2 from __future__ import unicode_literals
17 from ..compat import (
20 compat_etree_fromstring,
26 compat_urllib_parse_unquote,
27 compat_urllib_parse_urlencode,
28 compat_urllib_request,
31 from ..downloader.f4m import remove_encrypted_media
55 parse_m3u8_attributes,
72 class InfoExtractor(object):
73 """Information Extractor class.
75 Information extractors are the classes that, given a URL, extract
76 information about the video (or videos) the URL refers to. This
77 information includes the real video URL, the video title, author and
78 others. The information is stored in a dictionary which is then
79 passed to the YoutubeDL. The YoutubeDL processes this
80 information possibly downloading the video to the file system, among
81 other possible outcomes.
83 The type field determines the type of the result.
84 By far the most common value (and the default if _type is missing) is
85 "video", which indicates a single video.
87 For a video, the dictionaries must include the following fields:
90 title: Video title, unescaped.
92 Additionally, it must contain either a formats entry or a url one:
94 formats: A list of dictionaries for each format available, ordered
95 from worst to best quality.
98 * url Mandatory. The URL of the video file
100 The URL of the manifest file in case of
101 fragmented media (DASH, hls, hds)
102 * ext Will be calculated from URL if missing
103 * format A human-readable description of the format
104 ("mp4 container with h264/opus").
105 Calculated from the format_id, width, height.
106 and format_note fields if missing.
107 * format_id A short description of the format
108 ("mp4_h264_opus" or "19").
109 Technically optional, but strongly recommended.
110 * format_note Additional info about the format
111 ("3D" or "DASH video")
112 * width Width of the video, if known
113 * height Height of the video, if known
114 * resolution Textual description of width and height
115 * tbr Average bitrate of audio and video in KBit/s
116 * abr Average audio bitrate in KBit/s
117 * acodec Name of the audio codec in use
118 * asr Audio sampling rate in Hertz
119 * vbr Average video bitrate in KBit/s
121 * vcodec Name of the video codec in use
122 * container Name of the container format
123 * filesize The number of bytes, if known in advance
124 * filesize_approx An estimate for the number of bytes
125 * player_url SWF Player URL (used for rtmpdump).
126 * protocol The protocol that will be used for the actual
127 download, lower-case.
128 "http", "https", "rtsp", "rtmp", "rtmpe",
129 "m3u8", "m3u8_native" or "http_dash_segments".
131 Base URL for fragments. Each fragment's path
132 value (if present) will be relative to
134 * fragments A list of fragments of a fragmented media.
135 Each fragment entry must contain either an url
136 or a path. If an url is present it should be
137 considered by a client. Otherwise both path and
138 fragment_base_url must be present. Here is
139 the list of all potential fields:
140 * "url" - fragment's URL
141 * "path" - fragment's path relative to
143 * "duration" (optional, int or float)
144 * "filesize" (optional, int)
145 * preference Order number of this format. If this field is
146 present and not None, the formats get sorted
147 by this field, regardless of all other values.
148 -1 for default (order by other properties),
149 -2 or smaller for less than default.
150 < -1000 to hide the format (if there is
151 another one which is strictly better)
152 * language Language code, e.g. "de" or "en-US".
153 * language_preference Is this in the language mentioned in
155 10 if it's what the URL is about,
156 -1 for default (don't know),
157 -10 otherwise, other values reserved for now.
158 * quality Order number of the video quality of this
159 format, irrespective of the file format.
160 -1 for default (order by other properties),
161 -2 or smaller for less than default.
162 * source_preference Order number for this video source
163 (quality takes higher priority)
164 -1 for default (order by other properties),
165 -2 or smaller for less than default.
166 * http_headers A dictionary of additional HTTP headers
167 to add to the request.
168 * stretched_ratio If given and not 1, indicates that the
169 video's pixels are not square.
170 width : height ratio as float.
171 * no_resume The server does not support resuming the
172 (HTTP or RTMP) download. Boolean.
174 url: Final video URL.
175 ext: Video filename extension.
176 format: The video format, defaults to ext (used for --get-format)
177 player_url: SWF Player URL (used for rtmpdump).
179 The following fields are optional:
181 alt_title: A secondary title of the video.
182 display_id An alternative identifier for the video, not necessarily
183 unique, but available before title. Typically, id is
184 something like "4234987", title "Dancing naked mole rats",
185 and display_id "dancing-naked-mole-rats"
186 thumbnails: A list of dictionaries, with the following entries:
187 * "id" (optional, string) - Thumbnail format ID
189 * "preference" (optional, int) - quality of the image
190 * "width" (optional, int)
191 * "height" (optional, int)
192 * "resolution" (optional, string "{width}x{height"},
194 * "filesize" (optional, int)
195 thumbnail: Full URL to a video thumbnail image.
196 description: Full video description.
197 uploader: Full name of the video uploader.
198 license: License name the video is licensed under.
199 creator: The creator of the video.
200 release_date: The date (YYYYMMDD) when the video was released.
201 timestamp: UNIX timestamp of the moment the video became available.
202 upload_date: Video upload date (YYYYMMDD).
203 If not explicitly set, calculated from timestamp.
204 uploader_id: Nickname or id of the video uploader.
205 uploader_url: Full URL to a personal webpage of the video uploader.
206 location: Physical location where the video was filmed.
207 subtitles: The available subtitles as a dictionary in the format
208 {tag: subformats}. "tag" is usually a language code, and
209 "subformats" is a list sorted from lower to higher
210 preference, each element is a dictionary with the "ext"
212 * "data": The subtitles file contents
213 * "url": A URL pointing to the subtitles file
214 "ext" will be calculated from URL if missing
215 automatic_captions: Like 'subtitles', used by the YoutubeIE for
216 automatically generated captions
217 duration: Length of the video in seconds, as an integer or float.
218 view_count: How many users have watched the video on the platform.
219 like_count: Number of positive ratings of the video
220 dislike_count: Number of negative ratings of the video
221 repost_count: Number of reposts of the video
222 average_rating: Average rating give by users, the scale used depends on the webpage
223 comment_count: Number of comments on the video
224 comments: A list of comments, each with one or more of the following
225 properties (all but one of text or html optional):
226 * "author" - human-readable name of the comment author
227 * "author_id" - user ID of the comment author
229 * "html" - Comment as HTML
230 * "text" - Plain text of the comment
231 * "timestamp" - UNIX timestamp of comment
232 * "parent" - ID of the comment this one is replying to.
233 Set to "root" to indicate that this is a
234 comment to the original video.
235 age_limit: Age restriction for the video, as an integer (years)
236 webpage_url: The URL to the video webpage, if given to youtube-dl it
237 should allow to get the same result again. (It will be set
238 by YoutubeDL if it's missing)
239 categories: A list of categories that the video falls in, for example
241 tags: A list of tags assigned to the video, e.g. ["sweden", "pop music"]
242 is_live: True, False, or None (=unknown). Whether this video is a
243 live stream that goes on instead of a fixed-length video.
244 start_time: Time in seconds where the reproduction should start, as
245 specified in the URL.
246 end_time: Time in seconds where the reproduction should end, as
247 specified in the URL.
249 The following fields should only be used when the video belongs to some logical
252 chapter: Name or title of the chapter the video belongs to.
253 chapter_number: Number of the chapter the video belongs to, as an integer.
254 chapter_id: Id of the chapter the video belongs to, as a unicode string.
256 The following fields should only be used when the video is an episode of some
257 series, programme or podcast:
259 series: Title of the series or programme the video episode belongs to.
260 season: Title of the season the video episode belongs to.
261 season_number: Number of the season the video episode belongs to, as an integer.
262 season_id: Id of the season the video episode belongs to, as a unicode string.
263 episode: Title of the video episode. Unlike mandatory video title field,
264 this field should denote the exact title of the video episode
265 without any kind of decoration.
266 episode_number: Number of the video episode within a season, as an integer.
267 episode_id: Id of the video episode, as a unicode string.
269 The following fields should only be used when the media is a track or a part of
272 track: Title of the track.
273 track_number: Number of the track within an album or a disc, as an integer.
274 track_id: Id of the track (useful in case of custom indexing, e.g. 6.iii),
276 artist: Artist(s) of the track.
277 genre: Genre(s) of the track.
278 album: Title of the album the track belongs to.
279 album_type: Type of the album (e.g. "Demo", "Full-length", "Split", "Compilation", etc).
280 album_artist: List of all artists appeared on the album (e.g.
281 "Ash Borer / Fell Voices" or "Various Artists", useful for splits
283 disc_number: Number of the disc or other physical medium the track belongs to,
285 release_year: Year (YYYY) when the album was released.
287 Unless mentioned otherwise, the fields should be Unicode strings.
289 Unless mentioned otherwise, None is equivalent to absence of information.
292 _type "playlist" indicates multiple videos.
293 There must be a key "entries", which is a list, an iterable, or a PagedList
294 object, each element of which is a valid dictionary by this specification.
296 Additionally, playlists can have "title", "description" and "id" attributes
297 with the same semantics as videos (see above).
300 _type "multi_video" indicates that there are multiple videos that
301 form a single show, for examples multiple acts of an opera or TV episode.
302 It must have an entries key like a playlist and contain all the keys
303 required for a video at the same time.
306 _type "url" indicates that the video must be extracted from another
307 location, possibly by a different extractor. Its only required key is:
308 "url" - the next URL to extract.
309 The key "ie_key" can be set to the class name (minus the trailing "IE",
310 e.g. "Youtube") if the extractor class is known in advance.
311 Additionally, the dictionary may have any properties of the resolved entity
312 known in advance, for example "title" if the title of the referred video is
316 _type "url_transparent" entities have the same specification as "url", but
317 indicate that the given additional information is more precise than the one
318 associated with the resolved URL.
319 This is useful when a site employs a video service that hosts the video and
320 its technical metadata, but that video service does not embed a useful
321 title, description etc.
324 Subclasses of this one should re-define the _real_initialize() and
325 _real_extract() methods and define a _VALID_URL regexp.
326 Probably, they should also be added to the list of extractors.
328 _GEO_BYPASS attribute may be set to False in order to disable
329 geo restriction bypass mechanisms for a particular extractor.
330 Though it won't disable explicit geo restriction bypass based on
331 country code provided with geo_bypass_country. (experimental)
333 _GEO_COUNTRIES attribute may contain a list of presumably geo unrestricted
334 countries for this extractor. One of these countries will be used by
335 geo restriction bypass mechanism right away in order to bypass
336 geo restriction, of course, if the mechanism is not disabled. (experimental)
338 NB: both these geo attributes are experimental and may change in future
339 or be completely removed.
341 Finally, the _WORKING attribute should be set to False for broken IEs
342 in order to warn the users and skip the tests.
347 _x_forwarded_for_ip = None
349 _GEO_COUNTRIES = None
352 def __init__(self, downloader=None):
353 """Constructor. Receives an optional downloader."""
355 self._x_forwarded_for_ip = None
356 self.set_downloader(downloader)
359 def suitable(cls, url):
360 """Receives a URL and returns True if suitable for this IE."""
362 # This does not use has/getattr intentionally - we want to know whether
363 # we have cached the regexp for *this* class, whereas getattr would also
364 # match the superclass
365 if '_VALID_URL_RE' not in cls.__dict__:
366 cls._VALID_URL_RE = re.compile(cls._VALID_URL)
367 return cls._VALID_URL_RE.match(url) is not None
370 def _match_id(cls, url):
371 if '_VALID_URL_RE' not in cls.__dict__:
372 cls._VALID_URL_RE = re.compile(cls._VALID_URL)
373 m = cls._VALID_URL_RE.match(url)
379 """Getter method for _WORKING."""
382 def initialize(self):
383 """Initializes an instance (authentication, etc)."""
384 self._initialize_geo_bypass(self._GEO_COUNTRIES)
386 self._real_initialize()
389 def _initialize_geo_bypass(self, countries):
391 Initialize geo restriction bypass mechanism.
393 This method is used to initialize geo bypass mechanism based on faking
394 X-Forwarded-For HTTP header. A random country from provided country list
395 is selected and a random IP belonging to this country is generated. This
396 IP will be passed as X-Forwarded-For HTTP header in all subsequent
399 This method will be used for initial geo bypass mechanism initialization
400 during the instance initialization with _GEO_COUNTRIES.
402 You may also manually call it from extractor's code if geo countries
403 information is not available beforehand (e.g. obtained during
404 extraction) or due to some another reason.
406 if not self._x_forwarded_for_ip:
407 country_code = self._downloader.params.get('geo_bypass_country', None)
408 # If there is no explicit country for geo bypass specified and
409 # the extractor is known to be geo restricted let's fake IP
410 # as X-Forwarded-For right away.
411 if (not country_code and
413 self._downloader.params.get('geo_bypass', True) and
415 country_code = random.choice(countries)
417 self._x_forwarded_for_ip = GeoUtils.random_ipv4(country_code)
418 if self._downloader.params.get('verbose', False):
419 self._downloader.to_stdout(
420 '[debug] Using fake IP %s (%s) as X-Forwarded-For.'
421 % (self._x_forwarded_for_ip, country_code.upper()))
423 def extract(self, url):
424 """Extracts URL information and returns it in list of dicts."""
429 ie_result = self._real_extract(url)
430 if self._x_forwarded_for_ip:
431 ie_result['__x_forwarded_for_ip'] = self._x_forwarded_for_ip
433 except GeoRestrictedError as e:
434 if self.__maybe_fake_ip_and_retry(e.countries):
437 except ExtractorError:
439 except compat_http_client.IncompleteRead as e:
440 raise ExtractorError('A network error has occurred.', cause=e, expected=True)
441 except (KeyError, StopIteration) as e:
442 raise ExtractorError('An extractor error has occurred.', cause=e)
444 def __maybe_fake_ip_and_retry(self, countries):
445 if (not self._downloader.params.get('geo_bypass_country', None) and
447 self._downloader.params.get('geo_bypass', True) and
448 not self._x_forwarded_for_ip and
450 country_code = random.choice(countries)
451 self._x_forwarded_for_ip = GeoUtils.random_ipv4(country_code)
452 if self._x_forwarded_for_ip:
454 'Video is geo restricted. Retrying extraction with fake IP %s (%s) as X-Forwarded-For.'
455 % (self._x_forwarded_for_ip, country_code.upper()))
459 def set_downloader(self, downloader):
460 """Sets the downloader for this IE."""
461 self._downloader = downloader
463 def _real_initialize(self):
464 """Real initialization process. Redefine in subclasses."""
467 def _real_extract(self, url):
468 """Real extraction process. Redefine in subclasses."""
473 """A string for getting the InfoExtractor with get_info_extractor"""
474 return compat_str(cls.__name__[:-2])
478 return compat_str(type(self).__name__[:-2])
480 def _request_webpage(self, url_or_request, video_id, note=None, errnote=None, fatal=True, data=None, headers={}, query={}):
481 """ Returns the response handle """
483 self.report_download_webpage(video_id)
484 elif note is not False:
486 self.to_screen('%s' % (note,))
488 self.to_screen('%s: %s' % (video_id, note))
489 if isinstance(url_or_request, compat_urllib_request.Request):
490 url_or_request = update_Request(
491 url_or_request, data=data, headers=headers, query=query)
494 url_or_request = update_url_query(url_or_request, query)
495 if data is not None or headers:
496 url_or_request = sanitized_Request(url_or_request, data, headers)
498 return self._downloader.urlopen(url_or_request)
499 except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err:
503 errnote = 'Unable to download webpage'
505 errmsg = '%s: %s' % (errnote, error_to_compat_str(err))
507 raise ExtractorError(errmsg, sys.exc_info()[2], cause=err)
509 self._downloader.report_warning(errmsg)
512 def _download_webpage_handle(self, url_or_request, video_id, note=None, errnote=None, fatal=True, encoding=None, data=None, headers={}, query={}):
513 """ Returns a tuple (page content as string, URL handle) """
514 # Strip hashes from the URL (#1038)
515 if isinstance(url_or_request, (compat_str, str)):
516 url_or_request = url_or_request.partition('#')[0]
518 # Some sites check X-Forwarded-For HTTP header in order to figure out
519 # the origin of the client behind proxy. This allows bypassing geo
520 # restriction by faking this header's value to IP that belongs to some
521 # geo unrestricted country. We will do so once we encounter any
522 # geo restriction error.
523 if self._x_forwarded_for_ip:
524 if 'X-Forwarded-For' not in headers:
525 headers['X-Forwarded-For'] = self._x_forwarded_for_ip
527 urlh = self._request_webpage(url_or_request, video_id, note, errnote, fatal, data=data, headers=headers, query=query)
531 content = self._webpage_read_content(urlh, url_or_request, video_id, note, errnote, fatal, encoding=encoding)
532 return (content, urlh)
535 def _guess_encoding_from_content(content_type, webpage_bytes):
536 m = re.match(r'[a-zA-Z0-9_.-]+/[a-zA-Z0-9_.-]+\s*;\s*charset=(.+)', content_type)
538 encoding = m.group(1)
540 m = re.search(br'<meta[^>]+charset=[\'"]?([^\'")]+)[ /\'">]',
541 webpage_bytes[:1024])
543 encoding = m.group(1).decode('ascii')
544 elif webpage_bytes.startswith(b'\xff\xfe'):
551 def __check_blocked(self, content):
552 first_block = content[:512]
553 if ('<title>Access to this site is blocked</title>' in content and
554 'Websense' in first_block):
555 msg = 'Access to this webpage has been blocked by Websense filtering software in your network.'
556 blocked_iframe = self._html_search_regex(
557 r'<iframe src="([^"]+)"', content,
558 'Websense information URL', default=None)
560 msg += ' Visit %s for more details' % blocked_iframe
561 raise ExtractorError(msg, expected=True)
562 if '<title>The URL you requested has been blocked</title>' in first_block:
564 'Access to this webpage has been blocked by Indian censorship. '
565 'Use a VPN or proxy server (with --proxy) to route around it.')
566 block_msg = self._html_search_regex(
567 r'</h1><p>(.*?)</p>',
568 content, 'block message', default=None)
570 msg += ' (Message: "%s")' % block_msg.replace('\n', ' ')
571 raise ExtractorError(msg, expected=True)
572 if ('<title>TTK :: Доступ к ресурсу ограничен</title>' in content and
573 'blocklist.rkn.gov.ru' in content):
574 raise ExtractorError(
575 'Access to this webpage has been blocked by decision of the Russian government. '
576 'Visit http://blocklist.rkn.gov.ru/ for a block reason.',
579 def _webpage_read_content(self, urlh, url_or_request, video_id, note=None, errnote=None, fatal=True, prefix=None, encoding=None):
580 content_type = urlh.headers.get('Content-Type', '')
581 webpage_bytes = urlh.read()
582 if prefix is not None:
583 webpage_bytes = prefix + webpage_bytes
585 encoding = self._guess_encoding_from_content(content_type, webpage_bytes)
586 if self._downloader.params.get('dump_intermediate_pages', False):
588 url = url_or_request.get_full_url()
589 except AttributeError:
591 self.to_screen('Dumping request to ' + url)
592 dump = base64.b64encode(webpage_bytes).decode('ascii')
593 self._downloader.to_screen(dump)
594 if self._downloader.params.get('write_pages', False):
596 url = url_or_request.get_full_url()
597 except AttributeError:
599 basen = '%s_%s' % (video_id, url)
601 h = '___' + hashlib.md5(basen.encode('utf-8')).hexdigest()
602 basen = basen[:240 - len(h)] + h
603 raw_filename = basen + '.dump'
604 filename = sanitize_filename(raw_filename, restricted=True)
605 self.to_screen('Saving request to ' + filename)
606 # Working around MAX_PATH limitation on Windows (see
607 # http://msdn.microsoft.com/en-us/library/windows/desktop/aa365247(v=vs.85).aspx)
608 if compat_os_name == 'nt':
609 absfilepath = os.path.abspath(filename)
610 if len(absfilepath) > 259:
611 filename = '\\\\?\\' + absfilepath
612 with open(filename, 'wb') as outf:
613 outf.write(webpage_bytes)
616 content = webpage_bytes.decode(encoding, 'replace')
618 content = webpage_bytes.decode('utf-8', 'replace')
620 self.__check_blocked(content)
624 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={}):
625 """ Returns the data of the page as a string """
628 while success is False:
630 res = self._download_webpage_handle(url_or_request, video_id, note, errnote, fatal, encoding=encoding, data=data, headers=headers, query=query)
632 except compat_http_client.IncompleteRead as e:
634 if try_count >= tries:
636 self._sleep(timeout, video_id)
643 def _download_xml(self, url_or_request, video_id,
644 note='Downloading XML', errnote='Unable to download XML',
645 transform_source=None, fatal=True, encoding=None, data=None, headers={}, query={}):
646 """Return the xml as an xml.etree.ElementTree.Element"""
647 xml_string = self._download_webpage(
648 url_or_request, video_id, note, errnote, fatal=fatal, encoding=encoding, data=data, headers=headers, query=query)
649 if xml_string is False:
652 xml_string = transform_source(xml_string)
653 return compat_etree_fromstring(xml_string.encode('utf-8'))
655 def _download_json(self, url_or_request, video_id,
656 note='Downloading JSON metadata',
657 errnote='Unable to download JSON metadata',
658 transform_source=None,
659 fatal=True, encoding=None, data=None, headers={}, query={}):
660 json_string = self._download_webpage(
661 url_or_request, video_id, note, errnote, fatal=fatal,
662 encoding=encoding, data=data, headers=headers, query=query)
663 if (not fatal) and json_string is False:
665 return self._parse_json(
666 json_string, video_id, transform_source=transform_source, fatal=fatal)
668 def _parse_json(self, json_string, video_id, transform_source=None, fatal=True):
670 json_string = transform_source(json_string)
672 return json.loads(json_string)
673 except ValueError as ve:
674 errmsg = '%s: Failed to parse JSON ' % video_id
676 raise ExtractorError(errmsg, cause=ve)
678 self.report_warning(errmsg + str(ve))
680 def report_warning(self, msg, video_id=None):
681 idstr = '' if video_id is None else '%s: ' % video_id
682 self._downloader.report_warning(
683 '[%s] %s%s' % (self.IE_NAME, idstr, msg))
685 def to_screen(self, msg):
686 """Print msg to screen, prefixing it with '[ie_name]'"""
687 self._downloader.to_screen('[%s] %s' % (self.IE_NAME, msg))
689 def report_extraction(self, id_or_name):
690 """Report information extraction."""
691 self.to_screen('%s: Extracting information' % id_or_name)
693 def report_download_webpage(self, video_id):
694 """Report webpage download."""
695 self.to_screen('%s: Downloading webpage' % video_id)
697 def report_age_confirmation(self):
698 """Report attempt to confirm age."""
699 self.to_screen('Confirming age')
701 def report_login(self):
702 """Report attempt to log in."""
703 self.to_screen('Logging in')
706 def raise_login_required(msg='This video is only available for registered users'):
707 raise ExtractorError(
708 '%s. Use --username and --password or --netrc to provide account credentials.' % msg,
712 def raise_geo_restricted(msg='This video is not available from your location due to geo restriction', countries=None):
713 raise GeoRestrictedError(msg, countries=countries)
715 # Methods for following #608
717 def url_result(url, ie=None, video_id=None, video_title=None):
718 """Returns a URL that points to a page that should be processed"""
719 # TODO: ie should be the class used for getting the info
720 video_info = {'_type': 'url',
723 if video_id is not None:
724 video_info['id'] = video_id
725 if video_title is not None:
726 video_info['title'] = video_title
729 def playlist_from_matches(self, matches, video_id, video_title, getter=None, ie=None):
731 self.url_result(self._proto_relative_url(getter(m) if getter else m), ie)
733 return self.playlist_result(
734 urlrs, playlist_id=video_id, playlist_title=video_title)
737 def playlist_result(entries, playlist_id=None, playlist_title=None, playlist_description=None):
738 """Returns a playlist"""
739 video_info = {'_type': 'playlist',
742 video_info['id'] = playlist_id
744 video_info['title'] = playlist_title
745 if playlist_description:
746 video_info['description'] = playlist_description
749 def _search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
751 Perform a regex search on the given string, using a single or a list of
752 patterns returning the first matching group.
753 In case of failure return a default value or raise a WARNING or a
754 RegexNotFoundError, depending on fatal, specifying the field name.
756 if isinstance(pattern, (str, compat_str, compiled_regex_type)):
757 mobj = re.search(pattern, string, flags)
760 mobj = re.search(p, string, flags)
764 if not self._downloader.params.get('no_color') and compat_os_name != 'nt' and sys.stderr.isatty():
765 _name = '\033[0;34m%s\033[0m' % name
771 # return the first matching group
772 return next(g for g in mobj.groups() if g is not None)
774 return mobj.group(group)
775 elif default is not NO_DEFAULT:
778 raise RegexNotFoundError('Unable to extract %s' % _name)
780 self._downloader.report_warning('unable to extract %s' % _name + bug_reports_message())
783 def _html_search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
785 Like _search_regex, but strips HTML tags and unescapes entities.
787 res = self._search_regex(pattern, string, name, default, fatal, flags, group)
789 return clean_html(res).strip()
793 def _get_netrc_login_info(self, netrc_machine=None):
796 netrc_machine = netrc_machine or self._NETRC_MACHINE
798 if self._downloader.params.get('usenetrc', False):
800 info = netrc.netrc().authenticators(netrc_machine)
805 raise netrc.NetrcParseError(
806 'No authenticators for %s' % netrc_machine)
807 except (IOError, netrc.NetrcParseError) as err:
808 self._downloader.report_warning(
809 'parsing .netrc: %s' % error_to_compat_str(err))
811 return username, password
813 def _get_login_info(self, username_option='username', password_option='password', netrc_machine=None):
815 Get the login info as (username, password)
816 First look for the manually specified credentials using username_option
817 and password_option as keys in params dictionary. If no such credentials
818 available look in the netrc file using the netrc_machine or _NETRC_MACHINE
820 If there's no info available, return (None, None)
822 if self._downloader is None:
825 downloader_params = self._downloader.params
827 # Attempt to use provided username and password or .netrc data
828 if downloader_params.get(username_option) is not None:
829 username = downloader_params[username_option]
830 password = downloader_params[password_option]
832 username, password = self._get_netrc_login_info(netrc_machine)
834 return username, password
836 def _get_tfa_info(self, note='two-factor verification code'):
838 Get the two-factor authentication info
839 TODO - asking the user will be required for sms/phone verify
840 currently just uses the command line option
841 If there's no info available, return None
843 if self._downloader is None:
845 downloader_params = self._downloader.params
847 if downloader_params.get('twofactor') is not None:
848 return downloader_params['twofactor']
850 return compat_getpass('Type %s and press [Return]: ' % note)
852 # Helper functions for extracting OpenGraph info
854 def _og_regexes(prop):
855 content_re = r'content=(?:"([^"]+?)"|\'([^\']+?)\'|\s*([^\s"\'=<>`]+?))'
856 property_re = (r'(?:name|property)=(?:\'og:%(prop)s\'|"og:%(prop)s"|\s*og:%(prop)s\b)'
857 % {'prop': re.escape(prop)})
858 template = r'<meta[^>]+?%s[^>]+?%s'
860 template % (property_re, content_re),
861 template % (content_re, property_re),
865 def _meta_regex(prop):
866 return r'''(?isx)<meta
867 (?=[^>]+(?:itemprop|name|property|id|http-equiv)=(["\']?)%s\1)
868 [^>]+?content=(["\'])(?P<content>.*?)\2''' % re.escape(prop)
870 def _og_search_property(self, prop, html, name=None, **kargs):
871 if not isinstance(prop, (list, tuple)):
874 name = 'OpenGraph %s' % prop[0]
877 og_regexes.extend(self._og_regexes(p))
878 escaped = self._search_regex(og_regexes, html, name, flags=re.DOTALL, **kargs)
881 return unescapeHTML(escaped)
883 def _og_search_thumbnail(self, html, **kargs):
884 return self._og_search_property('image', html, 'thumbnail URL', fatal=False, **kargs)
886 def _og_search_description(self, html, **kargs):
887 return self._og_search_property('description', html, fatal=False, **kargs)
889 def _og_search_title(self, html, **kargs):
890 return self._og_search_property('title', html, **kargs)
892 def _og_search_video_url(self, html, name='video url', secure=True, **kargs):
893 regexes = self._og_regexes('video') + self._og_regexes('video:url')
895 regexes = self._og_regexes('video:secure_url') + regexes
896 return self._html_search_regex(regexes, html, name, **kargs)
898 def _og_search_url(self, html, **kargs):
899 return self._og_search_property('url', html, **kargs)
901 def _html_search_meta(self, name, html, display_name=None, fatal=False, **kwargs):
902 if not isinstance(name, (list, tuple)):
904 if display_name is None:
905 display_name = name[0]
906 return self._html_search_regex(
907 [self._meta_regex(n) for n in name],
908 html, display_name, fatal=fatal, group='content', **kwargs)
910 def _dc_search_uploader(self, html):
911 return self._html_search_meta('dc.creator', html, 'uploader')
913 def _rta_search(self, html):
914 # See http://www.rtalabel.org/index.php?content=howtofaq#single
915 if re.search(r'(?ix)<meta\s+name="rating"\s+'
916 r' content="RTA-5042-1996-1400-1577-RTA"',
921 def _media_rating_search(self, html):
922 # See http://www.tjg-designs.com/WP/metadata-code-examples-adding-metadata-to-your-web-pages/
923 rating = self._html_search_meta('rating', html)
935 return RATING_TABLE.get(rating.lower())
937 def _family_friendly_search(self, html):
938 # See http://schema.org/VideoObject
939 family_friendly = self._html_search_meta('isFamilyFriendly', html)
941 if not family_friendly:
950 return RATING_TABLE.get(family_friendly.lower())
952 def _twitter_search_player(self, html):
953 return self._html_search_meta('twitter:player', html,
954 'twitter card player')
956 def _search_json_ld(self, html, video_id, expected_type=None, **kwargs):
957 json_ld = self._search_regex(
958 r'(?s)<script[^>]+type=(["\'])application/ld\+json\1[^>]*>(?P<json_ld>.+?)</script>',
959 html, 'JSON-LD', group='json_ld', **kwargs)
960 default = kwargs.get('default', NO_DEFAULT)
962 return default if default is not NO_DEFAULT else {}
963 # JSON-LD may be malformed and thus `fatal` should be respected.
964 # At the same time `default` may be passed that assumes `fatal=False`
965 # for _search_regex. Let's simulate the same behavior here as well.
966 fatal = kwargs.get('fatal', True) if default == NO_DEFAULT else False
967 return self._json_ld(json_ld, video_id, fatal=fatal, expected_type=expected_type)
969 def _json_ld(self, json_ld, video_id, fatal=True, expected_type=None):
970 if isinstance(json_ld, compat_str):
971 json_ld = self._parse_json(json_ld, video_id, fatal=fatal)
975 if not isinstance(json_ld, (list, tuple, dict)):
977 if isinstance(json_ld, dict):
980 def extract_video_object(e):
981 assert e['@type'] == 'VideoObject'
983 'url': e.get('contentUrl'),
984 'title': unescapeHTML(e.get('name')),
985 'description': unescapeHTML(e.get('description')),
986 'thumbnail': e.get('thumbnailUrl') or e.get('thumbnailURL'),
987 'duration': parse_duration(e.get('duration')),
988 'timestamp': unified_timestamp(e.get('uploadDate')),
989 'filesize': float_or_none(e.get('contentSize')),
990 'tbr': int_or_none(e.get('bitrate')),
991 'width': int_or_none(e.get('width')),
992 'height': int_or_none(e.get('height')),
993 'view_count': int_or_none(e.get('interactionCount')),
997 if e.get('@context') == 'http://schema.org':
998 item_type = e.get('@type')
999 if expected_type is not None and expected_type != item_type:
1001 if item_type == 'TVEpisode':
1003 'episode': unescapeHTML(e.get('name')),
1004 'episode_number': int_or_none(e.get('episodeNumber')),
1005 'description': unescapeHTML(e.get('description')),
1007 part_of_season = e.get('partOfSeason')
1008 if isinstance(part_of_season, dict) and part_of_season.get('@type') == 'TVSeason':
1009 info['season_number'] = int_or_none(part_of_season.get('seasonNumber'))
1010 part_of_series = e.get('partOfSeries') or e.get('partOfTVSeries')
1011 if isinstance(part_of_series, dict) and part_of_series.get('@type') == 'TVSeries':
1012 info['series'] = unescapeHTML(part_of_series.get('name'))
1013 elif item_type == 'Article':
1015 'timestamp': parse_iso8601(e.get('datePublished')),
1016 'title': unescapeHTML(e.get('headline')),
1017 'description': unescapeHTML(e.get('articleBody')),
1019 elif item_type == 'VideoObject':
1020 extract_video_object(e)
1021 elif item_type == 'WebPage':
1022 video = e.get('video')
1023 if isinstance(video, dict) and video.get('@type') == 'VideoObject':
1024 extract_video_object(video)
1026 return dict((k, v) for k, v in info.items() if v is not None)
1029 def _hidden_inputs(html):
1030 html = re.sub(r'<!--(?:(?!<!--).)*-->', '', html)
1032 for input in re.findall(r'(?i)(<input[^>]+>)', html):
1033 attrs = extract_attributes(input)
1036 if attrs.get('type') not in ('hidden', 'submit'):
1038 name = attrs.get('name') or attrs.get('id')
1039 value = attrs.get('value')
1040 if name and value is not None:
1041 hidden_inputs[name] = value
1042 return hidden_inputs
1044 def _form_hidden_inputs(self, form_id, html):
1045 form = self._search_regex(
1046 r'(?is)<form[^>]+?id=(["\'])%s\1[^>]*>(?P<form>.+?)</form>' % form_id,
1047 html, '%s form' % form_id, group='form')
1048 return self._hidden_inputs(form)
1050 def _sort_formats(self, formats, field_preference=None):
1052 raise ExtractorError('No video formats found')
1055 # Automatically determine tbr when missing based on abr and vbr (improves
1056 # formats sorting in some cases)
1057 if 'tbr' not in f and f.get('abr') is not None and f.get('vbr') is not None:
1058 f['tbr'] = f['abr'] + f['vbr']
1060 def _formats_key(f):
1061 # TODO remove the following workaround
1062 from ..utils import determine_ext
1063 if not f.get('ext') and 'url' in f:
1064 f['ext'] = determine_ext(f['url'])
1066 if isinstance(field_preference, (list, tuple)):
1069 if f.get(field) is not None
1070 else ('' if field == 'format_id' else -1)
1071 for field in field_preference)
1073 preference = f.get('preference')
1074 if preference is None:
1076 if f.get('ext') in ['f4f', 'f4m']: # Not yet supported
1079 protocol = f.get('protocol') or determine_protocol(f)
1080 proto_preference = 0 if protocol in ['http', 'https'] else (-0.5 if protocol == 'rtsp' else -0.1)
1082 if f.get('vcodec') == 'none': # audio only
1084 if self._downloader.params.get('prefer_free_formats'):
1085 ORDER = ['aac', 'mp3', 'm4a', 'webm', 'ogg', 'opus']
1087 ORDER = ['webm', 'opus', 'ogg', 'mp3', 'aac', 'm4a']
1090 audio_ext_preference = ORDER.index(f['ext'])
1092 audio_ext_preference = -1
1094 if f.get('acodec') == 'none': # video only
1096 if self._downloader.params.get('prefer_free_formats'):
1097 ORDER = ['flv', 'mp4', 'webm']
1099 ORDER = ['webm', 'flv', 'mp4']
1101 ext_preference = ORDER.index(f['ext'])
1104 audio_ext_preference = 0
1108 f.get('language_preference') if f.get('language_preference') is not None else -1,
1109 f.get('quality') if f.get('quality') is not None else -1,
1110 f.get('tbr') if f.get('tbr') is not None else -1,
1111 f.get('filesize') if f.get('filesize') is not None else -1,
1112 f.get('vbr') if f.get('vbr') is not None else -1,
1113 f.get('height') if f.get('height') is not None else -1,
1114 f.get('width') if f.get('width') is not None else -1,
1117 f.get('abr') if f.get('abr') is not None else -1,
1118 audio_ext_preference,
1119 f.get('fps') if f.get('fps') is not None else -1,
1120 f.get('filesize_approx') if f.get('filesize_approx') is not None else -1,
1121 f.get('source_preference') if f.get('source_preference') is not None else -1,
1122 f.get('format_id') if f.get('format_id') is not None else '',
1124 formats.sort(key=_formats_key)
1126 def _check_formats(self, formats, video_id):
1128 formats[:] = filter(
1129 lambda f: self._is_valid_url(
1131 item='%s video format' % f.get('format_id') if f.get('format_id') else 'video'),
1135 def _remove_duplicate_formats(formats):
1139 if f['url'] not in format_urls:
1140 format_urls.add(f['url'])
1141 unique_formats.append(f)
1142 formats[:] = unique_formats
1144 def _is_valid_url(self, url, video_id, item='video', headers={}):
1145 url = self._proto_relative_url(url, scheme='http:')
1146 # For now assume non HTTP(S) URLs always valid
1147 if not (url.startswith('http://') or url.startswith('https://')):
1150 self._request_webpage(url, video_id, 'Checking %s URL' % item, headers=headers)
1152 except ExtractorError as e:
1153 if isinstance(e.cause, compat_urllib_error.URLError):
1155 '%s: %s URL is invalid, skipping' % (video_id, item))
1159 def http_scheme(self):
1160 """ Either "http:" or "https:", depending on the user's preferences """
1163 if self._downloader.params.get('prefer_insecure', False)
1166 def _proto_relative_url(self, url, scheme=None):
1169 if url.startswith('//'):
1171 scheme = self.http_scheme()
1176 def _sleep(self, timeout, video_id, msg_template=None):
1177 if msg_template is None:
1178 msg_template = '%(video_id)s: Waiting for %(timeout)s seconds'
1179 msg = msg_template % {'video_id': video_id, 'timeout': timeout}
1183 def _extract_f4m_formats(self, manifest_url, video_id, preference=None, f4m_id=None,
1184 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1185 fatal=True, m3u8_id=None):
1186 manifest = self._download_xml(
1187 manifest_url, video_id, 'Downloading f4m manifest',
1188 'Unable to download f4m manifest',
1189 # Some manifests may be malformed, e.g. prosiebensat1 generated manifests
1190 # (see https://github.com/rg3/youtube-dl/issues/6215#issuecomment-121704244)
1191 transform_source=transform_source,
1194 if manifest is False:
1197 return self._parse_f4m_formats(
1198 manifest, manifest_url, video_id, preference=preference, f4m_id=f4m_id,
1199 transform_source=transform_source, fatal=fatal, m3u8_id=m3u8_id)
1201 def _parse_f4m_formats(self, manifest, manifest_url, video_id, preference=None, f4m_id=None,
1202 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1203 fatal=True, m3u8_id=None):
1204 # currently youtube-dl cannot decode the playerVerificationChallenge as Akamai uses Adobe Alchemy
1205 akamai_pv = manifest.find('{http://ns.adobe.com/f4m/1.0}pv-2.0')
1206 if akamai_pv is not None and ';' in akamai_pv.text:
1207 playerVerificationChallenge = akamai_pv.text.split(';')[0]
1208 if playerVerificationChallenge.strip() != '':
1212 manifest_version = '1.0'
1213 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/1.0}media')
1215 manifest_version = '2.0'
1216 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/2.0}media')
1217 # Remove unsupported DRM protected media from final formats
1218 # rendition (see https://github.com/rg3/youtube-dl/issues/8573).
1219 media_nodes = remove_encrypted_media(media_nodes)
1222 base_url = xpath_text(
1223 manifest, ['{http://ns.adobe.com/f4m/1.0}baseURL', '{http://ns.adobe.com/f4m/2.0}baseURL'],
1224 'base URL', default=None)
1226 base_url = base_url.strip()
1228 bootstrap_info = xpath_element(
1229 manifest, ['{http://ns.adobe.com/f4m/1.0}bootstrapInfo', '{http://ns.adobe.com/f4m/2.0}bootstrapInfo'],
1230 'bootstrap info', default=None)
1233 mime_type = xpath_text(
1234 manifest, ['{http://ns.adobe.com/f4m/1.0}mimeType', '{http://ns.adobe.com/f4m/2.0}mimeType'],
1235 'base URL', default=None)
1236 if mime_type and mime_type.startswith('audio/'):
1239 for i, media_el in enumerate(media_nodes):
1240 tbr = int_or_none(media_el.attrib.get('bitrate'))
1241 width = int_or_none(media_el.attrib.get('width'))
1242 height = int_or_none(media_el.attrib.get('height'))
1243 format_id = '-'.join(filter(None, [f4m_id, compat_str(i if tbr is None else tbr)]))
1244 # If <bootstrapInfo> is present, the specified f4m is a
1245 # stream-level manifest, and only set-level manifests may refer to
1246 # external resources. See section 11.4 and section 4 of F4M spec
1247 if bootstrap_info is None:
1249 # @href is introduced in 2.0, see section 11.6 of F4M spec
1250 if manifest_version == '2.0':
1251 media_url = media_el.attrib.get('href')
1252 if media_url is None:
1253 media_url = media_el.attrib.get('url')
1257 media_url if media_url.startswith('http://') or media_url.startswith('https://')
1258 else ((base_url or '/'.join(manifest_url.split('/')[:-1])) + '/' + media_url))
1259 # If media_url is itself a f4m manifest do the recursive extraction
1260 # since bitrates in parent manifest (this one) and media_url manifest
1261 # may differ leading to inability to resolve the format by requested
1262 # bitrate in f4m downloader
1263 ext = determine_ext(manifest_url)
1265 f4m_formats = self._extract_f4m_formats(
1266 manifest_url, video_id, preference=preference, f4m_id=f4m_id,
1267 transform_source=transform_source, fatal=fatal)
1268 # Sometimes stream-level manifest contains single media entry that
1269 # does not contain any quality metadata (e.g. http://matchtv.ru/#live-player).
1270 # At the same time parent's media entry in set-level manifest may
1271 # contain it. We will copy it from parent in such cases.
1272 if len(f4m_formats) == 1:
1275 'tbr': f.get('tbr') or tbr,
1276 'width': f.get('width') or width,
1277 'height': f.get('height') or height,
1278 'format_id': f.get('format_id') if not tbr else format_id,
1281 formats.extend(f4m_formats)
1284 formats.extend(self._extract_m3u8_formats(
1285 manifest_url, video_id, 'mp4', preference=preference,
1286 m3u8_id=m3u8_id, fatal=fatal))
1289 'format_id': format_id,
1290 'url': manifest_url,
1291 'manifest_url': manifest_url,
1292 'ext': 'flv' if bootstrap_info is not None else None,
1297 'preference': preference,
1301 def _m3u8_meta_format(self, m3u8_url, ext=None, preference=None, m3u8_id=None):
1303 'format_id': '-'.join(filter(None, [m3u8_id, 'meta'])),
1307 'preference': preference - 100 if preference else -100,
1308 'resolution': 'multiple',
1309 'format_note': 'Quality selection URL',
1312 def _extract_m3u8_formats(self, m3u8_url, video_id, ext=None,
1313 entry_protocol='m3u8', preference=None,
1314 m3u8_id=None, note=None, errnote=None,
1315 fatal=True, live=False):
1316 res = self._download_webpage_handle(
1318 note=note or 'Downloading m3u8 information',
1319 errnote=errnote or 'Failed to download m3u8 information',
1325 m3u8_doc, urlh = res
1326 m3u8_url = urlh.geturl()
1328 return self._parse_m3u8_formats(
1329 m3u8_doc, m3u8_url, ext=ext, entry_protocol=entry_protocol,
1330 preference=preference, m3u8_id=m3u8_id, live=live)
1332 def _parse_m3u8_formats(self, m3u8_doc, m3u8_url, ext=None,
1333 entry_protocol='m3u8', preference=None,
1334 m3u8_id=None, live=False):
1335 if '#EXT-X-FAXS-CM:' in m3u8_doc: # Adobe Flash Access
1340 format_url = lambda u: (
1342 if re.match(r'^https?://', u)
1343 else compat_urlparse.urljoin(m3u8_url, u))
1346 # 1. https://tools.ietf.org/html/draft-pantos-http-live-streaming-21
1347 # 2. https://github.com/rg3/youtube-dl/issues/12211
1349 # We should try extracting formats only from master playlists [1, 4.3.4],
1350 # i.e. playlists that describe available qualities. On the other hand
1351 # media playlists [1, 4.3.3] should be returned as is since they contain
1352 # just the media without qualities renditions.
1353 # Fortunately, master playlist can be easily distinguished from media
1354 # playlist based on particular tags availability. As of [1, 4.3.3, 4.3.4]
1355 # master playlist tags MUST NOT appear in a media playist and vice versa.
1356 # As of [1, 4.3.3.1] #EXT-X-TARGETDURATION tag is REQUIRED for every
1357 # media playlist and MUST NOT appear in master playlist thus we can
1358 # clearly detect media playlist with this criterion.
1360 if '#EXT-X-TARGETDURATION' in m3u8_doc: # media playlist, return as is
1363 'format_id': m3u8_id,
1365 'protocol': entry_protocol,
1366 'preference': preference,
1370 last_stream_inf = {}
1372 def extract_media(x_media_line):
1373 media = parse_m3u8_attributes(x_media_line)
1374 # As per [1, 4.3.4.1] TYPE, GROUP-ID and NAME are REQUIRED
1375 media_type, group_id, name = media.get('TYPE'), media.get('GROUP-ID'), media.get('NAME')
1376 if not (media_type and group_id and name):
1378 groups.setdefault(group_id, []).append(media)
1379 if media_type not in ('VIDEO', 'AUDIO'):
1381 media_url = media.get('URI')
1384 for v in (group_id, name):
1388 'format_id': '-'.join(format_id),
1389 'url': format_url(media_url),
1390 'manifest_url': m3u8_url,
1391 'language': media.get('LANGUAGE'),
1393 'protocol': entry_protocol,
1394 'preference': preference,
1396 if media_type == 'AUDIO':
1397 f['vcodec'] = 'none'
1400 def build_stream_name():
1401 # Despite specification does not mention NAME attribute for
1402 # EXT-X-STREAM-INF tag it still sometimes may be present (see [1]
1403 # or vidio test in TestInfoExtractor.test_parse_m3u8_formats)
1404 # 1. http://www.vidio.com/watch/165683-dj_ambred-booyah-live-2015
1405 stream_name = last_stream_inf.get('NAME')
1408 # If there is no NAME in EXT-X-STREAM-INF it will be obtained
1409 # from corresponding rendition group
1410 stream_group_id = last_stream_inf.get('VIDEO')
1411 if not stream_group_id:
1413 stream_group = groups.get(stream_group_id)
1414 if not stream_group:
1415 return stream_group_id
1416 rendition = stream_group[0]
1417 return rendition.get('NAME') or stream_group_id
1419 for line in m3u8_doc.splitlines():
1420 if line.startswith('#EXT-X-STREAM-INF:'):
1421 last_stream_inf = parse_m3u8_attributes(line)
1422 elif line.startswith('#EXT-X-MEDIA:'):
1424 elif line.startswith('#') or not line.strip():
1427 tbr = float_or_none(
1428 last_stream_inf.get('AVERAGE-BANDWIDTH') or
1429 last_stream_inf.get('BANDWIDTH'), scale=1000)
1432 format_id.append(m3u8_id)
1433 stream_name = build_stream_name()
1434 # Bandwidth of live streams may differ over time thus making
1435 # format_id unpredictable. So it's better to keep provided
1438 format_id.append(stream_name if stream_name else '%d' % (tbr if tbr else len(formats)))
1439 manifest_url = format_url(line.strip())
1441 'format_id': '-'.join(format_id),
1442 'url': manifest_url,
1443 'manifest_url': m3u8_url,
1446 'fps': float_or_none(last_stream_inf.get('FRAME-RATE')),
1447 'protocol': entry_protocol,
1448 'preference': preference,
1450 resolution = last_stream_inf.get('RESOLUTION')
1452 mobj = re.search(r'(?P<width>\d+)[xX](?P<height>\d+)', resolution)
1454 f['width'] = int(mobj.group('width'))
1455 f['height'] = int(mobj.group('height'))
1456 # Unified Streaming Platform
1458 r'audio.*?(?:%3D|=)(\d+)(?:-video.*?(?:%3D|=)(\d+))?', f['url'])
1460 abr, vbr = mobj.groups()
1461 abr, vbr = float_or_none(abr, 1000), float_or_none(vbr, 1000)
1466 codecs = parse_codecs(last_stream_inf.get('CODECS'))
1468 audio_group_id = last_stream_inf.get('AUDIO')
1469 # As per [1, 4.3.4.1.1] any EXT-X-STREAM-INF tag which
1470 # references a rendition group MUST have a CODECS attribute.
1471 # However, this is not always respected, for example, [2]
1472 # contains EXT-X-STREAM-INF tag which references AUDIO
1473 # rendition group but does not have CODECS and despite
1474 # referencing audio group an audio group, it represents
1475 # a complete (with audio and video) format. So, for such cases
1476 # we will ignore references to rendition groups and treat them
1477 # as complete formats.
1478 if audio_group_id and codecs and f.get('vcodec') != 'none':
1479 audio_group = groups.get(audio_group_id)
1480 if audio_group and audio_group[0].get('URI'):
1481 # TODO: update acodec for audio only formats with
1483 f['acodec'] = 'none'
1485 last_stream_inf = {}
1489 def _xpath_ns(path, namespace=None):
1493 for c in path.split('/'):
1494 if not c or c == '.':
1497 out.append('{%s}%s' % (namespace, c))
1498 return '/'.join(out)
1500 def _extract_smil_formats(self, smil_url, video_id, fatal=True, f4m_params=None, transform_source=None):
1501 smil = self._download_smil(smil_url, video_id, fatal=fatal, transform_source=transform_source)
1507 namespace = self._parse_smil_namespace(smil)
1509 return self._parse_smil_formats(
1510 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
1512 def _extract_smil_info(self, smil_url, video_id, fatal=True, f4m_params=None):
1513 smil = self._download_smil(smil_url, video_id, fatal=fatal)
1516 return self._parse_smil(smil, smil_url, video_id, f4m_params=f4m_params)
1518 def _download_smil(self, smil_url, video_id, fatal=True, transform_source=None):
1519 return self._download_xml(
1520 smil_url, video_id, 'Downloading SMIL file',
1521 'Unable to download SMIL file', fatal=fatal, transform_source=transform_source)
1523 def _parse_smil(self, smil, smil_url, video_id, f4m_params=None):
1524 namespace = self._parse_smil_namespace(smil)
1526 formats = self._parse_smil_formats(
1527 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
1528 subtitles = self._parse_smil_subtitles(smil, namespace=namespace)
1530 video_id = os.path.splitext(url_basename(smil_url))[0]
1534 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
1535 name = meta.attrib.get('name')
1536 content = meta.attrib.get('content')
1537 if not name or not content:
1539 if not title and name == 'title':
1541 elif not description and name in ('description', 'abstract'):
1542 description = content
1543 elif not upload_date and name == 'date':
1544 upload_date = unified_strdate(content)
1547 'id': image.get('type'),
1548 'url': image.get('src'),
1549 'width': int_or_none(image.get('width')),
1550 'height': int_or_none(image.get('height')),
1551 } for image in smil.findall(self._xpath_ns('.//image', namespace)) if image.get('src')]
1555 'title': title or video_id,
1556 'description': description,
1557 'upload_date': upload_date,
1558 'thumbnails': thumbnails,
1560 'subtitles': subtitles,
1563 def _parse_smil_namespace(self, smil):
1564 return self._search_regex(
1565 r'(?i)^{([^}]+)?}smil$', smil.tag, 'namespace', default=None)
1567 def _parse_smil_formats(self, smil, smil_url, video_id, namespace=None, f4m_params=None, transform_rtmp_url=None):
1569 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
1570 b = meta.get('base') or meta.get('httpBase')
1581 media = smil.findall(self._xpath_ns('.//video', namespace)) + smil.findall(self._xpath_ns('.//audio', namespace))
1582 for medium in media:
1583 src = medium.get('src')
1584 if not src or src in srcs:
1588 bitrate = float_or_none(medium.get('system-bitrate') or medium.get('systemBitrate'), 1000)
1589 filesize = int_or_none(medium.get('size') or medium.get('fileSize'))
1590 width = int_or_none(medium.get('width'))
1591 height = int_or_none(medium.get('height'))
1592 proto = medium.get('proto')
1593 ext = medium.get('ext')
1594 src_ext = determine_ext(src)
1595 streamer = medium.get('streamer') or base
1597 if proto == 'rtmp' or streamer.startswith('rtmp'):
1603 'format_id': 'rtmp-%d' % (rtmp_count if bitrate is None else bitrate),
1605 'filesize': filesize,
1609 if transform_rtmp_url:
1610 streamer, src = transform_rtmp_url(streamer, src)
1611 formats[-1].update({
1617 src_url = src if src.startswith('http') else compat_urlparse.urljoin(base, src)
1618 src_url = src_url.strip()
1620 if proto == 'm3u8' or src_ext == 'm3u8':
1621 m3u8_formats = self._extract_m3u8_formats(
1622 src_url, video_id, ext or 'mp4', m3u8_id='hls', fatal=False)
1623 if len(m3u8_formats) == 1:
1625 m3u8_formats[0].update({
1626 'format_id': 'hls-%d' % (m3u8_count if bitrate is None else bitrate),
1631 formats.extend(m3u8_formats)
1634 if src_ext == 'f4m':
1639 'plugin': 'flowplayer-3.2.0.1',
1641 f4m_url += '&' if '?' in f4m_url else '?'
1642 f4m_url += compat_urllib_parse_urlencode(f4m_params)
1643 formats.extend(self._extract_f4m_formats(f4m_url, video_id, f4m_id='hds', fatal=False))
1646 if src_url.startswith('http') and self._is_valid_url(src, video_id):
1650 'ext': ext or src_ext or 'flv',
1651 'format_id': 'http-%d' % (bitrate or http_count),
1653 'filesize': filesize,
1661 def _parse_smil_subtitles(self, smil, namespace=None, subtitles_lang='en'):
1664 for num, textstream in enumerate(smil.findall(self._xpath_ns('.//textstream', namespace))):
1665 src = textstream.get('src')
1666 if not src or src in urls:
1669 ext = textstream.get('ext') or mimetype2ext(textstream.get('type')) or determine_ext(src)
1670 lang = textstream.get('systemLanguage') or textstream.get('systemLanguageName') or textstream.get('lang') or subtitles_lang
1671 subtitles.setdefault(lang, []).append({
1677 def _extract_xspf_playlist(self, playlist_url, playlist_id, fatal=True):
1678 xspf = self._download_xml(
1679 playlist_url, playlist_id, 'Downloading xpsf playlist',
1680 'Unable to download xspf manifest', fatal=fatal)
1683 return self._parse_xspf(xspf, playlist_id)
1685 def _parse_xspf(self, playlist, playlist_id):
1687 'xspf': 'http://xspf.org/ns/0/',
1688 's1': 'http://static.streamone.nl/player/ns/0',
1692 for track in playlist.findall(xpath_with_ns('./xspf:trackList/xspf:track', NS_MAP)):
1694 track, xpath_with_ns('./xspf:title', NS_MAP), 'title', default=playlist_id)
1695 description = xpath_text(
1696 track, xpath_with_ns('./xspf:annotation', NS_MAP), 'description')
1697 thumbnail = xpath_text(
1698 track, xpath_with_ns('./xspf:image', NS_MAP), 'thumbnail')
1699 duration = float_or_none(
1700 xpath_text(track, xpath_with_ns('./xspf:duration', NS_MAP), 'duration'), 1000)
1703 'url': location.text,
1704 'format_id': location.get(xpath_with_ns('s1:label', NS_MAP)),
1705 'width': int_or_none(location.get(xpath_with_ns('s1:width', NS_MAP))),
1706 'height': int_or_none(location.get(xpath_with_ns('s1:height', NS_MAP))),
1707 } for location in track.findall(xpath_with_ns('./xspf:location', NS_MAP))]
1708 self._sort_formats(formats)
1713 'description': description,
1714 'thumbnail': thumbnail,
1715 'duration': duration,
1720 def _extract_mpd_formats(self, mpd_url, video_id, mpd_id=None, note=None, errnote=None, fatal=True, formats_dict={}):
1721 res = self._download_webpage_handle(
1723 note=note or 'Downloading MPD manifest',
1724 errnote=errnote or 'Failed to download MPD manifest',
1729 mpd_base_url = base_url(urlh.geturl())
1731 return self._parse_mpd_formats(
1732 compat_etree_fromstring(mpd.encode('utf-8')), mpd_id, mpd_base_url,
1733 formats_dict=formats_dict, mpd_url=mpd_url)
1735 def _parse_mpd_formats(self, mpd_doc, mpd_id=None, mpd_base_url='', formats_dict={}, mpd_url=None):
1737 Parse formats from MPD manifest.
1739 1. MPEG-DASH Standard, ISO/IEC 23009-1:2014(E),
1740 http://standards.iso.org/ittf/PubliclyAvailableStandards/c065274_ISO_IEC_23009-1_2014.zip
1741 2. https://en.wikipedia.org/wiki/Dynamic_Adaptive_Streaming_over_HTTP
1743 if mpd_doc.get('type') == 'dynamic':
1746 namespace = self._search_regex(r'(?i)^{([^}]+)?}MPD$', mpd_doc.tag, 'namespace', default=None)
1749 return self._xpath_ns(path, namespace)
1751 def is_drm_protected(element):
1752 return element.find(_add_ns('ContentProtection')) is not None
1754 def extract_multisegment_info(element, ms_parent_info):
1755 ms_info = ms_parent_info.copy()
1757 # As per [1, 5.3.9.2.2] SegmentList and SegmentTemplate share some
1758 # common attributes and elements. We will only extract relevant
1760 def extract_common(source):
1761 segment_timeline = source.find(_add_ns('SegmentTimeline'))
1762 if segment_timeline is not None:
1763 s_e = segment_timeline.findall(_add_ns('S'))
1765 ms_info['total_number'] = 0
1768 r = int(s.get('r', 0))
1769 ms_info['total_number'] += 1 + r
1770 ms_info['s'].append({
1771 't': int(s.get('t', 0)),
1772 # @d is mandatory (see [1, 5.3.9.6.2, Table 17, page 60])
1773 'd': int(s.attrib['d']),
1776 start_number = source.get('startNumber')
1778 ms_info['start_number'] = int(start_number)
1779 timescale = source.get('timescale')
1781 ms_info['timescale'] = int(timescale)
1782 segment_duration = source.get('duration')
1783 if segment_duration:
1784 ms_info['segment_duration'] = int(segment_duration)
1786 def extract_Initialization(source):
1787 initialization = source.find(_add_ns('Initialization'))
1788 if initialization is not None:
1789 ms_info['initialization_url'] = initialization.attrib['sourceURL']
1791 segment_list = element.find(_add_ns('SegmentList'))
1792 if segment_list is not None:
1793 extract_common(segment_list)
1794 extract_Initialization(segment_list)
1795 segment_urls_e = segment_list.findall(_add_ns('SegmentURL'))
1797 ms_info['segment_urls'] = [segment.attrib['media'] for segment in segment_urls_e]
1799 segment_template = element.find(_add_ns('SegmentTemplate'))
1800 if segment_template is not None:
1801 extract_common(segment_template)
1802 media = segment_template.get('media')
1804 ms_info['media'] = media
1805 initialization = segment_template.get('initialization')
1807 ms_info['initialization'] = initialization
1809 extract_Initialization(segment_template)
1812 mpd_duration = parse_duration(mpd_doc.get('mediaPresentationDuration'))
1814 for period in mpd_doc.findall(_add_ns('Period')):
1815 period_duration = parse_duration(period.get('duration')) or mpd_duration
1816 period_ms_info = extract_multisegment_info(period, {
1820 for adaptation_set in period.findall(_add_ns('AdaptationSet')):
1821 if is_drm_protected(adaptation_set):
1823 adaption_set_ms_info = extract_multisegment_info(adaptation_set, period_ms_info)
1824 for representation in adaptation_set.findall(_add_ns('Representation')):
1825 if is_drm_protected(representation):
1827 representation_attrib = adaptation_set.attrib.copy()
1828 representation_attrib.update(representation.attrib)
1829 # According to [1, 5.3.7.2, Table 9, page 41], @mimeType is mandatory
1830 mime_type = representation_attrib['mimeType']
1831 content_type = mime_type.split('/')[0]
1832 if content_type == 'text':
1833 # TODO implement WebVTT downloading
1835 elif content_type in ('video', 'audio'):
1837 for element in (representation, adaptation_set, period, mpd_doc):
1838 base_url_e = element.find(_add_ns('BaseURL'))
1839 if base_url_e is not None:
1840 base_url = base_url_e.text + base_url
1841 if re.match(r'^https?://', base_url):
1843 if mpd_base_url and not re.match(r'^https?://', base_url):
1844 if not mpd_base_url.endswith('/') and not base_url.startswith('/'):
1846 base_url = mpd_base_url + base_url
1847 representation_id = representation_attrib.get('id')
1848 lang = representation_attrib.get('lang')
1849 url_el = representation.find(_add_ns('BaseURL'))
1850 filesize = int_or_none(url_el.attrib.get('{http://youtube.com/yt/2012/10/10}contentLength') if url_el is not None else None)
1851 bandwidth = int_or_none(representation_attrib.get('bandwidth'))
1853 'format_id': '%s-%s' % (mpd_id, representation_id) if mpd_id else representation_id,
1855 'manifest_url': mpd_url,
1856 'ext': mimetype2ext(mime_type),
1857 'width': int_or_none(representation_attrib.get('width')),
1858 'height': int_or_none(representation_attrib.get('height')),
1859 'tbr': float_or_none(bandwidth, 1000),
1860 'asr': int_or_none(representation_attrib.get('audioSamplingRate')),
1861 'fps': int_or_none(representation_attrib.get('frameRate')),
1862 'language': lang if lang not in ('mul', 'und', 'zxx', 'mis') else None,
1863 'format_note': 'DASH %s' % content_type,
1864 'filesize': filesize,
1866 f.update(parse_codecs(representation_attrib.get('codecs')))
1867 representation_ms_info = extract_multisegment_info(representation, adaption_set_ms_info)
1869 def prepare_template(template_name, identifiers):
1870 t = representation_ms_info[template_name]
1871 t = t.replace('$RepresentationID$', representation_id)
1872 t = re.sub(r'\$(%s)\$' % '|'.join(identifiers), r'%(\1)d', t)
1873 t = re.sub(r'\$(%s)%%([^$]+)\$' % '|'.join(identifiers), r'%(\1)\2', t)
1874 t.replace('$$', '$')
1877 # @initialization is a regular template like @media one
1878 # so it should be handled just the same way (see
1879 # https://github.com/rg3/youtube-dl/issues/11605)
1880 if 'initialization' in representation_ms_info:
1881 initialization_template = prepare_template(
1883 # As per [1, 5.3.9.4.2, Table 15, page 54] $Number$ and
1884 # $Time$ shall not be included for @initialization thus
1885 # only $Bandwidth$ remains
1887 representation_ms_info['initialization_url'] = initialization_template % {
1888 'Bandwidth': bandwidth,
1891 if 'segment_urls' not in representation_ms_info and 'media' in representation_ms_info:
1893 media_template = prepare_template('media', ('Number', 'Bandwidth', 'Time'))
1895 # As per [1, 5.3.9.4.4, Table 16, page 55] $Number$ and $Time$
1896 # can't be used at the same time
1897 if '%(Number' in media_template and 's' not in representation_ms_info:
1898 segment_duration = None
1899 if 'total_number' not in representation_ms_info and 'segment_duration':
1900 segment_duration = float_or_none(representation_ms_info['segment_duration'], representation_ms_info['timescale'])
1901 representation_ms_info['total_number'] = int(math.ceil(float(period_duration) / segment_duration))
1902 representation_ms_info['fragments'] = [{
1903 'url': media_template % {
1904 'Number': segment_number,
1905 'Bandwidth': bandwidth,
1907 'duration': segment_duration,
1908 } for segment_number in range(
1909 representation_ms_info['start_number'],
1910 representation_ms_info['total_number'] + representation_ms_info['start_number'])]
1912 # $Number*$ or $Time$ in media template with S list available
1913 # Example $Number*$: http://www.svtplay.se/klipp/9023742/stopptid-om-bjorn-borg
1914 # Example $Time$: https://play.arkena.com/embed/avp/v2/player/media/b41dda37-d8e7-4d3f-b1b5-9a9db578bdfe/1/129411
1915 representation_ms_info['fragments'] = []
1918 segment_number = representation_ms_info['start_number']
1920 def add_segment_url():
1921 segment_url = media_template % {
1922 'Time': segment_time,
1923 'Bandwidth': bandwidth,
1924 'Number': segment_number,
1926 representation_ms_info['fragments'].append({
1928 'duration': float_or_none(segment_d, representation_ms_info['timescale']),
1931 for num, s in enumerate(representation_ms_info['s']):
1932 segment_time = s.get('t') or segment_time
1936 for r in range(s.get('r', 0)):
1937 segment_time += segment_d
1940 segment_time += segment_d
1941 elif 'segment_urls' in representation_ms_info and 's' in representation_ms_info:
1943 # Example: https://www.youtube.com/watch?v=iXZV5uAYMJI
1944 # or any YouTube dashsegments video
1947 timescale = representation_ms_info['timescale']
1948 for s in representation_ms_info['s']:
1949 duration = float_or_none(s['d'], timescale)
1950 for r in range(s.get('r', 0) + 1):
1952 'url': representation_ms_info['segment_urls'][segment_index],
1953 'duration': duration,
1956 representation_ms_info['fragments'] = fragments
1957 # NB: MPD manifest may contain direct URLs to unfragmented media.
1958 # No fragments key is present in this case.
1959 if 'fragments' in representation_ms_info:
1962 'protocol': 'http_dash_segments',
1964 if 'initialization_url' in representation_ms_info:
1965 initialization_url = representation_ms_info['initialization_url']
1966 if not f.get('url'):
1967 f['url'] = initialization_url
1968 f['fragments'].append({'url': initialization_url})
1969 f['fragments'].extend(representation_ms_info['fragments'])
1970 for fragment in f['fragments']:
1971 fragment['url'] = urljoin(base_url, fragment['url'])
1973 existing_format = next(
1974 fo for fo in formats
1975 if fo['format_id'] == representation_id)
1976 except StopIteration:
1977 full_info = formats_dict.get(representation_id, {}).copy()
1979 formats.append(full_info)
1981 existing_format.update(f)
1983 self.report_warning('Unknown MIME type %s in DASH manifest' % mime_type)
1986 def _extract_ism_formats(self, ism_url, video_id, ism_id=None, note=None, errnote=None, fatal=True):
1987 res = self._download_webpage_handle(
1989 note=note or 'Downloading ISM manifest',
1990 errnote=errnote or 'Failed to download ISM manifest',
1996 return self._parse_ism_formats(
1997 compat_etree_fromstring(ism.encode('utf-8')), urlh.geturl(), ism_id)
1999 def _parse_ism_formats(self, ism_doc, ism_url, ism_id=None):
2000 if ism_doc.get('IsLive') == 'TRUE' or ism_doc.find('Protection') is not None:
2003 duration = int(ism_doc.attrib['Duration'])
2004 timescale = int_or_none(ism_doc.get('TimeScale')) or 10000000
2007 for stream in ism_doc.findall('StreamIndex'):
2008 stream_type = stream.get('Type')
2009 if stream_type not in ('video', 'audio'):
2011 url_pattern = stream.attrib['Url']
2012 stream_timescale = int_or_none(stream.get('TimeScale')) or timescale
2013 stream_name = stream.get('Name')
2014 for track in stream.findall('QualityLevel'):
2015 fourcc = track.get('FourCC')
2016 # TODO: add support for WVC1 and WMAP
2017 if fourcc not in ('H264', 'AVC1', 'AACL'):
2018 self.report_warning('%s is not a supported codec' % fourcc)
2020 tbr = int(track.attrib['Bitrate']) // 1000
2021 width = int_or_none(track.get('MaxWidth'))
2022 height = int_or_none(track.get('MaxHeight'))
2023 sampling_rate = int_or_none(track.get('SamplingRate'))
2025 track_url_pattern = re.sub(r'{[Bb]itrate}', track.attrib['Bitrate'], url_pattern)
2026 track_url_pattern = compat_urlparse.urljoin(ism_url, track_url_pattern)
2032 stream_fragments = stream.findall('c')
2033 for stream_fragment_index, stream_fragment in enumerate(stream_fragments):
2034 fragment_ctx['time'] = int_or_none(stream_fragment.get('t')) or fragment_ctx['time']
2035 fragment_repeat = int_or_none(stream_fragment.get('r')) or 1
2036 fragment_ctx['duration'] = int_or_none(stream_fragment.get('d'))
2037 if not fragment_ctx['duration']:
2039 next_fragment_time = int(stream_fragment[stream_fragment_index + 1].attrib['t'])
2041 next_fragment_time = duration
2042 fragment_ctx['duration'] = (next_fragment_time - fragment_ctx['time']) / fragment_repeat
2043 for _ in range(fragment_repeat):
2045 'url': re.sub(r'{start[ _]time}', compat_str(fragment_ctx['time']), track_url_pattern),
2046 'duration': fragment_ctx['duration'] / stream_timescale,
2048 fragment_ctx['time'] += fragment_ctx['duration']
2052 format_id.append(ism_id)
2054 format_id.append(stream_name)
2055 format_id.append(compat_str(tbr))
2058 'format_id': '-'.join(format_id),
2060 'manifest_url': ism_url,
2061 'ext': 'ismv' if stream_type == 'video' else 'isma',
2065 'asr': sampling_rate,
2066 'vcodec': 'none' if stream_type == 'audio' else fourcc,
2067 'acodec': 'none' if stream_type == 'video' else fourcc,
2069 'fragments': fragments,
2070 '_download_params': {
2071 'duration': duration,
2072 'timescale': stream_timescale,
2073 'width': width or 0,
2074 'height': height or 0,
2076 'codec_private_data': track.get('CodecPrivateData'),
2077 'sampling_rate': sampling_rate,
2078 'channels': int_or_none(track.get('Channels', 2)),
2079 'bits_per_sample': int_or_none(track.get('BitsPerSample', 16)),
2080 'nal_unit_length_field': int_or_none(track.get('NALUnitLengthField', 4)),
2085 def _parse_html5_media_entries(self, base_url, webpage, video_id, m3u8_id=None, m3u8_entry_protocol='m3u8', mpd_id=None, preference=None):
2086 def absolute_url(video_url):
2087 return compat_urlparse.urljoin(base_url, video_url)
2089 def parse_content_type(content_type):
2090 if not content_type:
2092 ctr = re.search(r'(?P<mimetype>[^/]+/[^;]+)(?:;\s*codecs="?(?P<codecs>[^"]+))?', content_type)
2094 mimetype, codecs = ctr.groups()
2095 f = parse_codecs(codecs)
2096 f['ext'] = mimetype2ext(mimetype)
2100 def _media_formats(src, cur_media_type):
2101 full_url = absolute_url(src)
2102 ext = determine_ext(full_url)
2104 is_plain_url = False
2105 formats = self._extract_m3u8_formats(
2106 full_url, video_id, ext='mp4',
2107 entry_protocol=m3u8_entry_protocol, m3u8_id=m3u8_id,
2108 preference=preference)
2110 is_plain_url = False
2111 formats = self._extract_mpd_formats(
2112 full_url, video_id, mpd_id=mpd_id)
2117 'vcodec': 'none' if cur_media_type == 'audio' else None,
2119 return is_plain_url, formats
2122 media_tags = [(media_tag, media_type, '')
2123 for media_tag, media_type
2124 in re.findall(r'(?s)(<(video|audio)[^>]*/>)', webpage)]
2125 media_tags.extend(re.findall(
2126 # We only allow video|audio followed by a whitespace or '>'.
2127 # Allowing more characters may end up in significant slow down (see
2128 # https://github.com/rg3/youtube-dl/issues/11979, example URL:
2129 # http://www.porntrex.com/maps/videositemap.xml).
2130 r'(?s)(<(?P<tag>video|audio)(?:\s+[^>]*)?>)(.*?)</(?P=tag)>', webpage))
2131 for media_tag, media_type, media_content in media_tags:
2136 media_attributes = extract_attributes(media_tag)
2137 src = media_attributes.get('src')
2139 _, formats = _media_formats(src, media_type)
2140 media_info['formats'].extend(formats)
2141 media_info['thumbnail'] = media_attributes.get('poster')
2143 for source_tag in re.findall(r'<source[^>]+>', media_content):
2144 source_attributes = extract_attributes(source_tag)
2145 src = source_attributes.get('src')
2148 is_plain_url, formats = _media_formats(src, media_type)
2150 f = parse_content_type(source_attributes.get('type'))
2151 f.update(formats[0])
2152 media_info['formats'].append(f)
2154 media_info['formats'].extend(formats)
2155 for track_tag in re.findall(r'<track[^>]+>', media_content):
2156 track_attributes = extract_attributes(track_tag)
2157 kind = track_attributes.get('kind')
2158 if not kind or kind in ('subtitles', 'captions'):
2159 src = track_attributes.get('src')
2162 lang = track_attributes.get('srclang') or track_attributes.get('lang') or track_attributes.get('label')
2163 media_info['subtitles'].setdefault(lang, []).append({
2164 'url': absolute_url(src),
2166 if media_info['formats'] or media_info['subtitles']:
2167 entries.append(media_info)
2170 def _extract_akamai_formats(self, manifest_url, video_id, hosts={}):
2172 hdcore_sign = 'hdcore=3.7.0'
2173 f4m_url = re.sub(r'(https?://[^/+])/i/', r'\1/z/', manifest_url).replace('/master.m3u8', '/manifest.f4m')
2174 hds_host = hosts.get('hds')
2176 f4m_url = re.sub(r'(https?://)[^/]+', r'\1' + hds_host, f4m_url)
2177 if 'hdcore=' not in f4m_url:
2178 f4m_url += ('&' if '?' in f4m_url else '?') + hdcore_sign
2179 f4m_formats = self._extract_f4m_formats(
2180 f4m_url, video_id, f4m_id='hds', fatal=False)
2181 for entry in f4m_formats:
2182 entry.update({'extra_param_to_segment_url': hdcore_sign})
2183 formats.extend(f4m_formats)
2184 m3u8_url = re.sub(r'(https?://[^/]+)/z/', r'\1/i/', manifest_url).replace('/manifest.f4m', '/master.m3u8')
2185 hls_host = hosts.get('hls')
2187 m3u8_url = re.sub(r'(https?://)[^/]+', r'\1' + hls_host, m3u8_url)
2188 formats.extend(self._extract_m3u8_formats(
2189 m3u8_url, video_id, 'mp4', 'm3u8_native',
2190 m3u8_id='hls', fatal=False))
2193 def _extract_wowza_formats(self, url, video_id, m3u8_entry_protocol='m3u8_native', skip_protocols=[]):
2194 url = re.sub(r'/(?:manifest|playlist|jwplayer)\.(?:m3u8|f4m|mpd|smil)', '', url)
2195 url_base = self._search_regex(r'(?:https?|rtmp|rtsp)(://[^?]+)', url, 'format url')
2196 http_base_url = 'http' + url_base
2198 if 'm3u8' not in skip_protocols:
2199 formats.extend(self._extract_m3u8_formats(
2200 http_base_url + '/playlist.m3u8', video_id, 'mp4',
2201 m3u8_entry_protocol, m3u8_id='hls', fatal=False))
2202 if 'f4m' not in skip_protocols:
2203 formats.extend(self._extract_f4m_formats(
2204 http_base_url + '/manifest.f4m',
2205 video_id, f4m_id='hds', fatal=False))
2206 if 'dash' not in skip_protocols:
2207 formats.extend(self._extract_mpd_formats(
2208 http_base_url + '/manifest.mpd',
2209 video_id, mpd_id='dash', fatal=False))
2210 if re.search(r'(?:/smil:|\.smil)', url_base):
2211 if 'smil' not in skip_protocols:
2212 rtmp_formats = self._extract_smil_formats(
2213 http_base_url + '/jwplayer.smil',
2214 video_id, fatal=False)
2215 for rtmp_format in rtmp_formats:
2216 rtsp_format = rtmp_format.copy()
2217 rtsp_format['url'] = '%s/%s' % (rtmp_format['url'], rtmp_format['play_path'])
2218 del rtsp_format['play_path']
2219 del rtsp_format['ext']
2220 rtsp_format.update({
2221 'url': rtsp_format['url'].replace('rtmp://', 'rtsp://'),
2222 'format_id': rtmp_format['format_id'].replace('rtmp', 'rtsp'),
2225 formats.extend([rtmp_format, rtsp_format])
2227 for protocol in ('rtmp', 'rtsp'):
2228 if protocol not in skip_protocols:
2230 'url': protocol + url_base,
2231 'format_id': protocol,
2232 'protocol': protocol,
2236 def _find_jwplayer_data(self, webpage, video_id=None, transform_source=js_to_json):
2238 r'(?s)jwplayer\((?P<quote>[\'"])[^\'" ]+(?P=quote)\)(?!</script>).*?\.setup\s*\((?P<options>[^)]+)\)',
2242 jwplayer_data = self._parse_json(mobj.group('options'),
2244 transform_source=transform_source)
2245 except ExtractorError:
2248 if isinstance(jwplayer_data, dict):
2249 return jwplayer_data
2251 def _extract_jwplayer_data(self, webpage, video_id, *args, **kwargs):
2252 jwplayer_data = self._find_jwplayer_data(
2253 webpage, video_id, transform_source=js_to_json)
2254 return self._parse_jwplayer_data(
2255 jwplayer_data, video_id, *args, **kwargs)
2257 def _parse_jwplayer_data(self, jwplayer_data, video_id=None, require_title=True,
2258 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
2259 # JWPlayer backward compatibility: flattened playlists
2260 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/api/config.js#L81-L96
2261 if 'playlist' not in jwplayer_data:
2262 jwplayer_data = {'playlist': [jwplayer_data]}
2266 # JWPlayer backward compatibility: single playlist item
2267 # https://github.com/jwplayer/jwplayer/blob/v7.7.0/src/js/playlist/playlist.js#L10
2268 if not isinstance(jwplayer_data['playlist'], list):
2269 jwplayer_data['playlist'] = [jwplayer_data['playlist']]
2271 for video_data in jwplayer_data['playlist']:
2272 # JWPlayer backward compatibility: flattened sources
2273 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/playlist/item.js#L29-L35
2274 if 'sources' not in video_data:
2275 video_data['sources'] = [video_data]
2277 this_video_id = video_id or video_data['mediaid']
2279 formats = self._parse_jwplayer_formats(
2280 video_data['sources'], video_id=this_video_id, m3u8_id=m3u8_id,
2281 mpd_id=mpd_id, rtmp_params=rtmp_params, base_url=base_url)
2282 self._sort_formats(formats)
2285 tracks = video_data.get('tracks')
2286 if tracks and isinstance(tracks, list):
2287 for track in tracks:
2288 if track.get('kind') != 'captions':
2290 track_url = urljoin(base_url, track.get('file'))
2293 subtitles.setdefault(track.get('label') or 'en', []).append({
2294 'url': self._proto_relative_url(track_url)
2298 'id': this_video_id,
2299 'title': video_data['title'] if require_title else video_data.get('title'),
2300 'description': video_data.get('description'),
2301 'thumbnail': self._proto_relative_url(video_data.get('image')),
2302 'timestamp': int_or_none(video_data.get('pubdate')),
2303 'duration': float_or_none(jwplayer_data.get('duration') or video_data.get('duration')),
2304 'subtitles': subtitles,
2307 if len(entries) == 1:
2310 return self.playlist_result(entries)
2312 def _parse_jwplayer_formats(self, jwplayer_sources_data, video_id=None,
2313 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
2316 for source in jwplayer_sources_data:
2317 source_url = self._proto_relative_url(source.get('file'))
2321 source_url = compat_urlparse.urljoin(base_url, source_url)
2322 if source_url in urls:
2324 urls.append(source_url)
2325 source_type = source.get('type') or ''
2326 ext = mimetype2ext(source_type) or determine_ext(source_url)
2327 if source_type == 'hls' or ext == 'm3u8':
2328 formats.extend(self._extract_m3u8_formats(
2329 source_url, video_id, 'mp4', entry_protocol='m3u8_native',
2330 m3u8_id=m3u8_id, fatal=False))
2332 formats.extend(self._extract_mpd_formats(
2333 source_url, video_id, mpd_id=mpd_id, fatal=False))
2335 formats.extend(self._extract_smil_formats(
2336 source_url, video_id, fatal=False))
2337 # https://github.com/jwplayer/jwplayer/blob/master/src/js/providers/default.js#L67
2338 elif source_type.startswith('audio') or ext in (
2339 'oga', 'aac', 'mp3', 'mpeg', 'vorbis'):
2346 height = int_or_none(source.get('height'))
2348 # Often no height is provided but there is a label in
2349 # format like "1080p", "720p SD", or 1080.
2350 height = int_or_none(self._search_regex(
2351 r'^(\d{3,4})[pP]?(?:\b|$)', compat_str(source.get('label') or ''),
2352 'height', default=None))
2355 'width': int_or_none(source.get('width')),
2357 'tbr': int_or_none(source.get('bitrate')),
2360 if source_url.startswith('rtmp'):
2361 a_format['ext'] = 'flv'
2362 # See com/longtailvideo/jwplayer/media/RTMPMediaProvider.as
2363 # of jwplayer.flash.swf
2364 rtmp_url_parts = re.split(
2365 r'((?:mp4|mp3|flv):)', source_url, 1)
2366 if len(rtmp_url_parts) == 3:
2367 rtmp_url, prefix, play_path = rtmp_url_parts
2370 'play_path': prefix + play_path,
2373 a_format.update(rtmp_params)
2374 formats.append(a_format)
2377 def _live_title(self, name):
2378 """ Generate the title for a live video """
2379 now = datetime.datetime.now()
2380 now_str = now.strftime('%Y-%m-%d %H:%M')
2381 return name + ' ' + now_str
2383 def _int(self, v, name, fatal=False, **kwargs):
2384 res = int_or_none(v, **kwargs)
2385 if 'get_attr' in kwargs:
2386 print(getattr(v, kwargs['get_attr']))
2388 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
2390 raise ExtractorError(msg)
2392 self._downloader.report_warning(msg)
2395 def _float(self, v, name, fatal=False, **kwargs):
2396 res = float_or_none(v, **kwargs)
2398 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
2400 raise ExtractorError(msg)
2402 self._downloader.report_warning(msg)
2405 def _set_cookie(self, domain, name, value, expire_time=None):
2406 cookie = compat_cookiejar.Cookie(
2407 0, name, value, None, None, domain, None,
2408 None, '/', True, False, expire_time, '', None, None, None)
2409 self._downloader.cookiejar.set_cookie(cookie)
2411 def _get_cookies(self, url):
2412 """ Return a compat_cookies.SimpleCookie with the cookies for the url """
2413 req = sanitized_Request(url)
2414 self._downloader.cookiejar.add_cookie_header(req)
2415 return compat_cookies.SimpleCookie(req.get_header('Cookie'))
2417 def get_testcases(self, include_onlymatching=False):
2418 t = getattr(self, '_TEST', None)
2420 assert not hasattr(self, '_TESTS'), \
2421 '%s has _TEST and _TESTS' % type(self).__name__
2424 tests = getattr(self, '_TESTS', [])
2426 if not include_onlymatching and t.get('only_matching', False):
2428 t['name'] = type(self).__name__[:-len('IE')]
2431 def is_suitable(self, age_limit):
2432 """ Test whether the extractor is generally suitable for the given
2433 age limit (i.e. pornographic sites are not, all others usually are) """
2435 any_restricted = False
2436 for tc in self.get_testcases(include_onlymatching=False):
2437 if tc.get('playlist', []):
2438 tc = tc['playlist'][0]
2439 is_restricted = age_restricted(
2440 tc.get('info_dict', {}).get('age_limit'), age_limit)
2441 if not is_restricted:
2443 any_restricted = any_restricted or is_restricted
2444 return not any_restricted
2446 def extract_subtitles(self, *args, **kwargs):
2447 if (self._downloader.params.get('writesubtitles', False) or
2448 self._downloader.params.get('listsubtitles')):
2449 return self._get_subtitles(*args, **kwargs)
2452 def _get_subtitles(self, *args, **kwargs):
2453 raise NotImplementedError('This method must be implemented by subclasses')
2456 def _merge_subtitle_items(subtitle_list1, subtitle_list2):
2457 """ Merge subtitle items for one language. Items with duplicated URLs
2458 will be dropped. """
2459 list1_urls = set([item['url'] for item in subtitle_list1])
2460 ret = list(subtitle_list1)
2461 ret.extend([item for item in subtitle_list2 if item['url'] not in list1_urls])
2465 def _merge_subtitles(cls, subtitle_dict1, subtitle_dict2):
2466 """ Merge two subtitle dictionaries, language by language. """
2467 ret = dict(subtitle_dict1)
2468 for lang in subtitle_dict2:
2469 ret[lang] = cls._merge_subtitle_items(subtitle_dict1.get(lang, []), subtitle_dict2[lang])
2472 def extract_automatic_captions(self, *args, **kwargs):
2473 if (self._downloader.params.get('writeautomaticsub', False) or
2474 self._downloader.params.get('listsubtitles')):
2475 return self._get_automatic_captions(*args, **kwargs)
2478 def _get_automatic_captions(self, *args, **kwargs):
2479 raise NotImplementedError('This method must be implemented by subclasses')
2481 def mark_watched(self, *args, **kwargs):
2482 if (self._downloader.params.get('mark_watched', False) and
2483 (self._get_login_info()[0] is not None or
2484 self._downloader.params.get('cookiefile') is not None)):
2485 self._mark_watched(*args, **kwargs)
2487 def _mark_watched(self, *args, **kwargs):
2488 raise NotImplementedError('This method must be implemented by subclasses')
2490 def geo_verification_headers(self):
2492 geo_verification_proxy = self._downloader.params.get('geo_verification_proxy')
2493 if geo_verification_proxy:
2494 headers['Ytdl-request-proxy'] = geo_verification_proxy
2497 def _generic_id(self, url):
2498 return compat_urllib_parse_unquote(os.path.splitext(url.rstrip('/').split('/')[-1])[0])
2500 def _generic_title(self, url):
2501 return compat_urllib_parse_unquote(os.path.splitext(url_basename(url))[0])
2504 class SearchInfoExtractor(InfoExtractor):
2506 Base class for paged search queries extractors.
2507 They accept URLs in the format _SEARCH_KEY(|all|[0-9]):{query}
2508 Instances should define _SEARCH_KEY and _MAX_RESULTS.
2512 def _make_valid_url(cls):
2513 return r'%s(?P<prefix>|[1-9][0-9]*|all):(?P<query>[\s\S]+)' % cls._SEARCH_KEY
2516 def suitable(cls, url):
2517 return re.match(cls._make_valid_url(), url) is not None
2519 def _real_extract(self, query):
2520 mobj = re.match(self._make_valid_url(), query)
2522 raise ExtractorError('Invalid search query "%s"' % query)
2524 prefix = mobj.group('prefix')
2525 query = mobj.group('query')
2527 return self._get_n_results(query, 1)
2528 elif prefix == 'all':
2529 return self._get_n_results(query, self._MAX_RESULTS)
2533 raise ExtractorError('invalid download number %s for query "%s"' % (n, query))
2534 elif n > self._MAX_RESULTS:
2535 self._downloader.report_warning('%s returns max %i results (you requested %i)' % (self._SEARCH_KEY, self._MAX_RESULTS, n))
2536 n = self._MAX_RESULTS
2537 return self._get_n_results(query, n)
2539 def _get_n_results(self, query, n):
2540 """Get a specified number of results for a query"""
2541 raise NotImplementedError('This method must be implemented by subclasses')
2544 def SEARCH_KEY(self):
2545 return self._SEARCH_KEY