1 from __future__ import unicode_literals
15 from ..compat import (
18 compat_etree_fromstring,
24 compat_urllib_parse_unquote,
25 compat_urllib_parse_urlencode,
26 compat_urllib_request,
29 from ..downloader.f4m import remove_encrypted_media
60 parse_m3u8_attributes,
67 class InfoExtractor(object):
68 """Information Extractor class.
70 Information extractors are the classes that, given a URL, extract
71 information about the video (or videos) the URL refers to. This
72 information includes the real video URL, the video title, author and
73 others. The information is stored in a dictionary which is then
74 passed to the YoutubeDL. The YoutubeDL processes this
75 information possibly downloading the video to the file system, among
76 other possible outcomes.
78 The type field determines the type of the result.
79 By far the most common value (and the default if _type is missing) is
80 "video", which indicates a single video.
82 For a video, the dictionaries must include the following fields:
85 title: Video title, unescaped.
87 Additionally, it must contain either a formats entry or a url one:
89 formats: A list of dictionaries for each format available, ordered
90 from worst to best quality.
93 * url Mandatory. The URL of the video file
95 The URL of the manifest file in case of
96 fragmented media (DASH, hls, hds)
97 * ext Will be calculated from URL if missing
98 * format A human-readable description of the format
99 ("mp4 container with h264/opus").
100 Calculated from the format_id, width, height.
101 and format_note fields if missing.
102 * format_id A short description of the format
103 ("mp4_h264_opus" or "19").
104 Technically optional, but strongly recommended.
105 * format_note Additional info about the format
106 ("3D" or "DASH video")
107 * width Width of the video, if known
108 * height Height of the video, if known
109 * resolution Textual description of width and height
110 * tbr Average bitrate of audio and video in KBit/s
111 * abr Average audio bitrate in KBit/s
112 * acodec Name of the audio codec in use
113 * asr Audio sampling rate in Hertz
114 * vbr Average video bitrate in KBit/s
116 * vcodec Name of the video codec in use
117 * container Name of the container format
118 * filesize The number of bytes, if known in advance
119 * filesize_approx An estimate for the number of bytes
120 * player_url SWF Player URL (used for rtmpdump).
121 * protocol The protocol that will be used for the actual
122 download, lower-case.
123 "http", "https", "rtsp", "rtmp", "rtmpe",
124 "m3u8", "m3u8_native" or "http_dash_segments".
126 Base URL for fragments. Each fragment's path
127 value (if present) will be relative to
129 * fragments A list of fragments of a fragmented media.
130 Each fragment entry must contain either an url
131 or a path. If an url is present it should be
132 considered by a client. Otherwise both path and
133 fragment_base_url must be present. Here is
134 the list of all potential fields:
135 * "url" - fragment's URL
136 * "path" - fragment's path relative to
138 * "duration" (optional, int or float)
139 * "filesize" (optional, int)
140 * preference Order number of this format. If this field is
141 present and not None, the formats get sorted
142 by this field, regardless of all other values.
143 -1 for default (order by other properties),
144 -2 or smaller for less than default.
145 < -1000 to hide the format (if there is
146 another one which is strictly better)
147 * language Language code, e.g. "de" or "en-US".
148 * language_preference Is this in the language mentioned in
150 10 if it's what the URL is about,
151 -1 for default (don't know),
152 -10 otherwise, other values reserved for now.
153 * quality Order number of the video quality of this
154 format, irrespective of the file format.
155 -1 for default (order by other properties),
156 -2 or smaller for less than default.
157 * source_preference Order number for this video source
158 (quality takes higher priority)
159 -1 for default (order by other properties),
160 -2 or smaller for less than default.
161 * http_headers A dictionary of additional HTTP headers
162 to add to the request.
163 * stretched_ratio If given and not 1, indicates that the
164 video's pixels are not square.
165 width : height ratio as float.
166 * no_resume The server does not support resuming the
167 (HTTP or RTMP) download. Boolean.
169 url: Final video URL.
170 ext: Video filename extension.
171 format: The video format, defaults to ext (used for --get-format)
172 player_url: SWF Player URL (used for rtmpdump).
174 The following fields are optional:
176 alt_title: A secondary title of the video.
177 display_id An alternative identifier for the video, not necessarily
178 unique, but available before title. Typically, id is
179 something like "4234987", title "Dancing naked mole rats",
180 and display_id "dancing-naked-mole-rats"
181 thumbnails: A list of dictionaries, with the following entries:
182 * "id" (optional, string) - Thumbnail format ID
184 * "preference" (optional, int) - quality of the image
185 * "width" (optional, int)
186 * "height" (optional, int)
187 * "resolution" (optional, string "{width}x{height"},
189 * "filesize" (optional, int)
190 thumbnail: Full URL to a video thumbnail image.
191 description: Full video description.
192 uploader: Full name of the video uploader.
193 license: License name the video is licensed under.
194 creator: The creator of the video.
195 release_date: The date (YYYYMMDD) when the video was released.
196 timestamp: UNIX timestamp of the moment the video became available.
197 upload_date: Video upload date (YYYYMMDD).
198 If not explicitly set, calculated from timestamp.
199 uploader_id: Nickname or id of the video uploader.
200 uploader_url: Full URL to a personal webpage of the video uploader.
201 location: Physical location where the video was filmed.
202 subtitles: The available subtitles as a dictionary in the format
203 {tag: subformats}. "tag" is usually a language code, and
204 "subformats" is a list sorted from lower to higher
205 preference, each element is a dictionary with the "ext"
207 * "data": The subtitles file contents
208 * "url": A URL pointing to the subtitles file
209 "ext" will be calculated from URL if missing
210 automatic_captions: Like 'subtitles', used by the YoutubeIE for
211 automatically generated captions
212 duration: Length of the video in seconds, as an integer or float.
213 view_count: How many users have watched the video on the platform.
214 like_count: Number of positive ratings of the video
215 dislike_count: Number of negative ratings of the video
216 repost_count: Number of reposts of the video
217 average_rating: Average rating give by users, the scale used depends on the webpage
218 comment_count: Number of comments on the video
219 comments: A list of comments, each with one or more of the following
220 properties (all but one of text or html optional):
221 * "author" - human-readable name of the comment author
222 * "author_id" - user ID of the comment author
224 * "html" - Comment as HTML
225 * "text" - Plain text of the comment
226 * "timestamp" - UNIX timestamp of comment
227 * "parent" - ID of the comment this one is replying to.
228 Set to "root" to indicate that this is a
229 comment to the original video.
230 age_limit: Age restriction for the video, as an integer (years)
231 webpage_url: The URL to the video webpage, if given to youtube-dl it
232 should allow to get the same result again. (It will be set
233 by YoutubeDL if it's missing)
234 categories: A list of categories that the video falls in, for example
236 tags: A list of tags assigned to the video, e.g. ["sweden", "pop music"]
237 is_live: True, False, or None (=unknown). Whether this video is a
238 live stream that goes on instead of a fixed-length video.
239 start_time: Time in seconds where the reproduction should start, as
240 specified in the URL.
241 end_time: Time in seconds where the reproduction should end, as
242 specified in the URL.
244 The following fields should only be used when the video belongs to some logical
247 chapter: Name or title of the chapter the video belongs to.
248 chapter_number: Number of the chapter the video belongs to, as an integer.
249 chapter_id: Id of the chapter the video belongs to, as a unicode string.
251 The following fields should only be used when the video is an episode of some
252 series, programme or podcast:
254 series: Title of the series or programme the video episode belongs to.
255 season: Title of the season the video episode belongs to.
256 season_number: Number of the season the video episode belongs to, as an integer.
257 season_id: Id of the season the video episode belongs to, as a unicode string.
258 episode: Title of the video episode. Unlike mandatory video title field,
259 this field should denote the exact title of the video episode
260 without any kind of decoration.
261 episode_number: Number of the video episode within a season, as an integer.
262 episode_id: Id of the video episode, as a unicode string.
264 The following fields should only be used when the media is a track or a part of
267 track: Title of the track.
268 track_number: Number of the track within an album or a disc, as an integer.
269 track_id: Id of the track (useful in case of custom indexing, e.g. 6.iii),
271 artist: Artist(s) of the track.
272 genre: Genre(s) of the track.
273 album: Title of the album the track belongs to.
274 album_type: Type of the album (e.g. "Demo", "Full-length", "Split", "Compilation", etc).
275 album_artist: List of all artists appeared on the album (e.g.
276 "Ash Borer / Fell Voices" or "Various Artists", useful for splits
278 disc_number: Number of the disc or other physical medium the track belongs to,
280 release_year: Year (YYYY) when the album was released.
282 Unless mentioned otherwise, the fields should be Unicode strings.
284 Unless mentioned otherwise, None is equivalent to absence of information.
287 _type "playlist" indicates multiple videos.
288 There must be a key "entries", which is a list, an iterable, or a PagedList
289 object, each element of which is a valid dictionary by this specification.
291 Additionally, playlists can have "title", "description" and "id" attributes
292 with the same semantics as videos (see above).
295 _type "multi_video" indicates that there are multiple videos that
296 form a single show, for examples multiple acts of an opera or TV episode.
297 It must have an entries key like a playlist and contain all the keys
298 required for a video at the same time.
301 _type "url" indicates that the video must be extracted from another
302 location, possibly by a different extractor. Its only required key is:
303 "url" - the next URL to extract.
304 The key "ie_key" can be set to the class name (minus the trailing "IE",
305 e.g. "Youtube") if the extractor class is known in advance.
306 Additionally, the dictionary may have any properties of the resolved entity
307 known in advance, for example "title" if the title of the referred video is
311 _type "url_transparent" entities have the same specification as "url", but
312 indicate that the given additional information is more precise than the one
313 associated with the resolved URL.
314 This is useful when a site employs a video service that hosts the video and
315 its technical metadata, but that video service does not embed a useful
316 title, description etc.
319 Subclasses of this one should re-define the _real_initialize() and
320 _real_extract() methods and define a _VALID_URL regexp.
321 Probably, they should also be added to the list of extractors.
323 Finally, the _WORKING attribute should be set to False for broken IEs
324 in order to warn the users and skip the tests.
331 def __init__(self, downloader=None):
332 """Constructor. Receives an optional downloader."""
334 self.set_downloader(downloader)
337 def suitable(cls, url):
338 """Receives a URL and returns True if suitable for this IE."""
340 # This does not use has/getattr intentionally - we want to know whether
341 # we have cached the regexp for *this* class, whereas getattr would also
342 # match the superclass
343 if '_VALID_URL_RE' not in cls.__dict__:
344 cls._VALID_URL_RE = re.compile(cls._VALID_URL)
345 return cls._VALID_URL_RE.match(url) is not None
348 def _match_id(cls, url):
349 if '_VALID_URL_RE' not in cls.__dict__:
350 cls._VALID_URL_RE = re.compile(cls._VALID_URL)
351 m = cls._VALID_URL_RE.match(url)
357 """Getter method for _WORKING."""
360 def initialize(self):
361 """Initializes an instance (authentication, etc)."""
363 self._real_initialize()
366 def extract(self, url):
367 """Extracts URL information and returns it in list of dicts."""
370 return self._real_extract(url)
371 except ExtractorError:
373 except compat_http_client.IncompleteRead as e:
374 raise ExtractorError('A network error has occurred.', cause=e, expected=True)
375 except (KeyError, StopIteration) as e:
376 raise ExtractorError('An extractor error has occurred.', cause=e)
378 def set_downloader(self, downloader):
379 """Sets the downloader for this IE."""
380 self._downloader = downloader
382 def _real_initialize(self):
383 """Real initialization process. Redefine in subclasses."""
386 def _real_extract(self, url):
387 """Real extraction process. Redefine in subclasses."""
392 """A string for getting the InfoExtractor with get_info_extractor"""
393 return compat_str(cls.__name__[:-2])
397 return compat_str(type(self).__name__[:-2])
399 def _request_webpage(self, url_or_request, video_id, note=None, errnote=None, fatal=True, data=None, headers={}, query={}):
400 """ Returns the response handle """
402 self.report_download_webpage(video_id)
403 elif note is not False:
405 self.to_screen('%s' % (note,))
407 self.to_screen('%s: %s' % (video_id, note))
408 if isinstance(url_or_request, compat_urllib_request.Request):
409 url_or_request = update_Request(
410 url_or_request, data=data, headers=headers, query=query)
413 url_or_request = update_url_query(url_or_request, query)
414 if data is not None or headers:
415 url_or_request = sanitized_Request(url_or_request, data, headers)
417 return self._downloader.urlopen(url_or_request)
418 except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err:
422 errnote = 'Unable to download webpage'
424 errmsg = '%s: %s' % (errnote, error_to_compat_str(err))
426 raise ExtractorError(errmsg, sys.exc_info()[2], cause=err)
428 self._downloader.report_warning(errmsg)
431 def _download_webpage_handle(self, url_or_request, video_id, note=None, errnote=None, fatal=True, encoding=None, data=None, headers={}, query={}):
432 """ Returns a tuple (page content as string, URL handle) """
433 # Strip hashes from the URL (#1038)
434 if isinstance(url_or_request, (compat_str, str)):
435 url_or_request = url_or_request.partition('#')[0]
437 urlh = self._request_webpage(url_or_request, video_id, note, errnote, fatal, data=data, headers=headers, query=query)
441 content = self._webpage_read_content(urlh, url_or_request, video_id, note, errnote, fatal, encoding=encoding)
442 return (content, urlh)
445 def _guess_encoding_from_content(content_type, webpage_bytes):
446 m = re.match(r'[a-zA-Z0-9_.-]+/[a-zA-Z0-9_.-]+\s*;\s*charset=(.+)', content_type)
448 encoding = m.group(1)
450 m = re.search(br'<meta[^>]+charset=[\'"]?([^\'")]+)[ /\'">]',
451 webpage_bytes[:1024])
453 encoding = m.group(1).decode('ascii')
454 elif webpage_bytes.startswith(b'\xff\xfe'):
461 def _webpage_read_content(self, urlh, url_or_request, video_id, note=None, errnote=None, fatal=True, prefix=None, encoding=None):
462 content_type = urlh.headers.get('Content-Type', '')
463 webpage_bytes = urlh.read()
464 if prefix is not None:
465 webpage_bytes = prefix + webpage_bytes
467 encoding = self._guess_encoding_from_content(content_type, webpage_bytes)
468 if self._downloader.params.get('dump_intermediate_pages', False):
470 url = url_or_request.get_full_url()
471 except AttributeError:
473 self.to_screen('Dumping request to ' + url)
474 dump = base64.b64encode(webpage_bytes).decode('ascii')
475 self._downloader.to_screen(dump)
476 if self._downloader.params.get('write_pages', False):
478 url = url_or_request.get_full_url()
479 except AttributeError:
481 basen = '%s_%s' % (video_id, url)
483 h = '___' + hashlib.md5(basen.encode('utf-8')).hexdigest()
484 basen = basen[:240 - len(h)] + h
485 raw_filename = basen + '.dump'
486 filename = sanitize_filename(raw_filename, restricted=True)
487 self.to_screen('Saving request to ' + filename)
488 # Working around MAX_PATH limitation on Windows (see
489 # http://msdn.microsoft.com/en-us/library/windows/desktop/aa365247(v=vs.85).aspx)
490 if compat_os_name == 'nt':
491 absfilepath = os.path.abspath(filename)
492 if len(absfilepath) > 259:
493 filename = '\\\\?\\' + absfilepath
494 with open(filename, 'wb') as outf:
495 outf.write(webpage_bytes)
498 content = webpage_bytes.decode(encoding, 'replace')
500 content = webpage_bytes.decode('utf-8', 'replace')
502 if ('<title>Access to this site is blocked</title>' in content and
503 'Websense' in content[:512]):
504 msg = 'Access to this webpage has been blocked by Websense filtering software in your network.'
505 blocked_iframe = self._html_search_regex(
506 r'<iframe src="([^"]+)"', content,
507 'Websense information URL', default=None)
509 msg += ' Visit %s for more details' % blocked_iframe
510 raise ExtractorError(msg, expected=True)
511 if '<title>The URL you requested has been blocked</title>' in content[:512]:
513 'Access to this webpage has been blocked by Indian censorship. '
514 'Use a VPN or proxy server (with --proxy) to route around it.')
515 block_msg = self._html_search_regex(
516 r'</h1><p>(.*?)</p>',
517 content, 'block message', default=None)
519 msg += ' (Message: "%s")' % block_msg.replace('\n', ' ')
520 raise ExtractorError(msg, expected=True)
524 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={}):
525 """ Returns the data of the page as a string """
528 while success is False:
530 res = self._download_webpage_handle(url_or_request, video_id, note, errnote, fatal, encoding=encoding, data=data, headers=headers, query=query)
532 except compat_http_client.IncompleteRead as e:
534 if try_count >= tries:
536 self._sleep(timeout, video_id)
543 def _download_xml(self, url_or_request, video_id,
544 note='Downloading XML', errnote='Unable to download XML',
545 transform_source=None, fatal=True, encoding=None, data=None, headers={}, query={}):
546 """Return the xml as an xml.etree.ElementTree.Element"""
547 xml_string = self._download_webpage(
548 url_or_request, video_id, note, errnote, fatal=fatal, encoding=encoding, data=data, headers=headers, query=query)
549 if xml_string is False:
552 xml_string = transform_source(xml_string)
553 return compat_etree_fromstring(xml_string.encode('utf-8'))
555 def _download_json(self, url_or_request, video_id,
556 note='Downloading JSON metadata',
557 errnote='Unable to download JSON metadata',
558 transform_source=None,
559 fatal=True, encoding=None, data=None, headers={}, query={}):
560 json_string = self._download_webpage(
561 url_or_request, video_id, note, errnote, fatal=fatal,
562 encoding=encoding, data=data, headers=headers, query=query)
563 if (not fatal) and json_string is False:
565 return self._parse_json(
566 json_string, video_id, transform_source=transform_source, fatal=fatal)
568 def _parse_json(self, json_string, video_id, transform_source=None, fatal=True):
570 json_string = transform_source(json_string)
572 return json.loads(json_string)
573 except ValueError as ve:
574 errmsg = '%s: Failed to parse JSON ' % video_id
576 raise ExtractorError(errmsg, cause=ve)
578 self.report_warning(errmsg + str(ve))
580 def report_warning(self, msg, video_id=None):
581 idstr = '' if video_id is None else '%s: ' % video_id
582 self._downloader.report_warning(
583 '[%s] %s%s' % (self.IE_NAME, idstr, msg))
585 def to_screen(self, msg):
586 """Print msg to screen, prefixing it with '[ie_name]'"""
587 self._downloader.to_screen('[%s] %s' % (self.IE_NAME, msg))
589 def report_extraction(self, id_or_name):
590 """Report information extraction."""
591 self.to_screen('%s: Extracting information' % id_or_name)
593 def report_download_webpage(self, video_id):
594 """Report webpage download."""
595 self.to_screen('%s: Downloading webpage' % video_id)
597 def report_age_confirmation(self):
598 """Report attempt to confirm age."""
599 self.to_screen('Confirming age')
601 def report_login(self):
602 """Report attempt to log in."""
603 self.to_screen('Logging in')
606 def raise_login_required(msg='This video is only available for registered users'):
607 raise ExtractorError(
608 '%s. Use --username and --password or --netrc to provide account credentials.' % msg,
612 def raise_geo_restricted(msg='This video is not available from your location due to geo restriction'):
613 raise ExtractorError(
614 '%s. You might want to use --proxy to workaround.' % msg,
617 # Methods for following #608
619 def url_result(url, ie=None, video_id=None, video_title=None):
620 """Returns a URL that points to a page that should be processed"""
621 # TODO: ie should be the class used for getting the info
622 video_info = {'_type': 'url',
625 if video_id is not None:
626 video_info['id'] = video_id
627 if video_title is not None:
628 video_info['title'] = video_title
632 def playlist_result(entries, playlist_id=None, playlist_title=None, playlist_description=None):
633 """Returns a playlist"""
634 video_info = {'_type': 'playlist',
637 video_info['id'] = playlist_id
639 video_info['title'] = playlist_title
640 if playlist_description:
641 video_info['description'] = playlist_description
644 def _search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
646 Perform a regex search on the given string, using a single or a list of
647 patterns returning the first matching group.
648 In case of failure return a default value or raise a WARNING or a
649 RegexNotFoundError, depending on fatal, specifying the field name.
651 if isinstance(pattern, (str, compat_str, compiled_regex_type)):
652 mobj = re.search(pattern, string, flags)
655 mobj = re.search(p, string, flags)
659 if not self._downloader.params.get('no_color') and compat_os_name != 'nt' and sys.stderr.isatty():
660 _name = '\033[0;34m%s\033[0m' % name
666 # return the first matching group
667 return next(g for g in mobj.groups() if g is not None)
669 return mobj.group(group)
670 elif default is not NO_DEFAULT:
673 raise RegexNotFoundError('Unable to extract %s' % _name)
675 self._downloader.report_warning('unable to extract %s' % _name + bug_reports_message())
678 def _html_search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
680 Like _search_regex, but strips HTML tags and unescapes entities.
682 res = self._search_regex(pattern, string, name, default, fatal, flags, group)
684 return clean_html(res).strip()
688 def _get_netrc_login_info(self, netrc_machine=None):
691 netrc_machine = netrc_machine or self._NETRC_MACHINE
693 if self._downloader.params.get('usenetrc', False):
695 info = netrc.netrc().authenticators(netrc_machine)
700 raise netrc.NetrcParseError(
701 'No authenticators for %s' % netrc_machine)
702 except (IOError, netrc.NetrcParseError) as err:
703 self._downloader.report_warning(
704 'parsing .netrc: %s' % error_to_compat_str(err))
706 return username, password
708 def _get_login_info(self, username_option='username', password_option='password', netrc_machine=None):
710 Get the login info as (username, password)
711 First look for the manually specified credentials using username_option
712 and password_option as keys in params dictionary. If no such credentials
713 available look in the netrc file using the netrc_machine or _NETRC_MACHINE
715 If there's no info available, return (None, None)
717 if self._downloader is None:
720 downloader_params = self._downloader.params
722 # Attempt to use provided username and password or .netrc data
723 if downloader_params.get(username_option) is not None:
724 username = downloader_params[username_option]
725 password = downloader_params[password_option]
727 username, password = self._get_netrc_login_info(netrc_machine)
729 return username, password
731 def _get_tfa_info(self, note='two-factor verification code'):
733 Get the two-factor authentication info
734 TODO - asking the user will be required for sms/phone verify
735 currently just uses the command line option
736 If there's no info available, return None
738 if self._downloader is None:
740 downloader_params = self._downloader.params
742 if downloader_params.get('twofactor') is not None:
743 return downloader_params['twofactor']
745 return compat_getpass('Type %s and press [Return]: ' % note)
747 # Helper functions for extracting OpenGraph info
749 def _og_regexes(prop):
750 content_re = r'content=(?:"([^"]+?)"|\'([^\']+?)\'|\s*([^\s"\'=<>`]+?))'
751 property_re = (r'(?:name|property)=(?:\'og:%(prop)s\'|"og:%(prop)s"|\s*og:%(prop)s\b)'
752 % {'prop': re.escape(prop)})
753 template = r'<meta[^>]+?%s[^>]+?%s'
755 template % (property_re, content_re),
756 template % (content_re, property_re),
760 def _meta_regex(prop):
761 return r'''(?isx)<meta
762 (?=[^>]+(?:itemprop|name|property|id|http-equiv)=(["\']?)%s\1)
763 [^>]+?content=(["\'])(?P<content>.*?)\2''' % re.escape(prop)
765 def _og_search_property(self, prop, html, name=None, **kargs):
766 if not isinstance(prop, (list, tuple)):
769 name = 'OpenGraph %s' % prop[0]
772 og_regexes.extend(self._og_regexes(p))
773 escaped = self._search_regex(og_regexes, html, name, flags=re.DOTALL, **kargs)
776 return unescapeHTML(escaped)
778 def _og_search_thumbnail(self, html, **kargs):
779 return self._og_search_property('image', html, 'thumbnail URL', fatal=False, **kargs)
781 def _og_search_description(self, html, **kargs):
782 return self._og_search_property('description', html, fatal=False, **kargs)
784 def _og_search_title(self, html, **kargs):
785 return self._og_search_property('title', html, **kargs)
787 def _og_search_video_url(self, html, name='video url', secure=True, **kargs):
788 regexes = self._og_regexes('video') + self._og_regexes('video:url')
790 regexes = self._og_regexes('video:secure_url') + regexes
791 return self._html_search_regex(regexes, html, name, **kargs)
793 def _og_search_url(self, html, **kargs):
794 return self._og_search_property('url', html, **kargs)
796 def _html_search_meta(self, name, html, display_name=None, fatal=False, **kwargs):
797 if not isinstance(name, (list, tuple)):
799 if display_name is None:
800 display_name = name[0]
801 return self._html_search_regex(
802 [self._meta_regex(n) for n in name],
803 html, display_name, fatal=fatal, group='content', **kwargs)
805 def _dc_search_uploader(self, html):
806 return self._html_search_meta('dc.creator', html, 'uploader')
808 def _rta_search(self, html):
809 # See http://www.rtalabel.org/index.php?content=howtofaq#single
810 if re.search(r'(?ix)<meta\s+name="rating"\s+'
811 r' content="RTA-5042-1996-1400-1577-RTA"',
816 def _media_rating_search(self, html):
817 # See http://www.tjg-designs.com/WP/metadata-code-examples-adding-metadata-to-your-web-pages/
818 rating = self._html_search_meta('rating', html)
830 return RATING_TABLE.get(rating.lower())
832 def _family_friendly_search(self, html):
833 # See http://schema.org/VideoObject
834 family_friendly = self._html_search_meta('isFamilyFriendly', html)
836 if not family_friendly:
845 return RATING_TABLE.get(family_friendly.lower())
847 def _twitter_search_player(self, html):
848 return self._html_search_meta('twitter:player', html,
849 'twitter card player')
851 def _search_json_ld(self, html, video_id, expected_type=None, **kwargs):
852 json_ld = self._search_regex(
853 r'(?s)<script[^>]+type=(["\'])application/ld\+json\1[^>]*>(?P<json_ld>.+?)</script>',
854 html, 'JSON-LD', group='json_ld', **kwargs)
855 default = kwargs.get('default', NO_DEFAULT)
857 return default if default is not NO_DEFAULT else {}
858 # JSON-LD may be malformed and thus `fatal` should be respected.
859 # At the same time `default` may be passed that assumes `fatal=False`
860 # for _search_regex. Let's simulate the same behavior here as well.
861 fatal = kwargs.get('fatal', True) if default == NO_DEFAULT else False
862 return self._json_ld(json_ld, video_id, fatal=fatal, expected_type=expected_type)
864 def _json_ld(self, json_ld, video_id, fatal=True, expected_type=None):
865 if isinstance(json_ld, compat_str):
866 json_ld = self._parse_json(json_ld, video_id, fatal=fatal)
870 if not isinstance(json_ld, (list, tuple, dict)):
872 if isinstance(json_ld, dict):
875 if e.get('@context') == 'http://schema.org':
876 item_type = e.get('@type')
877 if expected_type is not None and expected_type != item_type:
879 if item_type == 'TVEpisode':
881 'episode': unescapeHTML(e.get('name')),
882 'episode_number': int_or_none(e.get('episodeNumber')),
883 'description': unescapeHTML(e.get('description')),
885 part_of_season = e.get('partOfSeason')
886 if isinstance(part_of_season, dict) and part_of_season.get('@type') == 'TVSeason':
887 info['season_number'] = int_or_none(part_of_season.get('seasonNumber'))
888 part_of_series = e.get('partOfSeries') or e.get('partOfTVSeries')
889 if isinstance(part_of_series, dict) and part_of_series.get('@type') == 'TVSeries':
890 info['series'] = unescapeHTML(part_of_series.get('name'))
891 elif item_type == 'Article':
893 'timestamp': parse_iso8601(e.get('datePublished')),
894 'title': unescapeHTML(e.get('headline')),
895 'description': unescapeHTML(e.get('articleBody')),
897 elif item_type == 'VideoObject':
899 'url': e.get('contentUrl'),
900 'title': unescapeHTML(e.get('name')),
901 'description': unescapeHTML(e.get('description')),
902 'thumbnail': e.get('thumbnailUrl') or e.get('thumbnailURL'),
903 'duration': parse_duration(e.get('duration')),
904 'timestamp': unified_timestamp(e.get('uploadDate')),
905 'filesize': float_or_none(e.get('contentSize')),
906 'tbr': int_or_none(e.get('bitrate')),
907 'width': int_or_none(e.get('width')),
908 'height': int_or_none(e.get('height')),
911 return dict((k, v) for k, v in info.items() if v is not None)
914 def _hidden_inputs(html):
915 html = re.sub(r'<!--(?:(?!<!--).)*-->', '', html)
917 for input in re.findall(r'(?i)(<input[^>]+>)', html):
918 attrs = extract_attributes(input)
921 if attrs.get('type') not in ('hidden', 'submit'):
923 name = attrs.get('name') or attrs.get('id')
924 value = attrs.get('value')
925 if name and value is not None:
926 hidden_inputs[name] = value
929 def _form_hidden_inputs(self, form_id, html):
930 form = self._search_regex(
931 r'(?is)<form[^>]+?id=(["\'])%s\1[^>]*>(?P<form>.+?)</form>' % form_id,
932 html, '%s form' % form_id, group='form')
933 return self._hidden_inputs(form)
935 def _sort_formats(self, formats, field_preference=None):
937 raise ExtractorError('No video formats found')
940 # Automatically determine tbr when missing based on abr and vbr (improves
941 # formats sorting in some cases)
942 if 'tbr' not in f and f.get('abr') is not None and f.get('vbr') is not None:
943 f['tbr'] = f['abr'] + f['vbr']
946 # TODO remove the following workaround
947 from ..utils import determine_ext
948 if not f.get('ext') and 'url' in f:
949 f['ext'] = determine_ext(f['url'])
951 if isinstance(field_preference, (list, tuple)):
954 if f.get(field) is not None
955 else ('' if field == 'format_id' else -1)
956 for field in field_preference)
958 preference = f.get('preference')
959 if preference is None:
961 if f.get('ext') in ['f4f', 'f4m']: # Not yet supported
964 protocol = f.get('protocol') or determine_protocol(f)
965 proto_preference = 0 if protocol in ['http', 'https'] else (-0.5 if protocol == 'rtsp' else -0.1)
967 if f.get('vcodec') == 'none': # audio only
969 if self._downloader.params.get('prefer_free_formats'):
970 ORDER = ['aac', 'mp3', 'm4a', 'webm', 'ogg', 'opus']
972 ORDER = ['webm', 'opus', 'ogg', 'mp3', 'aac', 'm4a']
975 audio_ext_preference = ORDER.index(f['ext'])
977 audio_ext_preference = -1
979 if f.get('acodec') == 'none': # video only
981 if self._downloader.params.get('prefer_free_formats'):
982 ORDER = ['flv', 'mp4', 'webm']
984 ORDER = ['webm', 'flv', 'mp4']
986 ext_preference = ORDER.index(f['ext'])
989 audio_ext_preference = 0
993 f.get('language_preference') if f.get('language_preference') is not None else -1,
994 f.get('quality') if f.get('quality') is not None else -1,
995 f.get('tbr') if f.get('tbr') is not None else -1,
996 f.get('filesize') if f.get('filesize') is not None else -1,
997 f.get('vbr') if f.get('vbr') is not None else -1,
998 f.get('height') if f.get('height') is not None else -1,
999 f.get('width') if f.get('width') is not None else -1,
1002 f.get('abr') if f.get('abr') is not None else -1,
1003 audio_ext_preference,
1004 f.get('fps') if f.get('fps') is not None else -1,
1005 f.get('filesize_approx') if f.get('filesize_approx') is not None else -1,
1006 f.get('source_preference') if f.get('source_preference') is not None else -1,
1007 f.get('format_id') if f.get('format_id') is not None else '',
1009 formats.sort(key=_formats_key)
1011 def _check_formats(self, formats, video_id):
1013 formats[:] = filter(
1014 lambda f: self._is_valid_url(
1016 item='%s video format' % f.get('format_id') if f.get('format_id') else 'video'),
1020 def _remove_duplicate_formats(formats):
1024 if f['url'] not in format_urls:
1025 format_urls.add(f['url'])
1026 unique_formats.append(f)
1027 formats[:] = unique_formats
1029 def _is_valid_url(self, url, video_id, item='video', headers={}):
1030 url = self._proto_relative_url(url, scheme='http:')
1031 # For now assume non HTTP(S) URLs always valid
1032 if not (url.startswith('http://') or url.startswith('https://')):
1035 self._request_webpage(url, video_id, 'Checking %s URL' % item, headers=headers)
1037 except ExtractorError as e:
1038 if isinstance(e.cause, compat_urllib_error.URLError):
1040 '%s: %s URL is invalid, skipping' % (video_id, item))
1044 def http_scheme(self):
1045 """ Either "http:" or "https:", depending on the user's preferences """
1048 if self._downloader.params.get('prefer_insecure', False)
1051 def _proto_relative_url(self, url, scheme=None):
1054 if url.startswith('//'):
1056 scheme = self.http_scheme()
1061 def _sleep(self, timeout, video_id, msg_template=None):
1062 if msg_template is None:
1063 msg_template = '%(video_id)s: Waiting for %(timeout)s seconds'
1064 msg = msg_template % {'video_id': video_id, 'timeout': timeout}
1068 def _extract_f4m_formats(self, manifest_url, video_id, preference=None, f4m_id=None,
1069 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1070 fatal=True, m3u8_id=None):
1071 manifest = self._download_xml(
1072 manifest_url, video_id, 'Downloading f4m manifest',
1073 'Unable to download f4m manifest',
1074 # Some manifests may be malformed, e.g. prosiebensat1 generated manifests
1075 # (see https://github.com/rg3/youtube-dl/issues/6215#issuecomment-121704244)
1076 transform_source=transform_source,
1079 if manifest is False:
1082 return self._parse_f4m_formats(
1083 manifest, manifest_url, video_id, preference=preference, f4m_id=f4m_id,
1084 transform_source=transform_source, fatal=fatal, m3u8_id=m3u8_id)
1086 def _parse_f4m_formats(self, manifest, manifest_url, video_id, preference=None, f4m_id=None,
1087 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1088 fatal=True, m3u8_id=None):
1089 # currently youtube-dl cannot decode the playerVerificationChallenge as Akamai uses Adobe Alchemy
1090 akamai_pv = manifest.find('{http://ns.adobe.com/f4m/1.0}pv-2.0')
1091 if akamai_pv is not None and ';' in akamai_pv.text:
1092 playerVerificationChallenge = akamai_pv.text.split(';')[0]
1093 if playerVerificationChallenge.strip() != '':
1097 manifest_version = '1.0'
1098 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/1.0}media')
1100 manifest_version = '2.0'
1101 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/2.0}media')
1102 # Remove unsupported DRM protected media from final formats
1103 # rendition (see https://github.com/rg3/youtube-dl/issues/8573).
1104 media_nodes = remove_encrypted_media(media_nodes)
1107 base_url = xpath_text(
1108 manifest, ['{http://ns.adobe.com/f4m/1.0}baseURL', '{http://ns.adobe.com/f4m/2.0}baseURL'],
1109 'base URL', default=None)
1111 base_url = base_url.strip()
1113 bootstrap_info = xpath_element(
1114 manifest, ['{http://ns.adobe.com/f4m/1.0}bootstrapInfo', '{http://ns.adobe.com/f4m/2.0}bootstrapInfo'],
1115 'bootstrap info', default=None)
1118 mime_type = xpath_text(
1119 manifest, ['{http://ns.adobe.com/f4m/1.0}mimeType', '{http://ns.adobe.com/f4m/2.0}mimeType'],
1120 'base URL', default=None)
1121 if mime_type and mime_type.startswith('audio/'):
1124 for i, media_el in enumerate(media_nodes):
1125 tbr = int_or_none(media_el.attrib.get('bitrate'))
1126 width = int_or_none(media_el.attrib.get('width'))
1127 height = int_or_none(media_el.attrib.get('height'))
1128 format_id = '-'.join(filter(None, [f4m_id, compat_str(i if tbr is None else tbr)]))
1129 # If <bootstrapInfo> is present, the specified f4m is a
1130 # stream-level manifest, and only set-level manifests may refer to
1131 # external resources. See section 11.4 and section 4 of F4M spec
1132 if bootstrap_info is None:
1134 # @href is introduced in 2.0, see section 11.6 of F4M spec
1135 if manifest_version == '2.0':
1136 media_url = media_el.attrib.get('href')
1137 if media_url is None:
1138 media_url = media_el.attrib.get('url')
1142 media_url if media_url.startswith('http://') or media_url.startswith('https://')
1143 else ((base_url or '/'.join(manifest_url.split('/')[:-1])) + '/' + media_url))
1144 # If media_url is itself a f4m manifest do the recursive extraction
1145 # since bitrates in parent manifest (this one) and media_url manifest
1146 # may differ leading to inability to resolve the format by requested
1147 # bitrate in f4m downloader
1148 ext = determine_ext(manifest_url)
1150 f4m_formats = self._extract_f4m_formats(
1151 manifest_url, video_id, preference=preference, f4m_id=f4m_id,
1152 transform_source=transform_source, fatal=fatal)
1153 # Sometimes stream-level manifest contains single media entry that
1154 # does not contain any quality metadata (e.g. http://matchtv.ru/#live-player).
1155 # At the same time parent's media entry in set-level manifest may
1156 # contain it. We will copy it from parent in such cases.
1157 if len(f4m_formats) == 1:
1160 'tbr': f.get('tbr') or tbr,
1161 'width': f.get('width') or width,
1162 'height': f.get('height') or height,
1163 'format_id': f.get('format_id') if not tbr else format_id,
1166 formats.extend(f4m_formats)
1169 formats.extend(self._extract_m3u8_formats(
1170 manifest_url, video_id, 'mp4', preference=preference,
1171 m3u8_id=m3u8_id, fatal=fatal))
1174 'format_id': format_id,
1175 'url': manifest_url,
1176 'manifest_url': manifest_url,
1177 'ext': 'flv' if bootstrap_info is not None else None,
1182 'preference': preference,
1186 def _m3u8_meta_format(self, m3u8_url, ext=None, preference=None, m3u8_id=None):
1188 'format_id': '-'.join(filter(None, [m3u8_id, 'meta'])),
1192 'preference': preference - 100 if preference else -100,
1193 'resolution': 'multiple',
1194 'format_note': 'Quality selection URL',
1197 def _extract_m3u8_formats(self, m3u8_url, video_id, ext=None,
1198 entry_protocol='m3u8', preference=None,
1199 m3u8_id=None, note=None, errnote=None,
1200 fatal=True, live=False):
1202 res = self._download_webpage_handle(
1204 note=note or 'Downloading m3u8 information',
1205 errnote=errnote or 'Failed to download m3u8 information',
1209 m3u8_doc, urlh = res
1210 m3u8_url = urlh.geturl()
1212 if '#EXT-X-FAXS-CM:' in m3u8_doc: # Adobe Flash Access
1215 formats = [self._m3u8_meta_format(m3u8_url, ext, preference, m3u8_id)]
1217 format_url = lambda u: (
1219 if re.match(r'^https?://', u)
1220 else compat_urlparse.urljoin(m3u8_url, u))
1222 # We should try extracting formats only from master playlists [1], i.e.
1223 # playlists that describe available qualities. On the other hand media
1224 # playlists [2] should be returned as is since they contain just the media
1225 # without qualities renditions.
1226 # Fortunately, master playlist can be easily distinguished from media
1227 # playlist based on particular tags availability. As of [1, 2] master
1228 # playlist tags MUST NOT appear in a media playist and vice versa.
1229 # As of [3] #EXT-X-TARGETDURATION tag is REQUIRED for every media playlist
1230 # and MUST NOT appear in master playlist thus we can clearly detect media
1231 # playlist with this criterion.
1232 # 1. https://tools.ietf.org/html/draft-pantos-http-live-streaming-17#section-4.3.4
1233 # 2. https://tools.ietf.org/html/draft-pantos-http-live-streaming-17#section-4.3.3
1234 # 3. https://tools.ietf.org/html/draft-pantos-http-live-streaming-17#section-4.3.3.1
1235 if '#EXT-X-TARGETDURATION' in m3u8_doc: # media playlist, return as is
1238 'format_id': m3u8_id,
1240 'protocol': entry_protocol,
1241 'preference': preference,
1243 audio_in_video_stream = {}
1246 for line in m3u8_doc.splitlines():
1247 if line.startswith('#EXT-X-STREAM-INF:'):
1248 last_info = parse_m3u8_attributes(line)
1249 elif line.startswith('#EXT-X-MEDIA:'):
1250 media = parse_m3u8_attributes(line)
1251 media_type = media.get('TYPE')
1252 if media_type in ('VIDEO', 'AUDIO'):
1253 group_id = media.get('GROUP-ID')
1254 media_url = media.get('URI')
1257 for v in (group_id, media.get('NAME')):
1261 'format_id': '-'.join(format_id),
1262 'url': format_url(media_url),
1263 'language': media.get('LANGUAGE'),
1265 'protocol': entry_protocol,
1266 'preference': preference,
1268 if media_type == 'AUDIO':
1269 f['vcodec'] = 'none'
1270 if group_id and not audio_in_video_stream.get(group_id):
1271 audio_in_video_stream[group_id] = False
1274 # When there is no URI in EXT-X-MEDIA let this tag's
1275 # data be used by regular URI lines below
1277 if media_type == 'AUDIO' and group_id:
1278 audio_in_video_stream[group_id] = True
1279 elif line.startswith('#') or not line.strip():
1282 tbr = int_or_none(last_info.get('AVERAGE-BANDWIDTH') or last_info.get('BANDWIDTH'), scale=1000)
1285 format_id.append(m3u8_id)
1286 # Despite specification does not mention NAME attribute for
1287 # EXT-X-STREAM-INF it still sometimes may be present
1288 stream_name = last_info.get('NAME') or last_media.get('NAME')
1289 # Bandwidth of live streams may differ over time thus making
1290 # format_id unpredictable. So it's better to keep provided
1293 format_id.append(stream_name if stream_name else '%d' % (tbr if tbr else len(formats)))
1294 manifest_url = format_url(line.strip())
1296 'format_id': '-'.join(format_id),
1297 'url': manifest_url,
1298 'manifest_url': manifest_url,
1301 'fps': float_or_none(last_info.get('FRAME-RATE')),
1302 'protocol': entry_protocol,
1303 'preference': preference,
1305 resolution = last_info.get('RESOLUTION')
1307 mobj = re.search(r'(?P<width>\d+)[xX](?P<height>\d+)', resolution)
1309 f['width'] = int(mobj.group('width'))
1310 f['height'] = int(mobj.group('height'))
1311 # Unified Streaming Platform
1313 r'audio.*?(?:%3D|=)(\d+)(?:-video.*?(?:%3D|=)(\d+))?', f['url'])
1315 abr, vbr = mobj.groups()
1316 abr, vbr = float_or_none(abr, 1000), float_or_none(vbr, 1000)
1321 f.update(parse_codecs(last_info.get('CODECS')))
1322 if audio_in_video_stream.get(last_info.get('AUDIO')) is False and f['vcodec'] != 'none':
1323 # TODO: update acodec for audio only formats with the same GROUP-ID
1324 f['acodec'] = 'none'
1331 def _xpath_ns(path, namespace=None):
1335 for c in path.split('/'):
1336 if not c or c == '.':
1339 out.append('{%s}%s' % (namespace, c))
1340 return '/'.join(out)
1342 def _extract_smil_formats(self, smil_url, video_id, fatal=True, f4m_params=None, transform_source=None):
1343 smil = self._download_smil(smil_url, video_id, fatal=fatal, transform_source=transform_source)
1349 namespace = self._parse_smil_namespace(smil)
1351 return self._parse_smil_formats(
1352 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
1354 def _extract_smil_info(self, smil_url, video_id, fatal=True, f4m_params=None):
1355 smil = self._download_smil(smil_url, video_id, fatal=fatal)
1358 return self._parse_smil(smil, smil_url, video_id, f4m_params=f4m_params)
1360 def _download_smil(self, smil_url, video_id, fatal=True, transform_source=None):
1361 return self._download_xml(
1362 smil_url, video_id, 'Downloading SMIL file',
1363 'Unable to download SMIL file', fatal=fatal, transform_source=transform_source)
1365 def _parse_smil(self, smil, smil_url, video_id, f4m_params=None):
1366 namespace = self._parse_smil_namespace(smil)
1368 formats = self._parse_smil_formats(
1369 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
1370 subtitles = self._parse_smil_subtitles(smil, namespace=namespace)
1372 video_id = os.path.splitext(url_basename(smil_url))[0]
1376 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
1377 name = meta.attrib.get('name')
1378 content = meta.attrib.get('content')
1379 if not name or not content:
1381 if not title and name == 'title':
1383 elif not description and name in ('description', 'abstract'):
1384 description = content
1385 elif not upload_date and name == 'date':
1386 upload_date = unified_strdate(content)
1389 'id': image.get('type'),
1390 'url': image.get('src'),
1391 'width': int_or_none(image.get('width')),
1392 'height': int_or_none(image.get('height')),
1393 } for image in smil.findall(self._xpath_ns('.//image', namespace)) if image.get('src')]
1397 'title': title or video_id,
1398 'description': description,
1399 'upload_date': upload_date,
1400 'thumbnails': thumbnails,
1402 'subtitles': subtitles,
1405 def _parse_smil_namespace(self, smil):
1406 return self._search_regex(
1407 r'(?i)^{([^}]+)?}smil$', smil.tag, 'namespace', default=None)
1409 def _parse_smil_formats(self, smil, smil_url, video_id, namespace=None, f4m_params=None, transform_rtmp_url=None):
1411 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
1412 b = meta.get('base') or meta.get('httpBase')
1423 media = smil.findall(self._xpath_ns('.//video', namespace)) + smil.findall(self._xpath_ns('.//audio', namespace))
1424 for medium in media:
1425 src = medium.get('src')
1426 if not src or src in srcs:
1430 bitrate = float_or_none(medium.get('system-bitrate') or medium.get('systemBitrate'), 1000)
1431 filesize = int_or_none(medium.get('size') or medium.get('fileSize'))
1432 width = int_or_none(medium.get('width'))
1433 height = int_or_none(medium.get('height'))
1434 proto = medium.get('proto')
1435 ext = medium.get('ext')
1436 src_ext = determine_ext(src)
1437 streamer = medium.get('streamer') or base
1439 if proto == 'rtmp' or streamer.startswith('rtmp'):
1445 'format_id': 'rtmp-%d' % (rtmp_count if bitrate is None else bitrate),
1447 'filesize': filesize,
1451 if transform_rtmp_url:
1452 streamer, src = transform_rtmp_url(streamer, src)
1453 formats[-1].update({
1459 src_url = src if src.startswith('http') else compat_urlparse.urljoin(base, src)
1460 src_url = src_url.strip()
1462 if proto == 'm3u8' or src_ext == 'm3u8':
1463 m3u8_formats = self._extract_m3u8_formats(
1464 src_url, video_id, ext or 'mp4', m3u8_id='hls', fatal=False)
1465 if len(m3u8_formats) == 1:
1467 m3u8_formats[0].update({
1468 'format_id': 'hls-%d' % (m3u8_count if bitrate is None else bitrate),
1473 formats.extend(m3u8_formats)
1476 if src_ext == 'f4m':
1481 'plugin': 'flowplayer-3.2.0.1',
1483 f4m_url += '&' if '?' in f4m_url else '?'
1484 f4m_url += compat_urllib_parse_urlencode(f4m_params)
1485 formats.extend(self._extract_f4m_formats(f4m_url, video_id, f4m_id='hds', fatal=False))
1488 if src_url.startswith('http') and self._is_valid_url(src, video_id):
1492 'ext': ext or src_ext or 'flv',
1493 'format_id': 'http-%d' % (bitrate or http_count),
1495 'filesize': filesize,
1503 def _parse_smil_subtitles(self, smil, namespace=None, subtitles_lang='en'):
1506 for num, textstream in enumerate(smil.findall(self._xpath_ns('.//textstream', namespace))):
1507 src = textstream.get('src')
1508 if not src or src in urls:
1511 ext = textstream.get('ext') or mimetype2ext(textstream.get('type')) or determine_ext(src)
1512 lang = textstream.get('systemLanguage') or textstream.get('systemLanguageName') or textstream.get('lang') or subtitles_lang
1513 subtitles.setdefault(lang, []).append({
1519 def _extract_xspf_playlist(self, playlist_url, playlist_id, fatal=True):
1520 xspf = self._download_xml(
1521 playlist_url, playlist_id, 'Downloading xpsf playlist',
1522 'Unable to download xspf manifest', fatal=fatal)
1525 return self._parse_xspf(xspf, playlist_id)
1527 def _parse_xspf(self, playlist, playlist_id):
1529 'xspf': 'http://xspf.org/ns/0/',
1530 's1': 'http://static.streamone.nl/player/ns/0',
1534 for track in playlist.findall(xpath_with_ns('./xspf:trackList/xspf:track', NS_MAP)):
1536 track, xpath_with_ns('./xspf:title', NS_MAP), 'title', default=playlist_id)
1537 description = xpath_text(
1538 track, xpath_with_ns('./xspf:annotation', NS_MAP), 'description')
1539 thumbnail = xpath_text(
1540 track, xpath_with_ns('./xspf:image', NS_MAP), 'thumbnail')
1541 duration = float_or_none(
1542 xpath_text(track, xpath_with_ns('./xspf:duration', NS_MAP), 'duration'), 1000)
1545 'url': location.text,
1546 'format_id': location.get(xpath_with_ns('s1:label', NS_MAP)),
1547 'width': int_or_none(location.get(xpath_with_ns('s1:width', NS_MAP))),
1548 'height': int_or_none(location.get(xpath_with_ns('s1:height', NS_MAP))),
1549 } for location in track.findall(xpath_with_ns('./xspf:location', NS_MAP))]
1550 self._sort_formats(formats)
1555 'description': description,
1556 'thumbnail': thumbnail,
1557 'duration': duration,
1562 def _extract_mpd_formats(self, mpd_url, video_id, mpd_id=None, note=None, errnote=None, fatal=True, formats_dict={}):
1563 res = self._download_webpage_handle(
1565 note=note or 'Downloading MPD manifest',
1566 errnote=errnote or 'Failed to download MPD manifest',
1571 mpd_base_url = base_url(urlh.geturl())
1573 return self._parse_mpd_formats(
1574 compat_etree_fromstring(mpd.encode('utf-8')), mpd_id, mpd_base_url,
1575 formats_dict=formats_dict, mpd_url=mpd_url)
1577 def _parse_mpd_formats(self, mpd_doc, mpd_id=None, mpd_base_url='', formats_dict={}, mpd_url=None):
1579 Parse formats from MPD manifest.
1581 1. MPEG-DASH Standard, ISO/IEC 23009-1:2014(E),
1582 http://standards.iso.org/ittf/PubliclyAvailableStandards/c065274_ISO_IEC_23009-1_2014.zip
1583 2. https://en.wikipedia.org/wiki/Dynamic_Adaptive_Streaming_over_HTTP
1585 if mpd_doc.get('type') == 'dynamic':
1588 namespace = self._search_regex(r'(?i)^{([^}]+)?}MPD$', mpd_doc.tag, 'namespace', default=None)
1591 return self._xpath_ns(path, namespace)
1593 def is_drm_protected(element):
1594 return element.find(_add_ns('ContentProtection')) is not None
1596 def extract_multisegment_info(element, ms_parent_info):
1597 ms_info = ms_parent_info.copy()
1599 # As per [1, 5.3.9.2.2] SegmentList and SegmentTemplate share some
1600 # common attributes and elements. We will only extract relevant
1602 def extract_common(source):
1603 segment_timeline = source.find(_add_ns('SegmentTimeline'))
1604 if segment_timeline is not None:
1605 s_e = segment_timeline.findall(_add_ns('S'))
1607 ms_info['total_number'] = 0
1610 r = int(s.get('r', 0))
1611 ms_info['total_number'] += 1 + r
1612 ms_info['s'].append({
1613 't': int(s.get('t', 0)),
1614 # @d is mandatory (see [1, 5.3.9.6.2, Table 17, page 60])
1615 'd': int(s.attrib['d']),
1618 start_number = source.get('startNumber')
1620 ms_info['start_number'] = int(start_number)
1621 timescale = source.get('timescale')
1623 ms_info['timescale'] = int(timescale)
1624 segment_duration = source.get('duration')
1625 if segment_duration:
1626 ms_info['segment_duration'] = int(segment_duration)
1628 def extract_Initialization(source):
1629 initialization = source.find(_add_ns('Initialization'))
1630 if initialization is not None:
1631 ms_info['initialization_url'] = initialization.attrib['sourceURL']
1633 segment_list = element.find(_add_ns('SegmentList'))
1634 if segment_list is not None:
1635 extract_common(segment_list)
1636 extract_Initialization(segment_list)
1637 segment_urls_e = segment_list.findall(_add_ns('SegmentURL'))
1639 ms_info['segment_urls'] = [segment.attrib['media'] for segment in segment_urls_e]
1641 segment_template = element.find(_add_ns('SegmentTemplate'))
1642 if segment_template is not None:
1643 extract_common(segment_template)
1644 media = segment_template.get('media')
1646 ms_info['media'] = media
1647 initialization = segment_template.get('initialization')
1649 ms_info['initialization'] = initialization
1651 extract_Initialization(segment_template)
1654 mpd_duration = parse_duration(mpd_doc.get('mediaPresentationDuration'))
1656 for period in mpd_doc.findall(_add_ns('Period')):
1657 period_duration = parse_duration(period.get('duration')) or mpd_duration
1658 period_ms_info = extract_multisegment_info(period, {
1662 for adaptation_set in period.findall(_add_ns('AdaptationSet')):
1663 if is_drm_protected(adaptation_set):
1665 adaption_set_ms_info = extract_multisegment_info(adaptation_set, period_ms_info)
1666 for representation in adaptation_set.findall(_add_ns('Representation')):
1667 if is_drm_protected(representation):
1669 representation_attrib = adaptation_set.attrib.copy()
1670 representation_attrib.update(representation.attrib)
1671 # According to [1, 5.3.7.2, Table 9, page 41], @mimeType is mandatory
1672 mime_type = representation_attrib['mimeType']
1673 content_type = mime_type.split('/')[0]
1674 if content_type == 'text':
1675 # TODO implement WebVTT downloading
1677 elif content_type == 'video' or content_type == 'audio':
1679 for element in (representation, adaptation_set, period, mpd_doc):
1680 base_url_e = element.find(_add_ns('BaseURL'))
1681 if base_url_e is not None:
1682 base_url = base_url_e.text + base_url
1683 if re.match(r'^https?://', base_url):
1685 if mpd_base_url and not re.match(r'^https?://', base_url):
1686 if not mpd_base_url.endswith('/') and not base_url.startswith('/'):
1688 base_url = mpd_base_url + base_url
1689 representation_id = representation_attrib.get('id')
1690 lang = representation_attrib.get('lang')
1691 url_el = representation.find(_add_ns('BaseURL'))
1692 filesize = int_or_none(url_el.attrib.get('{http://youtube.com/yt/2012/10/10}contentLength') if url_el is not None else None)
1693 bandwidth = int_or_none(representation_attrib.get('bandwidth'))
1695 'format_id': '%s-%s' % (mpd_id, representation_id) if mpd_id else representation_id,
1697 'manifest_url': mpd_url,
1698 'ext': mimetype2ext(mime_type),
1699 'width': int_or_none(representation_attrib.get('width')),
1700 'height': int_or_none(representation_attrib.get('height')),
1701 'tbr': int_or_none(bandwidth, 1000),
1702 'asr': int_or_none(representation_attrib.get('audioSamplingRate')),
1703 'fps': int_or_none(representation_attrib.get('frameRate')),
1704 'language': lang if lang not in ('mul', 'und', 'zxx', 'mis') else None,
1705 'format_note': 'DASH %s' % content_type,
1706 'filesize': filesize,
1708 f.update(parse_codecs(representation_attrib.get('codecs')))
1709 representation_ms_info = extract_multisegment_info(representation, adaption_set_ms_info)
1711 def prepare_template(template_name, identifiers):
1712 t = representation_ms_info[template_name]
1713 t = t.replace('$RepresentationID$', representation_id)
1714 t = re.sub(r'\$(%s)\$' % '|'.join(identifiers), r'%(\1)d', t)
1715 t = re.sub(r'\$(%s)%%([^$]+)\$' % '|'.join(identifiers), r'%(\1)\2', t)
1716 t.replace('$$', '$')
1719 # @initialization is a regular template like @media one
1720 # so it should be handled just the same way (see
1721 # https://github.com/rg3/youtube-dl/issues/11605)
1722 if 'initialization' in representation_ms_info:
1723 initialization_template = prepare_template(
1725 # As per [1, 5.3.9.4.2, Table 15, page 54] $Number$ and
1726 # $Time$ shall not be included for @initialization thus
1727 # only $Bandwidth$ remains
1729 representation_ms_info['initialization_url'] = initialization_template % {
1730 'Bandwidth': bandwidth,
1733 if 'segment_urls' not in representation_ms_info and 'media' in representation_ms_info:
1735 media_template = prepare_template('media', ('Number', 'Bandwidth', 'Time'))
1737 # As per [1, 5.3.9.4.4, Table 16, page 55] $Number$ and $Time$
1738 # can't be used at the same time
1739 if '%(Number' in media_template and 's' not in representation_ms_info:
1740 segment_duration = None
1741 if 'total_number' not in representation_ms_info and 'segment_duration':
1742 segment_duration = float_or_none(representation_ms_info['segment_duration'], representation_ms_info['timescale'])
1743 representation_ms_info['total_number'] = int(math.ceil(float(period_duration) / segment_duration))
1744 representation_ms_info['fragments'] = [{
1745 'url': media_template % {
1746 'Number': segment_number,
1747 'Bandwidth': bandwidth,
1749 'duration': segment_duration,
1750 } for segment_number in range(
1751 representation_ms_info['start_number'],
1752 representation_ms_info['total_number'] + representation_ms_info['start_number'])]
1754 # $Number*$ or $Time$ in media template with S list available
1755 # Example $Number*$: http://www.svtplay.se/klipp/9023742/stopptid-om-bjorn-borg
1756 # Example $Time$: https://play.arkena.com/embed/avp/v2/player/media/b41dda37-d8e7-4d3f-b1b5-9a9db578bdfe/1/129411
1757 representation_ms_info['fragments'] = []
1760 segment_number = representation_ms_info['start_number']
1762 def add_segment_url():
1763 segment_url = media_template % {
1764 'Time': segment_time,
1765 'Bandwidth': bandwidth,
1766 'Number': segment_number,
1768 representation_ms_info['fragments'].append({
1770 'duration': float_or_none(segment_d, representation_ms_info['timescale']),
1773 for num, s in enumerate(representation_ms_info['s']):
1774 segment_time = s.get('t') or segment_time
1778 for r in range(s.get('r', 0)):
1779 segment_time += segment_d
1782 segment_time += segment_d
1783 elif 'segment_urls' in representation_ms_info and 's' in representation_ms_info:
1785 # Example: https://www.youtube.com/watch?v=iXZV5uAYMJI
1786 # or any YouTube dashsegments video
1789 timescale = representation_ms_info['timescale']
1790 for s in representation_ms_info['s']:
1791 duration = float_or_none(s['d'], timescale)
1792 for r in range(s.get('r', 0) + 1):
1794 'url': representation_ms_info['segment_urls'][segment_index],
1795 'duration': duration,
1798 representation_ms_info['fragments'] = fragments
1799 # NB: MPD manifest may contain direct URLs to unfragmented media.
1800 # No fragments key is present in this case.
1801 if 'fragments' in representation_ms_info:
1804 'protocol': 'http_dash_segments',
1806 if 'initialization_url' in representation_ms_info:
1807 initialization_url = representation_ms_info['initialization_url']
1808 if not f.get('url'):
1809 f['url'] = initialization_url
1810 f['fragments'].append({'url': initialization_url})
1811 f['fragments'].extend(representation_ms_info['fragments'])
1812 for fragment in f['fragments']:
1813 fragment['url'] = urljoin(base_url, fragment['url'])
1815 existing_format = next(
1816 fo for fo in formats
1817 if fo['format_id'] == representation_id)
1818 except StopIteration:
1819 full_info = formats_dict.get(representation_id, {}).copy()
1821 formats.append(full_info)
1823 existing_format.update(f)
1825 self.report_warning('Unknown MIME type %s in DASH manifest' % mime_type)
1828 def _extract_ism_formats(self, ism_url, video_id, ism_id=None, note=None, errnote=None, fatal=True):
1829 res = self._download_webpage_handle(
1831 note=note or 'Downloading ISM manifest',
1832 errnote=errnote or 'Failed to download ISM manifest',
1838 return self._parse_ism_formats(
1839 compat_etree_fromstring(ism.encode('utf-8')), urlh.geturl(), ism_id)
1841 def _parse_ism_formats(self, ism_doc, ism_url, ism_id=None):
1842 if ism_doc.get('IsLive') == 'TRUE' or ism_doc.find('Protection') is not None:
1845 duration = int(ism_doc.attrib['Duration'])
1846 timescale = int_or_none(ism_doc.get('TimeScale')) or 10000000
1849 for stream in ism_doc.findall('StreamIndex'):
1850 stream_type = stream.get('Type')
1851 if stream_type not in ('video', 'audio'):
1853 url_pattern = stream.attrib['Url']
1854 stream_timescale = int_or_none(stream.get('TimeScale')) or timescale
1855 stream_name = stream.get('Name')
1856 for track in stream.findall('QualityLevel'):
1857 fourcc = track.get('FourCC')
1858 # TODO: add support for WVC1 and WMAP
1859 if fourcc not in ('H264', 'AVC1', 'AACL'):
1860 self.report_warning('%s is not a supported codec' % fourcc)
1862 tbr = int(track.attrib['Bitrate']) // 1000
1863 width = int_or_none(track.get('MaxWidth'))
1864 height = int_or_none(track.get('MaxHeight'))
1865 sampling_rate = int_or_none(track.get('SamplingRate'))
1867 track_url_pattern = re.sub(r'{[Bb]itrate}', track.attrib['Bitrate'], url_pattern)
1868 track_url_pattern = compat_urlparse.urljoin(ism_url, track_url_pattern)
1874 stream_fragments = stream.findall('c')
1875 for stream_fragment_index, stream_fragment in enumerate(stream_fragments):
1876 fragment_ctx['time'] = int_or_none(stream_fragment.get('t')) or fragment_ctx['time']
1877 fragment_repeat = int_or_none(stream_fragment.get('r')) or 1
1878 fragment_ctx['duration'] = int_or_none(stream_fragment.get('d'))
1879 if not fragment_ctx['duration']:
1881 next_fragment_time = int(stream_fragment[stream_fragment_index + 1].attrib['t'])
1883 next_fragment_time = duration
1884 fragment_ctx['duration'] = (next_fragment_time - fragment_ctx['time']) / fragment_repeat
1885 for _ in range(fragment_repeat):
1887 'url': re.sub(r'{start[ _]time}', compat_str(fragment_ctx['time']), track_url_pattern),
1888 'duration': fragment_ctx['duration'] / stream_timescale,
1890 fragment_ctx['time'] += fragment_ctx['duration']
1894 format_id.append(ism_id)
1896 format_id.append(stream_name)
1897 format_id.append(compat_str(tbr))
1900 'format_id': '-'.join(format_id),
1902 'manifest_url': ism_url,
1903 'ext': 'ismv' if stream_type == 'video' else 'isma',
1907 'asr': sampling_rate,
1908 'vcodec': 'none' if stream_type == 'audio' else fourcc,
1909 'acodec': 'none' if stream_type == 'video' else fourcc,
1911 'fragments': fragments,
1912 '_download_params': {
1913 'duration': duration,
1914 'timescale': stream_timescale,
1915 'width': width or 0,
1916 'height': height or 0,
1918 'codec_private_data': track.get('CodecPrivateData'),
1919 'sampling_rate': sampling_rate,
1920 'channels': int_or_none(track.get('Channels', 2)),
1921 'bits_per_sample': int_or_none(track.get('BitsPerSample', 16)),
1922 'nal_unit_length_field': int_or_none(track.get('NALUnitLengthField', 4)),
1927 def _parse_html5_media_entries(self, base_url, webpage, video_id, m3u8_id=None, m3u8_entry_protocol='m3u8', mpd_id=None):
1928 def absolute_url(video_url):
1929 return compat_urlparse.urljoin(base_url, video_url)
1931 def parse_content_type(content_type):
1932 if not content_type:
1934 ctr = re.search(r'(?P<mimetype>[^/]+/[^;]+)(?:;\s*codecs="?(?P<codecs>[^"]+))?', content_type)
1936 mimetype, codecs = ctr.groups()
1937 f = parse_codecs(codecs)
1938 f['ext'] = mimetype2ext(mimetype)
1942 def _media_formats(src, cur_media_type):
1943 full_url = absolute_url(src)
1944 ext = determine_ext(full_url)
1946 is_plain_url = False
1947 formats = self._extract_m3u8_formats(
1948 full_url, video_id, ext='mp4',
1949 entry_protocol=m3u8_entry_protocol, m3u8_id=m3u8_id)
1951 is_plain_url = False
1952 formats = self._extract_mpd_formats(
1953 full_url, video_id, mpd_id=mpd_id)
1958 'vcodec': 'none' if cur_media_type == 'audio' else None,
1960 return is_plain_url, formats
1963 media_tags = [(media_tag, media_type, '')
1964 for media_tag, media_type
1965 in re.findall(r'(?s)(<(video|audio)[^>]*/>)', webpage)]
1966 media_tags.extend(re.findall(
1967 # We only allow video|audio followed by a whitespace or '>'.
1968 # Allowing more characters may end up in significant slow down (see
1969 # https://github.com/rg3/youtube-dl/issues/11979, example URL:
1970 # http://www.porntrex.com/maps/videositemap.xml).
1971 r'(?s)(<(?P<tag>video|audio)(?:\s+[^>]*)?>)(.*?)</(?P=tag)>', webpage))
1972 for media_tag, media_type, media_content in media_tags:
1977 media_attributes = extract_attributes(media_tag)
1978 src = media_attributes.get('src')
1980 _, formats = _media_formats(src, media_type)
1981 media_info['formats'].extend(formats)
1982 media_info['thumbnail'] = media_attributes.get('poster')
1984 for source_tag in re.findall(r'<source[^>]+>', media_content):
1985 source_attributes = extract_attributes(source_tag)
1986 src = source_attributes.get('src')
1989 is_plain_url, formats = _media_formats(src, media_type)
1991 f = parse_content_type(source_attributes.get('type'))
1992 f.update(formats[0])
1993 media_info['formats'].append(f)
1995 media_info['formats'].extend(formats)
1996 for track_tag in re.findall(r'<track[^>]+>', media_content):
1997 track_attributes = extract_attributes(track_tag)
1998 kind = track_attributes.get('kind')
1999 if not kind or kind in ('subtitles', 'captions'):
2000 src = track_attributes.get('src')
2003 lang = track_attributes.get('srclang') or track_attributes.get('lang') or track_attributes.get('label')
2004 media_info['subtitles'].setdefault(lang, []).append({
2005 'url': absolute_url(src),
2007 if media_info['formats'] or media_info['subtitles']:
2008 entries.append(media_info)
2011 def _extract_akamai_formats(self, manifest_url, video_id, hosts={}):
2013 hdcore_sign = 'hdcore=3.7.0'
2014 f4m_url = re.sub(r'(https?://[^/+])/i/', r'\1/z/', manifest_url).replace('/master.m3u8', '/manifest.f4m')
2015 hds_host = hosts.get('hds')
2017 f4m_url = re.sub(r'(https?://)[^/]+', r'\1' + hds_host, f4m_url)
2018 if 'hdcore=' not in f4m_url:
2019 f4m_url += ('&' if '?' in f4m_url else '?') + hdcore_sign
2020 f4m_formats = self._extract_f4m_formats(
2021 f4m_url, video_id, f4m_id='hds', fatal=False)
2022 for entry in f4m_formats:
2023 entry.update({'extra_param_to_segment_url': hdcore_sign})
2024 formats.extend(f4m_formats)
2025 m3u8_url = re.sub(r'(https?://[^/]+)/z/', r'\1/i/', manifest_url).replace('/manifest.f4m', '/master.m3u8')
2026 hls_host = hosts.get('hls')
2028 m3u8_url = re.sub(r'(https?://)[^/]+', r'\1' + hls_host, m3u8_url)
2029 formats.extend(self._extract_m3u8_formats(
2030 m3u8_url, video_id, 'mp4', 'm3u8_native',
2031 m3u8_id='hls', fatal=False))
2034 def _extract_wowza_formats(self, url, video_id, m3u8_entry_protocol='m3u8_native', skip_protocols=[]):
2035 url = re.sub(r'/(?:manifest|playlist|jwplayer)\.(?:m3u8|f4m|mpd|smil)', '', url)
2036 url_base = self._search_regex(r'(?:https?|rtmp|rtsp)(://[^?]+)', url, 'format url')
2037 http_base_url = 'http' + url_base
2039 if 'm3u8' not in skip_protocols:
2040 formats.extend(self._extract_m3u8_formats(
2041 http_base_url + '/playlist.m3u8', video_id, 'mp4',
2042 m3u8_entry_protocol, m3u8_id='hls', fatal=False))
2043 if 'f4m' not in skip_protocols:
2044 formats.extend(self._extract_f4m_formats(
2045 http_base_url + '/manifest.f4m',
2046 video_id, f4m_id='hds', fatal=False))
2047 if 'dash' not in skip_protocols:
2048 formats.extend(self._extract_mpd_formats(
2049 http_base_url + '/manifest.mpd',
2050 video_id, mpd_id='dash', fatal=False))
2051 if re.search(r'(?:/smil:|\.smil)', url_base):
2052 if 'smil' not in skip_protocols:
2053 rtmp_formats = self._extract_smil_formats(
2054 http_base_url + '/jwplayer.smil',
2055 video_id, fatal=False)
2056 for rtmp_format in rtmp_formats:
2057 rtsp_format = rtmp_format.copy()
2058 rtsp_format['url'] = '%s/%s' % (rtmp_format['url'], rtmp_format['play_path'])
2059 del rtsp_format['play_path']
2060 del rtsp_format['ext']
2061 rtsp_format.update({
2062 'url': rtsp_format['url'].replace('rtmp://', 'rtsp://'),
2063 'format_id': rtmp_format['format_id'].replace('rtmp', 'rtsp'),
2066 formats.extend([rtmp_format, rtsp_format])
2068 for protocol in ('rtmp', 'rtsp'):
2069 if protocol not in skip_protocols:
2071 'url': protocol + url_base,
2072 'format_id': protocol,
2073 'protocol': protocol,
2078 def _find_jwplayer_data(webpage):
2080 r'jwplayer\((?P<quote>[\'"])[^\'" ]+(?P=quote)\)\.setup\s*\((?P<options>[^)]+)\)',
2083 return mobj.group('options')
2085 def _extract_jwplayer_data(self, webpage, video_id, *args, **kwargs):
2086 jwplayer_data = self._parse_json(
2087 self._find_jwplayer_data(webpage), video_id,
2088 transform_source=js_to_json)
2089 return self._parse_jwplayer_data(
2090 jwplayer_data, video_id, *args, **kwargs)
2092 def _parse_jwplayer_data(self, jwplayer_data, video_id=None, require_title=True,
2093 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
2094 # JWPlayer backward compatibility: flattened playlists
2095 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/api/config.js#L81-L96
2096 if 'playlist' not in jwplayer_data:
2097 jwplayer_data = {'playlist': [jwplayer_data]}
2101 # JWPlayer backward compatibility: single playlist item
2102 # https://github.com/jwplayer/jwplayer/blob/v7.7.0/src/js/playlist/playlist.js#L10
2103 if not isinstance(jwplayer_data['playlist'], list):
2104 jwplayer_data['playlist'] = [jwplayer_data['playlist']]
2106 for video_data in jwplayer_data['playlist']:
2107 # JWPlayer backward compatibility: flattened sources
2108 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/playlist/item.js#L29-L35
2109 if 'sources' not in video_data:
2110 video_data['sources'] = [video_data]
2112 this_video_id = video_id or video_data['mediaid']
2115 for source in video_data['sources']:
2116 source_url = self._proto_relative_url(source['file'])
2118 source_url = compat_urlparse.urljoin(base_url, source_url)
2119 source_type = source.get('type') or ''
2120 ext = mimetype2ext(source_type) or determine_ext(source_url)
2121 if source_type == 'hls' or ext == 'm3u8':
2122 formats.extend(self._extract_m3u8_formats(
2123 source_url, this_video_id, 'mp4', 'm3u8_native', m3u8_id=m3u8_id, fatal=False))
2125 formats.extend(self._extract_mpd_formats(
2126 source_url, this_video_id, mpd_id=mpd_id, fatal=False))
2127 # https://github.com/jwplayer/jwplayer/blob/master/src/js/providers/default.js#L67
2128 elif source_type.startswith('audio') or ext in ('oga', 'aac', 'mp3', 'mpeg', 'vorbis'):
2135 height = int_or_none(source.get('height'))
2137 # Often no height is provided but there is a label in
2138 # format like 1080p.
2139 height = int_or_none(self._search_regex(
2140 r'^(\d{3,})[pP]$', source.get('label') or '',
2141 'height', default=None))
2144 'width': int_or_none(source.get('width')),
2148 if source_url.startswith('rtmp'):
2149 a_format['ext'] = 'flv'
2151 # See com/longtailvideo/jwplayer/media/RTMPMediaProvider.as
2152 # of jwplayer.flash.swf
2153 rtmp_url_parts = re.split(
2154 r'((?:mp4|mp3|flv):)', source_url, 1)
2155 if len(rtmp_url_parts) == 3:
2156 rtmp_url, prefix, play_path = rtmp_url_parts
2159 'play_path': prefix + play_path,
2162 a_format.update(rtmp_params)
2163 formats.append(a_format)
2164 self._sort_formats(formats)
2167 tracks = video_data.get('tracks')
2168 if tracks and isinstance(tracks, list):
2169 for track in tracks:
2170 if track.get('kind') != 'captions':
2172 track_url = urljoin(base_url, track.get('file'))
2175 subtitles.setdefault(track.get('label') or 'en', []).append({
2176 'url': self._proto_relative_url(track_url)
2180 'id': this_video_id,
2181 'title': video_data['title'] if require_title else video_data.get('title'),
2182 'description': video_data.get('description'),
2183 'thumbnail': self._proto_relative_url(video_data.get('image')),
2184 'timestamp': int_or_none(video_data.get('pubdate')),
2185 'duration': float_or_none(jwplayer_data.get('duration') or video_data.get('duration')),
2186 'subtitles': subtitles,
2189 if len(entries) == 1:
2192 return self.playlist_result(entries)
2194 def _live_title(self, name):
2195 """ Generate the title for a live video """
2196 now = datetime.datetime.now()
2197 now_str = now.strftime('%Y-%m-%d %H:%M')
2198 return name + ' ' + now_str
2200 def _int(self, v, name, fatal=False, **kwargs):
2201 res = int_or_none(v, **kwargs)
2202 if 'get_attr' in kwargs:
2203 print(getattr(v, kwargs['get_attr']))
2205 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
2207 raise ExtractorError(msg)
2209 self._downloader.report_warning(msg)
2212 def _float(self, v, name, fatal=False, **kwargs):
2213 res = float_or_none(v, **kwargs)
2215 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
2217 raise ExtractorError(msg)
2219 self._downloader.report_warning(msg)
2222 def _set_cookie(self, domain, name, value, expire_time=None):
2223 cookie = compat_cookiejar.Cookie(
2224 0, name, value, None, None, domain, None,
2225 None, '/', True, False, expire_time, '', None, None, None)
2226 self._downloader.cookiejar.set_cookie(cookie)
2228 def _get_cookies(self, url):
2229 """ Return a compat_cookies.SimpleCookie with the cookies for the url """
2230 req = sanitized_Request(url)
2231 self._downloader.cookiejar.add_cookie_header(req)
2232 return compat_cookies.SimpleCookie(req.get_header('Cookie'))
2234 def get_testcases(self, include_onlymatching=False):
2235 t = getattr(self, '_TEST', None)
2237 assert not hasattr(self, '_TESTS'), \
2238 '%s has _TEST and _TESTS' % type(self).__name__
2241 tests = getattr(self, '_TESTS', [])
2243 if not include_onlymatching and t.get('only_matching', False):
2245 t['name'] = type(self).__name__[:-len('IE')]
2248 def is_suitable(self, age_limit):
2249 """ Test whether the extractor is generally suitable for the given
2250 age limit (i.e. pornographic sites are not, all others usually are) """
2252 any_restricted = False
2253 for tc in self.get_testcases(include_onlymatching=False):
2254 if tc.get('playlist', []):
2255 tc = tc['playlist'][0]
2256 is_restricted = age_restricted(
2257 tc.get('info_dict', {}).get('age_limit'), age_limit)
2258 if not is_restricted:
2260 any_restricted = any_restricted or is_restricted
2261 return not any_restricted
2263 def extract_subtitles(self, *args, **kwargs):
2264 if (self._downloader.params.get('writesubtitles', False) or
2265 self._downloader.params.get('listsubtitles')):
2266 return self._get_subtitles(*args, **kwargs)
2269 def _get_subtitles(self, *args, **kwargs):
2270 raise NotImplementedError('This method must be implemented by subclasses')
2273 def _merge_subtitle_items(subtitle_list1, subtitle_list2):
2274 """ Merge subtitle items for one language. Items with duplicated URLs
2275 will be dropped. """
2276 list1_urls = set([item['url'] for item in subtitle_list1])
2277 ret = list(subtitle_list1)
2278 ret.extend([item for item in subtitle_list2 if item['url'] not in list1_urls])
2282 def _merge_subtitles(cls, subtitle_dict1, subtitle_dict2):
2283 """ Merge two subtitle dictionaries, language by language. """
2284 ret = dict(subtitle_dict1)
2285 for lang in subtitle_dict2:
2286 ret[lang] = cls._merge_subtitle_items(subtitle_dict1.get(lang, []), subtitle_dict2[lang])
2289 def extract_automatic_captions(self, *args, **kwargs):
2290 if (self._downloader.params.get('writeautomaticsub', False) or
2291 self._downloader.params.get('listsubtitles')):
2292 return self._get_automatic_captions(*args, **kwargs)
2295 def _get_automatic_captions(self, *args, **kwargs):
2296 raise NotImplementedError('This method must be implemented by subclasses')
2298 def mark_watched(self, *args, **kwargs):
2299 if (self._downloader.params.get('mark_watched', False) and
2300 (self._get_login_info()[0] is not None or
2301 self._downloader.params.get('cookiefile') is not None)):
2302 self._mark_watched(*args, **kwargs)
2304 def _mark_watched(self, *args, **kwargs):
2305 raise NotImplementedError('This method must be implemented by subclasses')
2307 def geo_verification_headers(self):
2309 geo_verification_proxy = self._downloader.params.get('geo_verification_proxy')
2310 if geo_verification_proxy:
2311 headers['Ytdl-request-proxy'] = geo_verification_proxy
2314 def _generic_id(self, url):
2315 return compat_urllib_parse_unquote(os.path.splitext(url.rstrip('/').split('/')[-1])[0])
2317 def _generic_title(self, url):
2318 return compat_urllib_parse_unquote(os.path.splitext(url_basename(url))[0])
2321 class SearchInfoExtractor(InfoExtractor):
2323 Base class for paged search queries extractors.
2324 They accept URLs in the format _SEARCH_KEY(|all|[0-9]):{query}
2325 Instances should define _SEARCH_KEY and _MAX_RESULTS.
2329 def _make_valid_url(cls):
2330 return r'%s(?P<prefix>|[1-9][0-9]*|all):(?P<query>[\s\S]+)' % cls._SEARCH_KEY
2333 def suitable(cls, url):
2334 return re.match(cls._make_valid_url(), url) is not None
2336 def _real_extract(self, query):
2337 mobj = re.match(self._make_valid_url(), query)
2339 raise ExtractorError('Invalid search query "%s"' % query)
2341 prefix = mobj.group('prefix')
2342 query = mobj.group('query')
2344 return self._get_n_results(query, 1)
2345 elif prefix == 'all':
2346 return self._get_n_results(query, self._MAX_RESULTS)
2350 raise ExtractorError('invalid download number %s for query "%s"' % (n, query))
2351 elif n > self._MAX_RESULTS:
2352 self._downloader.report_warning('%s returns max %i results (you requested %i)' % (self._SEARCH_KEY, self._MAX_RESULTS, n))
2353 n = self._MAX_RESULTS
2354 return self._get_n_results(query, n)
2356 def _get_n_results(self, query, n):
2357 """Get a specified number of results for a query"""
2358 raise NotImplementedError('This method must be implemented by subclasses')
2361 def SEARCH_KEY(self):
2362 return self._SEARCH_KEY