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')),
996 if e.get('@context') == 'http://schema.org':
997 item_type = e.get('@type')
998 if expected_type is not None and expected_type != item_type:
1000 if item_type == 'TVEpisode':
1002 'episode': unescapeHTML(e.get('name')),
1003 'episode_number': int_or_none(e.get('episodeNumber')),
1004 'description': unescapeHTML(e.get('description')),
1006 part_of_season = e.get('partOfSeason')
1007 if isinstance(part_of_season, dict) and part_of_season.get('@type') == 'TVSeason':
1008 info['season_number'] = int_or_none(part_of_season.get('seasonNumber'))
1009 part_of_series = e.get('partOfSeries') or e.get('partOfTVSeries')
1010 if isinstance(part_of_series, dict) and part_of_series.get('@type') == 'TVSeries':
1011 info['series'] = unescapeHTML(part_of_series.get('name'))
1012 elif item_type == 'Article':
1014 'timestamp': parse_iso8601(e.get('datePublished')),
1015 'title': unescapeHTML(e.get('headline')),
1016 'description': unescapeHTML(e.get('articleBody')),
1018 elif item_type == 'VideoObject':
1019 extract_video_object(e)
1020 elif item_type == 'WebPage':
1021 video = e.get('video')
1022 if isinstance(video, dict) and video.get('@type') == 'VideoObject':
1023 extract_video_object(video)
1025 return dict((k, v) for k, v in info.items() if v is not None)
1028 def _hidden_inputs(html):
1029 html = re.sub(r'<!--(?:(?!<!--).)*-->', '', html)
1031 for input in re.findall(r'(?i)(<input[^>]+>)', html):
1032 attrs = extract_attributes(input)
1035 if attrs.get('type') not in ('hidden', 'submit'):
1037 name = attrs.get('name') or attrs.get('id')
1038 value = attrs.get('value')
1039 if name and value is not None:
1040 hidden_inputs[name] = value
1041 return hidden_inputs
1043 def _form_hidden_inputs(self, form_id, html):
1044 form = self._search_regex(
1045 r'(?is)<form[^>]+?id=(["\'])%s\1[^>]*>(?P<form>.+?)</form>' % form_id,
1046 html, '%s form' % form_id, group='form')
1047 return self._hidden_inputs(form)
1049 def _sort_formats(self, formats, field_preference=None):
1051 raise ExtractorError('No video formats found')
1054 # Automatically determine tbr when missing based on abr and vbr (improves
1055 # formats sorting in some cases)
1056 if 'tbr' not in f and f.get('abr') is not None and f.get('vbr') is not None:
1057 f['tbr'] = f['abr'] + f['vbr']
1059 def _formats_key(f):
1060 # TODO remove the following workaround
1061 from ..utils import determine_ext
1062 if not f.get('ext') and 'url' in f:
1063 f['ext'] = determine_ext(f['url'])
1065 if isinstance(field_preference, (list, tuple)):
1068 if f.get(field) is not None
1069 else ('' if field == 'format_id' else -1)
1070 for field in field_preference)
1072 preference = f.get('preference')
1073 if preference is None:
1075 if f.get('ext') in ['f4f', 'f4m']: # Not yet supported
1078 protocol = f.get('protocol') or determine_protocol(f)
1079 proto_preference = 0 if protocol in ['http', 'https'] else (-0.5 if protocol == 'rtsp' else -0.1)
1081 if f.get('vcodec') == 'none': # audio only
1083 if self._downloader.params.get('prefer_free_formats'):
1084 ORDER = ['aac', 'mp3', 'm4a', 'webm', 'ogg', 'opus']
1086 ORDER = ['webm', 'opus', 'ogg', 'mp3', 'aac', 'm4a']
1089 audio_ext_preference = ORDER.index(f['ext'])
1091 audio_ext_preference = -1
1093 if f.get('acodec') == 'none': # video only
1095 if self._downloader.params.get('prefer_free_formats'):
1096 ORDER = ['flv', 'mp4', 'webm']
1098 ORDER = ['webm', 'flv', 'mp4']
1100 ext_preference = ORDER.index(f['ext'])
1103 audio_ext_preference = 0
1107 f.get('language_preference') if f.get('language_preference') is not None else -1,
1108 f.get('quality') if f.get('quality') is not None else -1,
1109 f.get('tbr') if f.get('tbr') is not None else -1,
1110 f.get('filesize') if f.get('filesize') is not None else -1,
1111 f.get('vbr') if f.get('vbr') is not None else -1,
1112 f.get('height') if f.get('height') is not None else -1,
1113 f.get('width') if f.get('width') is not None else -1,
1116 f.get('abr') if f.get('abr') is not None else -1,
1117 audio_ext_preference,
1118 f.get('fps') if f.get('fps') is not None else -1,
1119 f.get('filesize_approx') if f.get('filesize_approx') is not None else -1,
1120 f.get('source_preference') if f.get('source_preference') is not None else -1,
1121 f.get('format_id') if f.get('format_id') is not None else '',
1123 formats.sort(key=_formats_key)
1125 def _check_formats(self, formats, video_id):
1127 formats[:] = filter(
1128 lambda f: self._is_valid_url(
1130 item='%s video format' % f.get('format_id') if f.get('format_id') else 'video'),
1134 def _remove_duplicate_formats(formats):
1138 if f['url'] not in format_urls:
1139 format_urls.add(f['url'])
1140 unique_formats.append(f)
1141 formats[:] = unique_formats
1143 def _is_valid_url(self, url, video_id, item='video', headers={}):
1144 url = self._proto_relative_url(url, scheme='http:')
1145 # For now assume non HTTP(S) URLs always valid
1146 if not (url.startswith('http://') or url.startswith('https://')):
1149 self._request_webpage(url, video_id, 'Checking %s URL' % item, headers=headers)
1151 except ExtractorError as e:
1152 if isinstance(e.cause, compat_urllib_error.URLError):
1154 '%s: %s URL is invalid, skipping' % (video_id, item))
1158 def http_scheme(self):
1159 """ Either "http:" or "https:", depending on the user's preferences """
1162 if self._downloader.params.get('prefer_insecure', False)
1165 def _proto_relative_url(self, url, scheme=None):
1168 if url.startswith('//'):
1170 scheme = self.http_scheme()
1175 def _sleep(self, timeout, video_id, msg_template=None):
1176 if msg_template is None:
1177 msg_template = '%(video_id)s: Waiting for %(timeout)s seconds'
1178 msg = msg_template % {'video_id': video_id, 'timeout': timeout}
1182 def _extract_f4m_formats(self, manifest_url, video_id, preference=None, f4m_id=None,
1183 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1184 fatal=True, m3u8_id=None):
1185 manifest = self._download_xml(
1186 manifest_url, video_id, 'Downloading f4m manifest',
1187 'Unable to download f4m manifest',
1188 # Some manifests may be malformed, e.g. prosiebensat1 generated manifests
1189 # (see https://github.com/rg3/youtube-dl/issues/6215#issuecomment-121704244)
1190 transform_source=transform_source,
1193 if manifest is False:
1196 return self._parse_f4m_formats(
1197 manifest, manifest_url, video_id, preference=preference, f4m_id=f4m_id,
1198 transform_source=transform_source, fatal=fatal, m3u8_id=m3u8_id)
1200 def _parse_f4m_formats(self, manifest, manifest_url, video_id, preference=None, f4m_id=None,
1201 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1202 fatal=True, m3u8_id=None):
1203 # currently youtube-dl cannot decode the playerVerificationChallenge as Akamai uses Adobe Alchemy
1204 akamai_pv = manifest.find('{http://ns.adobe.com/f4m/1.0}pv-2.0')
1205 if akamai_pv is not None and ';' in akamai_pv.text:
1206 playerVerificationChallenge = akamai_pv.text.split(';')[0]
1207 if playerVerificationChallenge.strip() != '':
1211 manifest_version = '1.0'
1212 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/1.0}media')
1214 manifest_version = '2.0'
1215 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/2.0}media')
1216 # Remove unsupported DRM protected media from final formats
1217 # rendition (see https://github.com/rg3/youtube-dl/issues/8573).
1218 media_nodes = remove_encrypted_media(media_nodes)
1221 base_url = xpath_text(
1222 manifest, ['{http://ns.adobe.com/f4m/1.0}baseURL', '{http://ns.adobe.com/f4m/2.0}baseURL'],
1223 'base URL', default=None)
1225 base_url = base_url.strip()
1227 bootstrap_info = xpath_element(
1228 manifest, ['{http://ns.adobe.com/f4m/1.0}bootstrapInfo', '{http://ns.adobe.com/f4m/2.0}bootstrapInfo'],
1229 'bootstrap info', default=None)
1232 mime_type = xpath_text(
1233 manifest, ['{http://ns.adobe.com/f4m/1.0}mimeType', '{http://ns.adobe.com/f4m/2.0}mimeType'],
1234 'base URL', default=None)
1235 if mime_type and mime_type.startswith('audio/'):
1238 for i, media_el in enumerate(media_nodes):
1239 tbr = int_or_none(media_el.attrib.get('bitrate'))
1240 width = int_or_none(media_el.attrib.get('width'))
1241 height = int_or_none(media_el.attrib.get('height'))
1242 format_id = '-'.join(filter(None, [f4m_id, compat_str(i if tbr is None else tbr)]))
1243 # If <bootstrapInfo> is present, the specified f4m is a
1244 # stream-level manifest, and only set-level manifests may refer to
1245 # external resources. See section 11.4 and section 4 of F4M spec
1246 if bootstrap_info is None:
1248 # @href is introduced in 2.0, see section 11.6 of F4M spec
1249 if manifest_version == '2.0':
1250 media_url = media_el.attrib.get('href')
1251 if media_url is None:
1252 media_url = media_el.attrib.get('url')
1256 media_url if media_url.startswith('http://') or media_url.startswith('https://')
1257 else ((base_url or '/'.join(manifest_url.split('/')[:-1])) + '/' + media_url))
1258 # If media_url is itself a f4m manifest do the recursive extraction
1259 # since bitrates in parent manifest (this one) and media_url manifest
1260 # may differ leading to inability to resolve the format by requested
1261 # bitrate in f4m downloader
1262 ext = determine_ext(manifest_url)
1264 f4m_formats = self._extract_f4m_formats(
1265 manifest_url, video_id, preference=preference, f4m_id=f4m_id,
1266 transform_source=transform_source, fatal=fatal)
1267 # Sometimes stream-level manifest contains single media entry that
1268 # does not contain any quality metadata (e.g. http://matchtv.ru/#live-player).
1269 # At the same time parent's media entry in set-level manifest may
1270 # contain it. We will copy it from parent in such cases.
1271 if len(f4m_formats) == 1:
1274 'tbr': f.get('tbr') or tbr,
1275 'width': f.get('width') or width,
1276 'height': f.get('height') or height,
1277 'format_id': f.get('format_id') if not tbr else format_id,
1280 formats.extend(f4m_formats)
1283 formats.extend(self._extract_m3u8_formats(
1284 manifest_url, video_id, 'mp4', preference=preference,
1285 m3u8_id=m3u8_id, fatal=fatal))
1288 'format_id': format_id,
1289 'url': manifest_url,
1290 'manifest_url': manifest_url,
1291 'ext': 'flv' if bootstrap_info is not None else None,
1296 'preference': preference,
1300 def _m3u8_meta_format(self, m3u8_url, ext=None, preference=None, m3u8_id=None):
1302 'format_id': '-'.join(filter(None, [m3u8_id, 'meta'])),
1306 'preference': preference - 100 if preference else -100,
1307 'resolution': 'multiple',
1308 'format_note': 'Quality selection URL',
1311 def _extract_m3u8_formats(self, m3u8_url, video_id, ext=None,
1312 entry_protocol='m3u8', preference=None,
1313 m3u8_id=None, note=None, errnote=None,
1314 fatal=True, live=False):
1315 res = self._download_webpage_handle(
1317 note=note or 'Downloading m3u8 information',
1318 errnote=errnote or 'Failed to download m3u8 information',
1324 m3u8_doc, urlh = res
1325 m3u8_url = urlh.geturl()
1327 return self._parse_m3u8_formats(
1328 m3u8_doc, m3u8_url, ext=ext, entry_protocol=entry_protocol,
1329 preference=preference, m3u8_id=m3u8_id, live=live)
1331 def _parse_m3u8_formats(self, m3u8_doc, m3u8_url, ext=None,
1332 entry_protocol='m3u8', preference=None,
1333 m3u8_id=None, live=False):
1334 if '#EXT-X-FAXS-CM:' in m3u8_doc: # Adobe Flash Access
1339 format_url = lambda u: (
1341 if re.match(r'^https?://', u)
1342 else compat_urlparse.urljoin(m3u8_url, u))
1345 # 1. https://tools.ietf.org/html/draft-pantos-http-live-streaming-21
1346 # 2. https://github.com/rg3/youtube-dl/issues/12211
1348 # We should try extracting formats only from master playlists [1, 4.3.4],
1349 # i.e. playlists that describe available qualities. On the other hand
1350 # media playlists [1, 4.3.3] should be returned as is since they contain
1351 # just the media without qualities renditions.
1352 # Fortunately, master playlist can be easily distinguished from media
1353 # playlist based on particular tags availability. As of [1, 4.3.3, 4.3.4]
1354 # master playlist tags MUST NOT appear in a media playist and vice versa.
1355 # As of [1, 4.3.3.1] #EXT-X-TARGETDURATION tag is REQUIRED for every
1356 # media playlist and MUST NOT appear in master playlist thus we can
1357 # clearly detect media playlist with this criterion.
1359 if '#EXT-X-TARGETDURATION' in m3u8_doc: # media playlist, return as is
1362 'format_id': m3u8_id,
1364 'protocol': entry_protocol,
1365 'preference': preference,
1369 last_stream_inf = {}
1371 def extract_media(x_media_line):
1372 media = parse_m3u8_attributes(x_media_line)
1373 # As per [1, 4.3.4.1] TYPE, GROUP-ID and NAME are REQUIRED
1374 media_type, group_id, name = media.get('TYPE'), media.get('GROUP-ID'), media.get('NAME')
1375 if not (media_type and group_id and name):
1377 groups.setdefault(group_id, []).append(media)
1378 if media_type not in ('VIDEO', 'AUDIO'):
1380 media_url = media.get('URI')
1383 for v in (group_id, name):
1387 'format_id': '-'.join(format_id),
1388 'url': format_url(media_url),
1389 'manifest_url': m3u8_url,
1390 'language': media.get('LANGUAGE'),
1392 'protocol': entry_protocol,
1393 'preference': preference,
1395 if media_type == 'AUDIO':
1396 f['vcodec'] = 'none'
1399 def build_stream_name():
1400 # Despite specification does not mention NAME attribute for
1401 # EXT-X-STREAM-INF tag it still sometimes may be present (see [1]
1402 # or vidio test in TestInfoExtractor.test_parse_m3u8_formats)
1403 # 1. http://www.vidio.com/watch/165683-dj_ambred-booyah-live-2015
1404 stream_name = last_stream_inf.get('NAME')
1407 # If there is no NAME in EXT-X-STREAM-INF it will be obtained
1408 # from corresponding rendition group
1409 stream_group_id = last_stream_inf.get('VIDEO')
1410 if not stream_group_id:
1412 stream_group = groups.get(stream_group_id)
1413 if not stream_group:
1414 return stream_group_id
1415 rendition = stream_group[0]
1416 return rendition.get('NAME') or stream_group_id
1418 for line in m3u8_doc.splitlines():
1419 if line.startswith('#EXT-X-STREAM-INF:'):
1420 last_stream_inf = parse_m3u8_attributes(line)
1421 elif line.startswith('#EXT-X-MEDIA:'):
1423 elif line.startswith('#') or not line.strip():
1426 tbr = float_or_none(
1427 last_stream_inf.get('AVERAGE-BANDWIDTH') or
1428 last_stream_inf.get('BANDWIDTH'), scale=1000)
1431 format_id.append(m3u8_id)
1432 stream_name = build_stream_name()
1433 # Bandwidth of live streams may differ over time thus making
1434 # format_id unpredictable. So it's better to keep provided
1437 format_id.append(stream_name if stream_name else '%d' % (tbr if tbr else len(formats)))
1438 manifest_url = format_url(line.strip())
1440 'format_id': '-'.join(format_id),
1441 'url': manifest_url,
1442 'manifest_url': m3u8_url,
1445 'fps': float_or_none(last_stream_inf.get('FRAME-RATE')),
1446 'protocol': entry_protocol,
1447 'preference': preference,
1449 resolution = last_stream_inf.get('RESOLUTION')
1451 mobj = re.search(r'(?P<width>\d+)[xX](?P<height>\d+)', resolution)
1453 f['width'] = int(mobj.group('width'))
1454 f['height'] = int(mobj.group('height'))
1455 # Unified Streaming Platform
1457 r'audio.*?(?:%3D|=)(\d+)(?:-video.*?(?:%3D|=)(\d+))?', f['url'])
1459 abr, vbr = mobj.groups()
1460 abr, vbr = float_or_none(abr, 1000), float_or_none(vbr, 1000)
1465 codecs = parse_codecs(last_stream_inf.get('CODECS'))
1467 audio_group_id = last_stream_inf.get('AUDIO')
1468 # As per [1, 4.3.4.1.1] any EXT-X-STREAM-INF tag which
1469 # references a rendition group MUST have a CODECS attribute.
1470 # However, this is not always respected, for example, [2]
1471 # contains EXT-X-STREAM-INF tag which references AUDIO
1472 # rendition group but does not have CODECS and despite
1473 # referencing audio group an audio group, it represents
1474 # a complete (with audio and video) format. So, for such cases
1475 # we will ignore references to rendition groups and treat them
1476 # as complete formats.
1477 if audio_group_id and codecs and f.get('vcodec') != 'none':
1478 audio_group = groups.get(audio_group_id)
1479 if audio_group and audio_group[0].get('URI'):
1480 # TODO: update acodec for audio only formats with
1482 f['acodec'] = 'none'
1484 last_stream_inf = {}
1488 def _xpath_ns(path, namespace=None):
1492 for c in path.split('/'):
1493 if not c or c == '.':
1496 out.append('{%s}%s' % (namespace, c))
1497 return '/'.join(out)
1499 def _extract_smil_formats(self, smil_url, video_id, fatal=True, f4m_params=None, transform_source=None):
1500 smil = self._download_smil(smil_url, video_id, fatal=fatal, transform_source=transform_source)
1506 namespace = self._parse_smil_namespace(smil)
1508 return self._parse_smil_formats(
1509 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
1511 def _extract_smil_info(self, smil_url, video_id, fatal=True, f4m_params=None):
1512 smil = self._download_smil(smil_url, video_id, fatal=fatal)
1515 return self._parse_smil(smil, smil_url, video_id, f4m_params=f4m_params)
1517 def _download_smil(self, smil_url, video_id, fatal=True, transform_source=None):
1518 return self._download_xml(
1519 smil_url, video_id, 'Downloading SMIL file',
1520 'Unable to download SMIL file', fatal=fatal, transform_source=transform_source)
1522 def _parse_smil(self, smil, smil_url, video_id, f4m_params=None):
1523 namespace = self._parse_smil_namespace(smil)
1525 formats = self._parse_smil_formats(
1526 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
1527 subtitles = self._parse_smil_subtitles(smil, namespace=namespace)
1529 video_id = os.path.splitext(url_basename(smil_url))[0]
1533 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
1534 name = meta.attrib.get('name')
1535 content = meta.attrib.get('content')
1536 if not name or not content:
1538 if not title and name == 'title':
1540 elif not description and name in ('description', 'abstract'):
1541 description = content
1542 elif not upload_date and name == 'date':
1543 upload_date = unified_strdate(content)
1546 'id': image.get('type'),
1547 'url': image.get('src'),
1548 'width': int_or_none(image.get('width')),
1549 'height': int_or_none(image.get('height')),
1550 } for image in smil.findall(self._xpath_ns('.//image', namespace)) if image.get('src')]
1554 'title': title or video_id,
1555 'description': description,
1556 'upload_date': upload_date,
1557 'thumbnails': thumbnails,
1559 'subtitles': subtitles,
1562 def _parse_smil_namespace(self, smil):
1563 return self._search_regex(
1564 r'(?i)^{([^}]+)?}smil$', smil.tag, 'namespace', default=None)
1566 def _parse_smil_formats(self, smil, smil_url, video_id, namespace=None, f4m_params=None, transform_rtmp_url=None):
1568 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
1569 b = meta.get('base') or meta.get('httpBase')
1580 media = smil.findall(self._xpath_ns('.//video', namespace)) + smil.findall(self._xpath_ns('.//audio', namespace))
1581 for medium in media:
1582 src = medium.get('src')
1583 if not src or src in srcs:
1587 bitrate = float_or_none(medium.get('system-bitrate') or medium.get('systemBitrate'), 1000)
1588 filesize = int_or_none(medium.get('size') or medium.get('fileSize'))
1589 width = int_or_none(medium.get('width'))
1590 height = int_or_none(medium.get('height'))
1591 proto = medium.get('proto')
1592 ext = medium.get('ext')
1593 src_ext = determine_ext(src)
1594 streamer = medium.get('streamer') or base
1596 if proto == 'rtmp' or streamer.startswith('rtmp'):
1602 'format_id': 'rtmp-%d' % (rtmp_count if bitrate is None else bitrate),
1604 'filesize': filesize,
1608 if transform_rtmp_url:
1609 streamer, src = transform_rtmp_url(streamer, src)
1610 formats[-1].update({
1616 src_url = src if src.startswith('http') else compat_urlparse.urljoin(base, src)
1617 src_url = src_url.strip()
1619 if proto == 'm3u8' or src_ext == 'm3u8':
1620 m3u8_formats = self._extract_m3u8_formats(
1621 src_url, video_id, ext or 'mp4', m3u8_id='hls', fatal=False)
1622 if len(m3u8_formats) == 1:
1624 m3u8_formats[0].update({
1625 'format_id': 'hls-%d' % (m3u8_count if bitrate is None else bitrate),
1630 formats.extend(m3u8_formats)
1633 if src_ext == 'f4m':
1638 'plugin': 'flowplayer-3.2.0.1',
1640 f4m_url += '&' if '?' in f4m_url else '?'
1641 f4m_url += compat_urllib_parse_urlencode(f4m_params)
1642 formats.extend(self._extract_f4m_formats(f4m_url, video_id, f4m_id='hds', fatal=False))
1645 if src_url.startswith('http') and self._is_valid_url(src, video_id):
1649 'ext': ext or src_ext or 'flv',
1650 'format_id': 'http-%d' % (bitrate or http_count),
1652 'filesize': filesize,
1660 def _parse_smil_subtitles(self, smil, namespace=None, subtitles_lang='en'):
1663 for num, textstream in enumerate(smil.findall(self._xpath_ns('.//textstream', namespace))):
1664 src = textstream.get('src')
1665 if not src or src in urls:
1668 ext = textstream.get('ext') or mimetype2ext(textstream.get('type')) or determine_ext(src)
1669 lang = textstream.get('systemLanguage') or textstream.get('systemLanguageName') or textstream.get('lang') or subtitles_lang
1670 subtitles.setdefault(lang, []).append({
1676 def _extract_xspf_playlist(self, playlist_url, playlist_id, fatal=True):
1677 xspf = self._download_xml(
1678 playlist_url, playlist_id, 'Downloading xpsf playlist',
1679 'Unable to download xspf manifest', fatal=fatal)
1682 return self._parse_xspf(xspf, playlist_id)
1684 def _parse_xspf(self, playlist, playlist_id):
1686 'xspf': 'http://xspf.org/ns/0/',
1687 's1': 'http://static.streamone.nl/player/ns/0',
1691 for track in playlist.findall(xpath_with_ns('./xspf:trackList/xspf:track', NS_MAP)):
1693 track, xpath_with_ns('./xspf:title', NS_MAP), 'title', default=playlist_id)
1694 description = xpath_text(
1695 track, xpath_with_ns('./xspf:annotation', NS_MAP), 'description')
1696 thumbnail = xpath_text(
1697 track, xpath_with_ns('./xspf:image', NS_MAP), 'thumbnail')
1698 duration = float_or_none(
1699 xpath_text(track, xpath_with_ns('./xspf:duration', NS_MAP), 'duration'), 1000)
1702 'url': location.text,
1703 'format_id': location.get(xpath_with_ns('s1:label', NS_MAP)),
1704 'width': int_or_none(location.get(xpath_with_ns('s1:width', NS_MAP))),
1705 'height': int_or_none(location.get(xpath_with_ns('s1:height', NS_MAP))),
1706 } for location in track.findall(xpath_with_ns('./xspf:location', NS_MAP))]
1707 self._sort_formats(formats)
1712 'description': description,
1713 'thumbnail': thumbnail,
1714 'duration': duration,
1719 def _extract_mpd_formats(self, mpd_url, video_id, mpd_id=None, note=None, errnote=None, fatal=True, formats_dict={}):
1720 res = self._download_webpage_handle(
1722 note=note or 'Downloading MPD manifest',
1723 errnote=errnote or 'Failed to download MPD manifest',
1728 mpd_base_url = base_url(urlh.geturl())
1730 return self._parse_mpd_formats(
1731 compat_etree_fromstring(mpd.encode('utf-8')), mpd_id, mpd_base_url,
1732 formats_dict=formats_dict, mpd_url=mpd_url)
1734 def _parse_mpd_formats(self, mpd_doc, mpd_id=None, mpd_base_url='', formats_dict={}, mpd_url=None):
1736 Parse formats from MPD manifest.
1738 1. MPEG-DASH Standard, ISO/IEC 23009-1:2014(E),
1739 http://standards.iso.org/ittf/PubliclyAvailableStandards/c065274_ISO_IEC_23009-1_2014.zip
1740 2. https://en.wikipedia.org/wiki/Dynamic_Adaptive_Streaming_over_HTTP
1742 if mpd_doc.get('type') == 'dynamic':
1745 namespace = self._search_regex(r'(?i)^{([^}]+)?}MPD$', mpd_doc.tag, 'namespace', default=None)
1748 return self._xpath_ns(path, namespace)
1750 def is_drm_protected(element):
1751 return element.find(_add_ns('ContentProtection')) is not None
1753 def extract_multisegment_info(element, ms_parent_info):
1754 ms_info = ms_parent_info.copy()
1756 # As per [1, 5.3.9.2.2] SegmentList and SegmentTemplate share some
1757 # common attributes and elements. We will only extract relevant
1759 def extract_common(source):
1760 segment_timeline = source.find(_add_ns('SegmentTimeline'))
1761 if segment_timeline is not None:
1762 s_e = segment_timeline.findall(_add_ns('S'))
1764 ms_info['total_number'] = 0
1767 r = int(s.get('r', 0))
1768 ms_info['total_number'] += 1 + r
1769 ms_info['s'].append({
1770 't': int(s.get('t', 0)),
1771 # @d is mandatory (see [1, 5.3.9.6.2, Table 17, page 60])
1772 'd': int(s.attrib['d']),
1775 start_number = source.get('startNumber')
1777 ms_info['start_number'] = int(start_number)
1778 timescale = source.get('timescale')
1780 ms_info['timescale'] = int(timescale)
1781 segment_duration = source.get('duration')
1782 if segment_duration:
1783 ms_info['segment_duration'] = int(segment_duration)
1785 def extract_Initialization(source):
1786 initialization = source.find(_add_ns('Initialization'))
1787 if initialization is not None:
1788 ms_info['initialization_url'] = initialization.attrib['sourceURL']
1790 segment_list = element.find(_add_ns('SegmentList'))
1791 if segment_list is not None:
1792 extract_common(segment_list)
1793 extract_Initialization(segment_list)
1794 segment_urls_e = segment_list.findall(_add_ns('SegmentURL'))
1796 ms_info['segment_urls'] = [segment.attrib['media'] for segment in segment_urls_e]
1798 segment_template = element.find(_add_ns('SegmentTemplate'))
1799 if segment_template is not None:
1800 extract_common(segment_template)
1801 media = segment_template.get('media')
1803 ms_info['media'] = media
1804 initialization = segment_template.get('initialization')
1806 ms_info['initialization'] = initialization
1808 extract_Initialization(segment_template)
1811 mpd_duration = parse_duration(mpd_doc.get('mediaPresentationDuration'))
1813 for period in mpd_doc.findall(_add_ns('Period')):
1814 period_duration = parse_duration(period.get('duration')) or mpd_duration
1815 period_ms_info = extract_multisegment_info(period, {
1819 for adaptation_set in period.findall(_add_ns('AdaptationSet')):
1820 if is_drm_protected(adaptation_set):
1822 adaption_set_ms_info = extract_multisegment_info(adaptation_set, period_ms_info)
1823 for representation in adaptation_set.findall(_add_ns('Representation')):
1824 if is_drm_protected(representation):
1826 representation_attrib = adaptation_set.attrib.copy()
1827 representation_attrib.update(representation.attrib)
1828 # According to [1, 5.3.7.2, Table 9, page 41], @mimeType is mandatory
1829 mime_type = representation_attrib['mimeType']
1830 content_type = mime_type.split('/')[0]
1831 if content_type == 'text':
1832 # TODO implement WebVTT downloading
1834 elif content_type in ('video', 'audio'):
1836 for element in (representation, adaptation_set, period, mpd_doc):
1837 base_url_e = element.find(_add_ns('BaseURL'))
1838 if base_url_e is not None:
1839 base_url = base_url_e.text + base_url
1840 if re.match(r'^https?://', base_url):
1842 if mpd_base_url and not re.match(r'^https?://', base_url):
1843 if not mpd_base_url.endswith('/') and not base_url.startswith('/'):
1845 base_url = mpd_base_url + base_url
1846 representation_id = representation_attrib.get('id')
1847 lang = representation_attrib.get('lang')
1848 url_el = representation.find(_add_ns('BaseURL'))
1849 filesize = int_or_none(url_el.attrib.get('{http://youtube.com/yt/2012/10/10}contentLength') if url_el is not None else None)
1850 bandwidth = int_or_none(representation_attrib.get('bandwidth'))
1852 'format_id': '%s-%s' % (mpd_id, representation_id) if mpd_id else representation_id,
1854 'manifest_url': mpd_url,
1855 'ext': mimetype2ext(mime_type),
1856 'width': int_or_none(representation_attrib.get('width')),
1857 'height': int_or_none(representation_attrib.get('height')),
1858 'tbr': float_or_none(bandwidth, 1000),
1859 'asr': int_or_none(representation_attrib.get('audioSamplingRate')),
1860 'fps': int_or_none(representation_attrib.get('frameRate')),
1861 'language': lang if lang not in ('mul', 'und', 'zxx', 'mis') else None,
1862 'format_note': 'DASH %s' % content_type,
1863 'filesize': filesize,
1865 f.update(parse_codecs(representation_attrib.get('codecs')))
1866 representation_ms_info = extract_multisegment_info(representation, adaption_set_ms_info)
1868 def prepare_template(template_name, identifiers):
1869 t = representation_ms_info[template_name]
1870 t = t.replace('$RepresentationID$', representation_id)
1871 t = re.sub(r'\$(%s)\$' % '|'.join(identifiers), r'%(\1)d', t)
1872 t = re.sub(r'\$(%s)%%([^$]+)\$' % '|'.join(identifiers), r'%(\1)\2', t)
1873 t.replace('$$', '$')
1876 # @initialization is a regular template like @media one
1877 # so it should be handled just the same way (see
1878 # https://github.com/rg3/youtube-dl/issues/11605)
1879 if 'initialization' in representation_ms_info:
1880 initialization_template = prepare_template(
1882 # As per [1, 5.3.9.4.2, Table 15, page 54] $Number$ and
1883 # $Time$ shall not be included for @initialization thus
1884 # only $Bandwidth$ remains
1886 representation_ms_info['initialization_url'] = initialization_template % {
1887 'Bandwidth': bandwidth,
1890 if 'segment_urls' not in representation_ms_info and 'media' in representation_ms_info:
1892 media_template = prepare_template('media', ('Number', 'Bandwidth', 'Time'))
1894 # As per [1, 5.3.9.4.4, Table 16, page 55] $Number$ and $Time$
1895 # can't be used at the same time
1896 if '%(Number' in media_template and 's' not in representation_ms_info:
1897 segment_duration = None
1898 if 'total_number' not in representation_ms_info and 'segment_duration':
1899 segment_duration = float_or_none(representation_ms_info['segment_duration'], representation_ms_info['timescale'])
1900 representation_ms_info['total_number'] = int(math.ceil(float(period_duration) / segment_duration))
1901 representation_ms_info['fragments'] = [{
1902 'url': media_template % {
1903 'Number': segment_number,
1904 'Bandwidth': bandwidth,
1906 'duration': segment_duration,
1907 } for segment_number in range(
1908 representation_ms_info['start_number'],
1909 representation_ms_info['total_number'] + representation_ms_info['start_number'])]
1911 # $Number*$ or $Time$ in media template with S list available
1912 # Example $Number*$: http://www.svtplay.se/klipp/9023742/stopptid-om-bjorn-borg
1913 # Example $Time$: https://play.arkena.com/embed/avp/v2/player/media/b41dda37-d8e7-4d3f-b1b5-9a9db578bdfe/1/129411
1914 representation_ms_info['fragments'] = []
1917 segment_number = representation_ms_info['start_number']
1919 def add_segment_url():
1920 segment_url = media_template % {
1921 'Time': segment_time,
1922 'Bandwidth': bandwidth,
1923 'Number': segment_number,
1925 representation_ms_info['fragments'].append({
1927 'duration': float_or_none(segment_d, representation_ms_info['timescale']),
1930 for num, s in enumerate(representation_ms_info['s']):
1931 segment_time = s.get('t') or segment_time
1935 for r in range(s.get('r', 0)):
1936 segment_time += segment_d
1939 segment_time += segment_d
1940 elif 'segment_urls' in representation_ms_info and 's' in representation_ms_info:
1942 # Example: https://www.youtube.com/watch?v=iXZV5uAYMJI
1943 # or any YouTube dashsegments video
1946 timescale = representation_ms_info['timescale']
1947 for s in representation_ms_info['s']:
1948 duration = float_or_none(s['d'], timescale)
1949 for r in range(s.get('r', 0) + 1):
1951 'url': representation_ms_info['segment_urls'][segment_index],
1952 'duration': duration,
1955 representation_ms_info['fragments'] = fragments
1956 # NB: MPD manifest may contain direct URLs to unfragmented media.
1957 # No fragments key is present in this case.
1958 if 'fragments' in representation_ms_info:
1961 'protocol': 'http_dash_segments',
1963 if 'initialization_url' in representation_ms_info:
1964 initialization_url = representation_ms_info['initialization_url']
1965 if not f.get('url'):
1966 f['url'] = initialization_url
1967 f['fragments'].append({'url': initialization_url})
1968 f['fragments'].extend(representation_ms_info['fragments'])
1969 for fragment in f['fragments']:
1970 fragment['url'] = urljoin(base_url, fragment['url'])
1972 existing_format = next(
1973 fo for fo in formats
1974 if fo['format_id'] == representation_id)
1975 except StopIteration:
1976 full_info = formats_dict.get(representation_id, {}).copy()
1978 formats.append(full_info)
1980 existing_format.update(f)
1982 self.report_warning('Unknown MIME type %s in DASH manifest' % mime_type)
1985 def _extract_ism_formats(self, ism_url, video_id, ism_id=None, note=None, errnote=None, fatal=True):
1986 res = self._download_webpage_handle(
1988 note=note or 'Downloading ISM manifest',
1989 errnote=errnote or 'Failed to download ISM manifest',
1995 return self._parse_ism_formats(
1996 compat_etree_fromstring(ism.encode('utf-8')), urlh.geturl(), ism_id)
1998 def _parse_ism_formats(self, ism_doc, ism_url, ism_id=None):
1999 if ism_doc.get('IsLive') == 'TRUE' or ism_doc.find('Protection') is not None:
2002 duration = int(ism_doc.attrib['Duration'])
2003 timescale = int_or_none(ism_doc.get('TimeScale')) or 10000000
2006 for stream in ism_doc.findall('StreamIndex'):
2007 stream_type = stream.get('Type')
2008 if stream_type not in ('video', 'audio'):
2010 url_pattern = stream.attrib['Url']
2011 stream_timescale = int_or_none(stream.get('TimeScale')) or timescale
2012 stream_name = stream.get('Name')
2013 for track in stream.findall('QualityLevel'):
2014 fourcc = track.get('FourCC')
2015 # TODO: add support for WVC1 and WMAP
2016 if fourcc not in ('H264', 'AVC1', 'AACL'):
2017 self.report_warning('%s is not a supported codec' % fourcc)
2019 tbr = int(track.attrib['Bitrate']) // 1000
2020 width = int_or_none(track.get('MaxWidth'))
2021 height = int_or_none(track.get('MaxHeight'))
2022 sampling_rate = int_or_none(track.get('SamplingRate'))
2024 track_url_pattern = re.sub(r'{[Bb]itrate}', track.attrib['Bitrate'], url_pattern)
2025 track_url_pattern = compat_urlparse.urljoin(ism_url, track_url_pattern)
2031 stream_fragments = stream.findall('c')
2032 for stream_fragment_index, stream_fragment in enumerate(stream_fragments):
2033 fragment_ctx['time'] = int_or_none(stream_fragment.get('t')) or fragment_ctx['time']
2034 fragment_repeat = int_or_none(stream_fragment.get('r')) or 1
2035 fragment_ctx['duration'] = int_or_none(stream_fragment.get('d'))
2036 if not fragment_ctx['duration']:
2038 next_fragment_time = int(stream_fragment[stream_fragment_index + 1].attrib['t'])
2040 next_fragment_time = duration
2041 fragment_ctx['duration'] = (next_fragment_time - fragment_ctx['time']) / fragment_repeat
2042 for _ in range(fragment_repeat):
2044 'url': re.sub(r'{start[ _]time}', compat_str(fragment_ctx['time']), track_url_pattern),
2045 'duration': fragment_ctx['duration'] / stream_timescale,
2047 fragment_ctx['time'] += fragment_ctx['duration']
2051 format_id.append(ism_id)
2053 format_id.append(stream_name)
2054 format_id.append(compat_str(tbr))
2057 'format_id': '-'.join(format_id),
2059 'manifest_url': ism_url,
2060 'ext': 'ismv' if stream_type == 'video' else 'isma',
2064 'asr': sampling_rate,
2065 'vcodec': 'none' if stream_type == 'audio' else fourcc,
2066 'acodec': 'none' if stream_type == 'video' else fourcc,
2068 'fragments': fragments,
2069 '_download_params': {
2070 'duration': duration,
2071 'timescale': stream_timescale,
2072 'width': width or 0,
2073 'height': height or 0,
2075 'codec_private_data': track.get('CodecPrivateData'),
2076 'sampling_rate': sampling_rate,
2077 'channels': int_or_none(track.get('Channels', 2)),
2078 'bits_per_sample': int_or_none(track.get('BitsPerSample', 16)),
2079 'nal_unit_length_field': int_or_none(track.get('NALUnitLengthField', 4)),
2084 def _parse_html5_media_entries(self, base_url, webpage, video_id, m3u8_id=None, m3u8_entry_protocol='m3u8', mpd_id=None, preference=None):
2085 def absolute_url(video_url):
2086 return compat_urlparse.urljoin(base_url, video_url)
2088 def parse_content_type(content_type):
2089 if not content_type:
2091 ctr = re.search(r'(?P<mimetype>[^/]+/[^;]+)(?:;\s*codecs="?(?P<codecs>[^"]+))?', content_type)
2093 mimetype, codecs = ctr.groups()
2094 f = parse_codecs(codecs)
2095 f['ext'] = mimetype2ext(mimetype)
2099 def _media_formats(src, cur_media_type):
2100 full_url = absolute_url(src)
2101 ext = determine_ext(full_url)
2103 is_plain_url = False
2104 formats = self._extract_m3u8_formats(
2105 full_url, video_id, ext='mp4',
2106 entry_protocol=m3u8_entry_protocol, m3u8_id=m3u8_id,
2107 preference=preference)
2109 is_plain_url = False
2110 formats = self._extract_mpd_formats(
2111 full_url, video_id, mpd_id=mpd_id)
2116 'vcodec': 'none' if cur_media_type == 'audio' else None,
2118 return is_plain_url, formats
2121 media_tags = [(media_tag, media_type, '')
2122 for media_tag, media_type
2123 in re.findall(r'(?s)(<(video|audio)[^>]*/>)', webpage)]
2124 media_tags.extend(re.findall(
2125 # We only allow video|audio followed by a whitespace or '>'.
2126 # Allowing more characters may end up in significant slow down (see
2127 # https://github.com/rg3/youtube-dl/issues/11979, example URL:
2128 # http://www.porntrex.com/maps/videositemap.xml).
2129 r'(?s)(<(?P<tag>video|audio)(?:\s+[^>]*)?>)(.*?)</(?P=tag)>', webpage))
2130 for media_tag, media_type, media_content in media_tags:
2135 media_attributes = extract_attributes(media_tag)
2136 src = media_attributes.get('src')
2138 _, formats = _media_formats(src, media_type)
2139 media_info['formats'].extend(formats)
2140 media_info['thumbnail'] = media_attributes.get('poster')
2142 for source_tag in re.findall(r'<source[^>]+>', media_content):
2143 source_attributes = extract_attributes(source_tag)
2144 src = source_attributes.get('src')
2147 is_plain_url, formats = _media_formats(src, media_type)
2149 f = parse_content_type(source_attributes.get('type'))
2150 f.update(formats[0])
2151 media_info['formats'].append(f)
2153 media_info['formats'].extend(formats)
2154 for track_tag in re.findall(r'<track[^>]+>', media_content):
2155 track_attributes = extract_attributes(track_tag)
2156 kind = track_attributes.get('kind')
2157 if not kind or kind in ('subtitles', 'captions'):
2158 src = track_attributes.get('src')
2161 lang = track_attributes.get('srclang') or track_attributes.get('lang') or track_attributes.get('label')
2162 media_info['subtitles'].setdefault(lang, []).append({
2163 'url': absolute_url(src),
2165 if media_info['formats'] or media_info['subtitles']:
2166 entries.append(media_info)
2169 def _extract_akamai_formats(self, manifest_url, video_id, hosts={}):
2171 hdcore_sign = 'hdcore=3.7.0'
2172 f4m_url = re.sub(r'(https?://[^/+])/i/', r'\1/z/', manifest_url).replace('/master.m3u8', '/manifest.f4m')
2173 hds_host = hosts.get('hds')
2175 f4m_url = re.sub(r'(https?://)[^/]+', r'\1' + hds_host, f4m_url)
2176 if 'hdcore=' not in f4m_url:
2177 f4m_url += ('&' if '?' in f4m_url else '?') + hdcore_sign
2178 f4m_formats = self._extract_f4m_formats(
2179 f4m_url, video_id, f4m_id='hds', fatal=False)
2180 for entry in f4m_formats:
2181 entry.update({'extra_param_to_segment_url': hdcore_sign})
2182 formats.extend(f4m_formats)
2183 m3u8_url = re.sub(r'(https?://[^/]+)/z/', r'\1/i/', manifest_url).replace('/manifest.f4m', '/master.m3u8')
2184 hls_host = hosts.get('hls')
2186 m3u8_url = re.sub(r'(https?://)[^/]+', r'\1' + hls_host, m3u8_url)
2187 formats.extend(self._extract_m3u8_formats(
2188 m3u8_url, video_id, 'mp4', 'm3u8_native',
2189 m3u8_id='hls', fatal=False))
2192 def _extract_wowza_formats(self, url, video_id, m3u8_entry_protocol='m3u8_native', skip_protocols=[]):
2193 url = re.sub(r'/(?:manifest|playlist|jwplayer)\.(?:m3u8|f4m|mpd|smil)', '', url)
2194 url_base = self._search_regex(r'(?:https?|rtmp|rtsp)(://[^?]+)', url, 'format url')
2195 http_base_url = 'http' + url_base
2197 if 'm3u8' not in skip_protocols:
2198 formats.extend(self._extract_m3u8_formats(
2199 http_base_url + '/playlist.m3u8', video_id, 'mp4',
2200 m3u8_entry_protocol, m3u8_id='hls', fatal=False))
2201 if 'f4m' not in skip_protocols:
2202 formats.extend(self._extract_f4m_formats(
2203 http_base_url + '/manifest.f4m',
2204 video_id, f4m_id='hds', fatal=False))
2205 if 'dash' not in skip_protocols:
2206 formats.extend(self._extract_mpd_formats(
2207 http_base_url + '/manifest.mpd',
2208 video_id, mpd_id='dash', fatal=False))
2209 if re.search(r'(?:/smil:|\.smil)', url_base):
2210 if 'smil' not in skip_protocols:
2211 rtmp_formats = self._extract_smil_formats(
2212 http_base_url + '/jwplayer.smil',
2213 video_id, fatal=False)
2214 for rtmp_format in rtmp_formats:
2215 rtsp_format = rtmp_format.copy()
2216 rtsp_format['url'] = '%s/%s' % (rtmp_format['url'], rtmp_format['play_path'])
2217 del rtsp_format['play_path']
2218 del rtsp_format['ext']
2219 rtsp_format.update({
2220 'url': rtsp_format['url'].replace('rtmp://', 'rtsp://'),
2221 'format_id': rtmp_format['format_id'].replace('rtmp', 'rtsp'),
2224 formats.extend([rtmp_format, rtsp_format])
2226 for protocol in ('rtmp', 'rtsp'):
2227 if protocol not in skip_protocols:
2229 'url': protocol + url_base,
2230 'format_id': protocol,
2231 'protocol': protocol,
2235 def _find_jwplayer_data(self, webpage, video_id=None, transform_source=js_to_json):
2237 r'(?s)jwplayer\((?P<quote>[\'"])[^\'" ]+(?P=quote)\)(?!</script>).*?\.setup\s*\((?P<options>[^)]+)\)',
2241 jwplayer_data = self._parse_json(mobj.group('options'),
2243 transform_source=transform_source)
2244 except ExtractorError:
2247 if isinstance(jwplayer_data, dict):
2248 return jwplayer_data
2250 def _extract_jwplayer_data(self, webpage, video_id, *args, **kwargs):
2251 jwplayer_data = self._find_jwplayer_data(
2252 webpage, video_id, transform_source=js_to_json)
2253 return self._parse_jwplayer_data(
2254 jwplayer_data, video_id, *args, **kwargs)
2256 def _parse_jwplayer_data(self, jwplayer_data, video_id=None, require_title=True,
2257 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
2258 # JWPlayer backward compatibility: flattened playlists
2259 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/api/config.js#L81-L96
2260 if 'playlist' not in jwplayer_data:
2261 jwplayer_data = {'playlist': [jwplayer_data]}
2265 # JWPlayer backward compatibility: single playlist item
2266 # https://github.com/jwplayer/jwplayer/blob/v7.7.0/src/js/playlist/playlist.js#L10
2267 if not isinstance(jwplayer_data['playlist'], list):
2268 jwplayer_data['playlist'] = [jwplayer_data['playlist']]
2270 for video_data in jwplayer_data['playlist']:
2271 # JWPlayer backward compatibility: flattened sources
2272 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/playlist/item.js#L29-L35
2273 if 'sources' not in video_data:
2274 video_data['sources'] = [video_data]
2276 this_video_id = video_id or video_data['mediaid']
2278 formats = self._parse_jwplayer_formats(
2279 video_data['sources'], video_id=this_video_id, m3u8_id=m3u8_id,
2280 mpd_id=mpd_id, rtmp_params=rtmp_params, base_url=base_url)
2281 self._sort_formats(formats)
2284 tracks = video_data.get('tracks')
2285 if tracks and isinstance(tracks, list):
2286 for track in tracks:
2287 if track.get('kind') != 'captions':
2289 track_url = urljoin(base_url, track.get('file'))
2292 subtitles.setdefault(track.get('label') or 'en', []).append({
2293 'url': self._proto_relative_url(track_url)
2297 'id': this_video_id,
2298 'title': video_data['title'] if require_title else video_data.get('title'),
2299 'description': video_data.get('description'),
2300 'thumbnail': self._proto_relative_url(video_data.get('image')),
2301 'timestamp': int_or_none(video_data.get('pubdate')),
2302 'duration': float_or_none(jwplayer_data.get('duration') or video_data.get('duration')),
2303 'subtitles': subtitles,
2306 if len(entries) == 1:
2309 return self.playlist_result(entries)
2311 def _parse_jwplayer_formats(self, jwplayer_sources_data, video_id=None,
2312 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
2315 for source in jwplayer_sources_data:
2316 source_url = self._proto_relative_url(source.get('file'))
2320 source_url = compat_urlparse.urljoin(base_url, source_url)
2321 if source_url in urls:
2323 urls.append(source_url)
2324 source_type = source.get('type') or ''
2325 ext = mimetype2ext(source_type) or determine_ext(source_url)
2326 if source_type == 'hls' or ext == 'm3u8':
2327 formats.extend(self._extract_m3u8_formats(
2328 source_url, video_id, 'mp4', entry_protocol='m3u8_native',
2329 m3u8_id=m3u8_id, fatal=False))
2331 formats.extend(self._extract_mpd_formats(
2332 source_url, video_id, mpd_id=mpd_id, fatal=False))
2334 formats.extend(self._extract_smil_formats(
2335 source_url, video_id, fatal=False))
2336 # https://github.com/jwplayer/jwplayer/blob/master/src/js/providers/default.js#L67
2337 elif source_type.startswith('audio') or ext in (
2338 'oga', 'aac', 'mp3', 'mpeg', 'vorbis'):
2345 height = int_or_none(source.get('height'))
2347 # Often no height is provided but there is a label in
2348 # format like "1080p", "720p SD", or 1080.
2349 height = int_or_none(self._search_regex(
2350 r'^(\d{3,4})[pP]?(?:\b|$)', compat_str(source.get('label') or ''),
2351 'height', default=None))
2354 'width': int_or_none(source.get('width')),
2356 'tbr': int_or_none(source.get('bitrate')),
2359 if source_url.startswith('rtmp'):
2360 a_format['ext'] = 'flv'
2361 # See com/longtailvideo/jwplayer/media/RTMPMediaProvider.as
2362 # of jwplayer.flash.swf
2363 rtmp_url_parts = re.split(
2364 r'((?:mp4|mp3|flv):)', source_url, 1)
2365 if len(rtmp_url_parts) == 3:
2366 rtmp_url, prefix, play_path = rtmp_url_parts
2369 'play_path': prefix + play_path,
2372 a_format.update(rtmp_params)
2373 formats.append(a_format)
2376 def _live_title(self, name):
2377 """ Generate the title for a live video """
2378 now = datetime.datetime.now()
2379 now_str = now.strftime('%Y-%m-%d %H:%M')
2380 return name + ' ' + now_str
2382 def _int(self, v, name, fatal=False, **kwargs):
2383 res = int_or_none(v, **kwargs)
2384 if 'get_attr' in kwargs:
2385 print(getattr(v, kwargs['get_attr']))
2387 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
2389 raise ExtractorError(msg)
2391 self._downloader.report_warning(msg)
2394 def _float(self, v, name, fatal=False, **kwargs):
2395 res = float_or_none(v, **kwargs)
2397 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
2399 raise ExtractorError(msg)
2401 self._downloader.report_warning(msg)
2404 def _set_cookie(self, domain, name, value, expire_time=None):
2405 cookie = compat_cookiejar.Cookie(
2406 0, name, value, None, None, domain, None,
2407 None, '/', True, False, expire_time, '', None, None, None)
2408 self._downloader.cookiejar.set_cookie(cookie)
2410 def _get_cookies(self, url):
2411 """ Return a compat_cookies.SimpleCookie with the cookies for the url """
2412 req = sanitized_Request(url)
2413 self._downloader.cookiejar.add_cookie_header(req)
2414 return compat_cookies.SimpleCookie(req.get_header('Cookie'))
2416 def get_testcases(self, include_onlymatching=False):
2417 t = getattr(self, '_TEST', None)
2419 assert not hasattr(self, '_TESTS'), \
2420 '%s has _TEST and _TESTS' % type(self).__name__
2423 tests = getattr(self, '_TESTS', [])
2425 if not include_onlymatching and t.get('only_matching', False):
2427 t['name'] = type(self).__name__[:-len('IE')]
2430 def is_suitable(self, age_limit):
2431 """ Test whether the extractor is generally suitable for the given
2432 age limit (i.e. pornographic sites are not, all others usually are) """
2434 any_restricted = False
2435 for tc in self.get_testcases(include_onlymatching=False):
2436 if tc.get('playlist', []):
2437 tc = tc['playlist'][0]
2438 is_restricted = age_restricted(
2439 tc.get('info_dict', {}).get('age_limit'), age_limit)
2440 if not is_restricted:
2442 any_restricted = any_restricted or is_restricted
2443 return not any_restricted
2445 def extract_subtitles(self, *args, **kwargs):
2446 if (self._downloader.params.get('writesubtitles', False) or
2447 self._downloader.params.get('listsubtitles')):
2448 return self._get_subtitles(*args, **kwargs)
2451 def _get_subtitles(self, *args, **kwargs):
2452 raise NotImplementedError('This method must be implemented by subclasses')
2455 def _merge_subtitle_items(subtitle_list1, subtitle_list2):
2456 """ Merge subtitle items for one language. Items with duplicated URLs
2457 will be dropped. """
2458 list1_urls = set([item['url'] for item in subtitle_list1])
2459 ret = list(subtitle_list1)
2460 ret.extend([item for item in subtitle_list2 if item['url'] not in list1_urls])
2464 def _merge_subtitles(cls, subtitle_dict1, subtitle_dict2):
2465 """ Merge two subtitle dictionaries, language by language. """
2466 ret = dict(subtitle_dict1)
2467 for lang in subtitle_dict2:
2468 ret[lang] = cls._merge_subtitle_items(subtitle_dict1.get(lang, []), subtitle_dict2[lang])
2471 def extract_automatic_captions(self, *args, **kwargs):
2472 if (self._downloader.params.get('writeautomaticsub', False) or
2473 self._downloader.params.get('listsubtitles')):
2474 return self._get_automatic_captions(*args, **kwargs)
2477 def _get_automatic_captions(self, *args, **kwargs):
2478 raise NotImplementedError('This method must be implemented by subclasses')
2480 def mark_watched(self, *args, **kwargs):
2481 if (self._downloader.params.get('mark_watched', False) and
2482 (self._get_login_info()[0] is not None or
2483 self._downloader.params.get('cookiefile') is not None)):
2484 self._mark_watched(*args, **kwargs)
2486 def _mark_watched(self, *args, **kwargs):
2487 raise NotImplementedError('This method must be implemented by subclasses')
2489 def geo_verification_headers(self):
2491 geo_verification_proxy = self._downloader.params.get('geo_verification_proxy')
2492 if geo_verification_proxy:
2493 headers['Ytdl-request-proxy'] = geo_verification_proxy
2496 def _generic_id(self, url):
2497 return compat_urllib_parse_unquote(os.path.splitext(url.rstrip('/').split('/')[-1])[0])
2499 def _generic_title(self, url):
2500 return compat_urllib_parse_unquote(os.path.splitext(url_basename(url))[0])
2503 class SearchInfoExtractor(InfoExtractor):
2505 Base class for paged search queries extractors.
2506 They accept URLs in the format _SEARCH_KEY(|all|[0-9]):{query}
2507 Instances should define _SEARCH_KEY and _MAX_RESULTS.
2511 def _make_valid_url(cls):
2512 return r'%s(?P<prefix>|[1-9][0-9]*|all):(?P<query>[\s\S]+)' % cls._SEARCH_KEY
2515 def suitable(cls, url):
2516 return re.match(cls._make_valid_url(), url) is not None
2518 def _real_extract(self, query):
2519 mobj = re.match(self._make_valid_url(), query)
2521 raise ExtractorError('Invalid search query "%s"' % query)
2523 prefix = mobj.group('prefix')
2524 query = mobj.group('query')
2526 return self._get_n_results(query, 1)
2527 elif prefix == 'all':
2528 return self._get_n_results(query, self._MAX_RESULTS)
2532 raise ExtractorError('invalid download number %s for query "%s"' % (n, query))
2533 elif n > self._MAX_RESULTS:
2534 self._downloader.report_warning('%s returns max %i results (you requested %i)' % (self._SEARCH_KEY, self._MAX_RESULTS, n))
2535 n = self._MAX_RESULTS
2536 return self._get_n_results(query, n)
2538 def _get_n_results(self, query, n):
2539 """Get a specified number of results for a query"""
2540 raise NotImplementedError('This method must be implemented by subclasses')
2543 def SEARCH_KEY(self):
2544 return self._SEARCH_KEY