2 from __future__ import unicode_literals
18 from ..compat import (
19 compat_cookiejar_Cookie,
22 compat_etree_fromstring,
29 compat_urllib_parse_unquote,
30 compat_urllib_parse_urlencode,
31 compat_urllib_request,
33 compat_xml_parse_error,
35 from ..downloader.f4m import (
37 remove_encrypted_media,
65 parse_m3u8_attributes,
87 class InfoExtractor(object):
88 """Information Extractor class.
90 Information extractors are the classes that, given a URL, extract
91 information about the video (or videos) the URL refers to. This
92 information includes the real video URL, the video title, author and
93 others. The information is stored in a dictionary which is then
94 passed to the YoutubeDL. The YoutubeDL processes this
95 information possibly downloading the video to the file system, among
96 other possible outcomes.
98 The type field determines the type of the result.
99 By far the most common value (and the default if _type is missing) is
100 "video", which indicates a single video.
102 For a video, the dictionaries must include the following fields:
104 id: Video identifier.
105 title: Video title, unescaped.
107 Additionally, it must contain either a formats entry or a url one:
109 formats: A list of dictionaries for each format available, ordered
110 from worst to best quality.
113 * url The mandatory URL representing the media:
114 for plain file media - HTTP URL of this file,
116 for HLS - URL of the M3U8 media playlist,
117 for HDS - URL of the F4M manifest,
119 - HTTP URL to plain file media (in case of
121 - URL of the MPD manifest or base URL
122 representing the media if MPD manifest
123 is parsed from a string (in case of
125 for MSS - URL of the ISM manifest.
127 The URL of the manifest file in case of
129 for HLS - URL of the M3U8 master playlist,
130 for HDS - URL of the F4M manifest,
131 for DASH - URL of the MPD manifest,
132 for MSS - URL of the ISM manifest.
133 * ext Will be calculated from URL if missing
134 * format A human-readable description of the format
135 ("mp4 container with h264/opus").
136 Calculated from the format_id, width, height.
137 and format_note fields if missing.
138 * format_id A short description of the format
139 ("mp4_h264_opus" or "19").
140 Technically optional, but strongly recommended.
141 * format_note Additional info about the format
142 ("3D" or "DASH video")
143 * width Width of the video, if known
144 * height Height of the video, if known
145 * resolution Textual description of width and height
146 * tbr Average bitrate of audio and video in KBit/s
147 * abr Average audio bitrate in KBit/s
148 * acodec Name of the audio codec in use
149 * asr Audio sampling rate in Hertz
150 * vbr Average video bitrate in KBit/s
152 * vcodec Name of the video codec in use
153 * container Name of the container format
154 * filesize The number of bytes, if known in advance
155 * filesize_approx An estimate for the number of bytes
156 * player_url SWF Player URL (used for rtmpdump).
157 * protocol The protocol that will be used for the actual
158 download, lower-case.
159 "http", "https", "rtsp", "rtmp", "rtmpe",
160 "m3u8", "m3u8_native" or "http_dash_segments".
162 Base URL for fragments. Each fragment's path
163 value (if present) will be relative to
165 * fragments A list of fragments of a fragmented media.
166 Each fragment entry must contain either an url
167 or a path. If an url is present it should be
168 considered by a client. Otherwise both path and
169 fragment_base_url must be present. Here is
170 the list of all potential fields:
171 * "url" - fragment's URL
172 * "path" - fragment's path relative to
174 * "duration" (optional, int or float)
175 * "filesize" (optional, int)
176 * preference Order number of this format. If this field is
177 present and not None, the formats get sorted
178 by this field, regardless of all other values.
179 -1 for default (order by other properties),
180 -2 or smaller for less than default.
181 < -1000 to hide the format (if there is
182 another one which is strictly better)
183 * language Language code, e.g. "de" or "en-US".
184 * language_preference Is this in the language mentioned in
186 10 if it's what the URL is about,
187 -1 for default (don't know),
188 -10 otherwise, other values reserved for now.
189 * quality Order number of the video quality of this
190 format, irrespective of the file format.
191 -1 for default (order by other properties),
192 -2 or smaller for less than default.
193 * source_preference Order number for this video source
194 (quality takes higher priority)
195 -1 for default (order by other properties),
196 -2 or smaller for less than default.
197 * http_headers A dictionary of additional HTTP headers
198 to add to the request.
199 * stretched_ratio If given and not 1, indicates that the
200 video's pixels are not square.
201 width : height ratio as float.
202 * no_resume The server does not support resuming the
203 (HTTP or RTMP) download. Boolean.
204 * downloader_options A dictionary of downloader options as
205 described in FileDownloader
207 url: Final video URL.
208 ext: Video filename extension.
209 format: The video format, defaults to ext (used for --get-format)
210 player_url: SWF Player URL (used for rtmpdump).
212 The following fields are optional:
214 alt_title: A secondary title of the video.
215 display_id An alternative identifier for the video, not necessarily
216 unique, but available before title. Typically, id is
217 something like "4234987", title "Dancing naked mole rats",
218 and display_id "dancing-naked-mole-rats"
219 thumbnails: A list of dictionaries, with the following entries:
220 * "id" (optional, string) - Thumbnail format ID
222 * "preference" (optional, int) - quality of the image
223 * "width" (optional, int)
224 * "height" (optional, int)
225 * "resolution" (optional, string "{width}x{height}",
227 * "filesize" (optional, int)
228 thumbnail: Full URL to a video thumbnail image.
229 description: Full video description.
230 uploader: Full name of the video uploader.
231 license: License name the video is licensed under.
232 creator: The creator of the video.
233 release_date: The date (YYYYMMDD) when the video was released.
234 timestamp: UNIX timestamp of the moment the video became available.
235 upload_date: Video upload date (YYYYMMDD).
236 If not explicitly set, calculated from timestamp.
237 uploader_id: Nickname or id of the video uploader.
238 uploader_url: Full URL to a personal webpage of the video uploader.
239 channel: Full name of the channel the video is uploaded on.
240 Note that channel fields may or may not repeat uploader
241 fields. This depends on a particular extractor.
242 channel_id: Id of the channel.
243 channel_url: Full URL to a channel webpage.
244 location: Physical location where the video was filmed.
245 subtitles: The available subtitles as a dictionary in the format
246 {tag: subformats}. "tag" is usually a language code, and
247 "subformats" is a list sorted from lower to higher
248 preference, each element is a dictionary with the "ext"
250 * "data": The subtitles file contents
251 * "url": A URL pointing to the subtitles file
252 "ext" will be calculated from URL if missing
253 automatic_captions: Like 'subtitles', used by the YoutubeIE for
254 automatically generated captions
255 duration: Length of the video in seconds, as an integer or float.
256 view_count: How many users have watched the video on the platform.
257 like_count: Number of positive ratings of the video
258 dislike_count: Number of negative ratings of the video
259 repost_count: Number of reposts of the video
260 average_rating: Average rating give by users, the scale used depends on the webpage
261 comment_count: Number of comments on the video
262 comments: A list of comments, each with one or more of the following
263 properties (all but one of text or html optional):
264 * "author" - human-readable name of the comment author
265 * "author_id" - user ID of the comment author
267 * "html" - Comment as HTML
268 * "text" - Plain text of the comment
269 * "timestamp" - UNIX timestamp of comment
270 * "parent" - ID of the comment this one is replying to.
271 Set to "root" to indicate that this is a
272 comment to the original video.
273 age_limit: Age restriction for the video, as an integer (years)
274 webpage_url: The URL to the video webpage, if given to youtube-dl it
275 should allow to get the same result again. (It will be set
276 by YoutubeDL if it's missing)
277 categories: A list of categories that the video falls in, for example
279 tags: A list of tags assigned to the video, e.g. ["sweden", "pop music"]
280 is_live: True, False, or None (=unknown). Whether this video is a
281 live stream that goes on instead of a fixed-length video.
282 start_time: Time in seconds where the reproduction should start, as
283 specified in the URL.
284 end_time: Time in seconds where the reproduction should end, as
285 specified in the URL.
286 chapters: A list of dictionaries, with the following entries:
287 * "start_time" - The start time of the chapter in seconds
288 * "end_time" - The end time of the chapter in seconds
289 * "title" (optional, string)
291 The following fields should only be used when the video belongs to some logical
294 chapter: Name or title of the chapter the video belongs to.
295 chapter_number: Number of the chapter the video belongs to, as an integer.
296 chapter_id: Id of the chapter the video belongs to, as a unicode string.
298 The following fields should only be used when the video is an episode of some
299 series, programme or podcast:
301 series: Title of the series or programme the video episode belongs to.
302 season: Title of the season the video episode belongs to.
303 season_number: Number of the season the video episode belongs to, as an integer.
304 season_id: Id of the season the video episode belongs to, as a unicode string.
305 episode: Title of the video episode. Unlike mandatory video title field,
306 this field should denote the exact title of the video episode
307 without any kind of decoration.
308 episode_number: Number of the video episode within a season, as an integer.
309 episode_id: Id of the video episode, as a unicode string.
311 The following fields should only be used when the media is a track or a part of
314 track: Title of the track.
315 track_number: Number of the track within an album or a disc, as an integer.
316 track_id: Id of the track (useful in case of custom indexing, e.g. 6.iii),
318 artist: Artist(s) of the track.
319 genre: Genre(s) of the track.
320 album: Title of the album the track belongs to.
321 album_type: Type of the album (e.g. "Demo", "Full-length", "Split", "Compilation", etc).
322 album_artist: List of all artists appeared on the album (e.g.
323 "Ash Borer / Fell Voices" or "Various Artists", useful for splits
325 disc_number: Number of the disc or other physical medium the track belongs to,
327 release_year: Year (YYYY) when the album was released.
329 Unless mentioned otherwise, the fields should be Unicode strings.
331 Unless mentioned otherwise, None is equivalent to absence of information.
334 _type "playlist" indicates multiple videos.
335 There must be a key "entries", which is a list, an iterable, or a PagedList
336 object, each element of which is a valid dictionary by this specification.
338 Additionally, playlists can have "id", "title", "description", "uploader",
339 "uploader_id", "uploader_url" attributes with the same semantics as videos
343 _type "multi_video" indicates that there are multiple videos that
344 form a single show, for examples multiple acts of an opera or TV episode.
345 It must have an entries key like a playlist and contain all the keys
346 required for a video at the same time.
349 _type "url" indicates that the video must be extracted from another
350 location, possibly by a different extractor. Its only required key is:
351 "url" - the next URL to extract.
352 The key "ie_key" can be set to the class name (minus the trailing "IE",
353 e.g. "Youtube") if the extractor class is known in advance.
354 Additionally, the dictionary may have any properties of the resolved entity
355 known in advance, for example "title" if the title of the referred video is
359 _type "url_transparent" entities have the same specification as "url", but
360 indicate that the given additional information is more precise than the one
361 associated with the resolved URL.
362 This is useful when a site employs a video service that hosts the video and
363 its technical metadata, but that video service does not embed a useful
364 title, description etc.
367 Subclasses of this one should re-define the _real_initialize() and
368 _real_extract() methods and define a _VALID_URL regexp.
369 Probably, they should also be added to the list of extractors.
371 _GEO_BYPASS attribute may be set to False in order to disable
372 geo restriction bypass mechanisms for a particular extractor.
373 Though it won't disable explicit geo restriction bypass based on
374 country code provided with geo_bypass_country.
376 _GEO_COUNTRIES attribute may contain a list of presumably geo unrestricted
377 countries for this extractor. One of these countries will be used by
378 geo restriction bypass mechanism right away in order to bypass
379 geo restriction, of course, if the mechanism is not disabled.
381 _GEO_IP_BLOCKS attribute may contain a list of presumably geo unrestricted
382 IP blocks in CIDR notation for this extractor. One of these IP blocks
383 will be used by geo restriction bypass mechanism similarly
386 Finally, the _WORKING attribute should be set to False for broken IEs
387 in order to warn the users and skip the tests.
392 _x_forwarded_for_ip = None
394 _GEO_COUNTRIES = None
395 _GEO_IP_BLOCKS = None
398 def __init__(self, downloader=None):
399 """Constructor. Receives an optional downloader."""
401 self._x_forwarded_for_ip = None
402 self.set_downloader(downloader)
405 def suitable(cls, url):
406 """Receives a URL and returns True if suitable for this IE."""
408 # This does not use has/getattr intentionally - we want to know whether
409 # we have cached the regexp for *this* class, whereas getattr would also
410 # match the superclass
411 if '_VALID_URL_RE' not in cls.__dict__:
412 cls._VALID_URL_RE = re.compile(cls._VALID_URL)
413 return cls._VALID_URL_RE.match(url) is not None
416 def _match_id(cls, url):
417 if '_VALID_URL_RE' not in cls.__dict__:
418 cls._VALID_URL_RE = re.compile(cls._VALID_URL)
419 m = cls._VALID_URL_RE.match(url)
421 return compat_str(m.group('id'))
425 """Getter method for _WORKING."""
428 def initialize(self):
429 """Initializes an instance (authentication, etc)."""
430 self._initialize_geo_bypass({
431 'countries': self._GEO_COUNTRIES,
432 'ip_blocks': self._GEO_IP_BLOCKS,
435 self._real_initialize()
438 def _initialize_geo_bypass(self, geo_bypass_context):
440 Initialize geo restriction bypass mechanism.
442 This method is used to initialize geo bypass mechanism based on faking
443 X-Forwarded-For HTTP header. A random country from provided country list
444 is selected and a random IP belonging to this country is generated. This
445 IP will be passed as X-Forwarded-For HTTP header in all subsequent
448 This method will be used for initial geo bypass mechanism initialization
449 during the instance initialization with _GEO_COUNTRIES and
452 You may also manually call it from extractor's code if geo bypass
453 information is not available beforehand (e.g. obtained during
454 extraction) or due to some other reason. In this case you should pass
455 this information in geo bypass context passed as first argument. It may
456 contain following fields:
458 countries: List of geo unrestricted countries (similar
460 ip_blocks: List of geo unrestricted IP blocks in CIDR notation
461 (similar to _GEO_IP_BLOCKS)
464 if not self._x_forwarded_for_ip:
466 # Geo bypass mechanism is explicitly disabled by user
467 if not self._downloader.params.get('geo_bypass', True):
470 if not geo_bypass_context:
471 geo_bypass_context = {}
473 # Backward compatibility: previously _initialize_geo_bypass
474 # expected a list of countries, some 3rd party code may still use
476 if isinstance(geo_bypass_context, (list, tuple)):
477 geo_bypass_context = {
478 'countries': geo_bypass_context,
481 # The whole point of geo bypass mechanism is to fake IP
482 # as X-Forwarded-For HTTP header based on some IP block or
485 # Path 1: bypassing based on IP block in CIDR notation
487 # Explicit IP block specified by user, use it right away
488 # regardless of whether extractor is geo bypassable or not
489 ip_block = self._downloader.params.get('geo_bypass_ip_block', None)
491 # Otherwise use random IP block from geo bypass context but only
492 # if extractor is known as geo bypassable
494 ip_blocks = geo_bypass_context.get('ip_blocks')
495 if self._GEO_BYPASS and ip_blocks:
496 ip_block = random.choice(ip_blocks)
499 self._x_forwarded_for_ip = GeoUtils.random_ipv4(ip_block)
500 if self._downloader.params.get('verbose', False):
501 self._downloader.to_screen(
502 '[debug] Using fake IP %s as X-Forwarded-For.'
503 % self._x_forwarded_for_ip)
506 # Path 2: bypassing based on country code
508 # Explicit country code specified by user, use it right away
509 # regardless of whether extractor is geo bypassable or not
510 country = self._downloader.params.get('geo_bypass_country', None)
512 # Otherwise use random country code from geo bypass context but
513 # only if extractor is known as geo bypassable
515 countries = geo_bypass_context.get('countries')
516 if self._GEO_BYPASS and countries:
517 country = random.choice(countries)
520 self._x_forwarded_for_ip = GeoUtils.random_ipv4(country)
521 if self._downloader.params.get('verbose', False):
522 self._downloader.to_screen(
523 '[debug] Using fake IP %s (%s) as X-Forwarded-For.'
524 % (self._x_forwarded_for_ip, country.upper()))
526 def extract(self, url):
527 """Extracts URL information and returns it in list of dicts."""
532 ie_result = self._real_extract(url)
533 if self._x_forwarded_for_ip:
534 ie_result['__x_forwarded_for_ip'] = self._x_forwarded_for_ip
536 except GeoRestrictedError as e:
537 if self.__maybe_fake_ip_and_retry(e.countries):
540 except ExtractorError:
542 except compat_http_client.IncompleteRead as e:
543 raise ExtractorError('A network error has occurred.', cause=e, expected=True)
544 except (KeyError, StopIteration) as e:
545 raise ExtractorError('An extractor error has occurred.', cause=e)
547 def __maybe_fake_ip_and_retry(self, countries):
548 if (not self._downloader.params.get('geo_bypass_country', None)
550 and self._downloader.params.get('geo_bypass', True)
551 and not self._x_forwarded_for_ip
553 country_code = random.choice(countries)
554 self._x_forwarded_for_ip = GeoUtils.random_ipv4(country_code)
555 if self._x_forwarded_for_ip:
557 'Video is geo restricted. Retrying extraction with fake IP %s (%s) as X-Forwarded-For.'
558 % (self._x_forwarded_for_ip, country_code.upper()))
562 def set_downloader(self, downloader):
563 """Sets the downloader for this IE."""
564 self._downloader = downloader
566 def _real_initialize(self):
567 """Real initialization process. Redefine in subclasses."""
570 def _real_extract(self, url):
571 """Real extraction process. Redefine in subclasses."""
576 """A string for getting the InfoExtractor with get_info_extractor"""
577 return compat_str(cls.__name__[:-2])
581 return compat_str(type(self).__name__[:-2])
584 def __can_accept_status_code(err, expected_status):
585 assert isinstance(err, compat_urllib_error.HTTPError)
586 if expected_status is None:
588 if isinstance(expected_status, compat_integer_types):
589 return err.code == expected_status
590 elif isinstance(expected_status, (list, tuple)):
591 return err.code in expected_status
592 elif callable(expected_status):
593 return expected_status(err.code) is True
597 def _request_webpage(self, url_or_request, video_id, note=None, errnote=None, fatal=True, data=None, headers={}, query={}, expected_status=None):
599 Return the response handle.
601 See _download_webpage docstring for arguments specification.
604 self.report_download_webpage(video_id)
605 elif note is not False:
607 self.to_screen('%s' % (note,))
609 self.to_screen('%s: %s' % (video_id, note))
611 # Some sites check X-Forwarded-For HTTP header in order to figure out
612 # the origin of the client behind proxy. This allows bypassing geo
613 # restriction by faking this header's value to IP that belongs to some
614 # geo unrestricted country. We will do so once we encounter any
615 # geo restriction error.
616 if self._x_forwarded_for_ip:
617 if 'X-Forwarded-For' not in headers:
618 headers['X-Forwarded-For'] = self._x_forwarded_for_ip
620 if isinstance(url_or_request, compat_urllib_request.Request):
621 url_or_request = update_Request(
622 url_or_request, data=data, headers=headers, query=query)
625 url_or_request = update_url_query(url_or_request, query)
626 if data is not None or headers:
627 url_or_request = sanitized_Request(url_or_request, data, headers)
628 exceptions = [compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error]
629 if hasattr(ssl, 'CertificateError'):
630 exceptions.append(ssl.CertificateError)
632 return self._downloader.urlopen(url_or_request)
633 except tuple(exceptions) as err:
634 if isinstance(err, compat_urllib_error.HTTPError):
635 if self.__can_accept_status_code(err, expected_status):
636 # Retain reference to error to prevent file object from
637 # being closed before it can be read. Works around the
638 # effects of <https://bugs.python.org/issue15002>
639 # introduced in Python 3.4.1.
646 errnote = 'Unable to download webpage'
648 errmsg = '%s: %s' % (errnote, error_to_compat_str(err))
650 raise ExtractorError(errmsg, sys.exc_info()[2], cause=err)
652 self._downloader.report_warning(errmsg)
655 def _download_webpage_handle(self, url_or_request, video_id, note=None, errnote=None, fatal=True, encoding=None, data=None, headers={}, query={}, expected_status=None):
657 Return a tuple (page content as string, URL handle).
659 See _download_webpage docstring for arguments specification.
661 # Strip hashes from the URL (#1038)
662 if isinstance(url_or_request, (compat_str, str)):
663 url_or_request = url_or_request.partition('#')[0]
665 urlh = self._request_webpage(url_or_request, video_id, note, errnote, fatal, data=data, headers=headers, query=query, expected_status=expected_status)
669 content = self._webpage_read_content(urlh, url_or_request, video_id, note, errnote, fatal, encoding=encoding)
670 return (content, urlh)
673 def _guess_encoding_from_content(content_type, webpage_bytes):
674 m = re.match(r'[a-zA-Z0-9_.-]+/[a-zA-Z0-9_.-]+\s*;\s*charset=(.+)', content_type)
676 encoding = m.group(1)
678 m = re.search(br'<meta[^>]+charset=[\'"]?([^\'")]+)[ /\'">]',
679 webpage_bytes[:1024])
681 encoding = m.group(1).decode('ascii')
682 elif webpage_bytes.startswith(b'\xff\xfe'):
689 def __check_blocked(self, content):
690 first_block = content[:512]
691 if ('<title>Access to this site is blocked</title>' in content
692 and 'Websense' in first_block):
693 msg = 'Access to this webpage has been blocked by Websense filtering software in your network.'
694 blocked_iframe = self._html_search_regex(
695 r'<iframe src="([^"]+)"', content,
696 'Websense information URL', default=None)
698 msg += ' Visit %s for more details' % blocked_iframe
699 raise ExtractorError(msg, expected=True)
700 if '<title>The URL you requested has been blocked</title>' in first_block:
702 'Access to this webpage has been blocked by Indian censorship. '
703 'Use a VPN or proxy server (with --proxy) to route around it.')
704 block_msg = self._html_search_regex(
705 r'</h1><p>(.*?)</p>',
706 content, 'block message', default=None)
708 msg += ' (Message: "%s")' % block_msg.replace('\n', ' ')
709 raise ExtractorError(msg, expected=True)
710 if ('<title>TTK :: Доступ к ресурсу ограничен</title>' in content
711 and 'blocklist.rkn.gov.ru' in content):
712 raise ExtractorError(
713 'Access to this webpage has been blocked by decision of the Russian government. '
714 'Visit http://blocklist.rkn.gov.ru/ for a block reason.',
717 def _webpage_read_content(self, urlh, url_or_request, video_id, note=None, errnote=None, fatal=True, prefix=None, encoding=None):
718 content_type = urlh.headers.get('Content-Type', '')
719 webpage_bytes = urlh.read()
720 if prefix is not None:
721 webpage_bytes = prefix + webpage_bytes
723 encoding = self._guess_encoding_from_content(content_type, webpage_bytes)
724 if self._downloader.params.get('dump_intermediate_pages', False):
725 self.to_screen('Dumping request to ' + urlh.geturl())
726 dump = base64.b64encode(webpage_bytes).decode('ascii')
727 self._downloader.to_screen(dump)
728 if self._downloader.params.get('write_pages', False):
729 basen = '%s_%s' % (video_id, urlh.geturl())
731 h = '___' + hashlib.md5(basen.encode('utf-8')).hexdigest()
732 basen = basen[:240 - len(h)] + h
733 raw_filename = basen + '.dump'
734 filename = sanitize_filename(raw_filename, restricted=True)
735 self.to_screen('Saving request to ' + filename)
736 # Working around MAX_PATH limitation on Windows (see
737 # http://msdn.microsoft.com/en-us/library/windows/desktop/aa365247(v=vs.85).aspx)
738 if compat_os_name == 'nt':
739 absfilepath = os.path.abspath(filename)
740 if len(absfilepath) > 259:
741 filename = '\\\\?\\' + absfilepath
742 with open(filename, 'wb') as outf:
743 outf.write(webpage_bytes)
746 content = webpage_bytes.decode(encoding, 'replace')
748 content = webpage_bytes.decode('utf-8', 'replace')
750 self.__check_blocked(content)
754 def _download_webpage(
755 self, url_or_request, video_id, note=None, errnote=None,
756 fatal=True, tries=1, timeout=5, encoding=None, data=None,
757 headers={}, query={}, expected_status=None):
759 Return the data of the page as a string.
762 url_or_request -- plain text URL as a string or
763 a compat_urllib_request.Requestobject
764 video_id -- Video/playlist/item identifier (string)
767 note -- note printed before downloading (string)
768 errnote -- note printed in case of an error (string)
769 fatal -- flag denoting whether error should be considered fatal,
770 i.e. whether it should cause ExtractionError to be raised,
771 otherwise a warning will be reported and extraction continued
772 tries -- number of tries
773 timeout -- sleep interval between tries
774 encoding -- encoding for a page content decoding, guessed automatically
775 when not explicitly specified
776 data -- POST data (bytes)
777 headers -- HTTP headers (dict)
778 query -- URL query (dict)
779 expected_status -- allows to accept failed HTTP requests (non 2xx
780 status code) by explicitly specifying a set of accepted status
781 codes. Can be any of the following entities:
782 - an integer type specifying an exact failed status code to
784 - a list or a tuple of integer types specifying a list of
785 failed status codes to accept
786 - a callable accepting an actual failed status code and
787 returning True if it should be accepted
788 Note that this argument does not affect success status codes (2xx)
789 which are always accepted.
794 while success is False:
796 res = self._download_webpage_handle(
797 url_or_request, video_id, note, errnote, fatal,
798 encoding=encoding, data=data, headers=headers, query=query,
799 expected_status=expected_status)
801 except compat_http_client.IncompleteRead as e:
803 if try_count >= tries:
805 self._sleep(timeout, video_id)
812 def _download_xml_handle(
813 self, url_or_request, video_id, note='Downloading XML',
814 errnote='Unable to download XML', transform_source=None,
815 fatal=True, encoding=None, data=None, headers={}, query={},
816 expected_status=None):
818 Return a tuple (xml as an compat_etree_Element, URL handle).
820 See _download_webpage docstring for arguments specification.
822 res = self._download_webpage_handle(
823 url_or_request, video_id, note, errnote, fatal=fatal,
824 encoding=encoding, data=data, headers=headers, query=query,
825 expected_status=expected_status)
828 xml_string, urlh = res
829 return self._parse_xml(
830 xml_string, video_id, transform_source=transform_source,
834 self, url_or_request, video_id,
835 note='Downloading XML', errnote='Unable to download XML',
836 transform_source=None, fatal=True, encoding=None,
837 data=None, headers={}, query={}, expected_status=None):
839 Return the xml as an compat_etree_Element.
841 See _download_webpage docstring for arguments specification.
843 res = self._download_xml_handle(
844 url_or_request, video_id, note=note, errnote=errnote,
845 transform_source=transform_source, fatal=fatal, encoding=encoding,
846 data=data, headers=headers, query=query,
847 expected_status=expected_status)
848 return res if res is False else res[0]
850 def _parse_xml(self, xml_string, video_id, transform_source=None, fatal=True):
852 xml_string = transform_source(xml_string)
854 return compat_etree_fromstring(xml_string.encode('utf-8'))
855 except compat_xml_parse_error as ve:
856 errmsg = '%s: Failed to parse XML ' % video_id
858 raise ExtractorError(errmsg, cause=ve)
860 self.report_warning(errmsg + str(ve))
862 def _download_json_handle(
863 self, url_or_request, video_id, note='Downloading JSON metadata',
864 errnote='Unable to download JSON metadata', transform_source=None,
865 fatal=True, encoding=None, data=None, headers={}, query={},
866 expected_status=None):
868 Return a tuple (JSON object, URL handle).
870 See _download_webpage docstring for arguments specification.
872 res = self._download_webpage_handle(
873 url_or_request, video_id, note, errnote, fatal=fatal,
874 encoding=encoding, data=data, headers=headers, query=query,
875 expected_status=expected_status)
878 json_string, urlh = res
879 return self._parse_json(
880 json_string, video_id, transform_source=transform_source,
884 self, url_or_request, video_id, note='Downloading JSON metadata',
885 errnote='Unable to download JSON metadata', transform_source=None,
886 fatal=True, encoding=None, data=None, headers={}, query={},
887 expected_status=None):
889 Return the JSON object as a dict.
891 See _download_webpage docstring for arguments specification.
893 res = self._download_json_handle(
894 url_or_request, video_id, note=note, errnote=errnote,
895 transform_source=transform_source, fatal=fatal, encoding=encoding,
896 data=data, headers=headers, query=query,
897 expected_status=expected_status)
898 return res if res is False else res[0]
900 def _parse_json(self, json_string, video_id, transform_source=None, fatal=True):
902 json_string = transform_source(json_string)
904 return json.loads(json_string)
905 except ValueError as ve:
906 errmsg = '%s: Failed to parse JSON ' % video_id
908 raise ExtractorError(errmsg, cause=ve)
910 self.report_warning(errmsg + str(ve))
912 def report_warning(self, msg, video_id=None):
913 idstr = '' if video_id is None else '%s: ' % video_id
914 self._downloader.report_warning(
915 '[%s] %s%s' % (self.IE_NAME, idstr, msg))
917 def to_screen(self, msg):
918 """Print msg to screen, prefixing it with '[ie_name]'"""
919 self._downloader.to_screen('[%s] %s' % (self.IE_NAME, msg))
921 def report_extraction(self, id_or_name):
922 """Report information extraction."""
923 self.to_screen('%s: Extracting information' % id_or_name)
925 def report_download_webpage(self, video_id):
926 """Report webpage download."""
927 self.to_screen('%s: Downloading webpage' % video_id)
929 def report_age_confirmation(self):
930 """Report attempt to confirm age."""
931 self.to_screen('Confirming age')
933 def report_login(self):
934 """Report attempt to log in."""
935 self.to_screen('Logging in')
938 def raise_login_required(msg='This video is only available for registered users'):
939 raise ExtractorError(
940 '%s. Use --username and --password or --netrc to provide account credentials.' % msg,
944 def raise_geo_restricted(msg='This video is not available from your location due to geo restriction', countries=None):
945 raise GeoRestrictedError(msg, countries=countries)
947 # Methods for following #608
949 def url_result(url, ie=None, video_id=None, video_title=None):
950 """Returns a URL that points to a page that should be processed"""
951 # TODO: ie should be the class used for getting the info
952 video_info = {'_type': 'url',
955 if video_id is not None:
956 video_info['id'] = video_id
957 if video_title is not None:
958 video_info['title'] = video_title
961 def playlist_from_matches(self, matches, playlist_id=None, playlist_title=None, getter=None, ie=None):
963 self.url_result(self._proto_relative_url(getter(m) if getter else m), ie)
965 return self.playlist_result(
966 urls, playlist_id=playlist_id, playlist_title=playlist_title)
969 def playlist_result(entries, playlist_id=None, playlist_title=None, playlist_description=None):
970 """Returns a playlist"""
971 video_info = {'_type': 'playlist',
974 video_info['id'] = playlist_id
976 video_info['title'] = playlist_title
977 if playlist_description:
978 video_info['description'] = playlist_description
981 def _search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
983 Perform a regex search on the given string, using a single or a list of
984 patterns returning the first matching group.
985 In case of failure return a default value or raise a WARNING or a
986 RegexNotFoundError, depending on fatal, specifying the field name.
988 if isinstance(pattern, (str, compat_str, compiled_regex_type)):
989 mobj = re.search(pattern, string, flags)
992 mobj = re.search(p, string, flags)
996 if not self._downloader.params.get('no_color') and compat_os_name != 'nt' and sys.stderr.isatty():
997 _name = '\033[0;34m%s\033[0m' % name
1003 # return the first matching group
1004 return next(g for g in mobj.groups() if g is not None)
1006 return mobj.group(group)
1007 elif default is not NO_DEFAULT:
1010 raise RegexNotFoundError('Unable to extract %s' % _name)
1012 self._downloader.report_warning('unable to extract %s' % _name + bug_reports_message())
1015 def _html_search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
1017 Like _search_regex, but strips HTML tags and unescapes entities.
1019 res = self._search_regex(pattern, string, name, default, fatal, flags, group)
1021 return clean_html(res).strip()
1025 def _get_netrc_login_info(self, netrc_machine=None):
1028 netrc_machine = netrc_machine or self._NETRC_MACHINE
1030 if self._downloader.params.get('usenetrc', False):
1032 info = netrc.netrc().authenticators(netrc_machine)
1033 if info is not None:
1037 raise netrc.NetrcParseError(
1038 'No authenticators for %s' % netrc_machine)
1039 except (IOError, netrc.NetrcParseError) as err:
1040 self._downloader.report_warning(
1041 'parsing .netrc: %s' % error_to_compat_str(err))
1043 return username, password
1045 def _get_login_info(self, username_option='username', password_option='password', netrc_machine=None):
1047 Get the login info as (username, password)
1048 First look for the manually specified credentials using username_option
1049 and password_option as keys in params dictionary. If no such credentials
1050 available look in the netrc file using the netrc_machine or _NETRC_MACHINE
1052 If there's no info available, return (None, None)
1054 if self._downloader is None:
1057 downloader_params = self._downloader.params
1059 # Attempt to use provided username and password or .netrc data
1060 if downloader_params.get(username_option) is not None:
1061 username = downloader_params[username_option]
1062 password = downloader_params[password_option]
1064 username, password = self._get_netrc_login_info(netrc_machine)
1066 return username, password
1068 def _get_tfa_info(self, note='two-factor verification code'):
1070 Get the two-factor authentication info
1071 TODO - asking the user will be required for sms/phone verify
1072 currently just uses the command line option
1073 If there's no info available, return None
1075 if self._downloader is None:
1077 downloader_params = self._downloader.params
1079 if downloader_params.get('twofactor') is not None:
1080 return downloader_params['twofactor']
1082 return compat_getpass('Type %s and press [Return]: ' % note)
1084 # Helper functions for extracting OpenGraph info
1086 def _og_regexes(prop):
1087 content_re = r'content=(?:"([^"]+?)"|\'([^\']+?)\'|\s*([^\s"\'=<>`]+?))'
1088 property_re = (r'(?:name|property)=(?:\'og[:-]%(prop)s\'|"og[:-]%(prop)s"|\s*og[:-]%(prop)s\b)'
1089 % {'prop': re.escape(prop)})
1090 template = r'<meta[^>]+?%s[^>]+?%s'
1092 template % (property_re, content_re),
1093 template % (content_re, property_re),
1097 def _meta_regex(prop):
1098 return r'''(?isx)<meta
1099 (?=[^>]+(?:itemprop|name|property|id|http-equiv)=(["\']?)%s\1)
1100 [^>]+?content=(["\'])(?P<content>.*?)\2''' % re.escape(prop)
1102 def _og_search_property(self, prop, html, name=None, **kargs):
1103 if not isinstance(prop, (list, tuple)):
1106 name = 'OpenGraph %s' % prop[0]
1109 og_regexes.extend(self._og_regexes(p))
1110 escaped = self._search_regex(og_regexes, html, name, flags=re.DOTALL, **kargs)
1113 return unescapeHTML(escaped)
1115 def _og_search_thumbnail(self, html, **kargs):
1116 return self._og_search_property('image', html, 'thumbnail URL', fatal=False, **kargs)
1118 def _og_search_description(self, html, **kargs):
1119 return self._og_search_property('description', html, fatal=False, **kargs)
1121 def _og_search_title(self, html, **kargs):
1122 return self._og_search_property('title', html, **kargs)
1124 def _og_search_video_url(self, html, name='video url', secure=True, **kargs):
1125 regexes = self._og_regexes('video') + self._og_regexes('video:url')
1127 regexes = self._og_regexes('video:secure_url') + regexes
1128 return self._html_search_regex(regexes, html, name, **kargs)
1130 def _og_search_url(self, html, **kargs):
1131 return self._og_search_property('url', html, **kargs)
1133 def _html_search_meta(self, name, html, display_name=None, fatal=False, **kwargs):
1134 if not isinstance(name, (list, tuple)):
1136 if display_name is None:
1137 display_name = name[0]
1138 return self._html_search_regex(
1139 [self._meta_regex(n) for n in name],
1140 html, display_name, fatal=fatal, group='content', **kwargs)
1142 def _dc_search_uploader(self, html):
1143 return self._html_search_meta('dc.creator', html, 'uploader')
1145 def _rta_search(self, html):
1146 # See http://www.rtalabel.org/index.php?content=howtofaq#single
1147 if re.search(r'(?ix)<meta\s+name="rating"\s+'
1148 r' content="RTA-5042-1996-1400-1577-RTA"',
1153 def _media_rating_search(self, html):
1154 # See http://www.tjg-designs.com/WP/metadata-code-examples-adding-metadata-to-your-web-pages/
1155 rating = self._html_search_meta('rating', html)
1167 return RATING_TABLE.get(rating.lower())
1169 def _family_friendly_search(self, html):
1170 # See http://schema.org/VideoObject
1171 family_friendly = self._html_search_meta(
1172 'isFamilyFriendly', html, default=None)
1174 if not family_friendly:
1183 return RATING_TABLE.get(family_friendly.lower())
1185 def _twitter_search_player(self, html):
1186 return self._html_search_meta('twitter:player', html,
1187 'twitter card player')
1189 def _search_json_ld(self, html, video_id, expected_type=None, **kwargs):
1190 json_ld_list = list(re.finditer(JSON_LD_RE, html))
1191 default = kwargs.get('default', NO_DEFAULT)
1192 # JSON-LD may be malformed and thus `fatal` should be respected.
1193 # At the same time `default` may be passed that assumes `fatal=False`
1194 # for _search_regex. Let's simulate the same behavior here as well.
1195 fatal = kwargs.get('fatal', True) if default == NO_DEFAULT else False
1197 for mobj in json_ld_list:
1198 json_ld_item = self._parse_json(
1199 mobj.group('json_ld'), video_id, fatal=fatal)
1200 if not json_ld_item:
1202 if isinstance(json_ld_item, dict):
1203 json_ld.append(json_ld_item)
1204 elif isinstance(json_ld_item, (list, tuple)):
1205 json_ld.extend(json_ld_item)
1207 json_ld = self._json_ld(json_ld, video_id, fatal=fatal, expected_type=expected_type)
1210 if default is not NO_DEFAULT:
1213 raise RegexNotFoundError('Unable to extract JSON-LD')
1215 self._downloader.report_warning('unable to extract JSON-LD %s' % bug_reports_message())
1218 def _json_ld(self, json_ld, video_id, fatal=True, expected_type=None):
1219 if isinstance(json_ld, compat_str):
1220 json_ld = self._parse_json(json_ld, video_id, fatal=fatal)
1224 if not isinstance(json_ld, (list, tuple, dict)):
1226 if isinstance(json_ld, dict):
1229 INTERACTION_TYPE_MAP = {
1230 'CommentAction': 'comment',
1231 'AgreeAction': 'like',
1232 'DisagreeAction': 'dislike',
1233 'LikeAction': 'like',
1234 'DislikeAction': 'dislike',
1235 'ListenAction': 'view',
1236 'WatchAction': 'view',
1237 'ViewAction': 'view',
1240 def extract_interaction_statistic(e):
1241 interaction_statistic = e.get('interactionStatistic')
1242 if not isinstance(interaction_statistic, list):
1244 for is_e in interaction_statistic:
1245 if not isinstance(is_e, dict):
1247 if is_e.get('@type') != 'InteractionCounter':
1249 interaction_type = is_e.get('interactionType')
1250 if not isinstance(interaction_type, compat_str):
1252 # For interaction count some sites provide string instead of
1253 # an integer (as per spec) with non digit characters (e.g. ",")
1254 # so extracting count with more relaxed str_to_int
1255 interaction_count = str_to_int(is_e.get('userInteractionCount'))
1256 if interaction_count is None:
1258 count_kind = INTERACTION_TYPE_MAP.get(interaction_type.split('/')[-1])
1261 count_key = '%s_count' % count_kind
1262 if info.get(count_key) is not None:
1264 info[count_key] = interaction_count
1266 def extract_video_object(e):
1267 assert e['@type'] == 'VideoObject'
1269 'url': url_or_none(e.get('contentUrl')),
1270 'title': unescapeHTML(e.get('name')),
1271 'description': unescapeHTML(e.get('description')),
1272 'thumbnail': url_or_none(e.get('thumbnailUrl') or e.get('thumbnailURL')),
1273 'duration': parse_duration(e.get('duration')),
1274 'timestamp': unified_timestamp(e.get('uploadDate')),
1275 'uploader': str_or_none(e.get('author')),
1276 'filesize': float_or_none(e.get('contentSize')),
1277 'tbr': int_or_none(e.get('bitrate')),
1278 'width': int_or_none(e.get('width')),
1279 'height': int_or_none(e.get('height')),
1280 'view_count': int_or_none(e.get('interactionCount')),
1282 extract_interaction_statistic(e)
1286 item_type = e.get('@type')
1287 if expected_type is not None and expected_type != item_type:
1289 if item_type in ('TVEpisode', 'Episode'):
1290 episode_name = unescapeHTML(e.get('name'))
1292 'episode': episode_name,
1293 'episode_number': int_or_none(e.get('episodeNumber')),
1294 'description': unescapeHTML(e.get('description')),
1296 if not info.get('title') and episode_name:
1297 info['title'] = episode_name
1298 part_of_season = e.get('partOfSeason')
1299 if isinstance(part_of_season, dict) and part_of_season.get('@type') in ('TVSeason', 'Season', 'CreativeWorkSeason'):
1301 'season': unescapeHTML(part_of_season.get('name')),
1302 'season_number': int_or_none(part_of_season.get('seasonNumber')),
1304 part_of_series = e.get('partOfSeries') or e.get('partOfTVSeries')
1305 if isinstance(part_of_series, dict) and part_of_series.get('@type') in ('TVSeries', 'Series', 'CreativeWorkSeries'):
1306 info['series'] = unescapeHTML(part_of_series.get('name'))
1307 elif item_type == 'Movie':
1309 'title': unescapeHTML(e.get('name')),
1310 'description': unescapeHTML(e.get('description')),
1311 'duration': parse_duration(e.get('duration')),
1312 'timestamp': unified_timestamp(e.get('dateCreated')),
1314 elif item_type in ('Article', 'NewsArticle'):
1316 'timestamp': parse_iso8601(e.get('datePublished')),
1317 'title': unescapeHTML(e.get('headline')),
1318 'description': unescapeHTML(e.get('articleBody')),
1320 elif item_type == 'VideoObject':
1321 extract_video_object(e)
1322 if expected_type is None:
1326 video = e.get('video')
1327 if isinstance(video, dict) and video.get('@type') == 'VideoObject':
1328 extract_video_object(video)
1329 if expected_type is None:
1333 return dict((k, v) for k, v in info.items() if v is not None)
1336 def _hidden_inputs(html):
1337 html = re.sub(r'<!--(?:(?!<!--).)*-->', '', html)
1339 for input in re.findall(r'(?i)(<input[^>]+>)', html):
1340 attrs = extract_attributes(input)
1343 if attrs.get('type') not in ('hidden', 'submit'):
1345 name = attrs.get('name') or attrs.get('id')
1346 value = attrs.get('value')
1347 if name and value is not None:
1348 hidden_inputs[name] = value
1349 return hidden_inputs
1351 def _form_hidden_inputs(self, form_id, html):
1352 form = self._search_regex(
1353 r'(?is)<form[^>]+?id=(["\'])%s\1[^>]*>(?P<form>.+?)</form>' % form_id,
1354 html, '%s form' % form_id, group='form')
1355 return self._hidden_inputs(form)
1357 def _sort_formats(self, formats, field_preference=None):
1359 raise ExtractorError('No video formats found')
1362 # Automatically determine tbr when missing based on abr and vbr (improves
1363 # formats sorting in some cases)
1364 if 'tbr' not in f and f.get('abr') is not None and f.get('vbr') is not None:
1365 f['tbr'] = f['abr'] + f['vbr']
1367 def _formats_key(f):
1368 # TODO remove the following workaround
1369 from ..utils import determine_ext
1370 if not f.get('ext') and 'url' in f:
1371 f['ext'] = determine_ext(f['url'])
1373 if isinstance(field_preference, (list, tuple)):
1376 if f.get(field) is not None
1377 else ('' if field == 'format_id' else -1)
1378 for field in field_preference)
1380 preference = f.get('preference')
1381 if preference is None:
1383 if f.get('ext') in ['f4f', 'f4m']: # Not yet supported
1386 protocol = f.get('protocol') or determine_protocol(f)
1387 proto_preference = 0 if protocol in ['http', 'https'] else (-0.5 if protocol == 'rtsp' else -0.1)
1389 if f.get('vcodec') == 'none': # audio only
1391 if self._downloader.params.get('prefer_free_formats'):
1392 ORDER = ['aac', 'mp3', 'm4a', 'webm', 'ogg', 'opus']
1394 ORDER = ['webm', 'opus', 'ogg', 'mp3', 'aac', 'm4a']
1397 audio_ext_preference = ORDER.index(f['ext'])
1399 audio_ext_preference = -1
1401 if f.get('acodec') == 'none': # video only
1403 if self._downloader.params.get('prefer_free_formats'):
1404 ORDER = ['flv', 'mp4', 'webm']
1406 ORDER = ['webm', 'flv', 'mp4']
1408 ext_preference = ORDER.index(f['ext'])
1411 audio_ext_preference = 0
1415 f.get('language_preference') if f.get('language_preference') is not None else -1,
1416 f.get('quality') if f.get('quality') is not None else -1,
1417 f.get('tbr') if f.get('tbr') is not None else -1,
1418 f.get('filesize') if f.get('filesize') is not None else -1,
1419 f.get('vbr') if f.get('vbr') is not None else -1,
1420 f.get('height') if f.get('height') is not None else -1,
1421 f.get('width') if f.get('width') is not None else -1,
1424 f.get('abr') if f.get('abr') is not None else -1,
1425 audio_ext_preference,
1426 f.get('fps') if f.get('fps') is not None else -1,
1427 f.get('filesize_approx') if f.get('filesize_approx') is not None else -1,
1428 f.get('source_preference') if f.get('source_preference') is not None else -1,
1429 f.get('format_id') if f.get('format_id') is not None else '',
1431 formats.sort(key=_formats_key)
1433 def _check_formats(self, formats, video_id):
1435 formats[:] = filter(
1436 lambda f: self._is_valid_url(
1438 item='%s video format' % f.get('format_id') if f.get('format_id') else 'video'),
1442 def _remove_duplicate_formats(formats):
1446 if f['url'] not in format_urls:
1447 format_urls.add(f['url'])
1448 unique_formats.append(f)
1449 formats[:] = unique_formats
1451 def _is_valid_url(self, url, video_id, item='video', headers={}):
1452 url = self._proto_relative_url(url, scheme='http:')
1453 # For now assume non HTTP(S) URLs always valid
1454 if not (url.startswith('http://') or url.startswith('https://')):
1457 self._request_webpage(url, video_id, 'Checking %s URL' % item, headers=headers)
1459 except ExtractorError:
1461 '%s: %s URL is invalid, skipping' % (video_id, item))
1464 def http_scheme(self):
1465 """ Either "http:" or "https:", depending on the user's preferences """
1468 if self._downloader.params.get('prefer_insecure', False)
1471 def _proto_relative_url(self, url, scheme=None):
1474 if url.startswith('//'):
1476 scheme = self.http_scheme()
1481 def _sleep(self, timeout, video_id, msg_template=None):
1482 if msg_template is None:
1483 msg_template = '%(video_id)s: Waiting for %(timeout)s seconds'
1484 msg = msg_template % {'video_id': video_id, 'timeout': timeout}
1488 def _extract_f4m_formats(self, manifest_url, video_id, preference=None, f4m_id=None,
1489 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1490 fatal=True, m3u8_id=None, data=None, headers={}, query={}):
1491 manifest = self._download_xml(
1492 manifest_url, video_id, 'Downloading f4m manifest',
1493 'Unable to download f4m manifest',
1494 # Some manifests may be malformed, e.g. prosiebensat1 generated manifests
1495 # (see https://github.com/ytdl-org/youtube-dl/issues/6215#issuecomment-121704244)
1496 transform_source=transform_source,
1497 fatal=fatal, data=data, headers=headers, query=query)
1499 if manifest is False:
1502 return self._parse_f4m_formats(
1503 manifest, manifest_url, video_id, preference=preference, f4m_id=f4m_id,
1504 transform_source=transform_source, fatal=fatal, m3u8_id=m3u8_id)
1506 def _parse_f4m_formats(self, manifest, manifest_url, video_id, preference=None, f4m_id=None,
1507 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1508 fatal=True, m3u8_id=None):
1509 if not isinstance(manifest, compat_etree_Element) and not fatal:
1512 # currently youtube-dl cannot decode the playerVerificationChallenge as Akamai uses Adobe Alchemy
1513 akamai_pv = manifest.find('{http://ns.adobe.com/f4m/1.0}pv-2.0')
1514 if akamai_pv is not None and ';' in akamai_pv.text:
1515 playerVerificationChallenge = akamai_pv.text.split(';')[0]
1516 if playerVerificationChallenge.strip() != '':
1520 manifest_version = '1.0'
1521 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/1.0}media')
1523 manifest_version = '2.0'
1524 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/2.0}media')
1525 # Remove unsupported DRM protected media from final formats
1526 # rendition (see https://github.com/ytdl-org/youtube-dl/issues/8573).
1527 media_nodes = remove_encrypted_media(media_nodes)
1531 manifest_base_url = get_base_url(manifest)
1533 bootstrap_info = xpath_element(
1534 manifest, ['{http://ns.adobe.com/f4m/1.0}bootstrapInfo', '{http://ns.adobe.com/f4m/2.0}bootstrapInfo'],
1535 'bootstrap info', default=None)
1538 mime_type = xpath_text(
1539 manifest, ['{http://ns.adobe.com/f4m/1.0}mimeType', '{http://ns.adobe.com/f4m/2.0}mimeType'],
1540 'base URL', default=None)
1541 if mime_type and mime_type.startswith('audio/'):
1544 for i, media_el in enumerate(media_nodes):
1545 tbr = int_or_none(media_el.attrib.get('bitrate'))
1546 width = int_or_none(media_el.attrib.get('width'))
1547 height = int_or_none(media_el.attrib.get('height'))
1548 format_id = '-'.join(filter(None, [f4m_id, compat_str(i if tbr is None else tbr)]))
1549 # If <bootstrapInfo> is present, the specified f4m is a
1550 # stream-level manifest, and only set-level manifests may refer to
1551 # external resources. See section 11.4 and section 4 of F4M spec
1552 if bootstrap_info is None:
1554 # @href is introduced in 2.0, see section 11.6 of F4M spec
1555 if manifest_version == '2.0':
1556 media_url = media_el.attrib.get('href')
1557 if media_url is None:
1558 media_url = media_el.attrib.get('url')
1562 media_url if media_url.startswith('http://') or media_url.startswith('https://')
1563 else ((manifest_base_url or '/'.join(manifest_url.split('/')[:-1])) + '/' + media_url))
1564 # If media_url is itself a f4m manifest do the recursive extraction
1565 # since bitrates in parent manifest (this one) and media_url manifest
1566 # may differ leading to inability to resolve the format by requested
1567 # bitrate in f4m downloader
1568 ext = determine_ext(manifest_url)
1570 f4m_formats = self._extract_f4m_formats(
1571 manifest_url, video_id, preference=preference, f4m_id=f4m_id,
1572 transform_source=transform_source, fatal=fatal)
1573 # Sometimes stream-level manifest contains single media entry that
1574 # does not contain any quality metadata (e.g. http://matchtv.ru/#live-player).
1575 # At the same time parent's media entry in set-level manifest may
1576 # contain it. We will copy it from parent in such cases.
1577 if len(f4m_formats) == 1:
1580 'tbr': f.get('tbr') or tbr,
1581 'width': f.get('width') or width,
1582 'height': f.get('height') or height,
1583 'format_id': f.get('format_id') if not tbr else format_id,
1586 formats.extend(f4m_formats)
1589 formats.extend(self._extract_m3u8_formats(
1590 manifest_url, video_id, 'mp4', preference=preference,
1591 m3u8_id=m3u8_id, fatal=fatal))
1594 'format_id': format_id,
1595 'url': manifest_url,
1596 'manifest_url': manifest_url,
1597 'ext': 'flv' if bootstrap_info is not None else None,
1603 'preference': preference,
1607 def _m3u8_meta_format(self, m3u8_url, ext=None, preference=None, m3u8_id=None):
1609 'format_id': '-'.join(filter(None, [m3u8_id, 'meta'])),
1613 'preference': preference - 100 if preference else -100,
1614 'resolution': 'multiple',
1615 'format_note': 'Quality selection URL',
1618 def _extract_m3u8_formats(self, m3u8_url, video_id, ext=None,
1619 entry_protocol='m3u8', preference=None,
1620 m3u8_id=None, note=None, errnote=None,
1621 fatal=True, live=False, data=None, headers={},
1623 res = self._download_webpage_handle(
1625 note=note or 'Downloading m3u8 information',
1626 errnote=errnote or 'Failed to download m3u8 information',
1627 fatal=fatal, data=data, headers=headers, query=query)
1632 m3u8_doc, urlh = res
1633 m3u8_url = urlh.geturl()
1635 return self._parse_m3u8_formats(
1636 m3u8_doc, m3u8_url, ext=ext, entry_protocol=entry_protocol,
1637 preference=preference, m3u8_id=m3u8_id, live=live)
1639 def _parse_m3u8_formats(self, m3u8_doc, m3u8_url, ext=None,
1640 entry_protocol='m3u8', preference=None,
1641 m3u8_id=None, live=False):
1642 if '#EXT-X-FAXS-CM:' in m3u8_doc: # Adobe Flash Access
1645 if re.search(r'#EXT-X-SESSION-KEY:.*?URI="skd://', m3u8_doc): # Apple FairPlay
1650 format_url = lambda u: (
1652 if re.match(r'^https?://', u)
1653 else compat_urlparse.urljoin(m3u8_url, u))
1656 # 1. https://tools.ietf.org/html/draft-pantos-http-live-streaming-21
1657 # 2. https://github.com/ytdl-org/youtube-dl/issues/12211
1658 # 3. https://github.com/ytdl-org/youtube-dl/issues/18923
1660 # We should try extracting formats only from master playlists [1, 4.3.4],
1661 # i.e. playlists that describe available qualities. On the other hand
1662 # media playlists [1, 4.3.3] should be returned as is since they contain
1663 # just the media without qualities renditions.
1664 # Fortunately, master playlist can be easily distinguished from media
1665 # playlist based on particular tags availability. As of [1, 4.3.3, 4.3.4]
1666 # master playlist tags MUST NOT appear in a media playist and vice versa.
1667 # As of [1, 4.3.3.1] #EXT-X-TARGETDURATION tag is REQUIRED for every
1668 # media playlist and MUST NOT appear in master playlist thus we can
1669 # clearly detect media playlist with this criterion.
1671 if '#EXT-X-TARGETDURATION' in m3u8_doc: # media playlist, return as is
1674 'format_id': m3u8_id,
1676 'protocol': entry_protocol,
1677 'preference': preference,
1681 last_stream_inf = {}
1683 def extract_media(x_media_line):
1684 media = parse_m3u8_attributes(x_media_line)
1685 # As per [1, 4.3.4.1] TYPE, GROUP-ID and NAME are REQUIRED
1686 media_type, group_id, name = media.get('TYPE'), media.get('GROUP-ID'), media.get('NAME')
1687 if not (media_type and group_id and name):
1689 groups.setdefault(group_id, []).append(media)
1690 if media_type not in ('VIDEO', 'AUDIO'):
1692 media_url = media.get('URI')
1695 for v in (m3u8_id, group_id, name):
1699 'format_id': '-'.join(format_id),
1700 'url': format_url(media_url),
1701 'manifest_url': m3u8_url,
1702 'language': media.get('LANGUAGE'),
1704 'protocol': entry_protocol,
1705 'preference': preference,
1707 if media_type == 'AUDIO':
1708 f['vcodec'] = 'none'
1711 def build_stream_name():
1712 # Despite specification does not mention NAME attribute for
1713 # EXT-X-STREAM-INF tag it still sometimes may be present (see [1]
1714 # or vidio test in TestInfoExtractor.test_parse_m3u8_formats)
1715 # 1. http://www.vidio.com/watch/165683-dj_ambred-booyah-live-2015
1716 stream_name = last_stream_inf.get('NAME')
1719 # If there is no NAME in EXT-X-STREAM-INF it will be obtained
1720 # from corresponding rendition group
1721 stream_group_id = last_stream_inf.get('VIDEO')
1722 if not stream_group_id:
1724 stream_group = groups.get(stream_group_id)
1725 if not stream_group:
1726 return stream_group_id
1727 rendition = stream_group[0]
1728 return rendition.get('NAME') or stream_group_id
1730 # parse EXT-X-MEDIA tags before EXT-X-STREAM-INF in order to have the
1731 # chance to detect video only formats when EXT-X-STREAM-INF tags
1732 # precede EXT-X-MEDIA tags in HLS manifest such as [3].
1733 for line in m3u8_doc.splitlines():
1734 if line.startswith('#EXT-X-MEDIA:'):
1737 for line in m3u8_doc.splitlines():
1738 if line.startswith('#EXT-X-STREAM-INF:'):
1739 last_stream_inf = parse_m3u8_attributes(line)
1740 elif line.startswith('#') or not line.strip():
1743 tbr = float_or_none(
1744 last_stream_inf.get('AVERAGE-BANDWIDTH')
1745 or last_stream_inf.get('BANDWIDTH'), scale=1000)
1748 format_id.append(m3u8_id)
1749 stream_name = build_stream_name()
1750 # Bandwidth of live streams may differ over time thus making
1751 # format_id unpredictable. So it's better to keep provided
1754 format_id.append(stream_name if stream_name else '%d' % (tbr if tbr else len(formats)))
1755 manifest_url = format_url(line.strip())
1757 'format_id': '-'.join(format_id),
1758 'url': manifest_url,
1759 'manifest_url': m3u8_url,
1762 'fps': float_or_none(last_stream_inf.get('FRAME-RATE')),
1763 'protocol': entry_protocol,
1764 'preference': preference,
1766 resolution = last_stream_inf.get('RESOLUTION')
1768 mobj = re.search(r'(?P<width>\d+)[xX](?P<height>\d+)', resolution)
1770 f['width'] = int(mobj.group('width'))
1771 f['height'] = int(mobj.group('height'))
1772 # Unified Streaming Platform
1774 r'audio.*?(?:%3D|=)(\d+)(?:-video.*?(?:%3D|=)(\d+))?', f['url'])
1776 abr, vbr = mobj.groups()
1777 abr, vbr = float_or_none(abr, 1000), float_or_none(vbr, 1000)
1782 codecs = parse_codecs(last_stream_inf.get('CODECS'))
1784 audio_group_id = last_stream_inf.get('AUDIO')
1785 # As per [1, 4.3.4.1.1] any EXT-X-STREAM-INF tag which
1786 # references a rendition group MUST have a CODECS attribute.
1787 # However, this is not always respected, for example, [2]
1788 # contains EXT-X-STREAM-INF tag which references AUDIO
1789 # rendition group but does not have CODECS and despite
1790 # referencing an audio group it represents a complete
1791 # (with audio and video) format. So, for such cases we will
1792 # ignore references to rendition groups and treat them
1793 # as complete formats.
1794 if audio_group_id and codecs and f.get('vcodec') != 'none':
1795 audio_group = groups.get(audio_group_id)
1796 if audio_group and audio_group[0].get('URI'):
1797 # TODO: update acodec for audio only formats with
1799 f['acodec'] = 'none'
1803 progressive_uri = last_stream_inf.get('PROGRESSIVE-URI')
1806 del http_f['manifest_url']
1808 'format_id': f['format_id'].replace('hls-', 'http-'),
1810 'url': progressive_uri,
1812 formats.append(http_f)
1814 last_stream_inf = {}
1818 def _xpath_ns(path, namespace=None):
1822 for c in path.split('/'):
1823 if not c or c == '.':
1826 out.append('{%s}%s' % (namespace, c))
1827 return '/'.join(out)
1829 def _extract_smil_formats(self, smil_url, video_id, fatal=True, f4m_params=None, transform_source=None):
1830 smil = self._download_smil(smil_url, video_id, fatal=fatal, transform_source=transform_source)
1836 namespace = self._parse_smil_namespace(smil)
1838 return self._parse_smil_formats(
1839 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
1841 def _extract_smil_info(self, smil_url, video_id, fatal=True, f4m_params=None):
1842 smil = self._download_smil(smil_url, video_id, fatal=fatal)
1845 return self._parse_smil(smil, smil_url, video_id, f4m_params=f4m_params)
1847 def _download_smil(self, smil_url, video_id, fatal=True, transform_source=None):
1848 return self._download_xml(
1849 smil_url, video_id, 'Downloading SMIL file',
1850 'Unable to download SMIL file', fatal=fatal, transform_source=transform_source)
1852 def _parse_smil(self, smil, smil_url, video_id, f4m_params=None):
1853 namespace = self._parse_smil_namespace(smil)
1855 formats = self._parse_smil_formats(
1856 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
1857 subtitles = self._parse_smil_subtitles(smil, namespace=namespace)
1859 video_id = os.path.splitext(url_basename(smil_url))[0]
1863 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
1864 name = meta.attrib.get('name')
1865 content = meta.attrib.get('content')
1866 if not name or not content:
1868 if not title and name == 'title':
1870 elif not description and name in ('description', 'abstract'):
1871 description = content
1872 elif not upload_date and name == 'date':
1873 upload_date = unified_strdate(content)
1876 'id': image.get('type'),
1877 'url': image.get('src'),
1878 'width': int_or_none(image.get('width')),
1879 'height': int_or_none(image.get('height')),
1880 } for image in smil.findall(self._xpath_ns('.//image', namespace)) if image.get('src')]
1884 'title': title or video_id,
1885 'description': description,
1886 'upload_date': upload_date,
1887 'thumbnails': thumbnails,
1889 'subtitles': subtitles,
1892 def _parse_smil_namespace(self, smil):
1893 return self._search_regex(
1894 r'(?i)^{([^}]+)?}smil$', smil.tag, 'namespace', default=None)
1896 def _parse_smil_formats(self, smil, smil_url, video_id, namespace=None, f4m_params=None, transform_rtmp_url=None):
1898 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
1899 b = meta.get('base') or meta.get('httpBase')
1910 media = smil.findall(self._xpath_ns('.//video', namespace)) + smil.findall(self._xpath_ns('.//audio', namespace))
1911 for medium in media:
1912 src = medium.get('src')
1913 if not src or src in srcs:
1917 bitrate = float_or_none(medium.get('system-bitrate') or medium.get('systemBitrate'), 1000)
1918 filesize = int_or_none(medium.get('size') or medium.get('fileSize'))
1919 width = int_or_none(medium.get('width'))
1920 height = int_or_none(medium.get('height'))
1921 proto = medium.get('proto')
1922 ext = medium.get('ext')
1923 src_ext = determine_ext(src)
1924 streamer = medium.get('streamer') or base
1926 if proto == 'rtmp' or streamer.startswith('rtmp'):
1932 'format_id': 'rtmp-%d' % (rtmp_count if bitrate is None else bitrate),
1934 'filesize': filesize,
1938 if transform_rtmp_url:
1939 streamer, src = transform_rtmp_url(streamer, src)
1940 formats[-1].update({
1946 src_url = src if src.startswith('http') else compat_urlparse.urljoin(base, src)
1947 src_url = src_url.strip()
1949 if proto == 'm3u8' or src_ext == 'm3u8':
1950 m3u8_formats = self._extract_m3u8_formats(
1951 src_url, video_id, ext or 'mp4', m3u8_id='hls', fatal=False)
1952 if len(m3u8_formats) == 1:
1954 m3u8_formats[0].update({
1955 'format_id': 'hls-%d' % (m3u8_count if bitrate is None else bitrate),
1960 formats.extend(m3u8_formats)
1961 elif src_ext == 'f4m':
1966 'plugin': 'flowplayer-3.2.0.1',
1968 f4m_url += '&' if '?' in f4m_url else '?'
1969 f4m_url += compat_urllib_parse_urlencode(f4m_params)
1970 formats.extend(self._extract_f4m_formats(f4m_url, video_id, f4m_id='hds', fatal=False))
1971 elif src_ext == 'mpd':
1972 formats.extend(self._extract_mpd_formats(
1973 src_url, video_id, mpd_id='dash', fatal=False))
1974 elif re.search(r'\.ism/[Mm]anifest', src_url):
1975 formats.extend(self._extract_ism_formats(
1976 src_url, video_id, ism_id='mss', fatal=False))
1977 elif src_url.startswith('http') and self._is_valid_url(src, video_id):
1981 'ext': ext or src_ext or 'flv',
1982 'format_id': 'http-%d' % (bitrate or http_count),
1984 'filesize': filesize,
1991 def _parse_smil_subtitles(self, smil, namespace=None, subtitles_lang='en'):
1994 for num, textstream in enumerate(smil.findall(self._xpath_ns('.//textstream', namespace))):
1995 src = textstream.get('src')
1996 if not src or src in urls:
1999 ext = textstream.get('ext') or mimetype2ext(textstream.get('type')) or determine_ext(src)
2000 lang = textstream.get('systemLanguage') or textstream.get('systemLanguageName') or textstream.get('lang') or subtitles_lang
2001 subtitles.setdefault(lang, []).append({
2007 def _extract_xspf_playlist(self, xspf_url, playlist_id, fatal=True):
2008 xspf = self._download_xml(
2009 xspf_url, playlist_id, 'Downloading xpsf playlist',
2010 'Unable to download xspf manifest', fatal=fatal)
2013 return self._parse_xspf(
2014 xspf, playlist_id, xspf_url=xspf_url,
2015 xspf_base_url=base_url(xspf_url))
2017 def _parse_xspf(self, xspf_doc, playlist_id, xspf_url=None, xspf_base_url=None):
2019 'xspf': 'http://xspf.org/ns/0/',
2020 's1': 'http://static.streamone.nl/player/ns/0',
2024 for track in xspf_doc.findall(xpath_with_ns('./xspf:trackList/xspf:track', NS_MAP)):
2026 track, xpath_with_ns('./xspf:title', NS_MAP), 'title', default=playlist_id)
2027 description = xpath_text(
2028 track, xpath_with_ns('./xspf:annotation', NS_MAP), 'description')
2029 thumbnail = xpath_text(
2030 track, xpath_with_ns('./xspf:image', NS_MAP), 'thumbnail')
2031 duration = float_or_none(
2032 xpath_text(track, xpath_with_ns('./xspf:duration', NS_MAP), 'duration'), 1000)
2035 for location in track.findall(xpath_with_ns('./xspf:location', NS_MAP)):
2036 format_url = urljoin(xspf_base_url, location.text)
2041 'manifest_url': xspf_url,
2042 'format_id': location.get(xpath_with_ns('s1:label', NS_MAP)),
2043 'width': int_or_none(location.get(xpath_with_ns('s1:width', NS_MAP))),
2044 'height': int_or_none(location.get(xpath_with_ns('s1:height', NS_MAP))),
2046 self._sort_formats(formats)
2051 'description': description,
2052 'thumbnail': thumbnail,
2053 'duration': duration,
2058 def _extract_mpd_formats(self, mpd_url, video_id, mpd_id=None, note=None, errnote=None, fatal=True, formats_dict={}, data=None, headers={}, query={}):
2059 res = self._download_xml_handle(
2061 note=note or 'Downloading MPD manifest',
2062 errnote=errnote or 'Failed to download MPD manifest',
2063 fatal=fatal, data=data, headers=headers, query=query)
2069 mpd_base_url = base_url(urlh.geturl())
2071 return self._parse_mpd_formats(
2072 mpd_doc, mpd_id=mpd_id, mpd_base_url=mpd_base_url,
2073 formats_dict=formats_dict, mpd_url=mpd_url)
2075 def _parse_mpd_formats(self, mpd_doc, mpd_id=None, mpd_base_url='', formats_dict={}, mpd_url=None):
2077 Parse formats from MPD manifest.
2079 1. MPEG-DASH Standard, ISO/IEC 23009-1:2014(E),
2080 http://standards.iso.org/ittf/PubliclyAvailableStandards/c065274_ISO_IEC_23009-1_2014.zip
2081 2. https://en.wikipedia.org/wiki/Dynamic_Adaptive_Streaming_over_HTTP
2083 if mpd_doc.get('type') == 'dynamic':
2086 namespace = self._search_regex(r'(?i)^{([^}]+)?}MPD$', mpd_doc.tag, 'namespace', default=None)
2089 return self._xpath_ns(path, namespace)
2091 def is_drm_protected(element):
2092 return element.find(_add_ns('ContentProtection')) is not None
2094 def extract_multisegment_info(element, ms_parent_info):
2095 ms_info = ms_parent_info.copy()
2097 # As per [1, 5.3.9.2.2] SegmentList and SegmentTemplate share some
2098 # common attributes and elements. We will only extract relevant
2100 def extract_common(source):
2101 segment_timeline = source.find(_add_ns('SegmentTimeline'))
2102 if segment_timeline is not None:
2103 s_e = segment_timeline.findall(_add_ns('S'))
2105 ms_info['total_number'] = 0
2108 r = int(s.get('r', 0))
2109 ms_info['total_number'] += 1 + r
2110 ms_info['s'].append({
2111 't': int(s.get('t', 0)),
2112 # @d is mandatory (see [1, 5.3.9.6.2, Table 17, page 60])
2113 'd': int(s.attrib['d']),
2116 start_number = source.get('startNumber')
2118 ms_info['start_number'] = int(start_number)
2119 timescale = source.get('timescale')
2121 ms_info['timescale'] = int(timescale)
2122 segment_duration = source.get('duration')
2123 if segment_duration:
2124 ms_info['segment_duration'] = float(segment_duration)
2126 def extract_Initialization(source):
2127 initialization = source.find(_add_ns('Initialization'))
2128 if initialization is not None:
2129 ms_info['initialization_url'] = initialization.attrib['sourceURL']
2131 segment_list = element.find(_add_ns('SegmentList'))
2132 if segment_list is not None:
2133 extract_common(segment_list)
2134 extract_Initialization(segment_list)
2135 segment_urls_e = segment_list.findall(_add_ns('SegmentURL'))
2137 ms_info['segment_urls'] = [segment.attrib['media'] for segment in segment_urls_e]
2139 segment_template = element.find(_add_ns('SegmentTemplate'))
2140 if segment_template is not None:
2141 extract_common(segment_template)
2142 media = segment_template.get('media')
2144 ms_info['media'] = media
2145 initialization = segment_template.get('initialization')
2147 ms_info['initialization'] = initialization
2149 extract_Initialization(segment_template)
2152 mpd_duration = parse_duration(mpd_doc.get('mediaPresentationDuration'))
2154 for period in mpd_doc.findall(_add_ns('Period')):
2155 period_duration = parse_duration(period.get('duration')) or mpd_duration
2156 period_ms_info = extract_multisegment_info(period, {
2160 for adaptation_set in period.findall(_add_ns('AdaptationSet')):
2161 if is_drm_protected(adaptation_set):
2163 adaption_set_ms_info = extract_multisegment_info(adaptation_set, period_ms_info)
2164 for representation in adaptation_set.findall(_add_ns('Representation')):
2165 if is_drm_protected(representation):
2167 representation_attrib = adaptation_set.attrib.copy()
2168 representation_attrib.update(representation.attrib)
2169 # According to [1, 5.3.7.2, Table 9, page 41], @mimeType is mandatory
2170 mime_type = representation_attrib['mimeType']
2171 content_type = mime_type.split('/')[0]
2172 if content_type == 'text':
2173 # TODO implement WebVTT downloading
2175 elif content_type in ('video', 'audio'):
2177 for element in (representation, adaptation_set, period, mpd_doc):
2178 base_url_e = element.find(_add_ns('BaseURL'))
2179 if base_url_e is not None:
2180 base_url = base_url_e.text + base_url
2181 if re.match(r'^https?://', base_url):
2183 if mpd_base_url and not re.match(r'^https?://', base_url):
2184 if not mpd_base_url.endswith('/') and not base_url.startswith('/'):
2186 base_url = mpd_base_url + base_url
2187 representation_id = representation_attrib.get('id')
2188 lang = representation_attrib.get('lang')
2189 url_el = representation.find(_add_ns('BaseURL'))
2190 filesize = int_or_none(url_el.attrib.get('{http://youtube.com/yt/2012/10/10}contentLength') if url_el is not None else None)
2191 bandwidth = int_or_none(representation_attrib.get('bandwidth'))
2193 'format_id': '%s-%s' % (mpd_id, representation_id) if mpd_id else representation_id,
2194 'manifest_url': mpd_url,
2195 'ext': mimetype2ext(mime_type),
2196 'width': int_or_none(representation_attrib.get('width')),
2197 'height': int_or_none(representation_attrib.get('height')),
2198 'tbr': float_or_none(bandwidth, 1000),
2199 'asr': int_or_none(representation_attrib.get('audioSamplingRate')),
2200 'fps': int_or_none(representation_attrib.get('frameRate')),
2201 'language': lang if lang not in ('mul', 'und', 'zxx', 'mis') else None,
2202 'format_note': 'DASH %s' % content_type,
2203 'filesize': filesize,
2204 'container': mimetype2ext(mime_type) + '_dash',
2206 f.update(parse_codecs(representation_attrib.get('codecs')))
2207 representation_ms_info = extract_multisegment_info(representation, adaption_set_ms_info)
2209 def prepare_template(template_name, identifiers):
2210 tmpl = representation_ms_info[template_name]
2211 # First of, % characters outside $...$ templates
2212 # must be escaped by doubling for proper processing
2213 # by % operator string formatting used further (see
2214 # https://github.com/ytdl-org/youtube-dl/issues/16867).
2220 in_template = not in_template
2221 elif c == '%' and not in_template:
2223 # Next, $...$ templates are translated to their
2224 # %(...) counterparts to be used with % operator
2225 t = t.replace('$RepresentationID$', representation_id)
2226 t = re.sub(r'\$(%s)\$' % '|'.join(identifiers), r'%(\1)d', t)
2227 t = re.sub(r'\$(%s)%%([^$]+)\$' % '|'.join(identifiers), r'%(\1)\2', t)
2228 t.replace('$$', '$')
2231 # @initialization is a regular template like @media one
2232 # so it should be handled just the same way (see
2233 # https://github.com/ytdl-org/youtube-dl/issues/11605)
2234 if 'initialization' in representation_ms_info:
2235 initialization_template = prepare_template(
2237 # As per [1, 5.3.9.4.2, Table 15, page 54] $Number$ and
2238 # $Time$ shall not be included for @initialization thus
2239 # only $Bandwidth$ remains
2241 representation_ms_info['initialization_url'] = initialization_template % {
2242 'Bandwidth': bandwidth,
2245 def location_key(location):
2246 return 'url' if re.match(r'^https?://', location) else 'path'
2248 if 'segment_urls' not in representation_ms_info and 'media' in representation_ms_info:
2250 media_template = prepare_template('media', ('Number', 'Bandwidth', 'Time'))
2251 media_location_key = location_key(media_template)
2253 # As per [1, 5.3.9.4.4, Table 16, page 55] $Number$ and $Time$
2254 # can't be used at the same time
2255 if '%(Number' in media_template and 's' not in representation_ms_info:
2256 segment_duration = None
2257 if 'total_number' not in representation_ms_info and 'segment_duration' in representation_ms_info:
2258 segment_duration = float_or_none(representation_ms_info['segment_duration'], representation_ms_info['timescale'])
2259 representation_ms_info['total_number'] = int(math.ceil(float(period_duration) / segment_duration))
2260 representation_ms_info['fragments'] = [{
2261 media_location_key: media_template % {
2262 'Number': segment_number,
2263 'Bandwidth': bandwidth,
2265 'duration': segment_duration,
2266 } for segment_number in range(
2267 representation_ms_info['start_number'],
2268 representation_ms_info['total_number'] + representation_ms_info['start_number'])]
2270 # $Number*$ or $Time$ in media template with S list available
2271 # Example $Number*$: http://www.svtplay.se/klipp/9023742/stopptid-om-bjorn-borg
2272 # Example $Time$: https://play.arkena.com/embed/avp/v2/player/media/b41dda37-d8e7-4d3f-b1b5-9a9db578bdfe/1/129411
2273 representation_ms_info['fragments'] = []
2276 segment_number = representation_ms_info['start_number']
2278 def add_segment_url():
2279 segment_url = media_template % {
2280 'Time': segment_time,
2281 'Bandwidth': bandwidth,
2282 'Number': segment_number,
2284 representation_ms_info['fragments'].append({
2285 media_location_key: segment_url,
2286 'duration': float_or_none(segment_d, representation_ms_info['timescale']),
2289 for num, s in enumerate(representation_ms_info['s']):
2290 segment_time = s.get('t') or segment_time
2294 for r in range(s.get('r', 0)):
2295 segment_time += segment_d
2298 segment_time += segment_d
2299 elif 'segment_urls' in representation_ms_info and 's' in representation_ms_info:
2301 # Example: https://www.youtube.com/watch?v=iXZV5uAYMJI
2302 # or any YouTube dashsegments video
2305 timescale = representation_ms_info['timescale']
2306 for s in representation_ms_info['s']:
2307 duration = float_or_none(s['d'], timescale)
2308 for r in range(s.get('r', 0) + 1):
2309 segment_uri = representation_ms_info['segment_urls'][segment_index]
2311 location_key(segment_uri): segment_uri,
2312 'duration': duration,
2315 representation_ms_info['fragments'] = fragments
2316 elif 'segment_urls' in representation_ms_info:
2317 # Segment URLs with no SegmentTimeline
2318 # Example: https://www.seznam.cz/zpravy/clanek/cesko-zasahne-vitr-o-sile-vichrice-muze-byt-i-zivotu-nebezpecny-39091
2319 # https://github.com/ytdl-org/youtube-dl/pull/14844
2321 segment_duration = float_or_none(
2322 representation_ms_info['segment_duration'],
2323 representation_ms_info['timescale']) if 'segment_duration' in representation_ms_info else None
2324 for segment_url in representation_ms_info['segment_urls']:
2326 location_key(segment_url): segment_url,
2328 if segment_duration:
2329 fragment['duration'] = segment_duration
2330 fragments.append(fragment)
2331 representation_ms_info['fragments'] = fragments
2332 # If there is a fragments key available then we correctly recognized fragmented media.
2333 # Otherwise we will assume unfragmented media with direct access. Technically, such
2334 # assumption is not necessarily correct since we may simply have no support for
2335 # some forms of fragmented media renditions yet, but for now we'll use this fallback.
2336 if 'fragments' in representation_ms_info:
2338 # NB: mpd_url may be empty when MPD manifest is parsed from a string
2339 'url': mpd_url or base_url,
2340 'fragment_base_url': base_url,
2342 'protocol': 'http_dash_segments',
2344 if 'initialization_url' in representation_ms_info:
2345 initialization_url = representation_ms_info['initialization_url']
2346 if not f.get('url'):
2347 f['url'] = initialization_url
2348 f['fragments'].append({location_key(initialization_url): initialization_url})
2349 f['fragments'].extend(representation_ms_info['fragments'])
2351 # Assuming direct URL to unfragmented media.
2354 # According to [1, 5.3.5.2, Table 7, page 35] @id of Representation
2355 # is not necessarily unique within a Period thus formats with
2356 # the same `format_id` are quite possible. There are numerous examples
2357 # of such manifests (see https://github.com/ytdl-org/youtube-dl/issues/15111,
2358 # https://github.com/ytdl-org/youtube-dl/issues/13919)
2359 full_info = formats_dict.get(representation_id, {}).copy()
2361 formats.append(full_info)
2363 self.report_warning('Unknown MIME type %s in DASH manifest' % mime_type)
2366 def _extract_ism_formats(self, ism_url, video_id, ism_id=None, note=None, errnote=None, fatal=True, data=None, headers={}, query={}):
2367 res = self._download_xml_handle(
2369 note=note or 'Downloading ISM manifest',
2370 errnote=errnote or 'Failed to download ISM manifest',
2371 fatal=fatal, data=data, headers=headers, query=query)
2378 return self._parse_ism_formats(ism_doc, urlh.geturl(), ism_id)
2380 def _parse_ism_formats(self, ism_doc, ism_url, ism_id=None):
2382 Parse formats from ISM manifest.
2384 1. [MS-SSTR]: Smooth Streaming Protocol,
2385 https://msdn.microsoft.com/en-us/library/ff469518.aspx
2387 if ism_doc.get('IsLive') == 'TRUE' or ism_doc.find('Protection') is not None:
2390 duration = int(ism_doc.attrib['Duration'])
2391 timescale = int_or_none(ism_doc.get('TimeScale')) or 10000000
2394 for stream in ism_doc.findall('StreamIndex'):
2395 stream_type = stream.get('Type')
2396 if stream_type not in ('video', 'audio'):
2398 url_pattern = stream.attrib['Url']
2399 stream_timescale = int_or_none(stream.get('TimeScale')) or timescale
2400 stream_name = stream.get('Name')
2401 for track in stream.findall('QualityLevel'):
2402 fourcc = track.get('FourCC', 'AACL' if track.get('AudioTag') == '255' else None)
2403 # TODO: add support for WVC1 and WMAP
2404 if fourcc not in ('H264', 'AVC1', 'AACL'):
2405 self.report_warning('%s is not a supported codec' % fourcc)
2407 tbr = int(track.attrib['Bitrate']) // 1000
2408 # [1] does not mention Width and Height attributes. However,
2409 # they're often present while MaxWidth and MaxHeight are
2410 # missing, so should be used as fallbacks
2411 width = int_or_none(track.get('MaxWidth') or track.get('Width'))
2412 height = int_or_none(track.get('MaxHeight') or track.get('Height'))
2413 sampling_rate = int_or_none(track.get('SamplingRate'))
2415 track_url_pattern = re.sub(r'{[Bb]itrate}', track.attrib['Bitrate'], url_pattern)
2416 track_url_pattern = compat_urlparse.urljoin(ism_url, track_url_pattern)
2422 stream_fragments = stream.findall('c')
2423 for stream_fragment_index, stream_fragment in enumerate(stream_fragments):
2424 fragment_ctx['time'] = int_or_none(stream_fragment.get('t')) or fragment_ctx['time']
2425 fragment_repeat = int_or_none(stream_fragment.get('r')) or 1
2426 fragment_ctx['duration'] = int_or_none(stream_fragment.get('d'))
2427 if not fragment_ctx['duration']:
2429 next_fragment_time = int(stream_fragment[stream_fragment_index + 1].attrib['t'])
2431 next_fragment_time = duration
2432 fragment_ctx['duration'] = (next_fragment_time - fragment_ctx['time']) / fragment_repeat
2433 for _ in range(fragment_repeat):
2435 'url': re.sub(r'{start[ _]time}', compat_str(fragment_ctx['time']), track_url_pattern),
2436 'duration': fragment_ctx['duration'] / stream_timescale,
2438 fragment_ctx['time'] += fragment_ctx['duration']
2442 format_id.append(ism_id)
2444 format_id.append(stream_name)
2445 format_id.append(compat_str(tbr))
2448 'format_id': '-'.join(format_id),
2450 'manifest_url': ism_url,
2451 'ext': 'ismv' if stream_type == 'video' else 'isma',
2455 'asr': sampling_rate,
2456 'vcodec': 'none' if stream_type == 'audio' else fourcc,
2457 'acodec': 'none' if stream_type == 'video' else fourcc,
2459 'fragments': fragments,
2460 '_download_params': {
2461 'duration': duration,
2462 'timescale': stream_timescale,
2463 'width': width or 0,
2464 'height': height or 0,
2466 'codec_private_data': track.get('CodecPrivateData'),
2467 'sampling_rate': sampling_rate,
2468 'channels': int_or_none(track.get('Channels', 2)),
2469 'bits_per_sample': int_or_none(track.get('BitsPerSample', 16)),
2470 'nal_unit_length_field': int_or_none(track.get('NALUnitLengthField', 4)),
2475 def _parse_html5_media_entries(self, base_url, webpage, video_id, m3u8_id=None, m3u8_entry_protocol='m3u8', mpd_id=None, preference=None):
2476 def absolute_url(item_url):
2477 return urljoin(base_url, item_url)
2479 def parse_content_type(content_type):
2480 if not content_type:
2482 ctr = re.search(r'(?P<mimetype>[^/]+/[^;]+)(?:;\s*codecs="?(?P<codecs>[^"]+))?', content_type)
2484 mimetype, codecs = ctr.groups()
2485 f = parse_codecs(codecs)
2486 f['ext'] = mimetype2ext(mimetype)
2490 def _media_formats(src, cur_media_type, type_info={}):
2491 full_url = absolute_url(src)
2492 ext = type_info.get('ext') or determine_ext(full_url)
2494 is_plain_url = False
2495 formats = self._extract_m3u8_formats(
2496 full_url, video_id, ext='mp4',
2497 entry_protocol=m3u8_entry_protocol, m3u8_id=m3u8_id,
2498 preference=preference, fatal=False)
2500 is_plain_url = False
2501 formats = self._extract_mpd_formats(
2502 full_url, video_id, mpd_id=mpd_id, fatal=False)
2507 'vcodec': 'none' if cur_media_type == 'audio' else None,
2509 return is_plain_url, formats
2512 # amp-video and amp-audio are very similar to their HTML5 counterparts
2513 # so we wll include them right here (see
2514 # https://www.ampproject.org/docs/reference/components/amp-video)
2515 media_tags = [(media_tag, media_type, '')
2516 for media_tag, media_type
2517 in re.findall(r'(?s)(<(?:amp-)?(video|audio)[^>]*/>)', webpage)]
2518 media_tags.extend(re.findall(
2519 # We only allow video|audio followed by a whitespace or '>'.
2520 # Allowing more characters may end up in significant slow down (see
2521 # https://github.com/ytdl-org/youtube-dl/issues/11979, example URL:
2522 # http://www.porntrex.com/maps/videositemap.xml).
2523 r'(?s)(<(?P<tag>(?:amp-)?(?:video|audio))(?:\s+[^>]*)?>)(.*?)</(?P=tag)>', webpage))
2524 for media_tag, media_type, media_content in media_tags:
2529 media_attributes = extract_attributes(media_tag)
2530 src = strip_or_none(media_attributes.get('src'))
2532 _, formats = _media_formats(src, media_type)
2533 media_info['formats'].extend(formats)
2534 media_info['thumbnail'] = absolute_url(media_attributes.get('poster'))
2536 for source_tag in re.findall(r'<source[^>]+>', media_content):
2537 s_attr = extract_attributes(source_tag)
2538 # data-video-src and data-src are non standard but seen
2539 # several times in the wild
2540 src = strip_or_none(dict_get(s_attr, ('src', 'data-video-src', 'data-src')))
2543 f = parse_content_type(s_attr.get('type'))
2544 is_plain_url, formats = _media_formats(src, media_type, f)
2546 # width, height, res, label and title attributes are
2547 # all not standard but seen several times in the wild
2550 for lbl in ('label', 'title')
2551 if str_or_none(s_attr.get(lbl))
2553 width = int_or_none(s_attr.get('width'))
2554 height = (int_or_none(s_attr.get('height'))
2555 or int_or_none(s_attr.get('res')))
2556 if not width or not height:
2558 resolution = parse_resolution(lbl)
2561 width = width or resolution.get('width')
2562 height = height or resolution.get('height')
2564 tbr = parse_bitrate(lbl)
2573 'format_id': s_attr.get('label') or s_attr.get('title'),
2575 f.update(formats[0])
2576 media_info['formats'].append(f)
2578 media_info['formats'].extend(formats)
2579 for track_tag in re.findall(r'<track[^>]+>', media_content):
2580 track_attributes = extract_attributes(track_tag)
2581 kind = track_attributes.get('kind')
2582 if not kind or kind in ('subtitles', 'captions'):
2583 src = strip_or_none(track_attributes.get('src'))
2586 lang = track_attributes.get('srclang') or track_attributes.get('lang') or track_attributes.get('label')
2587 media_info['subtitles'].setdefault(lang, []).append({
2588 'url': absolute_url(src),
2590 for f in media_info['formats']:
2591 f.setdefault('http_headers', {})['Referer'] = base_url
2592 if media_info['formats'] or media_info['subtitles']:
2593 entries.append(media_info)
2596 def _extract_akamai_formats(self, manifest_url, video_id, hosts={}):
2598 hdcore_sign = 'hdcore=3.7.0'
2599 f4m_url = re.sub(r'(https?://[^/]+)/i/', r'\1/z/', manifest_url).replace('/master.m3u8', '/manifest.f4m')
2600 hds_host = hosts.get('hds')
2602 f4m_url = re.sub(r'(https?://)[^/]+', r'\1' + hds_host, f4m_url)
2603 if 'hdcore=' not in f4m_url:
2604 f4m_url += ('&' if '?' in f4m_url else '?') + hdcore_sign
2605 f4m_formats = self._extract_f4m_formats(
2606 f4m_url, video_id, f4m_id='hds', fatal=False)
2607 for entry in f4m_formats:
2608 entry.update({'extra_param_to_segment_url': hdcore_sign})
2609 formats.extend(f4m_formats)
2610 m3u8_url = re.sub(r'(https?://[^/]+)/z/', r'\1/i/', manifest_url).replace('/manifest.f4m', '/master.m3u8')
2611 hls_host = hosts.get('hls')
2613 m3u8_url = re.sub(r'(https?://)[^/]+', r'\1' + hls_host, m3u8_url)
2614 formats.extend(self._extract_m3u8_formats(
2615 m3u8_url, video_id, 'mp4', 'm3u8_native',
2616 m3u8_id='hls', fatal=False))
2619 def _extract_wowza_formats(self, url, video_id, m3u8_entry_protocol='m3u8_native', skip_protocols=[]):
2620 query = compat_urlparse.urlparse(url).query
2621 url = re.sub(r'/(?:manifest|playlist|jwplayer)\.(?:m3u8|f4m|mpd|smil)', '', url)
2623 r'(?:(?:http|rtmp|rtsp)(?P<s>s)?:)?(?P<url>//[^?]+)', url)
2624 url_base = mobj.group('url')
2625 http_base_url = '%s%s:%s' % ('http', mobj.group('s') or '', url_base)
2628 def manifest_url(manifest):
2629 m_url = '%s/%s' % (http_base_url, manifest)
2631 m_url += '?%s' % query
2634 if 'm3u8' not in skip_protocols:
2635 formats.extend(self._extract_m3u8_formats(
2636 manifest_url('playlist.m3u8'), video_id, 'mp4',
2637 m3u8_entry_protocol, m3u8_id='hls', fatal=False))
2638 if 'f4m' not in skip_protocols:
2639 formats.extend(self._extract_f4m_formats(
2640 manifest_url('manifest.f4m'),
2641 video_id, f4m_id='hds', fatal=False))
2642 if 'dash' not in skip_protocols:
2643 formats.extend(self._extract_mpd_formats(
2644 manifest_url('manifest.mpd'),
2645 video_id, mpd_id='dash', fatal=False))
2646 if re.search(r'(?:/smil:|\.smil)', url_base):
2647 if 'smil' not in skip_protocols:
2648 rtmp_formats = self._extract_smil_formats(
2649 manifest_url('jwplayer.smil'),
2650 video_id, fatal=False)
2651 for rtmp_format in rtmp_formats:
2652 rtsp_format = rtmp_format.copy()
2653 rtsp_format['url'] = '%s/%s' % (rtmp_format['url'], rtmp_format['play_path'])
2654 del rtsp_format['play_path']
2655 del rtsp_format['ext']
2656 rtsp_format.update({
2657 'url': rtsp_format['url'].replace('rtmp://', 'rtsp://'),
2658 'format_id': rtmp_format['format_id'].replace('rtmp', 'rtsp'),
2661 formats.extend([rtmp_format, rtsp_format])
2663 for protocol in ('rtmp', 'rtsp'):
2664 if protocol not in skip_protocols:
2666 'url': '%s:%s' % (protocol, url_base),
2667 'format_id': protocol,
2668 'protocol': protocol,
2672 def _find_jwplayer_data(self, webpage, video_id=None, transform_source=js_to_json):
2674 r'(?s)jwplayer\((?P<quote>[\'"])[^\'" ]+(?P=quote)\)(?!</script>).*?\.setup\s*\((?P<options>[^)]+)\)',
2678 jwplayer_data = self._parse_json(mobj.group('options'),
2680 transform_source=transform_source)
2681 except ExtractorError:
2684 if isinstance(jwplayer_data, dict):
2685 return jwplayer_data
2687 def _extract_jwplayer_data(self, webpage, video_id, *args, **kwargs):
2688 jwplayer_data = self._find_jwplayer_data(
2689 webpage, video_id, transform_source=js_to_json)
2690 return self._parse_jwplayer_data(
2691 jwplayer_data, video_id, *args, **kwargs)
2693 def _parse_jwplayer_data(self, jwplayer_data, video_id=None, require_title=True,
2694 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
2695 # JWPlayer backward compatibility: flattened playlists
2696 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/api/config.js#L81-L96
2697 if 'playlist' not in jwplayer_data:
2698 jwplayer_data = {'playlist': [jwplayer_data]}
2702 # JWPlayer backward compatibility: single playlist item
2703 # https://github.com/jwplayer/jwplayer/blob/v7.7.0/src/js/playlist/playlist.js#L10
2704 if not isinstance(jwplayer_data['playlist'], list):
2705 jwplayer_data['playlist'] = [jwplayer_data['playlist']]
2707 for video_data in jwplayer_data['playlist']:
2708 # JWPlayer backward compatibility: flattened sources
2709 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/playlist/item.js#L29-L35
2710 if 'sources' not in video_data:
2711 video_data['sources'] = [video_data]
2713 this_video_id = video_id or video_data['mediaid']
2715 formats = self._parse_jwplayer_formats(
2716 video_data['sources'], video_id=this_video_id, m3u8_id=m3u8_id,
2717 mpd_id=mpd_id, rtmp_params=rtmp_params, base_url=base_url)
2720 tracks = video_data.get('tracks')
2721 if tracks and isinstance(tracks, list):
2722 for track in tracks:
2723 if not isinstance(track, dict):
2725 track_kind = track.get('kind')
2726 if not track_kind or not isinstance(track_kind, compat_str):
2728 if track_kind.lower() not in ('captions', 'subtitles'):
2730 track_url = urljoin(base_url, track.get('file'))
2733 subtitles.setdefault(track.get('label') or 'en', []).append({
2734 'url': self._proto_relative_url(track_url)
2738 'id': this_video_id,
2739 'title': unescapeHTML(video_data['title'] if require_title else video_data.get('title')),
2740 'description': clean_html(video_data.get('description')),
2741 'thumbnail': urljoin(base_url, self._proto_relative_url(video_data.get('image'))),
2742 'timestamp': int_or_none(video_data.get('pubdate')),
2743 'duration': float_or_none(jwplayer_data.get('duration') or video_data.get('duration')),
2744 'subtitles': subtitles,
2746 # https://github.com/jwplayer/jwplayer/blob/master/src/js/utils/validator.js#L32
2747 if len(formats) == 1 and re.search(r'^(?:http|//).*(?:youtube\.com|youtu\.be)/.+', formats[0]['url']):
2749 '_type': 'url_transparent',
2750 'url': formats[0]['url'],
2753 self._sort_formats(formats)
2754 entry['formats'] = formats
2755 entries.append(entry)
2756 if len(entries) == 1:
2759 return self.playlist_result(entries)
2761 def _parse_jwplayer_formats(self, jwplayer_sources_data, video_id=None,
2762 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
2765 for source in jwplayer_sources_data:
2766 if not isinstance(source, dict):
2768 source_url = urljoin(
2769 base_url, self._proto_relative_url(source.get('file')))
2770 if not source_url or source_url in urls:
2772 urls.append(source_url)
2773 source_type = source.get('type') or ''
2774 ext = mimetype2ext(source_type) or determine_ext(source_url)
2775 if source_type == 'hls' or ext == 'm3u8':
2776 formats.extend(self._extract_m3u8_formats(
2777 source_url, video_id, 'mp4', entry_protocol='m3u8_native',
2778 m3u8_id=m3u8_id, fatal=False))
2779 elif source_type == 'dash' or ext == 'mpd':
2780 formats.extend(self._extract_mpd_formats(
2781 source_url, video_id, mpd_id=mpd_id, fatal=False))
2783 formats.extend(self._extract_smil_formats(
2784 source_url, video_id, fatal=False))
2785 # https://github.com/jwplayer/jwplayer/blob/master/src/js/providers/default.js#L67
2786 elif source_type.startswith('audio') or ext in (
2787 'oga', 'aac', 'mp3', 'mpeg', 'vorbis'):
2794 height = int_or_none(source.get('height'))
2796 # Often no height is provided but there is a label in
2797 # format like "1080p", "720p SD", or 1080.
2798 height = int_or_none(self._search_regex(
2799 r'^(\d{3,4})[pP]?(?:\b|$)', compat_str(source.get('label') or ''),
2800 'height', default=None))
2803 'width': int_or_none(source.get('width')),
2805 'tbr': int_or_none(source.get('bitrate')),
2808 if source_url.startswith('rtmp'):
2809 a_format['ext'] = 'flv'
2810 # See com/longtailvideo/jwplayer/media/RTMPMediaProvider.as
2811 # of jwplayer.flash.swf
2812 rtmp_url_parts = re.split(
2813 r'((?:mp4|mp3|flv):)', source_url, 1)
2814 if len(rtmp_url_parts) == 3:
2815 rtmp_url, prefix, play_path = rtmp_url_parts
2818 'play_path': prefix + play_path,
2821 a_format.update(rtmp_params)
2822 formats.append(a_format)
2825 def _live_title(self, name):
2826 """ Generate the title for a live video """
2827 now = datetime.datetime.now()
2828 now_str = now.strftime('%Y-%m-%d %H:%M')
2829 return name + ' ' + now_str
2831 def _int(self, v, name, fatal=False, **kwargs):
2832 res = int_or_none(v, **kwargs)
2833 if 'get_attr' in kwargs:
2834 print(getattr(v, kwargs['get_attr']))
2836 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
2838 raise ExtractorError(msg)
2840 self._downloader.report_warning(msg)
2843 def _float(self, v, name, fatal=False, **kwargs):
2844 res = float_or_none(v, **kwargs)
2846 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
2848 raise ExtractorError(msg)
2850 self._downloader.report_warning(msg)
2853 def _set_cookie(self, domain, name, value, expire_time=None, port=None,
2854 path='/', secure=False, discard=False, rest={}, **kwargs):
2855 cookie = compat_cookiejar_Cookie(
2856 0, name, value, port, port is not None, domain, True,
2857 domain.startswith('.'), path, True, secure, expire_time,
2858 discard, None, None, rest)
2859 self._downloader.cookiejar.set_cookie(cookie)
2861 def _get_cookies(self, url):
2862 """ Return a compat_cookies.SimpleCookie with the cookies for the url """
2863 req = sanitized_Request(url)
2864 self._downloader.cookiejar.add_cookie_header(req)
2865 return compat_cookies.SimpleCookie(req.get_header('Cookie'))
2867 def _apply_first_set_cookie_header(self, url_handle, cookie):
2869 Apply first Set-Cookie header instead of the last. Experimental.
2871 Some sites (e.g. [1-3]) may serve two cookies under the same name
2872 in Set-Cookie header and expect the first (old) one to be set rather
2873 than second (new). However, as of RFC6265 the newer one cookie
2874 should be set into cookie store what actually happens.
2875 We will workaround this issue by resetting the cookie to
2876 the first one manually.
2877 1. https://new.vk.com/
2878 2. https://github.com/ytdl-org/youtube-dl/issues/9841#issuecomment-227871201
2879 3. https://learning.oreilly.com/
2881 for header, cookies in url_handle.headers.items():
2882 if header.lower() != 'set-cookie':
2884 if sys.version_info[0] >= 3:
2885 cookies = cookies.encode('iso-8859-1')
2886 cookies = cookies.decode('utf-8')
2887 cookie_value = re.search(
2888 r'%s=(.+?);.*?\b[Dd]omain=(.+?)(?:[,;]|$)' % cookie, cookies)
2890 value, domain = cookie_value.groups()
2891 self._set_cookie(domain, cookie, value)
2894 def get_testcases(self, include_onlymatching=False):
2895 t = getattr(self, '_TEST', None)
2897 assert not hasattr(self, '_TESTS'), \
2898 '%s has _TEST and _TESTS' % type(self).__name__
2901 tests = getattr(self, '_TESTS', [])
2903 if not include_onlymatching and t.get('only_matching', False):
2905 t['name'] = type(self).__name__[:-len('IE')]
2908 def is_suitable(self, age_limit):
2909 """ Test whether the extractor is generally suitable for the given
2910 age limit (i.e. pornographic sites are not, all others usually are) """
2912 any_restricted = False
2913 for tc in self.get_testcases(include_onlymatching=False):
2914 if tc.get('playlist', []):
2915 tc = tc['playlist'][0]
2916 is_restricted = age_restricted(
2917 tc.get('info_dict', {}).get('age_limit'), age_limit)
2918 if not is_restricted:
2920 any_restricted = any_restricted or is_restricted
2921 return not any_restricted
2923 def extract_subtitles(self, *args, **kwargs):
2924 if (self._downloader.params.get('writesubtitles', False)
2925 or self._downloader.params.get('listsubtitles')):
2926 return self._get_subtitles(*args, **kwargs)
2929 def _get_subtitles(self, *args, **kwargs):
2930 raise NotImplementedError('This method must be implemented by subclasses')
2933 def _merge_subtitle_items(subtitle_list1, subtitle_list2):
2934 """ Merge subtitle items for one language. Items with duplicated URLs
2935 will be dropped. """
2936 list1_urls = set([item['url'] for item in subtitle_list1])
2937 ret = list(subtitle_list1)
2938 ret.extend([item for item in subtitle_list2 if item['url'] not in list1_urls])
2942 def _merge_subtitles(cls, subtitle_dict1, subtitle_dict2):
2943 """ Merge two subtitle dictionaries, language by language. """
2944 ret = dict(subtitle_dict1)
2945 for lang in subtitle_dict2:
2946 ret[lang] = cls._merge_subtitle_items(subtitle_dict1.get(lang, []), subtitle_dict2[lang])
2949 def extract_automatic_captions(self, *args, **kwargs):
2950 if (self._downloader.params.get('writeautomaticsub', False)
2951 or self._downloader.params.get('listsubtitles')):
2952 return self._get_automatic_captions(*args, **kwargs)
2955 def _get_automatic_captions(self, *args, **kwargs):
2956 raise NotImplementedError('This method must be implemented by subclasses')
2958 def mark_watched(self, *args, **kwargs):
2959 if (self._downloader.params.get('mark_watched', False)
2960 and (self._get_login_info()[0] is not None
2961 or self._downloader.params.get('cookiefile') is not None)):
2962 self._mark_watched(*args, **kwargs)
2964 def _mark_watched(self, *args, **kwargs):
2965 raise NotImplementedError('This method must be implemented by subclasses')
2967 def geo_verification_headers(self):
2969 geo_verification_proxy = self._downloader.params.get('geo_verification_proxy')
2970 if geo_verification_proxy:
2971 headers['Ytdl-request-proxy'] = geo_verification_proxy
2974 def _generic_id(self, url):
2975 return compat_urllib_parse_unquote(os.path.splitext(url.rstrip('/').split('/')[-1])[0])
2977 def _generic_title(self, url):
2978 return compat_urllib_parse_unquote(os.path.splitext(url_basename(url))[0])
2981 class SearchInfoExtractor(InfoExtractor):
2983 Base class for paged search queries extractors.
2984 They accept URLs in the format _SEARCH_KEY(|all|[0-9]):{query}
2985 Instances should define _SEARCH_KEY and _MAX_RESULTS.
2989 def _make_valid_url(cls):
2990 return r'%s(?P<prefix>|[1-9][0-9]*|all):(?P<query>[\s\S]+)' % cls._SEARCH_KEY
2993 def suitable(cls, url):
2994 return re.match(cls._make_valid_url(), url) is not None
2996 def _real_extract(self, query):
2997 mobj = re.match(self._make_valid_url(), query)
2999 raise ExtractorError('Invalid search query "%s"' % query)
3001 prefix = mobj.group('prefix')
3002 query = mobj.group('query')
3004 return self._get_n_results(query, 1)
3005 elif prefix == 'all':
3006 return self._get_n_results(query, self._MAX_RESULTS)
3010 raise ExtractorError('invalid download number %s for query "%s"' % (n, query))
3011 elif n > self._MAX_RESULTS:
3012 self._downloader.report_warning('%s returns max %i results (you requested %i)' % (self._SEARCH_KEY, self._MAX_RESULTS, n))
3013 n = self._MAX_RESULTS
3014 return self._get_n_results(query, n)
3016 def _get_n_results(self, query, n):
3017 """Get a specified number of results for a query"""
3018 raise NotImplementedError('This method must be implemented by subclasses')
3021 def SEARCH_KEY(self):
3022 return self._SEARCH_KEY