info
718 TODO - asking the user will be required for sms/phone verify
719 currently just uses the command line option
720 If there's no info available, return None
722 if self._downloader is None:
724 downloader_params = self._downloader.params
726 if downloader_params.get('twofactor') is not None:
727 return downloader_params['twofactor']
729 return compat_getpass('Type %s and press [Return]: ' % note)
731 # Helper functions for extracting OpenGraph info
733 def _og_regexes(prop):
734 content_re = r'content=(?:"([^"]+?)"|\'([^\']+?)\'|\s*([^\s"\'=<>`]+?))'
735 property_re = (r'(?:name|property)=(?:\'og:%(prop)s\'|"og:%(prop)s"|\s*og:%(prop)s\b)'
736 % {'prop': re.escape(prop)})
737 template = r'<meta[^>]+?%s[^>]+?%s'
739 template % (property_re, content_re),
740 template % (content_re, property_re),
744 def _meta_regex(prop):
745 return r'''(?isx)<meta
746 (?=[^>]+(?:itemprop|name|property|id|http-equiv)=(["\']?)%s\1)
747 [^>]+?content=(["\'])(?P<content>.*?)\2''' % re.escape(prop)
749 def _og_search_property(self, prop, html, name=None, **kargs):
750 if not isinstance(prop, (list, tuple)):
753 name = 'OpenGraph %s' % prop[0]
756 og_regexes.extend(self._og_regexes(p))
757 escaped = self._search_regex(og_regexes, html, name, flags=re.DOTALL, **kargs)
760 return unescapeHTML(escaped)
762 def _og_search_thumbnail(self, html, **kargs):
763 return self._og_search_property('image', html, 'thumbnail URL', fatal=False, **kargs)
765 def _og_search_description(self, html, **kargs):
766 return self._og_search_property('description', html, fatal=False, **kargs)
768 def _og_search_title(self, html, **kargs):
769 return self._og_search_property('title', html, **kargs)
771 def _og_search_video_url(self, html, name='video url', secure=True, **kargs):
772 regexes = self._og_regexes('video') + self._og_regexes('video:url')
774 regexes = self._og_regexes('video:secure_url') + regexes
775 return self._html_search_regex(regexes, html, name, **kargs)
777 def _og_search_url(self, html, **kargs):
778 return self._og_search_property('url', html, **kargs)
780 def _html_search_meta(self, name, html, display_name=None, fatal=False, **kwargs):
781 if not isinstance(name, (list, tuple)):
783 if display_name is None:
784 display_name = name[0]
785 return self._html_search_regex(
786 [self._meta_regex(n) for n in name],
787 html, display_name, fatal=fatal, group='content', **kwargs)
789 def _dc_search_uploader(self, html):
790 return self._html_search_meta('dc.creator', html, 'uploader')
792 def _rta_search(self, html):
793 # See http://www.rtalabel.org/index.php?content=howtofaq#single
794 if re.search(r'(?ix)<meta\s+name="rating"\s+'
795 r' content="RTA-5042-1996-1400-1577-RTA"',
800 def _media_rating_search(self, html):
801 # See http://www.tjg-designs.com/WP/metadata-code-examples-adding-metadata-to-your-web-pages/
802 rating = self._html_search_meta('rating', html)
814 return RATING_TABLE.get(rating.lower())
816 def _family_friendly_search(self, html):
817 # See http://schema.org/VideoObject
818 family_friendly = self._html_search_meta('isFamilyFriendly', html)
820 if not family_friendly:
829 return RATING_TABLE.get(family_friendly.lower())
831 def _twitter_search_player(self, html):
832 return self._html_search_meta('twitter:player', html,
833 'twitter card player')
835 def _search_json_ld(self, html, video_id, expected_type=None, **kwargs):
836 json_ld = self._search_regex(
837 r'(?s)<script[^>]+type=(["\'])application/ld\+json\1[^>]*>(?P<json_ld>.+?)</script>',
838 html, 'JSON-LD', group='json_ld', **kwargs)
839 default = kwargs.get('default', NO_DEFAULT)
841 return default if default is not NO_DEFAULT else {}
842 # JSON-LD may be malformed and thus `fatal` should be respected.
843 # At the same time `default` may be passed that assumes `fatal=False`
844 # for _search_regex. Let's simulate the same behavior here as well.
845 fatal = kwargs.get('fatal', True) if default == NO_DEFAULT else False
846 return self._json_ld(json_ld, video_id, fatal=fatal, expected_type=expected_type)
848 def _json_ld(self, json_ld, video_id, fatal=True, expected_type=None):
849 if isinstance(json_ld, compat_str):
850 json_ld = self._parse_json(json_ld, video_id, fatal=fatal)
854 if not isinstance(json_ld, (list, tuple, dict)):
856 if isinstance(json_ld, dict):
859 if e.get('@context') == 'http://schema.org':
860 item_type = e.get('@type')
861 if expected_type is not None and expected_type != item_type:
863 if item_type == 'TVEpisode':
865 'episode': unescapeHTML(e.get('name')),
866 'episode_number': int_or_none(e.get('episodeNumber')),
867 'description': unescapeHTML(e.get('description')),
869 part_of_season = e.get('partOfSeason')
870 if isinstance(part_of_season, dict) and part_of_season.get('@type') == 'TVSeason':
871 info['season_number'] = int_or_none(part_of_season.get('seasonNumber'))
872 part_of_series = e.get('partOfSeries') or e.get('partOfTVSeries')
873 if isinstance(part_of_series, dict) and part_of_series.get('@type') == 'TVSeries':
874 info['series'] = unescapeHTML(part_of_series.get('name'))
875 elif item_type == 'Article':
877 'timestamp': parse_iso8601(e.get('datePublished')),
878 'title': unescapeHTML(e.get('headline')),
879 'description': unescapeHTML(e.get('articleBody')),
881 elif item_type == 'VideoObject':
883 'url': e.get('contentUrl'),
884 'title': unescapeHTML(e.get('name')),
885 'description': unescapeHTML(e.get('description')),
886 'thumbnail': e.get('thumbnailUrl'),
887 'duration': parse_duration(e.get('duration')),
888 'timestamp': unified_timestamp(e.get('uploadDate')),
889 'filesize': float_or_none(e.get('contentSize')),
890 'tbr': int_or_none(e.get('bitrate')),
891 'width': int_or_none(e.get('width')),
892 'height': int_or_none(e.get('height')),
895 return dict((k, v) for k, v in info.items() if v is not None)
898 def _hidden_inputs(html):
899 html = re.sub(r'<!--(?:(?!<!--).)*-->', '', html)
901 for input in re.findall(r'(?i)(<input[^>]+>)', html):
902 attrs = extract_attributes(input)
905 if attrs.get('type') not in ('hidden', 'submit'):
907 name = attrs.get('name') or attrs.get('id')
908 value = attrs.get('value')
909 if name and value is not None:
910 hidden_inputs[name] = value
913 def _form_hidden_inputs(self, form_id, html):
914 form = self._search_regex(
915 r'(?is)<form[^>]+?id=(["\'])%s\1[^>]*>(?P<form>.+?)</form>' % form_id,
916 html, '%s form' % form_id, group='form')
917 return self._hidden_inputs(form)
919 def _sort_formats(self, formats, field_preference=None):
921 raise ExtractorError('No video formats found')
924 # Automatically determine tbr when missing based on abr and vbr (improves
925 # formats sorting in some cases)
926 if 'tbr' not in f and f.get('abr') is not None and f.get('vbr') is not None:
927 f['tbr'] = f['abr'] + f['vbr']
930 # TODO remove the following workaround
931 from ..utils import determine_ext
932 if not f.get('ext') and 'url' in f:
933 f['ext'] = determine_ext(f['url'])
935 if isinstance(field_preference, (list, tuple)):
938 if f.get(field) is not None
939 else ('' if field == 'format_id' else -1)
940 for field in field_preference)
942 preference = f.get('preference')
943 if preference is None:
945 if f.get('ext') in ['f4f', 'f4m']: # Not yet supported
948 protocol = f.get('protocol') or determine_protocol(f)
949 proto_preference = 0 if protocol in ['http', 'https'] else (-0.5 if protocol == 'rtsp' else -0.1)
951 if f.get('vcodec') == 'none': # audio only
953 if self._downloader.params.get('prefer_free_formats'):
954 ORDER = ['aac', 'mp3', 'm4a', 'webm', 'ogg', 'opus']
956 ORDER = ['webm', 'opus', 'ogg', 'mp3', 'aac', 'm4a']
959 audio_ext_preference = ORDER.index(f['ext'])
961 audio_ext_preference = -1
963 if f.get('acodec') == 'none': # video only
965 if self._downloader.params.get('prefer_free_formats'):
966 ORDER = ['flv', 'mp4', 'webm']
968 ORDER = ['webm', 'flv', 'mp4']
970 ext_preference = ORDER.index(f['ext'])
973 audio_ext_preference = 0
977 f.get('language_preference') if f.get('language_preference') is not None else -1,
978 f.get('quality') if f.get('quality') is not None else -1,
979 f.get('tbr') if f.get('tbr') is not None else -1,
980 f.get('filesize') if f.get('filesize') is not None else -1,
981 f.get('vbr') if f.get('vbr') is not None else -1,
982 f.get('height') if f.get('height') is not None else -1,
983 f.get('width') if f.get('width') is not None else -1,
986 f.get('abr') if f.get('abr') is not None else -1,
987 audio_ext_preference,
988 f.get('fps') if f.get('fps') is not None else -1,
989 f.get('filesize_approx') if f.get('filesize_approx') is not None else -1,
990 f.get('source_preference') if f.get('source_preference') is not None else -1,
991 f.get('format_id') if f.get('format_id') is not None else '',
993 formats.sort(key=_formats_key)
995 def _check_formats(self, formats, video_id):
998 lambda f: self._is_valid_url(
1000 item='%s video format' % f.get('format_id') if f.get('format_id') else 'video'),
1004 def _remove_duplicate_formats(formats):
1008 if f['url'] not in format_urls:
1009 format_urls.add(f['url'])
1010 unique_formats.append(f)
1011 formats[:] = unique_formats
1013 def _is_valid_url(self, url, video_id, item='video'):
1014 url = self._proto_relative_url(url, scheme='http:')
1015 # For now assume non HTTP(S) URLs always valid
1016 if not (url.startswith('http://') or url.startswith('https://')):
1019 self._request_webpage(url, video_id, 'Checking %s URL' % item)
1021 except ExtractorError as e:
1022 if isinstance(e.cause, compat_urllib_error.URLError):
1024 '%s: %s URL is invalid, skipping' % (video_id, item))
1028 def http_scheme(self):
1029 """ Either "http:" or "https:", depending on the user's preferences """
1032 if self._downloader.params.get('prefer_insecure', False)
1035 def _proto_relative_url(self, url, scheme=None):
1038 if url.startswith('//'):
1040 scheme = self.http_scheme()
1045 def _sleep(self, timeout, video_id, msg_template=None):
1046 if msg_template is None:
1047 msg_template = '%(video_id)s: Waiting for %(timeout)s seconds'
1048 msg = msg_template % {'video_id': video_id, 'timeout': timeout}
1052 def _extract_f4m_formats(self, manifest_url, video_id, preference=None, f4m_id=None,
1053 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1054 fatal=True, m3u8_id=None):
1055 manifest = self._download_xml(
1056 manifest_url, video_id, 'Downloading f4m manifest',
1057 'Unable to download f4m manifest',
1058 # Some manifests may be malformed, e.g. prosiebensat1 generated manifests
1059 # (see https://github.com/rg3/youtube-dl/issues/6215#issuecomment-121704244)
1060 transform_source=transform_source,
1063 if manifest is False:
1066 return self._parse_f4m_formats(
1067 manifest, manifest_url, video_id, preference=preference, f4m_id=f4m_id,
1068 transform_source=transform_source, fatal=fatal, m3u8_id=m3u8_id)
1070 def _parse_f4m_formats(self, manifest, manifest_url, video_id, preference=None, f4m_id=None,
1071 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1072 fatal=True, m3u8_id=None):
1073 # currently youtube-dl cannot decode the playerVerificationChallenge as Akamai uses Adobe Alchemy
1074 akamai_pv = manifest.find('{http://ns.adobe.com/f4m/1.0}pv-2.0')
1075 if akamai_pv is not None and ';' in akamai_pv.text:
1076 playerVerificationChallenge = akamai_pv.text.split(';')[0]
1077 if playerVerificationChallenge.strip() != '':
1081 manifest_version = '1.0'
1082 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/1.0}media')
1084 manifest_version = '2.0'
1085 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/2.0}media')
1086 # Remove unsupported DRM protected media from final formats
1087 # rendition (see https://github.com/rg3/youtube-dl/issues/8573).
1088 media_nodes = remove_encrypted_media(media_nodes)
1091 base_url = xpath_text(
1092 manifest, ['{http://ns.adobe.com/f4m/1.0}baseURL', '{http://ns.adobe.com/f4m/2.0}baseURL'],
1093 'base URL', default=None)
1095 base_url = base_url.strip()
1097 bootstrap_info = xpath_element(
1098 manifest, ['{http://ns.adobe.com/f4m/1.0}bootstrapInfo', '{http://ns.adobe.com/f4m/2.0}bootstrapInfo'],
1099 'bootstrap info', default=None)
1101 for i, media_el in enumerate(media_nodes):
1102 tbr = int_or_none(media_el.attrib.get('bitrate'))
1103 width = int_or_none(media_el.attrib.get('width'))
1104 height = int_or_none(media_el.attrib.get('height'))
1105 format_id = '-'.join(filter(None, [f4m_id, compat_str(i if tbr is None else tbr)]))
1106 # If <bootstrapInfo> is present, the specified f4m is a
1107 # stream-level manifest, and only set-level manifests may refer to
1108 # external resources. See section 11.4 and section 4 of F4M spec
1109 if bootstrap_info is None:
1111 # @href is introduced in 2.0, see section 11.6 of F4M spec
1112 if manifest_version == '2.0':
1113 media_url = media_el.attrib.get('href')
1114 if media_url is None:
1115 media_url = media_el.attrib.get('url')
1119 media_url if media_url.startswith('http://') or media_url.startswith('https://')
1120 else ((base_url or '/'.join(manifest_url.split('/')[:-1])) + '/' + media_url))
1121 # If media_url is itself a f4m manifest do the recursive extraction
1122 # since bitrates in parent manifest (this one) and media_url manifest
1123 # may differ leading to inability to resolve the format by requested
1124 # bitrate in f4m downloader
1125 ext = determine_ext(manifest_url)
1127 f4m_formats = self._extract_f4m_formats(
1128 manifest_url, video_id, preference=preference, f4m_id=f4m_id,
1129 transform_source=transform_source, fatal=fatal)
1130 # Sometimes stream-level manifest contains single media entry that
1131 # does not contain any quality metadata (e.g. http://matchtv.ru/#live-player).
1132 # At the same time parent's media entry in set-level manifest may
1133 # contain it. We will copy it from parent in such cases.
1134 if len(f4m_formats) == 1:
1137 'tbr': f.get('tbr') or tbr,
1138 'width': f.get('width') or width,
1139 'height': f.get('height') or height,
1140 'format_id': f.get('format_id') if not tbr else format_id,
1142 formats.extend(f4m_formats)
1145 formats.extend(self._extract_m3u8_formats(
1146 manifest_url, video_id, 'mp4', preference=preference,
1147 m3u8_id=m3u8_id, fatal=fatal))
1150 'format_id': format_id,
1151 'url': manifest_url,
1152 'ext': 'flv' if bootstrap_info is not None else None,
1156 'preference': preference,
1160 def _m3u8_meta_format(self, m3u8_url, ext=None, preference=None, m3u8_id=None):
1162 'format_id': '-'.join(filter(None, [m3u8_id, 'meta'])),
1166 'preference': preference - 100 if preference else -100,
1167 'resolution': 'multiple',
1168 'format_note': 'Quality selection URL',
1171 def _extract_m3u8_formats(self, m3u8_url, video_id, ext=None,
1172 entry_protocol='m3u8', preference=None,
1173 m3u8_id=None, note=None, errnote=None,
1174 fatal=True, live=False):
1176 res = self._download_webpage_handle(
1178 note=note or 'Downloading m3u8 information',
1179 errnote=errnote or 'Failed to download m3u8 information',
1183 m3u8_doc, urlh = res
1184 m3u8_url = urlh.geturl()
1186 formats = [self._m3u8_meta_format(m3u8_url, ext, preference, m3u8_id)]
1188 format_url = lambda u: (
1190 if re.match(r'^https?://', u)
1191 else compat_urlparse.urljoin(m3u8_url, u))
1193 # We should try extracting formats only from master playlists [1], i.e.
1194 # playlists that describe available qualities. On the other hand media
1195 # playlists [2] should be returned as is since they contain just the media
1196 # without qualities renditions.
1197 # Fortunately, master playlist can be easily distinguished from media
1198 # playlist based on particular tags availability. As of [1, 2] master
1199 # playlist tags MUST NOT appear in a media playist and vice versa.
1200 # As of [3] #EXT-X-TARGETDURATION tag is REQUIRED for every media playlist
1201 # and MUST NOT appear in master playlist thus we can clearly detect media
1202 # playlist with this criterion.
1203 # 1. https://tools.ietf.org/html/draft-pantos-http-live-streaming-17#section-4.3.4
1204 # 2. https://tools.ietf.org/html/draft-pantos-http-live-streaming-17#section-4.3.3
1205 # 3. https://tools.ietf.org/html/draft-pantos-http-live-streaming-17#section-4.3.3.1
1206 if '#EXT-X-TARGETDURATION' in m3u8_doc: # media playlist, return as is
1209 'format_id': m3u8_id,
1211 'protocol': entry_protocol,
1212 'preference': preference,
1216 for line in m3u8_doc.splitlines():
1217 if line.startswith('#EXT-X-STREAM-INF:'):
1218 last_info = parse_m3u8_attributes(line)
1219 elif line.startswith('#EXT-X-MEDIA:'):
1220 media = parse_m3u8_attributes(line)
1221 media_type = media.get('TYPE')
1222 if media_type in ('VIDEO', 'AUDIO'):
1223 media_url = media.get('URI')
1226 for v in (media.get('GROUP-ID'), media.get('NAME')):
1230 'format_id': '-'.join(format_id),
1231 'url': format_url(media_url),
1232 'language': media.get('LANGUAGE'),
1233 'vcodec': 'none' if media_type == 'AUDIO' else None,
1235 'protocol': entry_protocol,
1236 'preference': preference,
1239 # When there is no URI in EXT-X-MEDIA let this tag's
1240 # data be used by regular URI lines below
1242 elif line.startswith('#') or not line.strip():
1245 tbr = int_or_none(last_info.get('AVERAGE-BANDWIDTH') or last_info.get('BANDWIDTH'), scale=1000)
1248 format_id.append(m3u8_id)
1249 # Despite specification does not mention NAME attribute for
1250 # EXT-X-STREAM-INF it still sometimes may be present
1251 stream_name = last_info.get('NAME') or last_media.get('NAME')
1252 # Bandwidth of live streams may differ over time thus making
1253 # format_id unpredictable. So it's better to keep provided
1256 format_id.append(stream_name if stream_name else '%d' % (tbr if tbr else len(formats)))
1258 'format_id': '-'.join(format_id),
1259 'url': format_url(line.strip()),
1262 'fps': float_or_none(last_info.get('FRAME-RATE')),
1263 'protocol': entry_protocol,
1264 'preference': preference,
1266 resolution = last_info.get('RESOLUTION')
1268 width_str, height_str = resolution.split('x')
1269 f['width'] = int(width_str)
1270 f['height'] = int(height_str)
1271 # Unified Streaming Platform
1273 r'audio.*?(?:%3D|=)(\d+)(?:-video.*?(?:%3D|=)(\d+))?', f['url'])
1275 abr, vbr = mobj.groups()
1276 abr, vbr = float_or_none(abr, 1000), float_or_none(vbr, 1000)
1281 f.update(parse_codecs(last_info.get('CODECS')))
1288 def _xpath_ns(path, namespace=None):
1292 for c in path.split('/'):
1293 if not c or c == '.':
1296 out.append('{%s}%s' % (namespace, c))
1297 return '/'.join(out)
1299 def _extract_smil_formats(self, smil_url, video_id, fatal=True, f4m_params=None, transform_source=None):
1300 smil = self._download_smil(smil_url, video_id, fatal=fatal, transform_source=transform_source)
1306 namespace = self._parse_smil_namespace(smil)
1308 return self._parse_smil_formats(
1309 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
1311 def _extract_smil_info(self, smil_url, video_id, fatal=True, f4m_params=None):
1312 smil = self._download_smil(smil_url, video_id, fatal=fatal)
1315 return self._parse_smil(smil, smil_url, video_id, f4m_params=f4m_params)
1317 def _download_smil(self, smil_url, video_id, fatal=True, transform_source=None):
1318 return self._download_xml(
1319 smil_url, video_id, 'Downloading SMIL file',
1320 'Unable to download SMIL file', fatal=fatal, transform_source=transform_source)
1322 def _parse_smil(self, smil, smil_url, video_id, f4m_params=None):
1323 namespace = self._parse_smil_namespace(smil)
1325 formats = self._parse_smil_formats(
1326 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
1327 subtitles = self._parse_smil_subtitles(smil, namespace=namespace)
1329 video_id = os.path.splitext(url_basename(smil_url))[0]
1333 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
1334 name = meta.attrib.get('name')
1335 content = meta.attrib.get('content')
1336 if not name or not content:
1338 if not title and name == 'title':
1340 elif not description and name in ('description', 'abstract'):
1341 description = content
1342 elif not upload_date and name == 'date':
1343 upload_date = unified_strdate(content)
1346 'id': image.get('type'),
1347 'url': image.get('src'),
1348 'width': int_or_none(image.get('width')),
1349 'height': int_or_none(image.get('height')),
1350 } for image in smil.findall(self._xpath_ns('.//image', namespace)) if image.get('src')]
1354 'title': title or video_id,
1355 'description': description,
1356 'upload_date': upload_date,
1357 'thumbnails': thumbnails,
1359 'subtitles': subtitles,
1362 def _parse_smil_namespace(self, smil):
1363 return self._search_regex(
1364 r'(?i)^{([^}]+)?}smil$', smil.tag, 'namespace', default=None)
1366 def _parse_smil_formats(self, smil, smil_url, video_id, namespace=None, f4m_params=None, transform_rtmp_url=None):
1368 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
1369 b = meta.get('base') or meta.get('httpBase')
1380 media = smil.findall(self._xpath_ns('.//video', namespace)) + smil.findall(self._xpath_ns('.//audio', namespace))
1381 for medium in media:
1382 src = medium.get('src')
1383 if not src or src in srcs:
1387 bitrate = float_or_none(medium.get('system-bitrate') or medium.get('systemBitrate'), 1000)
1388 filesize = int_or_none(medium.get('size') or medium.get('fileSize'))
1389 width = int_or_none(medium.get('width'))
1390 height = int_or_none(medium.get('height'))
1391 proto = medium.get('proto')
1392 ext = medium.get('ext')
1393 src_ext = determine_ext(src)
1394 streamer = medium.get('streamer') or base
1396 if proto == 'rtmp' or streamer.startswith('rtmp'):
1402 'format_id': 'rtmp-%d' % (rtmp_count if bitrate is None else bitrate),
1404 'filesize': filesize,
1408 if transform_rtmp_url:
1409 streamer, src = transform_rtmp_url(streamer, src)
1410 formats[-1].update({
1416 src_url = src if src.startswith('http') else compat_urlparse.urljoin(base, src)
1417 src_url = src_url.strip()
1419 if proto == 'm3u8' or src_ext == 'm3u8':
1420 m3u8_formats = self._extract_m3u8_formats(
1421 src_url, video_id, ext or 'mp4', m3u8_id='hls', fatal=False)
1422 if len(m3u8_formats) == 1:
1424 m3u8_formats[0].update({
1425 'format_id': 'hls-%d' % (m3u8_count if bitrate is None else bitrate),
1430 formats.extend(m3u8_formats)
1433 if src_ext == 'f4m':
1438 'plugin': 'flowplayer-3.2.0.1',
1440 f4m_url += '&' if '?' in f4m_url else '?'
1441 f4m_url += compat_urllib_parse_urlencode(f4m_params)
1442 formats.extend(self._extract_f4m_formats(f4m_url, video_id, f4m_id='hds', fatal=False))
1445 if src_url.startswith('http') and self._is_valid_url(src, video_id):
1449 'ext': ext or src_ext or 'flv',
1450 'format_id': 'http-%d' % (bitrate or http_count),
1452 'filesize': filesize,
1460 def _parse_smil_subtitles(self, smil, namespace=None, subtitles_lang='en'):
1463 for num, textstream in enumerate(smil.findall(self._xpath_ns('.//textstream', namespace))):
1464 src = textstream.get('src')
1465 if not src or src in urls:
1468 ext = textstream.get('ext') or mimetype2ext(textstream.get('type')) or determine_ext(src)
1469 lang = textstream.get('systemLanguage') or textstream.get('systemLanguageName') or textstream.get('lang') or subtitles_lang
1470 subtitles.setdefault(lang, []).append({
1476 def _extract_xspf_playlist(self, playlist_url, playlist_id, fatal=True):
1477 xspf = self._download_xml(
1478 playlist_url, playlist_id, 'Downloading xpsf playlist',
1479 'Unable to download xspf manifest', fatal=fatal)
1482 return self._parse_xspf(xspf, playlist_id)
1484 def _parse_xspf(self, playlist, playlist_id):
1486 'xspf': 'http://xspf.org/ns/0/',
1487 's1': 'http://static.streamone.nl/player/ns/0',
1491 for track in playlist.findall(xpath_with_ns('./xspf:trackList/xspf:track', NS_MAP)):
1493 track, xpath_with_ns('./xspf:title', NS_MAP), 'title', default=playlist_id)
1494 description = xpath_text(
1495 track, xpath_with_ns('./xspf:annotation', NS_MAP), 'description')
1496 thumbnail = xpath_text(
1497 track, xpath_with_ns('./xspf:image', NS_MAP), 'thumbnail')
1498 duration = float_or_none(
1499 xpath_text(track, xpath_with_ns('./xspf:duration', NS_MAP), 'duration'), 1000)
1502 'url': location.text,
1503 'format_id': location.get(xpath_with_ns('s1:label', NS_MAP)),
1504 'width': int_or_none(location.get(xpath_with_ns('s1:width', NS_MAP))),
1505 'height': int_or_none(location.get(xpath_with_ns('s1:height', NS_MAP))),
1506 } for location in track.findall(xpath_with_ns('./xspf:location', NS_MAP))]
1507 self._sort_formats(formats)
1512 'description': description,
1513 'thumbnail': thumbnail,
1514 'duration': duration,
1519 def _extract_mpd_formats(self, mpd_url, video_id, mpd_id=None, note=None, errnote=None, fatal=True, formats_dict={}):
1520 res = self._download_webpage_handle(
1522 note=note or 'Downloading MPD manifest',
1523 errnote=errnote or 'Failed to download MPD manifest',
1528 mpd_base_url = re.match(r'https?://.+/', urlh.geturl()).group()
1530 return self._parse_mpd_formats(
1531 compat_etree_fromstring(mpd.encode('utf-8')), mpd_id, mpd_base_url, formats_dict=formats_dict)
1533 def _parse_mpd_formats(self, mpd_doc, mpd_id=None, mpd_base_url='', formats_dict={}):
1535 Parse formats from MPD manifest.
1537 1. MPEG-DASH Standard, ISO/IEC 23009-1:2014(E),
1538 http://standards.iso.org/ittf/PubliclyAvailableStandards/c065274_ISO_IEC_23009-1_2014.zip
1539 2. https://en.wikipedia.org/wiki/Dynamic_Adaptive_Streaming_over_HTTP
1541 if mpd_doc.get('type') == 'dynamic':
1544 namespace = self._search_regex(r'(?i)^{([^}]+)?}MPD$', mpd_doc.tag, 'namespace', default=None)
1547 return self._xpath_ns(path, namespace)
1549 def is_drm_protected(element):
1550 return element.find(_add_ns('ContentProtection')) is not None
1552 def extract_multisegment_info(element, ms_parent_info):
1553 ms_info = ms_parent_info.copy()
1554 segment_list = element.find(_add_ns('SegmentList'))
1555 if segment_list is not None:
1556 segment_urls_e = segment_list.findall(_add_ns('SegmentURL'))
1558 ms_info['segment_urls'] = [segment.attrib['media'] for segment in segment_urls_e]
1559 initialization = segment_list.find(_add_ns('Initialization'))
1560 if initialization is not None:
1561 ms_info['initialization_url'] = initialization.attrib['sourceURL']
1563 segment_template = element.find(_add_ns('SegmentTemplate'))
1564 if segment_template is not None:
1565 start_number = segment_template.get('startNumber')
1567 ms_info['start_number'] = int(start_number)
1568 segment_timeline = segment_template.find(_add_ns('SegmentTimeline'))
1569 if segment_timeline is not None:
1570 s_e = segment_timeline.findall(_add_ns('S'))
1572 ms_info['total_number'] = 0
1575 r = int(s.get('r', 0))
1576 ms_info['total_number'] += 1 + r
1577 ms_info['s'].append({
1578 't': int(s.get('t', 0)),
1579 # @d is mandatory (see [1, 5.3.9.6.2, Table 17, page 60])
1580 'd': int(s.attrib['d']),
1584 timescale = segment_template.get('timescale')
1586 ms_info['timescale'] = int(timescale)
1587 segment_duration = segment_template.get('duration')
1588 if segment_duration:
1589 ms_info['segment_duration'] = int(segment_duration)
1590 media_template = segment_template.get('media')
1592 ms_info['media_template'] = media_template
1593 initialization = segment_template.get('initialization')
1595 ms_info['initialization_url'] = initialization
1597 initialization = segment_template.find(_add_ns('Initialization'))
1598 if initialization is not None:
1599 ms_info['initialization_url'] = initialization.attrib['sourceURL']
1602 mpd_duration = parse_duration(mpd_doc.get('mediaPresentationDuration'))
1604 for period in mpd_doc.findall(_add_ns('Period')):
1605 period_duration = parse_duration(period.get('duration')) or mpd_duration
1606 period_ms_info = extract_multisegment_info(period, {
1610 for adaptation_set in period.findall(_add_ns('AdaptationSet')):
1611 if is_drm_protected(adaptation_set):
1613 adaption_set_ms_info = extract_multisegment_info(adaptation_set, period_ms_info)
1614 for representation in adaptation_set.findall(_add_ns('Representation')):
1615 if is_drm_protected(representation):
1617 representation_attrib = adaptation_set.attrib.copy()
1618 representation_attrib.update(representation.attrib)
1619 # According to [1, 5.3.7.2, Table 9, page 41], @mimeType is mandatory
1620 mime_type = representation_attrib['mimeType']
1621 content_type = mime_type.split('/')[0]
1622 if content_type == 'text':
1623 # TODO implement WebVTT downloading
1625 elif content_type == 'video' or content_type == 'audio':
1627 for element in (representation, adaptation_set, period, mpd_doc):
1628 base_url_e = element.find(_add_ns('BaseURL'))
1629 if base_url_e is not None:
1630 base_url = base_url_e.text + base_url
1631 if re.match(r'^https?://', base_url):
1633 if mpd_base_url and not re.match(r'^https?://', base_url):
1634 if not mpd_base_url.endswith('/') and not base_url.startswith('/'):
1636 base_url = mpd_base_url + base_url
1637 representation_id = representation_attrib.get('id')
1638 lang = representation_attrib.get('lang')
1639 url_el = representation.find(_add_ns('BaseURL'))
1640 filesize = int_or_none(url_el.attrib.get('{http://youtube.com/yt/2012/10/10}contentLength') if url_el is not None else None)
1642 'format_id': '%s-%s' % (mpd_id, representation_id) if mpd_id else representation_id,
1644 'ext': mimetype2ext(mime_type),
1645 'width': int_or_none(representation_attrib.get('width')),
1646 'height': int_or_none(representation_attrib.get('height')),
1647 'tbr': int_or_none(representation_attrib.get('bandwidth'), 1000),
1648 'asr': int_or_none(representation_attrib.get('audioSamplingRate')),
1649 'fps': int_or_none(representation_attrib.get('frameRate')),
1650 'vcodec': 'none' if content_type == 'audio' else representation_attrib.get('codecs'),
1651 'acodec': 'none' if content_type == 'video' else representation_attrib.get('codecs'),
1652 'language': lang if lang not in ('mul', 'und', 'zxx', 'mis') else None,
1653 'format_note': 'DASH %s' % content_type,
1654 'filesize': filesize,
1656 representation_ms_info = extract_multisegment_info(representation, adaption_set_ms_info)
1657 if 'segment_urls' not in representation_ms_info and 'media_template' in representation_ms_info:
1658 if 'total_number' not in representation_ms_info and 'segment_duration':
1659 segment_duration = float(representation_ms_info['segment_duration']) / float(representation_ms_info['timescale'])
1660 representation_ms_info['total_number'] = int(math.ceil(float(period_duration) / segment_duration))
1661 media_template = representation_ms_info['media_template']
1662 media_template = media_template.replace('$RepresentationID$', representation_id)
1663 media_template = re.sub(r'\$(Number|Bandwidth|Time)\$', r'%(\1)d', media_template)
1664 media_template = re.sub(r'\$(Number|Bandwidth|Time)%([^$]+)\$', r'%(\1)\2', media_template)
1665 media_template.replace('$$', '$')
1667 # As per [1, 5.3.9.4.4, Table 16, page 55] $Number$ and $Time$
1668 # can't be used at the same time
1669 if '%(Number' in media_template:
1670 representation_ms_info['segment_urls'] = [
1672 'Number': segment_number,
1673 'Bandwidth': representation_attrib.get('bandwidth'),
1675 for segment_number in range(
1676 representation_ms_info['start_number'],
1677 representation_ms_info['total_number'] + representation_ms_info['start_number'])]
1679 representation_ms_info['segment_urls'] = []
1682 def add_segment_url():
1683 representation_ms_info['segment_urls'].append(
1685 'Time': segment_time,
1686 'Bandwidth': representation_attrib.get('bandwidth'),
1690 for num, s in enumerate(representation_ms_info['s']):
1691 segment_time = s.get('t') or segment_time
1693 for r in range(s.get('r', 0)):
1694 segment_time += s['d']
1696 segment_time += s['d']
1697 if 'segment_urls' in representation_ms_info:
1699 'segment_urls': representation_ms_info['segment_urls'],
1700 'protocol': 'http_dash_segments',
1702 if 'initialization_url' in representation_ms_info:
1703 initialization_url = representation_ms_info['initialization_url'].replace('$RepresentationID$', representation_id)
1705 'initialization_url': initialization_url,
1707 if not f.get('url'):
1708 f['url'] = initialization_url
1710 existing_format = next(
1711 fo for fo in formats
1712 if fo['format_id'] == representation_id)
1713 except StopIteration:
1714 full_info = formats_dict.get(representation_id, {}).copy()
1716 formats.append(full_info)
1718 existing_format.update(f)
1720 self.report_warning('Unknown MIME type %s in DASH manifest' % mime_type)
1723 def _parse_html5_media_entries(self, base_url, webpage, video_id, m3u8_id=None, m3u8_entry_protocol='m3u8'):
1724 def absolute_url(video_url):
1725 return compat_urlparse.urljoin(base_url, video_url)
1727 def parse_content_type(content_type):
1728 if not content_type:
1730 ctr = re.search(r'(?P<mimetype>[^/]+/[^;]+)(?:;\s*codecs="?(?P<codecs>[^"]+))?', content_type)
1732 mimetype, codecs = ctr.groups()
1733 f = parse_codecs(codecs)
1734 f['ext'] = mimetype2ext(mimetype)
1738 def _media_formats(src, cur_media_type):
1739 full_url = absolute_url(src)
1740 if determine_ext(full_url) == 'm3u8':
1741 is_plain_url = False
1742 formats = self._extract_m3u8_formats(
1743 full_url, video_id, ext='mp4',
1744 entry_protocol=m3u8_entry_protocol, m3u8_id=m3u8_id)
1749 'vcodec': 'none' if cur_media_type == 'audio' else None,
1751 return is_plain_url, formats
1754 for media_tag, media_type, media_content in re.findall(r'(?s)(<(?P<tag>video|audio)[^>]*>)(.*?)</(?P=tag)>', webpage):
1759 media_attributes = extract_attributes(media_tag)
1760 src = media_attributes.get('src')
1762 _, formats = _media_formats(src, media_type)
1763 media_info['formats'].extend(formats)
1764 media_info['thumbnail'] = media_attributes.get('poster')
1766 for source_tag in re.findall(r'<source[^>]+>', media_content):
1767 source_attributes = extract_attributes(source_tag)
1768 src = source_attributes.get('src')
1771 is_plain_url, formats = _media_formats(src, media_type)
1773 f = parse_content_type(source_attributes.get('type'))
1774 f.update(formats[0])
1775 media_info['formats'].append(f)
1777 media_info['formats'].extend(formats)
1778 for track_tag in re.findall(r'<track[^>]+>', media_content):
1779 track_attributes = extract_attributes(track_tag)
1780 kind = track_attributes.get('kind')
1781 if not kind or kind == 'subtitles':
1782 src = track_attributes.get('src')
1785 lang = track_attributes.get('srclang') or track_attributes.get('lang') or track_attributes.get('label')
1786 media_info['subtitles'].setdefault(lang, []).append({
1787 'url': absolute_url(src),
1789 if media_info['formats']:
1790 entries.append(media_info)
1793 def _extract_akamai_formats(self, manifest_url, video_id):
1795 f4m_url = re.sub(r'(https?://.+?)/i/', r'\1/z/', manifest_url).replace('/master.m3u8', '/manifest.f4m')
1796 formats.extend(self._extract_f4m_formats(
1797 update_url_query(f4m_url, {'hdcore': '3.7.0'}),
1798 video_id, f4m_id='hds', fatal=False))
1799 m3u8_url = re.sub(r'(https?://.+?)/z/', r'\1/i/', manifest_url).replace('/manifest.f4m', '/master.m3u8')
1800 formats.extend(self._extract_m3u8_formats(
1801 m3u8_url, video_id, 'mp4', 'm3u8_native',
1802 m3u8_id='hls', fatal=False))
1805 def _extract_wowza_formats(self, url, video_id, m3u8_entry_protocol='m3u8_native', skip_protocols=[]):
1806 url = re.sub(r'/(?:manifest|playlist|jwplayer)\.(?:m3u8|f4m|mpd|smil)', '', url)
1807 url_base = self._search_regex(r'(?:https?|rtmp|rtsp)(://[^?]+)', url, 'format url')
1808 http_base_url = 'http' + url_base
1810 if 'm3u8' not in skip_protocols:
1811 formats.extend(self._extract_m3u8_formats(
1812 http_base_url + '/playlist.m3u8', video_id, 'mp4',
1813 m3u8_entry_protocol, m3u8_id='hls', fatal=False))
1814 if 'f4m' not in skip_protocols:
1815 formats.extend(self._extract_f4m_formats(
1816 http_base_url + '/manifest.f4m',
1817 video_id, f4m_id='hds', fatal=False))
1818 if re.search(r'(?:/smil:|\.smil)', url_base):
1819 if 'dash' not in skip_protocols:
1820 formats.extend(self._extract_mpd_formats(
1821 http_base_url + '/manifest.mpd',
1822 video_id, mpd_id='dash', fatal=False))
1823 if 'smil' not in skip_protocols:
1824 rtmp_formats = self._extract_smil_formats(
1825 http_base_url + '/jwplayer.smil',
1826 video_id, fatal=False)
1827 for rtmp_format in rtmp_formats:
1828 rtsp_format = rtmp_format.copy()
1829 rtsp_format['url'] = '%s/%s' % (rtmp_format['url'], rtmp_format['play_path'])
1830 del rtsp_format['play_path']
1831 del rtsp_format['ext']
1832 rtsp_format.update({
1833 'url': rtsp_format['url'].replace('rtmp://', 'rtsp://'),
1834 'format_id': rtmp_format['format_id'].replace('rtmp', 'rtsp'),
1837 formats.extend([rtmp_format, rtsp_format])
1839 for protocol in ('rtmp', 'rtsp'):
1840 if protocol not in skip_protocols:
1842 'url': protocol + url_base,
1843 'format_id': protocol,
1844 'protocol': protocol,
1848 def _live_title(self, name):
1849 """ Generate the title for a live video """
1850 now = datetime.datetime.now()
1851 now_str = now.strftime('%Y-%m-%d %H:%M')
1852 return name + ' ' + now_str
1854 def _int(self, v, name, fatal=False, **kwargs):
1855 res = int_or_none(v, **kwargs)
1856 if 'get_attr' in kwargs:
1857 print(getattr(v, kwargs['get_attr']))
1859 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
1861 raise ExtractorError(msg)
1863 self._downloader.report_warning(msg)
1866 def _float(self, v, name, fatal=False, **kwargs):
1867 res = float_or_none(v, **kwargs)
1869 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
1871 raise ExtractorError(msg)
1873 self._downloader.report_warning(msg)
1876 def _set_cookie(self, domain, name, value, expire_time=None):
1877 cookie = compat_cookiejar.Cookie(
1878 0, name, value, None, None, domain, None,
1879 None, '/', True, False, expire_time, '', None, None, None)
1880 self._downloader.cookiejar.set_cookie(cookie)
1882 def _get_cookies(self, url):
1883 """ Return a compat_cookies.SimpleCookie with the cookies for the url """
1884 req = sanitized_Request(url)
1885 self._downloader.cookiejar.add_cookie_header(req)
1886 return compat_cookies.SimpleCookie(req.get_header('Cookie'))
1888 def get_testcases(self, include_onlymatching=False):
1889 t = getattr(self, '_TEST', None)
1891 assert not hasattr(self, '_TESTS'), \
1892 '%s has _TEST and _TESTS' % type(self).__name__
1895 tests = getattr(self, '_TESTS', [])
1897 if not include_onlymatching and t.get('only_matching', False):
1899 t['name'] = type(self).__name__[:-len('IE')]
1902 def is_suitable(self, age_limit):
1903 """ Test whether the extractor is generally suitable for the given
1904 age limit (i.e. pornographic sites are not, all others usually are) """
1906 any_restricted = False
1907 for tc in self.get_testcases(include_onlymatching=False):
1908 if tc.get('playlist', []):
1909 tc = tc['playlist'][0]
1910 is_restricted = age_restricted(
1911 tc.get('info_dict', {}).get('age_limit'), age_limit)
1912 if not is_restricted:
1914 any_restricted = any_restricted or is_restricted
1915 return not any_restricted
1917 def extract_subtitles(self, *args, **kwargs):
1918 if (self._downloader.params.get('writesubtitles', False) or
1919 self._downloader.params.get('listsubtitles')):
1920 return self._get_subtitles(*args, **kwargs)
1923 def _get_subtitles(self, *args, **kwargs):
1924 raise NotImplementedError('This method must be implemented by subclasses')
1927 def _merge_subtitle_items(subtitle_list1, subtitle_list2):
1928 """ Merge subtitle items for one language. Items with duplicated URLs
1929 will be dropped. """
1930 list1_urls = set([item['url'] for item in subtitle_list1])
1931 ret = list(subtitle_list1)
1932 ret.extend([item for item in subtitle_list2 if item['url'] not in list1_urls])
1936 def _merge_subtitles(cls, subtitle_dict1, subtitle_dict2):
1937 """ Merge two subtitle dictionaries, language by language. """
1938 ret = dict(subtitle_dict1)
1939 for lang in subtitle_dict2:
1940 ret[lang] = cls._merge_subtitle_items(subtitle_dict1.get(lang, []), subtitle_dict2[lang])
1943 def extract_automatic_captions(self, *args, **kwargs):
1944 if (self._downloader.params.get('writeautomaticsub', False) or
1945 self._downloader.params.get('listsubtitles')):
1946 return self._get_automatic_captions(*args, **kwargs)
1949 def _get_automatic_captions(self, *args, **kwargs):
1950 raise NotImplementedError('This method must be implemented by subclasses')
1952 def mark_watched(self, *args, **kwargs):
1953 if (self._downloader.params.get('mark_watched', False) and
1954 (self._get_login_info()[0] is not None or
1955 self._downloader.params.get('cookiefile') is not None)):
1956 self._mark_watched(*args, **kwargs)
1958 def _mark_watched(self, *args, **kwargs):
1959 raise NotImplementedError('This method must be implemented by subclasses')
1961 def geo_verification_headers(self):
1963 geo_verification_proxy = self._downloader.params.get('geo_verification_proxy')
1964 if geo_verification_proxy:
1965 headers['Ytdl-request-proxy'] = geo_verification_proxy
1969 class SearchInfoExtractor(InfoExtractor):
1971 Base class for paged search queries extractors.
1972 They accept URLs in the format _SEARCH_KEY(|all|[0-9]):{query}
1973 Instances should define _SEARCH_KEY and _MAX_RESULTS.
1977 def _make_valid_url(cls):
1978 return r'%s(?P<prefix>|[1-9][0-9]*|all):(?P<query>[\s\S]+)' % cls._SEARCH_KEY
1981 def suitable(cls, url):
1982 return re.match(cls._make_valid_url(), url) is not None
1984 def _real_extract(self, query):
1985 mobj = re.match(self._make_valid_url(), query)
1987 raise ExtractorError('Invalid search query "%s"' % query)
1989 prefix = mobj.group('prefix')
1990 query = mobj.group('query')
1992 return self._get_n_results(query, 1)
1993 elif prefix == 'all':
1994 return self._get_n_results(query, self._MAX_RESULTS)
1998 raise ExtractorError('invalid download number %s for query "%s"' % (n, query))
1999 elif n > self._MAX_RESULTS:
2000 self._downloader.report_warning('%s returns max %i results (you requested %i)' % (self._SEARCH_KEY, self._MAX_RESULTS, n))
2001 n = self._MAX_RESULTS
2002 return self._get_n_results(query, n)
2004 def _get_n_results(self, query, n):
2005 """Get a specified number of results for a query"""
2006 raise NotImplementedError('This method must be implemented by subclasses')
2009 def SEARCH_KEY(self):
2010 return self._SEARCH_KEY