2 from __future__ import unicode_literals
18 from ..compat import (
21 compat_etree_fromstring,
28 compat_urllib_parse_unquote,
29 compat_urllib_parse_urlencode,
30 compat_urllib_request,
32 compat_xml_parse_error,
34 from ..downloader.f4m import (
36 remove_encrypted_media,
62 parse_m3u8_attributes,
80 class InfoExtractor(object):
81 """Information Extractor class.
83 Information extractors are the classes that, given a URL, extract
84 information about the video (or videos) the URL refers to. This
85 information includes the real video URL, the video title, author and
86 others. The information is stored in a dictionary which is then
87 passed to the YoutubeDL. The YoutubeDL processes this
88 information possibly downloading the video to the file system, among
89 other possible outcomes.
91 The type field determines the type of the result.
92 By far the most common value (and the default if _type is missing) is
93 "video", which indicates a single video.
95 For a video, the dictionaries must include the following fields:
98 title: Video title, unescaped.
100 Additionally, it must contain either a formats entry or a url one:
102 formats: A list of dictionaries for each format available, ordered
103 from worst to best quality.
106 * url The mandatory URL representing the media:
107 for plain file media - HTTP URL of this file,
109 for HLS - URL of the M3U8 media playlist,
110 for HDS - URL of the F4M manifest,
111 for DASH - URL of the MPD manifest,
112 for MSS - URL of the ISM manifest.
114 The URL of the manifest file in case of
116 for HLS - URL of the M3U8 master playlist,
117 for HDS - URL of the F4M manifest,
118 for DASH - URL of the MPD manifest,
119 for MSS - URL of the ISM manifest.
120 * ext Will be calculated from URL if missing
121 * format A human-readable description of the format
122 ("mp4 container with h264/opus").
123 Calculated from the format_id, width, height.
124 and format_note fields if missing.
125 * format_id A short description of the format
126 ("mp4_h264_opus" or "19").
127 Technically optional, but strongly recommended.
128 * format_note Additional info about the format
129 ("3D" or "DASH video")
130 * width Width of the video, if known
131 * height Height of the video, if known
132 * resolution Textual description of width and height
133 * tbr Average bitrate of audio and video in KBit/s
134 * abr Average audio bitrate in KBit/s
135 * acodec Name of the audio codec in use
136 * asr Audio sampling rate in Hertz
137 * vbr Average video bitrate in KBit/s
139 * vcodec Name of the video codec in use
140 * container Name of the container format
141 * filesize The number of bytes, if known in advance
142 * filesize_approx An estimate for the number of bytes
143 * player_url SWF Player URL (used for rtmpdump).
144 * protocol The protocol that will be used for the actual
145 download, lower-case.
146 "http", "https", "rtsp", "rtmp", "rtmpe",
147 "m3u8", "m3u8_native" or "http_dash_segments".
149 Base URL for fragments. Each fragment's path
150 value (if present) will be relative to
152 * fragments A list of fragments of a fragmented media.
153 Each fragment entry must contain either an url
154 or a path. If an url is present it should be
155 considered by a client. Otherwise both path and
156 fragment_base_url must be present. Here is
157 the list of all potential fields:
158 * "url" - fragment's URL
159 * "path" - fragment's path relative to
161 * "duration" (optional, int or float)
162 * "filesize" (optional, int)
163 * preference Order number of this format. If this field is
164 present and not None, the formats get sorted
165 by this field, regardless of all other values.
166 -1 for default (order by other properties),
167 -2 or smaller for less than default.
168 < -1000 to hide the format (if there is
169 another one which is strictly better)
170 * language Language code, e.g. "de" or "en-US".
171 * language_preference Is this in the language mentioned in
173 10 if it's what the URL is about,
174 -1 for default (don't know),
175 -10 otherwise, other values reserved for now.
176 * quality Order number of the video quality of this
177 format, irrespective of the file format.
178 -1 for default (order by other properties),
179 -2 or smaller for less than default.
180 * source_preference Order number for this video source
181 (quality takes higher priority)
182 -1 for default (order by other properties),
183 -2 or smaller for less than default.
184 * http_headers A dictionary of additional HTTP headers
185 to add to the request.
186 * stretched_ratio If given and not 1, indicates that the
187 video's pixels are not square.
188 width : height ratio as float.
189 * no_resume The server does not support resuming the
190 (HTTP or RTMP) download. Boolean.
191 * downloader_options A dictionary of downloader options as
192 described in FileDownloader
194 url: Final video URL.
195 ext: Video filename extension.
196 format: The video format, defaults to ext (used for --get-format)
197 player_url: SWF Player URL (used for rtmpdump).
199 The following fields are optional:
201 alt_title: A secondary title of the video.
202 display_id An alternative identifier for the video, not necessarily
203 unique, but available before title. Typically, id is
204 something like "4234987", title "Dancing naked mole rats",
205 and display_id "dancing-naked-mole-rats"
206 thumbnails: A list of dictionaries, with the following entries:
207 * "id" (optional, string) - Thumbnail format ID
209 * "preference" (optional, int) - quality of the image
210 * "width" (optional, int)
211 * "height" (optional, int)
212 * "resolution" (optional, string "{width}x{height"},
214 * "filesize" (optional, int)
215 thumbnail: Full URL to a video thumbnail image.
216 description: Full video description.
217 uploader: Full name of the video uploader.
218 license: License name the video is licensed under.
219 creator: The creator of the video.
220 release_date: The date (YYYYMMDD) when the video was released.
221 timestamp: UNIX timestamp of the moment the video became available.
222 upload_date: Video upload date (YYYYMMDD).
223 If not explicitly set, calculated from timestamp.
224 uploader_id: Nickname or id of the video uploader.
225 uploader_url: Full URL to a personal webpage of the video uploader.
226 channel: Full name of the channel the video is uploaded on.
227 Note that channel fields may or may not repeat uploader
228 fields. This depends on a particular extractor.
229 channel_id: Id of the channel.
230 channel_url: Full URL to a channel webpage.
231 location: Physical location where the video was filmed.
232 subtitles: The available subtitles as a dictionary in the format
233 {tag: subformats}. "tag" is usually a language code, and
234 "subformats" is a list sorted from lower to higher
235 preference, each element is a dictionary with the "ext"
237 * "data": The subtitles file contents
238 * "url": A URL pointing to the subtitles file
239 "ext" will be calculated from URL if missing
240 automatic_captions: Like 'subtitles', used by the YoutubeIE for
241 automatically generated captions
242 duration: Length of the video in seconds, as an integer or float.
243 view_count: How many users have watched the video on the platform.
244 like_count: Number of positive ratings of the video
245 dislike_count: Number of negative ratings of the video
246 repost_count: Number of reposts of the video
247 average_rating: Average rating give by users, the scale used depends on the webpage
248 comment_count: Number of comments on the video
249 comments: A list of comments, each with one or more of the following
250 properties (all but one of text or html optional):
251 * "author" - human-readable name of the comment author
252 * "author_id" - user ID of the comment author
254 * "html" - Comment as HTML
255 * "text" - Plain text of the comment
256 * "timestamp" - UNIX timestamp of comment
257 * "parent" - ID of the comment this one is replying to.
258 Set to "root" to indicate that this is a
259 comment to the original video.
260 age_limit: Age restriction for the video, as an integer (years)
261 webpage_url: The URL to the video webpage, if given to youtube-dl it
262 should allow to get the same result again. (It will be set
263 by YoutubeDL if it's missing)
264 categories: A list of categories that the video falls in, for example
266 tags: A list of tags assigned to the video, e.g. ["sweden", "pop music"]
267 is_live: True, False, or None (=unknown). Whether this video is a
268 live stream that goes on instead of a fixed-length video.
269 start_time: Time in seconds where the reproduction should start, as
270 specified in the URL.
271 end_time: Time in seconds where the reproduction should end, as
272 specified in the URL.
273 chapters: A list of dictionaries, with the following entries:
274 * "start_time" - The start time of the chapter in seconds
275 * "end_time" - The end time of the chapter in seconds
276 * "title" (optional, string)
278 The following fields should only be used when the video belongs to some logical
281 chapter: Name or title of the chapter the video belongs to.
282 chapter_number: Number of the chapter the video belongs to, as an integer.
283 chapter_id: Id of the chapter the video belongs to, as a unicode string.
285 The following fields should only be used when the video is an episode of some
286 series, programme or podcast:
288 series: Title of the series or programme the video episode belongs to.
289 season: Title of the season the video episode belongs to.
290 season_number: Number of the season the video episode belongs to, as an integer.
291 season_id: Id of the season the video episode belongs to, as a unicode string.
292 episode: Title of the video episode. Unlike mandatory video title field,
293 this field should denote the exact title of the video episode
294 without any kind of decoration.
295 episode_number: Number of the video episode within a season, as an integer.
296 episode_id: Id of the video episode, as a unicode string.
298 The following fields should only be used when the media is a track or a part of
301 track: Title of the track.
302 track_number: Number of the track within an album or a disc, as an integer.
303 track_id: Id of the track (useful in case of custom indexing, e.g. 6.iii),
305 artist: Artist(s) of the track.
306 genre: Genre(s) of the track.
307 album: Title of the album the track belongs to.
308 album_type: Type of the album (e.g. "Demo", "Full-length", "Split", "Compilation", etc).
309 album_artist: List of all artists appeared on the album (e.g.
310 "Ash Borer / Fell Voices" or "Various Artists", useful for splits
312 disc_number: Number of the disc or other physical medium the track belongs to,
314 release_year: Year (YYYY) when the album was released.
316 Unless mentioned otherwise, the fields should be Unicode strings.
318 Unless mentioned otherwise, None is equivalent to absence of information.
321 _type "playlist" indicates multiple videos.
322 There must be a key "entries", which is a list, an iterable, or a PagedList
323 object, each element of which is a valid dictionary by this specification.
325 Additionally, playlists can have "id", "title", "description", "uploader",
326 "uploader_id", "uploader_url" attributes with the same semantics as videos
330 _type "multi_video" indicates that there are multiple videos that
331 form a single show, for examples multiple acts of an opera or TV episode.
332 It must have an entries key like a playlist and contain all the keys
333 required for a video at the same time.
336 _type "url" indicates that the video must be extracted from another
337 location, possibly by a different extractor. Its only required key is:
338 "url" - the next URL to extract.
339 The key "ie_key" can be set to the class name (minus the trailing "IE",
340 e.g. "Youtube") if the extractor class is known in advance.
341 Additionally, the dictionary may have any properties of the resolved entity
342 known in advance, for example "title" if the title of the referred video is
346 _type "url_transparent" entities have the same specification as "url", but
347 indicate that the given additional information is more precise than the one
348 associated with the resolved URL.
349 This is useful when a site employs a video service that hosts the video and
350 its technical metadata, but that video service does not embed a useful
351 title, description etc.
354 Subclasses of this one should re-define the _real_initialize() and
355 _real_extract() methods and define a _VALID_URL regexp.
356 Probably, they should also be added to the list of extractors.
358 _GEO_BYPASS attribute may be set to False in order to disable
359 geo restriction bypass mechanisms for a particular extractor.
360 Though it won't disable explicit geo restriction bypass based on
361 country code provided with geo_bypass_country.
363 _GEO_COUNTRIES attribute may contain a list of presumably geo unrestricted
364 countries for this extractor. One of these countries will be used by
365 geo restriction bypass mechanism right away in order to bypass
366 geo restriction, of course, if the mechanism is not disabled.
368 _GEO_IP_BLOCKS attribute may contain a list of presumably geo unrestricted
369 IP blocks in CIDR notation for this extractor. One of these IP blocks
370 will be used by geo restriction bypass mechanism similarly
373 Finally, the _WORKING attribute should be set to False for broken IEs
374 in order to warn the users and skip the tests.
379 _x_forwarded_for_ip = None
381 _GEO_COUNTRIES = None
382 _GEO_IP_BLOCKS = None
385 def __init__(self, downloader=None):
386 """Constructor. Receives an optional downloader."""
388 self._x_forwarded_for_ip = None
389 self.set_downloader(downloader)
392 def suitable(cls, url):
393 """Receives a URL and returns True if suitable for this IE."""
395 # This does not use has/getattr intentionally - we want to know whether
396 # we have cached the regexp for *this* class, whereas getattr would also
397 # match the superclass
398 if '_VALID_URL_RE' not in cls.__dict__:
399 cls._VALID_URL_RE = re.compile(cls._VALID_URL)
400 return cls._VALID_URL_RE.match(url) is not None
403 def _match_id(cls, url):
404 if '_VALID_URL_RE' not in cls.__dict__:
405 cls._VALID_URL_RE = re.compile(cls._VALID_URL)
406 m = cls._VALID_URL_RE.match(url)
408 return compat_str(m.group('id'))
412 """Getter method for _WORKING."""
415 def initialize(self):
416 """Initializes an instance (authentication, etc)."""
417 self._initialize_geo_bypass({
418 'countries': self._GEO_COUNTRIES,
419 'ip_blocks': self._GEO_IP_BLOCKS,
422 self._real_initialize()
425 def _initialize_geo_bypass(self, geo_bypass_context):
427 Initialize geo restriction bypass mechanism.
429 This method is used to initialize geo bypass mechanism based on faking
430 X-Forwarded-For HTTP header. A random country from provided country list
431 is selected and a random IP belonging to this country is generated. This
432 IP will be passed as X-Forwarded-For HTTP header in all subsequent
435 This method will be used for initial geo bypass mechanism initialization
436 during the instance initialization with _GEO_COUNTRIES and
439 You may also manually call it from extractor's code if geo bypass
440 information is not available beforehand (e.g. obtained during
441 extraction) or due to some other reason. In this case you should pass
442 this information in geo bypass context passed as first argument. It may
443 contain following fields:
445 countries: List of geo unrestricted countries (similar
447 ip_blocks: List of geo unrestricted IP blocks in CIDR notation
448 (similar to _GEO_IP_BLOCKS)
451 if not self._x_forwarded_for_ip:
453 # Geo bypass mechanism is explicitly disabled by user
454 if not self._downloader.params.get('geo_bypass', True):
457 if not geo_bypass_context:
458 geo_bypass_context = {}
460 # Backward compatibility: previously _initialize_geo_bypass
461 # expected a list of countries, some 3rd party code may still use
463 if isinstance(geo_bypass_context, (list, tuple)):
464 geo_bypass_context = {
465 'countries': geo_bypass_context,
468 # The whole point of geo bypass mechanism is to fake IP
469 # as X-Forwarded-For HTTP header based on some IP block or
472 # Path 1: bypassing based on IP block in CIDR notation
474 # Explicit IP block specified by user, use it right away
475 # regardless of whether extractor is geo bypassable or not
476 ip_block = self._downloader.params.get('geo_bypass_ip_block', None)
478 # Otherwise use random IP block from geo bypass context but only
479 # if extractor is known as geo bypassable
481 ip_blocks = geo_bypass_context.get('ip_blocks')
482 if self._GEO_BYPASS and ip_blocks:
483 ip_block = random.choice(ip_blocks)
486 self._x_forwarded_for_ip = GeoUtils.random_ipv4(ip_block)
487 if self._downloader.params.get('verbose', False):
488 self._downloader.to_screen(
489 '[debug] Using fake IP %s as X-Forwarded-For.'
490 % self._x_forwarded_for_ip)
493 # Path 2: bypassing based on country code
495 # Explicit country code specified by user, use it right away
496 # regardless of whether extractor is geo bypassable or not
497 country = self._downloader.params.get('geo_bypass_country', None)
499 # Otherwise use random country code from geo bypass context but
500 # only if extractor is known as geo bypassable
502 countries = geo_bypass_context.get('countries')
503 if self._GEO_BYPASS and countries:
504 country = random.choice(countries)
507 self._x_forwarded_for_ip = GeoUtils.random_ipv4(country)
508 if self._downloader.params.get('verbose', False):
509 self._downloader.to_screen(
510 '[debug] Using fake IP %s (%s) as X-Forwarded-For.'
511 % (self._x_forwarded_for_ip, country.upper()))
513 def extract(self, url):
514 """Extracts URL information and returns it in list of dicts."""
519 ie_result = self._real_extract(url)
520 if self._x_forwarded_for_ip:
521 ie_result['__x_forwarded_for_ip'] = self._x_forwarded_for_ip
523 except GeoRestrictedError as e:
524 if self.__maybe_fake_ip_and_retry(e.countries):
527 except ExtractorError:
529 except compat_http_client.IncompleteRead as e:
530 raise ExtractorError('A network error has occurred.', cause=e, expected=True)
531 except (KeyError, StopIteration) as e:
532 raise ExtractorError('An extractor error has occurred.', cause=e)
534 def __maybe_fake_ip_and_retry(self, countries):
535 if (not self._downloader.params.get('geo_bypass_country', None) and
537 self._downloader.params.get('geo_bypass', True) and
538 not self._x_forwarded_for_ip and
540 country_code = random.choice(countries)
541 self._x_forwarded_for_ip = GeoUtils.random_ipv4(country_code)
542 if self._x_forwarded_for_ip:
544 'Video is geo restricted. Retrying extraction with fake IP %s (%s) as X-Forwarded-For.'
545 % (self._x_forwarded_for_ip, country_code.upper()))
549 def set_downloader(self, downloader):
550 """Sets the downloader for this IE."""
551 self._downloader = downloader
553 def _real_initialize(self):
554 """Real initialization process. Redefine in subclasses."""
557 def _real_extract(self, url):
558 """Real extraction process. Redefine in subclasses."""
563 """A string for getting the InfoExtractor with get_info_extractor"""
564 return compat_str(cls.__name__[:-2])
568 return compat_str(type(self).__name__[:-2])
571 def __can_accept_status_code(err, expected_status):
572 assert isinstance(err, compat_urllib_error.HTTPError)
573 if expected_status is None:
575 if isinstance(expected_status, compat_integer_types):
576 return err.code == expected_status
577 elif isinstance(expected_status, (list, tuple)):
578 return err.code in expected_status
579 elif callable(expected_status):
580 return expected_status(err.code) is True
584 def _request_webpage(self, url_or_request, video_id, note=None, errnote=None, fatal=True, data=None, headers={}, query={}, expected_status=None):
586 Return the response handle.
588 See _download_webpage docstring for arguments specification.
591 self.report_download_webpage(video_id)
592 elif note is not False:
594 self.to_screen('%s' % (note,))
596 self.to_screen('%s: %s' % (video_id, note))
598 # Some sites check X-Forwarded-For HTTP header in order to figure out
599 # the origin of the client behind proxy. This allows bypassing geo
600 # restriction by faking this header's value to IP that belongs to some
601 # geo unrestricted country. We will do so once we encounter any
602 # geo restriction error.
603 if self._x_forwarded_for_ip:
604 if 'X-Forwarded-For' not in headers:
605 headers['X-Forwarded-For'] = self._x_forwarded_for_ip
607 if isinstance(url_or_request, compat_urllib_request.Request):
608 url_or_request = update_Request(
609 url_or_request, data=data, headers=headers, query=query)
612 url_or_request = update_url_query(url_or_request, query)
613 if data is not None or headers:
614 url_or_request = sanitized_Request(url_or_request, data, headers)
616 return self._downloader.urlopen(url_or_request)
617 except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err:
618 if isinstance(err, compat_urllib_error.HTTPError):
619 if self.__can_accept_status_code(err, expected_status):
620 # Retain reference to error to prevent file object from
621 # being closed before it can be read. Works around the
622 # effects of <https://bugs.python.org/issue15002>
623 # introduced in Python 3.4.1.
630 errnote = 'Unable to download webpage'
632 errmsg = '%s: %s' % (errnote, error_to_compat_str(err))
634 raise ExtractorError(errmsg, sys.exc_info()[2], cause=err)
636 self._downloader.report_warning(errmsg)
639 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):
641 Return a tuple (page content as string, URL handle).
643 See _download_webpage docstring for arguments specification.
645 # Strip hashes from the URL (#1038)
646 if isinstance(url_or_request, (compat_str, str)):
647 url_or_request = url_or_request.partition('#')[0]
649 urlh = self._request_webpage(url_or_request, video_id, note, errnote, fatal, data=data, headers=headers, query=query, expected_status=expected_status)
653 content = self._webpage_read_content(urlh, url_or_request, video_id, note, errnote, fatal, encoding=encoding)
654 return (content, urlh)
657 def _guess_encoding_from_content(content_type, webpage_bytes):
658 m = re.match(r'[a-zA-Z0-9_.-]+/[a-zA-Z0-9_.-]+\s*;\s*charset=(.+)', content_type)
660 encoding = m.group(1)
662 m = re.search(br'<meta[^>]+charset=[\'"]?([^\'")]+)[ /\'">]',
663 webpage_bytes[:1024])
665 encoding = m.group(1).decode('ascii')
666 elif webpage_bytes.startswith(b'\xff\xfe'):
673 def __check_blocked(self, content):
674 first_block = content[:512]
675 if ('<title>Access to this site is blocked</title>' in content and
676 'Websense' in first_block):
677 msg = 'Access to this webpage has been blocked by Websense filtering software in your network.'
678 blocked_iframe = self._html_search_regex(
679 r'<iframe src="([^"]+)"', content,
680 'Websense information URL', default=None)
682 msg += ' Visit %s for more details' % blocked_iframe
683 raise ExtractorError(msg, expected=True)
684 if '<title>The URL you requested has been blocked</title>' in first_block:
686 'Access to this webpage has been blocked by Indian censorship. '
687 'Use a VPN or proxy server (with --proxy) to route around it.')
688 block_msg = self._html_search_regex(
689 r'</h1><p>(.*?)</p>',
690 content, 'block message', default=None)
692 msg += ' (Message: "%s")' % block_msg.replace('\n', ' ')
693 raise ExtractorError(msg, expected=True)
694 if ('<title>TTK :: Доступ к ресурсу ограничен</title>' in content and
695 'blocklist.rkn.gov.ru' in content):
696 raise ExtractorError(
697 'Access to this webpage has been blocked by decision of the Russian government. '
698 'Visit http://blocklist.rkn.gov.ru/ for a block reason.',
701 def _webpage_read_content(self, urlh, url_or_request, video_id, note=None, errnote=None, fatal=True, prefix=None, encoding=None):
702 content_type = urlh.headers.get('Content-Type', '')
703 webpage_bytes = urlh.read()
704 if prefix is not None:
705 webpage_bytes = prefix + webpage_bytes
707 encoding = self._guess_encoding_from_content(content_type, webpage_bytes)
708 if self._downloader.params.get('dump_intermediate_pages', False):
709 self.to_screen('Dumping request to ' + urlh.geturl())
710 dump = base64.b64encode(webpage_bytes).decode('ascii')
711 self._downloader.to_screen(dump)
712 if self._downloader.params.get('write_pages', False):
713 basen = '%s_%s' % (video_id, urlh.geturl())
715 h = '___' + hashlib.md5(basen.encode('utf-8')).hexdigest()
716 basen = basen[:240 - len(h)] + h
717 raw_filename = basen + '.dump'
718 filename = sanitize_filename(raw_filename, restricted=True)
719 self.to_screen('Saving request to ' + filename)
720 # Working around MAX_PATH limitation on Windows (see
721 # http://msdn.microsoft.com/en-us/library/windows/desktop/aa365247(v=vs.85).aspx)
722 if compat_os_name == 'nt':
723 absfilepath = os.path.abspath(filename)
724 if len(absfilepath) > 259:
725 filename = '\\\\?\\' + absfilepath
726 with open(filename, 'wb') as outf:
727 outf.write(webpage_bytes)
730 content = webpage_bytes.decode(encoding, 'replace')
732 content = webpage_bytes.decode('utf-8', 'replace')
734 self.__check_blocked(content)
738 def _download_webpage(
739 self, url_or_request, video_id, note=None, errnote=None,
740 fatal=True, tries=1, timeout=5, encoding=None, data=None,
741 headers={}, query={}, expected_status=None):
743 Return the data of the page as a string.
746 url_or_request -- plain text URL as a string or
747 a compat_urllib_request.Requestobject
748 video_id -- Video/playlist/item identifier (string)
751 note -- note printed before downloading (string)
752 errnote -- note printed in case of an error (string)
753 fatal -- flag denoting whether error should be considered fatal,
754 i.e. whether it should cause ExtractionError to be raised,
755 otherwise a warning will be reported and extraction continued
756 tries -- number of tries
757 timeout -- sleep interval between tries
758 encoding -- encoding for a page content decoding, guessed automatically
759 when not explicitly specified
760 data -- POST data (bytes)
761 headers -- HTTP headers (dict)
762 query -- URL query (dict)
763 expected_status -- allows to accept failed HTTP requests (non 2xx
764 status code) by explicitly specifying a set of accepted status
765 codes. Can be any of the following entities:
766 - an integer type specifying an exact failed status code to
768 - a list or a tuple of integer types specifying a list of
769 failed status codes to accept
770 - a callable accepting an actual failed status code and
771 returning True if it should be accepted
772 Note that this argument does not affect success status codes (2xx)
773 which are always accepted.
778 while success is False:
780 res = self._download_webpage_handle(
781 url_or_request, video_id, note, errnote, fatal,
782 encoding=encoding, data=data, headers=headers, query=query,
783 expected_status=expected_status)
785 except compat_http_client.IncompleteRead as e:
787 if try_count >= tries:
789 self._sleep(timeout, video_id)
796 def _download_xml_handle(
797 self, url_or_request, video_id, note='Downloading XML',
798 errnote='Unable to download XML', transform_source=None,
799 fatal=True, encoding=None, data=None, headers={}, query={},
800 expected_status=None):
802 Return a tuple (xml as an xml.etree.ElementTree.Element, URL handle).
804 See _download_webpage docstring for arguments specification.
806 res = self._download_webpage_handle(
807 url_or_request, video_id, note, errnote, fatal=fatal,
808 encoding=encoding, data=data, headers=headers, query=query,
809 expected_status=expected_status)
812 xml_string, urlh = res
813 return self._parse_xml(
814 xml_string, video_id, transform_source=transform_source,
818 self, url_or_request, video_id,
819 note='Downloading XML', errnote='Unable to download XML',
820 transform_source=None, fatal=True, encoding=None,
821 data=None, headers={}, query={}, expected_status=None):
823 Return the xml as an xml.etree.ElementTree.Element.
825 See _download_webpage docstring for arguments specification.
827 res = self._download_xml_handle(
828 url_or_request, video_id, note=note, errnote=errnote,
829 transform_source=transform_source, fatal=fatal, encoding=encoding,
830 data=data, headers=headers, query=query,
831 expected_status=expected_status)
832 return res if res is False else res[0]
834 def _parse_xml(self, xml_string, video_id, transform_source=None, fatal=True):
836 xml_string = transform_source(xml_string)
838 return compat_etree_fromstring(xml_string.encode('utf-8'))
839 except compat_xml_parse_error as ve:
840 errmsg = '%s: Failed to parse XML ' % video_id
842 raise ExtractorError(errmsg, cause=ve)
844 self.report_warning(errmsg + str(ve))
846 def _download_json_handle(
847 self, url_or_request, video_id, note='Downloading JSON metadata',
848 errnote='Unable to download JSON metadata', transform_source=None,
849 fatal=True, encoding=None, data=None, headers={}, query={},
850 expected_status=None):
852 Return a tuple (JSON object, URL handle).
854 See _download_webpage docstring for arguments specification.
856 res = self._download_webpage_handle(
857 url_or_request, video_id, note, errnote, fatal=fatal,
858 encoding=encoding, data=data, headers=headers, query=query,
859 expected_status=expected_status)
862 json_string, urlh = res
863 return self._parse_json(
864 json_string, video_id, transform_source=transform_source,
868 self, url_or_request, video_id, note='Downloading JSON metadata',
869 errnote='Unable to download JSON metadata', transform_source=None,
870 fatal=True, encoding=None, data=None, headers={}, query={},
871 expected_status=None):
873 Return the JSON object as a dict.
875 See _download_webpage docstring for arguments specification.
877 res = self._download_json_handle(
878 url_or_request, video_id, note=note, errnote=errnote,
879 transform_source=transform_source, fatal=fatal, encoding=encoding,
880 data=data, headers=headers, query=query,
881 expected_status=expected_status)
882 return res if res is False else res[0]
884 def _parse_json(self, json_string, video_id, transform_source=None, fatal=True):
886 json_string = transform_source(json_string)
888 return json.loads(json_string)
889 except ValueError as ve:
890 errmsg = '%s: Failed to parse JSON ' % video_id
892 raise ExtractorError(errmsg, cause=ve)
894 self.report_warning(errmsg + str(ve))
896 def report_warning(self, msg, video_id=None):
897 idstr = '' if video_id is None else '%s: ' % video_id
898 self._downloader.report_warning(
899 '[%s] %s%s' % (self.IE_NAME, idstr, msg))
901 def to_screen(self, msg):
902 """Print msg to screen, prefixing it with '[ie_name]'"""
903 self._downloader.to_screen('[%s] %s' % (self.IE_NAME, msg))
905 def report_extraction(self, id_or_name):
906 """Report information extraction."""
907 self.to_screen('%s: Extracting information' % id_or_name)
909 def report_download_webpage(self, video_id):
910 """Report webpage download."""
911 self.to_screen('%s: Downloading webpage' % video_id)
913 def report_age_confirmation(self):
914 """Report attempt to confirm age."""
915 self.to_screen('Confirming age')
917 def report_login(self):
918 """Report attempt to log in."""
919 self.to_screen('Logging in')
922 def raise_login_required(msg='This video is only available for registered users'):
923 raise ExtractorError(
924 '%s. Use --username and --password or --netrc to provide account credentials.' % msg,
928 def raise_geo_restricted(msg='This video is not available from your location due to geo restriction', countries=None):
929 raise GeoRestrictedError(msg, countries=countries)
931 # Methods for following #608
933 def url_result(url, ie=None, video_id=None, video_title=None):
934 """Returns a URL that points to a page that should be processed"""
935 # TODO: ie should be the class used for getting the info
936 video_info = {'_type': 'url',
939 if video_id is not None:
940 video_info['id'] = video_id
941 if video_title is not None:
942 video_info['title'] = video_title
945 def playlist_from_matches(self, matches, playlist_id=None, playlist_title=None, getter=None, ie=None):
947 self.url_result(self._proto_relative_url(getter(m) if getter else m), ie)
949 return self.playlist_result(
950 urls, playlist_id=playlist_id, playlist_title=playlist_title)
953 def playlist_result(entries, playlist_id=None, playlist_title=None, playlist_description=None):
954 """Returns a playlist"""
955 video_info = {'_type': 'playlist',
958 video_info['id'] = playlist_id
960 video_info['title'] = playlist_title
961 if playlist_description:
962 video_info['description'] = playlist_description
965 def _search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
967 Perform a regex search on the given string, using a single or a list of
968 patterns returning the first matching group.
969 In case of failure return a default value or raise a WARNING or a
970 RegexNotFoundError, depending on fatal, specifying the field name.
972 if isinstance(pattern, (str, compat_str, compiled_regex_type)):
973 mobj = re.search(pattern, string, flags)
976 mobj = re.search(p, string, flags)
980 if not self._downloader.params.get('no_color') and compat_os_name != 'nt' and sys.stderr.isatty():
981 _name = '\033[0;34m%s\033[0m' % name
987 # return the first matching group
988 return next(g for g in mobj.groups() if g is not None)
990 return mobj.group(group)
991 elif default is not NO_DEFAULT:
994 raise RegexNotFoundError('Unable to extract %s' % _name)
996 self._downloader.report_warning('unable to extract %s' % _name + bug_reports_message())
999 def _html_search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
1001 Like _search_regex, but strips HTML tags and unescapes entities.
1003 res = self._search_regex(pattern, string, name, default, fatal, flags, group)
1005 return clean_html(res).strip()
1009 def _get_netrc_login_info(self, netrc_machine=None):
1012 netrc_machine = netrc_machine or self._NETRC_MACHINE
1014 if self._downloader.params.get('usenetrc', False):
1016 info = netrc.netrc().authenticators(netrc_machine)
1017 if info is not None:
1021 raise netrc.NetrcParseError(
1022 'No authenticators for %s' % netrc_machine)
1023 except (IOError, netrc.NetrcParseError) as err:
1024 self._downloader.report_warning(
1025 'parsing .netrc: %s' % error_to_compat_str(err))
1027 return username, password
1029 def _get_login_info(self, username_option='username', password_option='password', netrc_machine=None):
1031 Get the login info as (username, password)
1032 First look for the manually specified credentials using username_option
1033 and password_option as keys in params dictionary. If no such credentials
1034 available look in the netrc file using the netrc_machine or _NETRC_MACHINE
1036 If there's no info available, return (None, None)
1038 if self._downloader is None:
1041 downloader_params = self._downloader.params
1043 # Attempt to use provided username and password or .netrc data
1044 if downloader_params.get(username_option) is not None:
1045 username = downloader_params[username_option]
1046 password = downloader_params[password_option]
1048 username, password = self._get_netrc_login_info(netrc_machine)
1050 return username, password
1052 def _get_tfa_info(self, note='two-factor verification code'):
1054 Get the two-factor authentication info
1055 TODO - asking the user will be required for sms/phone verify
1056 currently just uses the command line option
1057 If there's no info available, return None
1059 if self._downloader is None:
1061 downloader_params = self._downloader.params
1063 if downloader_params.get('twofactor') is not None:
1064 return downloader_params['twofactor']
1066 return compat_getpass('Type %s and press [Return]: ' % note)
1068 # Helper functions for extracting OpenGraph info
1070 def _og_regexes(prop):
1071 content_re = r'content=(?:"([^"]+?)"|\'([^\']+?)\'|\s*([^\s"\'=<>`]+?))'
1072 property_re = (r'(?:name|property)=(?:\'og[:-]%(prop)s\'|"og[:-]%(prop)s"|\s*og[:-]%(prop)s\b)'
1073 % {'prop': re.escape(prop)})
1074 template = r'<meta[^>]+?%s[^>]+?%s'
1076 template % (property_re, content_re),
1077 template % (content_re, property_re),
1081 def _meta_regex(prop):
1082 return r'''(?isx)<meta
1083 (?=[^>]+(?:itemprop|name|property|id|http-equiv)=(["\']?)%s\1)
1084 [^>]+?content=(["\'])(?P<content>.*?)\2''' % re.escape(prop)
1086 def _og_search_property(self, prop, html, name=None, **kargs):
1087 if not isinstance(prop, (list, tuple)):
1090 name = 'OpenGraph %s' % prop[0]
1093 og_regexes.extend(self._og_regexes(p))
1094 escaped = self._search_regex(og_regexes, html, name, flags=re.DOTALL, **kargs)
1097 return unescapeHTML(escaped)
1099 def _og_search_thumbnail(self, html, **kargs):
1100 return self._og_search_property('image', html, 'thumbnail URL', fatal=False, **kargs)
1102 def _og_search_description(self, html, **kargs):
1103 return self._og_search_property('description', html, fatal=False, **kargs)
1105 def _og_search_title(self, html, **kargs):
1106 return self._og_search_property('title', html, **kargs)
1108 def _og_search_video_url(self, html, name='video url', secure=True, **kargs):
1109 regexes = self._og_regexes('video') + self._og_regexes('video:url')
1111 regexes = self._og_regexes('video:secure_url') + regexes
1112 return self._html_search_regex(regexes, html, name, **kargs)
1114 def _og_search_url(self, html, **kargs):
1115 return self._og_search_property('url', html, **kargs)
1117 def _html_search_meta(self, name, html, display_name=None, fatal=False, **kwargs):
1118 if not isinstance(name, (list, tuple)):
1120 if display_name is None:
1121 display_name = name[0]
1122 return self._html_search_regex(
1123 [self._meta_regex(n) for n in name],
1124 html, display_name, fatal=fatal, group='content', **kwargs)
1126 def _dc_search_uploader(self, html):
1127 return self._html_search_meta('dc.creator', html, 'uploader')
1129 def _rta_search(self, html):
1130 # See http://www.rtalabel.org/index.php?content=howtofaq#single
1131 if re.search(r'(?ix)<meta\s+name="rating"\s+'
1132 r' content="RTA-5042-1996-1400-1577-RTA"',
1137 def _media_rating_search(self, html):
1138 # See http://www.tjg-designs.com/WP/metadata-code-examples-adding-metadata-to-your-web-pages/
1139 rating = self._html_search_meta('rating', html)
1151 return RATING_TABLE.get(rating.lower())
1153 def _family_friendly_search(self, html):
1154 # See http://schema.org/VideoObject
1155 family_friendly = self._html_search_meta(
1156 'isFamilyFriendly', html, default=None)
1158 if not family_friendly:
1167 return RATING_TABLE.get(family_friendly.lower())
1169 def _twitter_search_player(self, html):
1170 return self._html_search_meta('twitter:player', html,
1171 'twitter card player')
1173 def _search_json_ld(self, html, video_id, expected_type=None, **kwargs):
1174 json_ld = self._search_regex(
1175 JSON_LD_RE, html, 'JSON-LD', group='json_ld', **kwargs)
1176 default = kwargs.get('default', NO_DEFAULT)
1178 return default if default is not NO_DEFAULT else {}
1179 # JSON-LD may be malformed and thus `fatal` should be respected.
1180 # At the same time `default` may be passed that assumes `fatal=False`
1181 # for _search_regex. Let's simulate the same behavior here as well.
1182 fatal = kwargs.get('fatal', True) if default == NO_DEFAULT else False
1183 return self._json_ld(json_ld, video_id, fatal=fatal, expected_type=expected_type)
1185 def _json_ld(self, json_ld, video_id, fatal=True, expected_type=None):
1186 if isinstance(json_ld, compat_str):
1187 json_ld = self._parse_json(json_ld, video_id, fatal=fatal)
1191 if not isinstance(json_ld, (list, tuple, dict)):
1193 if isinstance(json_ld, dict):
1196 INTERACTION_TYPE_MAP = {
1197 'CommentAction': 'comment',
1198 'AgreeAction': 'like',
1199 'DisagreeAction': 'dislike',
1200 'LikeAction': 'like',
1201 'DislikeAction': 'dislike',
1202 'ListenAction': 'view',
1203 'WatchAction': 'view',
1204 'ViewAction': 'view',
1207 def extract_interaction_statistic(e):
1208 interaction_statistic = e.get('interactionStatistic')
1209 if not isinstance(interaction_statistic, list):
1211 for is_e in interaction_statistic:
1212 if not isinstance(is_e, dict):
1214 if is_e.get('@type') != 'InteractionCounter':
1216 interaction_type = is_e.get('interactionType')
1217 if not isinstance(interaction_type, compat_str):
1219 interaction_count = int_or_none(is_e.get('userInteractionCount'))
1220 if interaction_count is None:
1222 count_kind = INTERACTION_TYPE_MAP.get(interaction_type.split('/')[-1])
1225 count_key = '%s_count' % count_kind
1226 if info.get(count_key) is not None:
1228 info[count_key] = interaction_count
1230 def extract_video_object(e):
1231 assert e['@type'] == 'VideoObject'
1233 'url': url_or_none(e.get('contentUrl')),
1234 'title': unescapeHTML(e.get('name')),
1235 'description': unescapeHTML(e.get('description')),
1236 'thumbnail': url_or_none(e.get('thumbnailUrl') or e.get('thumbnailURL')),
1237 'duration': parse_duration(e.get('duration')),
1238 'timestamp': unified_timestamp(e.get('uploadDate')),
1239 'filesize': float_or_none(e.get('contentSize')),
1240 'tbr': int_or_none(e.get('bitrate')),
1241 'width': int_or_none(e.get('width')),
1242 'height': int_or_none(e.get('height')),
1243 'view_count': int_or_none(e.get('interactionCount')),
1245 extract_interaction_statistic(e)
1248 if isinstance(e.get('@context'), compat_str) and re.match(r'^https?://schema.org/?$', e.get('@context')):
1249 item_type = e.get('@type')
1250 if expected_type is not None and expected_type != item_type:
1252 if item_type in ('TVEpisode', 'Episode'):
1253 episode_name = unescapeHTML(e.get('name'))
1255 'episode': episode_name,
1256 'episode_number': int_or_none(e.get('episodeNumber')),
1257 'description': unescapeHTML(e.get('description')),
1259 if not info.get('title') and episode_name:
1260 info['title'] = episode_name
1261 part_of_season = e.get('partOfSeason')
1262 if isinstance(part_of_season, dict) and part_of_season.get('@type') in ('TVSeason', 'Season', 'CreativeWorkSeason'):
1264 'season': unescapeHTML(part_of_season.get('name')),
1265 'season_number': int_or_none(part_of_season.get('seasonNumber')),
1267 part_of_series = e.get('partOfSeries') or e.get('partOfTVSeries')
1268 if isinstance(part_of_series, dict) and part_of_series.get('@type') in ('TVSeries', 'Series', 'CreativeWorkSeries'):
1269 info['series'] = unescapeHTML(part_of_series.get('name'))
1270 elif item_type == 'Movie':
1272 'title': unescapeHTML(e.get('name')),
1273 'description': unescapeHTML(e.get('description')),
1274 'duration': parse_duration(e.get('duration')),
1275 'timestamp': unified_timestamp(e.get('dateCreated')),
1277 elif item_type in ('Article', 'NewsArticle'):
1279 'timestamp': parse_iso8601(e.get('datePublished')),
1280 'title': unescapeHTML(e.get('headline')),
1281 'description': unescapeHTML(e.get('articleBody')),
1283 elif item_type == 'VideoObject':
1284 extract_video_object(e)
1286 video = e.get('video')
1287 if isinstance(video, dict) and video.get('@type') == 'VideoObject':
1288 extract_video_object(video)
1290 return dict((k, v) for k, v in info.items() if v is not None)
1293 def _hidden_inputs(html):
1294 html = re.sub(r'<!--(?:(?!<!--).)*-->', '', html)
1296 for input in re.findall(r'(?i)(<input[^>]+>)', html):
1297 attrs = extract_attributes(input)
1300 if attrs.get('type') not in ('hidden', 'submit'):
1302 name = attrs.get('name') or attrs.get('id')
1303 value = attrs.get('value')
1304 if name and value is not None:
1305 hidden_inputs[name] = value
1306 return hidden_inputs
1308 def _form_hidden_inputs(self, form_id, html):
1309 form = self._search_regex(
1310 r'(?is)<form[^>]+?id=(["\'])%s\1[^>]*>(?P<form>.+?)</form>' % form_id,
1311 html, '%s form' % form_id, group='form')
1312 return self._hidden_inputs(form)
1314 def _sort_formats(self, formats, field_preference=None):
1316 raise ExtractorError('No video formats found')
1319 # Automatically determine tbr when missing based on abr and vbr (improves
1320 # formats sorting in some cases)
1321 if 'tbr' not in f and f.get('abr') is not None and f.get('vbr') is not None:
1322 f['tbr'] = f['abr'] + f['vbr']
1324 def _formats_key(f):
1325 # TODO remove the following workaround
1326 from ..utils import determine_ext
1327 if not f.get('ext') and 'url' in f:
1328 f['ext'] = determine_ext(f['url'])
1330 if isinstance(field_preference, (list, tuple)):
1333 if f.get(field) is not None
1334 else ('' if field == 'format_id' else -1)
1335 for field in field_preference)
1337 preference = f.get('preference')
1338 if preference is None:
1340 if f.get('ext') in ['f4f', 'f4m']: # Not yet supported
1343 protocol = f.get('protocol') or determine_protocol(f)
1344 proto_preference = 0 if protocol in ['http', 'https'] else (-0.5 if protocol == 'rtsp' else -0.1)
1346 if f.get('vcodec') == 'none': # audio only
1348 if self._downloader.params.get('prefer_free_formats'):
1349 ORDER = ['aac', 'mp3', 'm4a', 'webm', 'ogg', 'opus']
1351 ORDER = ['webm', 'opus', 'ogg', 'mp3', 'aac', 'm4a']
1354 audio_ext_preference = ORDER.index(f['ext'])
1356 audio_ext_preference = -1
1358 if f.get('acodec') == 'none': # video only
1360 if self._downloader.params.get('prefer_free_formats'):
1361 ORDER = ['flv', 'mp4', 'webm']
1363 ORDER = ['webm', 'flv', 'mp4']
1365 ext_preference = ORDER.index(f['ext'])
1368 audio_ext_preference = 0
1372 f.get('language_preference') if f.get('language_preference') is not None else -1,
1373 f.get('quality') if f.get('quality') is not None else -1,
1374 f.get('tbr') if f.get('tbr') is not None else -1,
1375 f.get('filesize') if f.get('filesize') is not None else -1,
1376 f.get('vbr') if f.get('vbr') is not None else -1,
1377 f.get('height') if f.get('height') is not None else -1,
1378 f.get('width') if f.get('width') is not None else -1,
1381 f.get('abr') if f.get('abr') is not None else -1,
1382 audio_ext_preference,
1383 f.get('fps') if f.get('fps') is not None else -1,
1384 f.get('filesize_approx') if f.get('filesize_approx') is not None else -1,
1385 f.get('source_preference') if f.get('source_preference') is not None else -1,
1386 f.get('format_id') if f.get('format_id') is not None else '',
1388 formats.sort(key=_formats_key)
1390 def _check_formats(self, formats, video_id):
1392 formats[:] = filter(
1393 lambda f: self._is_valid_url(
1395 item='%s video format' % f.get('format_id') if f.get('format_id') else 'video'),
1399 def _remove_duplicate_formats(formats):
1403 if f['url'] not in format_urls:
1404 format_urls.add(f['url'])
1405 unique_formats.append(f)
1406 formats[:] = unique_formats
1408 def _is_valid_url(self, url, video_id, item='video', headers={}):
1409 url = self._proto_relative_url(url, scheme='http:')
1410 # For now assume non HTTP(S) URLs always valid
1411 if not (url.startswith('http://') or url.startswith('https://')):
1414 self._request_webpage(url, video_id, 'Checking %s URL' % item, headers=headers)
1416 except ExtractorError as e:
1417 if isinstance(e.cause, compat_urllib_error.URLError):
1419 '%s: %s URL is invalid, skipping' % (video_id, item))
1423 def http_scheme(self):
1424 """ Either "http:" or "https:", depending on the user's preferences """
1427 if self._downloader.params.get('prefer_insecure', False)
1430 def _proto_relative_url(self, url, scheme=None):
1433 if url.startswith('//'):
1435 scheme = self.http_scheme()
1440 def _sleep(self, timeout, video_id, msg_template=None):
1441 if msg_template is None:
1442 msg_template = '%(video_id)s: Waiting for %(timeout)s seconds'
1443 msg = msg_template % {'video_id': video_id, 'timeout': timeout}
1447 def _extract_f4m_formats(self, manifest_url, video_id, preference=None, f4m_id=None,
1448 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1449 fatal=True, m3u8_id=None):
1450 manifest = self._download_xml(
1451 manifest_url, video_id, 'Downloading f4m manifest',
1452 'Unable to download f4m manifest',
1453 # Some manifests may be malformed, e.g. prosiebensat1 generated manifests
1454 # (see https://github.com/rg3/youtube-dl/issues/6215#issuecomment-121704244)
1455 transform_source=transform_source,
1458 if manifest is False:
1461 return self._parse_f4m_formats(
1462 manifest, manifest_url, video_id, preference=preference, f4m_id=f4m_id,
1463 transform_source=transform_source, fatal=fatal, m3u8_id=m3u8_id)
1465 def _parse_f4m_formats(self, manifest, manifest_url, video_id, preference=None, f4m_id=None,
1466 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1467 fatal=True, m3u8_id=None):
1468 if not isinstance(manifest, xml.etree.ElementTree.Element) and not fatal:
1471 # currently youtube-dl cannot decode the playerVerificationChallenge as Akamai uses Adobe Alchemy
1472 akamai_pv = manifest.find('{http://ns.adobe.com/f4m/1.0}pv-2.0')
1473 if akamai_pv is not None and ';' in akamai_pv.text:
1474 playerVerificationChallenge = akamai_pv.text.split(';')[0]
1475 if playerVerificationChallenge.strip() != '':
1479 manifest_version = '1.0'
1480 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/1.0}media')
1482 manifest_version = '2.0'
1483 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/2.0}media')
1484 # Remove unsupported DRM protected media from final formats
1485 # rendition (see https://github.com/rg3/youtube-dl/issues/8573).
1486 media_nodes = remove_encrypted_media(media_nodes)
1490 manifest_base_url = get_base_url(manifest)
1492 bootstrap_info = xpath_element(
1493 manifest, ['{http://ns.adobe.com/f4m/1.0}bootstrapInfo', '{http://ns.adobe.com/f4m/2.0}bootstrapInfo'],
1494 'bootstrap info', default=None)
1497 mime_type = xpath_text(
1498 manifest, ['{http://ns.adobe.com/f4m/1.0}mimeType', '{http://ns.adobe.com/f4m/2.0}mimeType'],
1499 'base URL', default=None)
1500 if mime_type and mime_type.startswith('audio/'):
1503 for i, media_el in enumerate(media_nodes):
1504 tbr = int_or_none(media_el.attrib.get('bitrate'))
1505 width = int_or_none(media_el.attrib.get('width'))
1506 height = int_or_none(media_el.attrib.get('height'))
1507 format_id = '-'.join(filter(None, [f4m_id, compat_str(i if tbr is None else tbr)]))
1508 # If <bootstrapInfo> is present, the specified f4m is a
1509 # stream-level manifest, and only set-level manifests may refer to
1510 # external resources. See section 11.4 and section 4 of F4M spec
1511 if bootstrap_info is None:
1513 # @href is introduced in 2.0, see section 11.6 of F4M spec
1514 if manifest_version == '2.0':
1515 media_url = media_el.attrib.get('href')
1516 if media_url is None:
1517 media_url = media_el.attrib.get('url')
1521 media_url if media_url.startswith('http://') or media_url.startswith('https://')
1522 else ((manifest_base_url or '/'.join(manifest_url.split('/')[:-1])) + '/' + media_url))
1523 # If media_url is itself a f4m manifest do the recursive extraction
1524 # since bitrates in parent manifest (this one) and media_url manifest
1525 # may differ leading to inability to resolve the format by requested
1526 # bitrate in f4m downloader
1527 ext = determine_ext(manifest_url)
1529 f4m_formats = self._extract_f4m_formats(
1530 manifest_url, video_id, preference=preference, f4m_id=f4m_id,
1531 transform_source=transform_source, fatal=fatal)
1532 # Sometimes stream-level manifest contains single media entry that
1533 # does not contain any quality metadata (e.g. http://matchtv.ru/#live-player).
1534 # At the same time parent's media entry in set-level manifest may
1535 # contain it. We will copy it from parent in such cases.
1536 if len(f4m_formats) == 1:
1539 'tbr': f.get('tbr') or tbr,
1540 'width': f.get('width') or width,
1541 'height': f.get('height') or height,
1542 'format_id': f.get('format_id') if not tbr else format_id,
1545 formats.extend(f4m_formats)
1548 formats.extend(self._extract_m3u8_formats(
1549 manifest_url, video_id, 'mp4', preference=preference,
1550 m3u8_id=m3u8_id, fatal=fatal))
1553 'format_id': format_id,
1554 'url': manifest_url,
1555 'manifest_url': manifest_url,
1556 'ext': 'flv' if bootstrap_info is not None else None,
1562 'preference': preference,
1566 def _m3u8_meta_format(self, m3u8_url, ext=None, preference=None, m3u8_id=None):
1568 'format_id': '-'.join(filter(None, [m3u8_id, 'meta'])),
1572 'preference': preference - 100 if preference else -100,
1573 'resolution': 'multiple',
1574 'format_note': 'Quality selection URL',
1577 def _extract_m3u8_formats(self, m3u8_url, video_id, ext=None,
1578 entry_protocol='m3u8', preference=None,
1579 m3u8_id=None, note=None, errnote=None,
1580 fatal=True, live=False):
1581 res = self._download_webpage_handle(
1583 note=note or 'Downloading m3u8 information',
1584 errnote=errnote or 'Failed to download m3u8 information',
1590 m3u8_doc, urlh = res
1591 m3u8_url = urlh.geturl()
1593 return self._parse_m3u8_formats(
1594 m3u8_doc, m3u8_url, ext=ext, entry_protocol=entry_protocol,
1595 preference=preference, m3u8_id=m3u8_id, live=live)
1597 def _parse_m3u8_formats(self, m3u8_doc, m3u8_url, ext=None,
1598 entry_protocol='m3u8', preference=None,
1599 m3u8_id=None, live=False):
1600 if '#EXT-X-FAXS-CM:' in m3u8_doc: # Adobe Flash Access
1603 if re.search(r'#EXT-X-SESSION-KEY:.*?URI="skd://', m3u8_doc): # Apple FairPlay
1608 format_url = lambda u: (
1610 if re.match(r'^https?://', u)
1611 else compat_urlparse.urljoin(m3u8_url, u))
1614 # 1. https://tools.ietf.org/html/draft-pantos-http-live-streaming-21
1615 # 2. https://github.com/rg3/youtube-dl/issues/12211
1616 # 3. https://github.com/rg3/youtube-dl/issues/18923
1618 # We should try extracting formats only from master playlists [1, 4.3.4],
1619 # i.e. playlists that describe available qualities. On the other hand
1620 # media playlists [1, 4.3.3] should be returned as is since they contain
1621 # just the media without qualities renditions.
1622 # Fortunately, master playlist can be easily distinguished from media
1623 # playlist based on particular tags availability. As of [1, 4.3.3, 4.3.4]
1624 # master playlist tags MUST NOT appear in a media playist and vice versa.
1625 # As of [1, 4.3.3.1] #EXT-X-TARGETDURATION tag is REQUIRED for every
1626 # media playlist and MUST NOT appear in master playlist thus we can
1627 # clearly detect media playlist with this criterion.
1629 if '#EXT-X-TARGETDURATION' in m3u8_doc: # media playlist, return as is
1632 'format_id': m3u8_id,
1634 'protocol': entry_protocol,
1635 'preference': preference,
1639 last_stream_inf = {}
1641 def extract_media(x_media_line):
1642 media = parse_m3u8_attributes(x_media_line)
1643 # As per [1, 4.3.4.1] TYPE, GROUP-ID and NAME are REQUIRED
1644 media_type, group_id, name = media.get('TYPE'), media.get('GROUP-ID'), media.get('NAME')
1645 if not (media_type and group_id and name):
1647 groups.setdefault(group_id, []).append(media)
1648 if media_type not in ('VIDEO', 'AUDIO'):
1650 media_url = media.get('URI')
1653 for v in (m3u8_id, group_id, name):
1657 'format_id': '-'.join(format_id),
1658 'url': format_url(media_url),
1659 'manifest_url': m3u8_url,
1660 'language': media.get('LANGUAGE'),
1662 'protocol': entry_protocol,
1663 'preference': preference,
1665 if media_type == 'AUDIO':
1666 f['vcodec'] = 'none'
1669 def build_stream_name():
1670 # Despite specification does not mention NAME attribute for
1671 # EXT-X-STREAM-INF tag it still sometimes may be present (see [1]
1672 # or vidio test in TestInfoExtractor.test_parse_m3u8_formats)
1673 # 1. http://www.vidio.com/watch/165683-dj_ambred-booyah-live-2015
1674 stream_name = last_stream_inf.get('NAME')
1677 # If there is no NAME in EXT-X-STREAM-INF it will be obtained
1678 # from corresponding rendition group
1679 stream_group_id = last_stream_inf.get('VIDEO')
1680 if not stream_group_id:
1682 stream_group = groups.get(stream_group_id)
1683 if not stream_group:
1684 return stream_group_id
1685 rendition = stream_group[0]
1686 return rendition.get('NAME') or stream_group_id
1688 # parse EXT-X-MEDIA tags before EXT-X-STREAM-INF in order to have the
1689 # chance to detect video only formats when EXT-X-STREAM-INF tags
1690 # precede EXT-X-MEDIA tags in HLS manifest such as [3].
1691 for line in m3u8_doc.splitlines():
1692 if line.startswith('#EXT-X-MEDIA:'):
1695 for line in m3u8_doc.splitlines():
1696 if line.startswith('#EXT-X-STREAM-INF:'):
1697 last_stream_inf = parse_m3u8_attributes(line)
1698 elif line.startswith('#') or not line.strip():
1701 tbr = float_or_none(
1702 last_stream_inf.get('AVERAGE-BANDWIDTH') or
1703 last_stream_inf.get('BANDWIDTH'), scale=1000)
1706 format_id.append(m3u8_id)
1707 stream_name = build_stream_name()
1708 # Bandwidth of live streams may differ over time thus making
1709 # format_id unpredictable. So it's better to keep provided
1712 format_id.append(stream_name if stream_name else '%d' % (tbr if tbr else len(formats)))
1713 manifest_url = format_url(line.strip())
1715 'format_id': '-'.join(format_id),
1716 'url': manifest_url,
1717 'manifest_url': m3u8_url,
1720 'fps': float_or_none(last_stream_inf.get('FRAME-RATE')),
1721 'protocol': entry_protocol,
1722 'preference': preference,
1724 resolution = last_stream_inf.get('RESOLUTION')
1726 mobj = re.search(r'(?P<width>\d+)[xX](?P<height>\d+)', resolution)
1728 f['width'] = int(mobj.group('width'))
1729 f['height'] = int(mobj.group('height'))
1730 # Unified Streaming Platform
1732 r'audio.*?(?:%3D|=)(\d+)(?:-video.*?(?:%3D|=)(\d+))?', f['url'])
1734 abr, vbr = mobj.groups()
1735 abr, vbr = float_or_none(abr, 1000), float_or_none(vbr, 1000)
1740 codecs = parse_codecs(last_stream_inf.get('CODECS'))
1742 audio_group_id = last_stream_inf.get('AUDIO')
1743 # As per [1, 4.3.4.1.1] any EXT-X-STREAM-INF tag which
1744 # references a rendition group MUST have a CODECS attribute.
1745 # However, this is not always respected, for example, [2]
1746 # contains EXT-X-STREAM-INF tag which references AUDIO
1747 # rendition group but does not have CODECS and despite
1748 # referencing an audio group it represents a complete
1749 # (with audio and video) format. So, for such cases we will
1750 # ignore references to rendition groups and treat them
1751 # as complete formats.
1752 if audio_group_id and codecs and f.get('vcodec') != 'none':
1753 audio_group = groups.get(audio_group_id)
1754 if audio_group and audio_group[0].get('URI'):
1755 # TODO: update acodec for audio only formats with
1757 f['acodec'] = 'none'
1759 last_stream_inf = {}
1763 def _xpath_ns(path, namespace=None):
1767 for c in path.split('/'):
1768 if not c or c == '.':
1771 out.append('{%s}%s' % (namespace, c))
1772 return '/'.join(out)
1774 def _extract_smil_formats(self, smil_url, video_id, fatal=True, f4m_params=None, transform_source=None):
1775 smil = self._download_smil(smil_url, video_id, fatal=fatal, transform_source=transform_source)
1781 namespace = self._parse_smil_namespace(smil)
1783 return self._parse_smil_formats(
1784 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
1786 def _extract_smil_info(self, smil_url, video_id, fatal=True, f4m_params=None):
1787 smil = self._download_smil(smil_url, video_id, fatal=fatal)
1790 return self._parse_smil(smil, smil_url, video_id, f4m_params=f4m_params)
1792 def _download_smil(self, smil_url, video_id, fatal=True, transform_source=None):
1793 return self._download_xml(
1794 smil_url, video_id, 'Downloading SMIL file',
1795 'Unable to download SMIL file', fatal=fatal, transform_source=transform_source)
1797 def _parse_smil(self, smil, smil_url, video_id, f4m_params=None):
1798 namespace = self._parse_smil_namespace(smil)
1800 formats = self._parse_smil_formats(
1801 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
1802 subtitles = self._parse_smil_subtitles(smil, namespace=namespace)
1804 video_id = os.path.splitext(url_basename(smil_url))[0]
1808 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
1809 name = meta.attrib.get('name')
1810 content = meta.attrib.get('content')
1811 if not name or not content:
1813 if not title and name == 'title':
1815 elif not description and name in ('description', 'abstract'):
1816 description = content
1817 elif not upload_date and name == 'date':
1818 upload_date = unified_strdate(content)
1821 'id': image.get('type'),
1822 'url': image.get('src'),
1823 'width': int_or_none(image.get('width')),
1824 'height': int_or_none(image.get('height')),
1825 } for image in smil.findall(self._xpath_ns('.//image', namespace)) if image.get('src')]
1829 'title': title or video_id,
1830 'description': description,
1831 'upload_date': upload_date,
1832 'thumbnails': thumbnails,
1834 'subtitles': subtitles,
1837 def _parse_smil_namespace(self, smil):
1838 return self._search_regex(
1839 r'(?i)^{([^}]+)?}smil$', smil.tag, 'namespace', default=None)
1841 def _parse_smil_formats(self, smil, smil_url, video_id, namespace=None, f4m_params=None, transform_rtmp_url=None):
1843 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
1844 b = meta.get('base') or meta.get('httpBase')
1855 media = smil.findall(self._xpath_ns('.//video', namespace)) + smil.findall(self._xpath_ns('.//audio', namespace))
1856 for medium in media:
1857 src = medium.get('src')
1858 if not src or src in srcs:
1862 bitrate = float_or_none(medium.get('system-bitrate') or medium.get('systemBitrate'), 1000)
1863 filesize = int_or_none(medium.get('size') or medium.get('fileSize'))
1864 width = int_or_none(medium.get('width'))
1865 height = int_or_none(medium.get('height'))
1866 proto = medium.get('proto')
1867 ext = medium.get('ext')
1868 src_ext = determine_ext(src)
1869 streamer = medium.get('streamer') or base
1871 if proto == 'rtmp' or streamer.startswith('rtmp'):
1877 'format_id': 'rtmp-%d' % (rtmp_count if bitrate is None else bitrate),
1879 'filesize': filesize,
1883 if transform_rtmp_url:
1884 streamer, src = transform_rtmp_url(streamer, src)
1885 formats[-1].update({
1891 src_url = src if src.startswith('http') else compat_urlparse.urljoin(base, src)
1892 src_url = src_url.strip()
1894 if proto == 'm3u8' or src_ext == 'm3u8':
1895 m3u8_formats = self._extract_m3u8_formats(
1896 src_url, video_id, ext or 'mp4', m3u8_id='hls', fatal=False)
1897 if len(m3u8_formats) == 1:
1899 m3u8_formats[0].update({
1900 'format_id': 'hls-%d' % (m3u8_count if bitrate is None else bitrate),
1905 formats.extend(m3u8_formats)
1906 elif src_ext == 'f4m':
1911 'plugin': 'flowplayer-3.2.0.1',
1913 f4m_url += '&' if '?' in f4m_url else '?'
1914 f4m_url += compat_urllib_parse_urlencode(f4m_params)
1915 formats.extend(self._extract_f4m_formats(f4m_url, video_id, f4m_id='hds', fatal=False))
1916 elif src_ext == 'mpd':
1917 formats.extend(self._extract_mpd_formats(
1918 src_url, video_id, mpd_id='dash', fatal=False))
1919 elif re.search(r'\.ism/[Mm]anifest', src_url):
1920 formats.extend(self._extract_ism_formats(
1921 src_url, video_id, ism_id='mss', fatal=False))
1922 elif src_url.startswith('http') and self._is_valid_url(src, video_id):
1926 'ext': ext or src_ext or 'flv',
1927 'format_id': 'http-%d' % (bitrate or http_count),
1929 'filesize': filesize,
1936 def _parse_smil_subtitles(self, smil, namespace=None, subtitles_lang='en'):
1939 for num, textstream in enumerate(smil.findall(self._xpath_ns('.//textstream', namespace))):
1940 src = textstream.get('src')
1941 if not src or src in urls:
1944 ext = textstream.get('ext') or mimetype2ext(textstream.get('type')) or determine_ext(src)
1945 lang = textstream.get('systemLanguage') or textstream.get('systemLanguageName') or textstream.get('lang') or subtitles_lang
1946 subtitles.setdefault(lang, []).append({
1952 def _extract_xspf_playlist(self, xspf_url, playlist_id, fatal=True):
1953 xspf = self._download_xml(
1954 xspf_url, playlist_id, 'Downloading xpsf playlist',
1955 'Unable to download xspf manifest', fatal=fatal)
1958 return self._parse_xspf(
1959 xspf, playlist_id, xspf_url=xspf_url,
1960 xspf_base_url=base_url(xspf_url))
1962 def _parse_xspf(self, xspf_doc, playlist_id, xspf_url=None, xspf_base_url=None):
1964 'xspf': 'http://xspf.org/ns/0/',
1965 's1': 'http://static.streamone.nl/player/ns/0',
1969 for track in xspf_doc.findall(xpath_with_ns('./xspf:trackList/xspf:track', NS_MAP)):
1971 track, xpath_with_ns('./xspf:title', NS_MAP), 'title', default=playlist_id)
1972 description = xpath_text(
1973 track, xpath_with_ns('./xspf:annotation', NS_MAP), 'description')
1974 thumbnail = xpath_text(
1975 track, xpath_with_ns('./xspf:image', NS_MAP), 'thumbnail')
1976 duration = float_or_none(
1977 xpath_text(track, xpath_with_ns('./xspf:duration', NS_MAP), 'duration'), 1000)
1980 for location in track.findall(xpath_with_ns('./xspf:location', NS_MAP)):
1981 format_url = urljoin(xspf_base_url, location.text)
1986 'manifest_url': xspf_url,
1987 'format_id': location.get(xpath_with_ns('s1:label', NS_MAP)),
1988 'width': int_or_none(location.get(xpath_with_ns('s1:width', NS_MAP))),
1989 'height': int_or_none(location.get(xpath_with_ns('s1:height', NS_MAP))),
1991 self._sort_formats(formats)
1996 'description': description,
1997 'thumbnail': thumbnail,
1998 'duration': duration,
2003 def _extract_mpd_formats(self, mpd_url, video_id, mpd_id=None, note=None, errnote=None, fatal=True, formats_dict={}):
2004 res = self._download_xml_handle(
2006 note=note or 'Downloading MPD manifest',
2007 errnote=errnote or 'Failed to download MPD manifest',
2012 mpd_base_url = base_url(urlh.geturl())
2014 return self._parse_mpd_formats(
2015 mpd_doc, mpd_id=mpd_id, mpd_base_url=mpd_base_url,
2016 formats_dict=formats_dict, mpd_url=mpd_url)
2018 def _parse_mpd_formats(self, mpd_doc, mpd_id=None, mpd_base_url='', formats_dict={}, mpd_url=None):
2020 Parse formats from MPD manifest.
2022 1. MPEG-DASH Standard, ISO/IEC 23009-1:2014(E),
2023 http://standards.iso.org/ittf/PubliclyAvailableStandards/c065274_ISO_IEC_23009-1_2014.zip
2024 2. https://en.wikipedia.org/wiki/Dynamic_Adaptive_Streaming_over_HTTP
2026 if mpd_doc.get('type') == 'dynamic':
2029 namespace = self._search_regex(r'(?i)^{([^}]+)?}MPD$', mpd_doc.tag, 'namespace', default=None)
2032 return self._xpath_ns(path, namespace)
2034 def is_drm_protected(element):
2035 return element.find(_add_ns('ContentProtection')) is not None
2037 def extract_multisegment_info(element, ms_parent_info):
2038 ms_info = ms_parent_info.copy()
2040 # As per [1, 5.3.9.2.2] SegmentList and SegmentTemplate share some
2041 # common attributes and elements. We will only extract relevant
2043 def extract_common(source):
2044 segment_timeline = source.find(_add_ns('SegmentTimeline'))
2045 if segment_timeline is not None:
2046 s_e = segment_timeline.findall(_add_ns('S'))
2048 ms_info['total_number'] = 0
2051 r = int(s.get('r', 0))
2052 ms_info['total_number'] += 1 + r
2053 ms_info['s'].append({
2054 't': int(s.get('t', 0)),
2055 # @d is mandatory (see [1, 5.3.9.6.2, Table 17, page 60])
2056 'd': int(s.attrib['d']),
2059 start_number = source.get('startNumber')
2061 ms_info['start_number'] = int(start_number)
2062 timescale = source.get('timescale')
2064 ms_info['timescale'] = int(timescale)
2065 segment_duration = source.get('duration')
2066 if segment_duration:
2067 ms_info['segment_duration'] = float(segment_duration)
2069 def extract_Initialization(source):
2070 initialization = source.find(_add_ns('Initialization'))
2071 if initialization is not None:
2072 ms_info['initialization_url'] = initialization.attrib['sourceURL']
2074 segment_list = element.find(_add_ns('SegmentList'))
2075 if segment_list is not None:
2076 extract_common(segment_list)
2077 extract_Initialization(segment_list)
2078 segment_urls_e = segment_list.findall(_add_ns('SegmentURL'))
2080 ms_info['segment_urls'] = [segment.attrib['media'] for segment in segment_urls_e]
2082 segment_template = element.find(_add_ns('SegmentTemplate'))
2083 if segment_template is not None:
2084 extract_common(segment_template)
2085 media = segment_template.get('media')
2087 ms_info['media'] = media
2088 initialization = segment_template.get('initialization')
2090 ms_info['initialization'] = initialization
2092 extract_Initialization(segment_template)
2095 mpd_duration = parse_duration(mpd_doc.get('mediaPresentationDuration'))
2097 for period in mpd_doc.findall(_add_ns('Period')):
2098 period_duration = parse_duration(period.get('duration')) or mpd_duration
2099 period_ms_info = extract_multisegment_info(period, {
2103 for adaptation_set in period.findall(_add_ns('AdaptationSet')):
2104 if is_drm_protected(adaptation_set):
2106 adaption_set_ms_info = extract_multisegment_info(adaptation_set, period_ms_info)
2107 for representation in adaptation_set.findall(_add_ns('Representation')):
2108 if is_drm_protected(representation):
2110 representation_attrib = adaptation_set.attrib.copy()
2111 representation_attrib.update(representation.attrib)
2112 # According to [1, 5.3.7.2, Table 9, page 41], @mimeType is mandatory
2113 mime_type = representation_attrib['mimeType']
2114 content_type = mime_type.split('/')[0]
2115 if content_type == 'text':
2116 # TODO implement WebVTT downloading
2118 elif content_type in ('video', 'audio'):
2120 for element in (representation, adaptation_set, period, mpd_doc):
2121 base_url_e = element.find(_add_ns('BaseURL'))
2122 if base_url_e is not None:
2123 base_url = base_url_e.text + base_url
2124 if re.match(r'^https?://', base_url):
2126 if mpd_base_url and not re.match(r'^https?://', base_url):
2127 if not mpd_base_url.endswith('/') and not base_url.startswith('/'):
2129 base_url = mpd_base_url + base_url
2130 representation_id = representation_attrib.get('id')
2131 lang = representation_attrib.get('lang')
2132 url_el = representation.find(_add_ns('BaseURL'))
2133 filesize = int_or_none(url_el.attrib.get('{http://youtube.com/yt/2012/10/10}contentLength') if url_el is not None else None)
2134 bandwidth = int_or_none(representation_attrib.get('bandwidth'))
2136 'format_id': '%s-%s' % (mpd_id, representation_id) if mpd_id else representation_id,
2138 'manifest_url': mpd_url,
2139 'ext': mimetype2ext(mime_type),
2140 'width': int_or_none(representation_attrib.get('width')),
2141 'height': int_or_none(representation_attrib.get('height')),
2142 'tbr': float_or_none(bandwidth, 1000),
2143 'asr': int_or_none(representation_attrib.get('audioSamplingRate')),
2144 'fps': int_or_none(representation_attrib.get('frameRate')),
2145 'language': lang if lang not in ('mul', 'und', 'zxx', 'mis') else None,
2146 'format_note': 'DASH %s' % content_type,
2147 'filesize': filesize,
2148 'container': mimetype2ext(mime_type) + '_dash',
2150 f.update(parse_codecs(representation_attrib.get('codecs')))
2151 representation_ms_info = extract_multisegment_info(representation, adaption_set_ms_info)
2153 def prepare_template(template_name, identifiers):
2154 tmpl = representation_ms_info[template_name]
2155 # First of, % characters outside $...$ templates
2156 # must be escaped by doubling for proper processing
2157 # by % operator string formatting used further (see
2158 # https://github.com/rg3/youtube-dl/issues/16867).
2164 in_template = not in_template
2165 elif c == '%' and not in_template:
2167 # Next, $...$ templates are translated to their
2168 # %(...) counterparts to be used with % operator
2169 t = t.replace('$RepresentationID$', representation_id)
2170 t = re.sub(r'\$(%s)\$' % '|'.join(identifiers), r'%(\1)d', t)
2171 t = re.sub(r'\$(%s)%%([^$]+)\$' % '|'.join(identifiers), r'%(\1)\2', t)
2172 t.replace('$$', '$')
2175 # @initialization is a regular template like @media one
2176 # so it should be handled just the same way (see
2177 # https://github.com/rg3/youtube-dl/issues/11605)
2178 if 'initialization' in representation_ms_info:
2179 initialization_template = prepare_template(
2181 # As per [1, 5.3.9.4.2, Table 15, page 54] $Number$ and
2182 # $Time$ shall not be included for @initialization thus
2183 # only $Bandwidth$ remains
2185 representation_ms_info['initialization_url'] = initialization_template % {
2186 'Bandwidth': bandwidth,
2189 def location_key(location):
2190 return 'url' if re.match(r'^https?://', location) else 'path'
2192 if 'segment_urls' not in representation_ms_info and 'media' in representation_ms_info:
2194 media_template = prepare_template('media', ('Number', 'Bandwidth', 'Time'))
2195 media_location_key = location_key(media_template)
2197 # As per [1, 5.3.9.4.4, Table 16, page 55] $Number$ and $Time$
2198 # can't be used at the same time
2199 if '%(Number' in media_template and 's' not in representation_ms_info:
2200 segment_duration = None
2201 if 'total_number' not in representation_ms_info and 'segment_duration' in representation_ms_info:
2202 segment_duration = float_or_none(representation_ms_info['segment_duration'], representation_ms_info['timescale'])
2203 representation_ms_info['total_number'] = int(math.ceil(float(period_duration) / segment_duration))
2204 representation_ms_info['fragments'] = [{
2205 media_location_key: media_template % {
2206 'Number': segment_number,
2207 'Bandwidth': bandwidth,
2209 'duration': segment_duration,
2210 } for segment_number in range(
2211 representation_ms_info['start_number'],
2212 representation_ms_info['total_number'] + representation_ms_info['start_number'])]
2214 # $Number*$ or $Time$ in media template with S list available
2215 # Example $Number*$: http://www.svtplay.se/klipp/9023742/stopptid-om-bjorn-borg
2216 # Example $Time$: https://play.arkena.com/embed/avp/v2/player/media/b41dda37-d8e7-4d3f-b1b5-9a9db578bdfe/1/129411
2217 representation_ms_info['fragments'] = []
2220 segment_number = representation_ms_info['start_number']
2222 def add_segment_url():
2223 segment_url = media_template % {
2224 'Time': segment_time,
2225 'Bandwidth': bandwidth,
2226 'Number': segment_number,
2228 representation_ms_info['fragments'].append({
2229 media_location_key: segment_url,
2230 'duration': float_or_none(segment_d, representation_ms_info['timescale']),
2233 for num, s in enumerate(representation_ms_info['s']):
2234 segment_time = s.get('t') or segment_time
2238 for r in range(s.get('r', 0)):
2239 segment_time += segment_d
2242 segment_time += segment_d
2243 elif 'segment_urls' in representation_ms_info and 's' in representation_ms_info:
2245 # Example: https://www.youtube.com/watch?v=iXZV5uAYMJI
2246 # or any YouTube dashsegments video
2249 timescale = representation_ms_info['timescale']
2250 for s in representation_ms_info['s']:
2251 duration = float_or_none(s['d'], timescale)
2252 for r in range(s.get('r', 0) + 1):
2253 segment_uri = representation_ms_info['segment_urls'][segment_index]
2255 location_key(segment_uri): segment_uri,
2256 'duration': duration,
2259 representation_ms_info['fragments'] = fragments
2260 elif 'segment_urls' in representation_ms_info:
2261 # Segment URLs with no SegmentTimeline
2262 # Example: https://www.seznam.cz/zpravy/clanek/cesko-zasahne-vitr-o-sile-vichrice-muze-byt-i-zivotu-nebezpecny-39091
2263 # https://github.com/rg3/youtube-dl/pull/14844
2265 segment_duration = float_or_none(
2266 representation_ms_info['segment_duration'],
2267 representation_ms_info['timescale']) if 'segment_duration' in representation_ms_info else None
2268 for segment_url in representation_ms_info['segment_urls']:
2270 location_key(segment_url): segment_url,
2272 if segment_duration:
2273 fragment['duration'] = segment_duration
2274 fragments.append(fragment)
2275 representation_ms_info['fragments'] = fragments
2276 # NB: MPD manifest may contain direct URLs to unfragmented media.
2277 # No fragments key is present in this case.
2278 if 'fragments' in representation_ms_info:
2280 'fragment_base_url': base_url,
2282 'protocol': 'http_dash_segments',
2284 if 'initialization_url' in representation_ms_info:
2285 initialization_url = representation_ms_info['initialization_url']
2286 if not f.get('url'):
2287 f['url'] = initialization_url
2288 f['fragments'].append({location_key(initialization_url): initialization_url})
2289 f['fragments'].extend(representation_ms_info['fragments'])
2290 # According to [1, 5.3.5.2, Table 7, page 35] @id of Representation
2291 # is not necessarily unique within a Period thus formats with
2292 # the same `format_id` are quite possible. There are numerous examples
2293 # of such manifests (see https://github.com/rg3/youtube-dl/issues/15111,
2294 # https://github.com/rg3/youtube-dl/issues/13919)
2295 full_info = formats_dict.get(representation_id, {}).copy()
2297 formats.append(full_info)
2299 self.report_warning('Unknown MIME type %s in DASH manifest' % mime_type)
2302 def _extract_ism_formats(self, ism_url, video_id, ism_id=None, note=None, errnote=None, fatal=True):
2303 res = self._download_xml_handle(
2305 note=note or 'Downloading ISM manifest',
2306 errnote=errnote or 'Failed to download ISM manifest',
2312 return self._parse_ism_formats(ism_doc, urlh.geturl(), ism_id)
2314 def _parse_ism_formats(self, ism_doc, ism_url, ism_id=None):
2316 Parse formats from ISM manifest.
2318 1. [MS-SSTR]: Smooth Streaming Protocol,
2319 https://msdn.microsoft.com/en-us/library/ff469518.aspx
2321 if ism_doc.get('IsLive') == 'TRUE' or ism_doc.find('Protection') is not None:
2324 duration = int(ism_doc.attrib['Duration'])
2325 timescale = int_or_none(ism_doc.get('TimeScale')) or 10000000
2328 for stream in ism_doc.findall('StreamIndex'):
2329 stream_type = stream.get('Type')
2330 if stream_type not in ('video', 'audio'):
2332 url_pattern = stream.attrib['Url']
2333 stream_timescale = int_or_none(stream.get('TimeScale')) or timescale
2334 stream_name = stream.get('Name')
2335 for track in stream.findall('QualityLevel'):
2336 fourcc = track.get('FourCC', 'AACL' if track.get('AudioTag') == '255' else None)
2337 # TODO: add support for WVC1 and WMAP
2338 if fourcc not in ('H264', 'AVC1', 'AACL'):
2339 self.report_warning('%s is not a supported codec' % fourcc)
2341 tbr = int(track.attrib['Bitrate']) // 1000
2342 # [1] does not mention Width and Height attributes. However,
2343 # they're often present while MaxWidth and MaxHeight are
2344 # missing, so should be used as fallbacks
2345 width = int_or_none(track.get('MaxWidth') or track.get('Width'))
2346 height = int_or_none(track.get('MaxHeight') or track.get('Height'))
2347 sampling_rate = int_or_none(track.get('SamplingRate'))
2349 track_url_pattern = re.sub(r'{[Bb]itrate}', track.attrib['Bitrate'], url_pattern)
2350 track_url_pattern = compat_urlparse.urljoin(ism_url, track_url_pattern)
2356 stream_fragments = stream.findall('c')
2357 for stream_fragment_index, stream_fragment in enumerate(stream_fragments):
2358 fragment_ctx['time'] = int_or_none(stream_fragment.get('t')) or fragment_ctx['time']
2359 fragment_repeat = int_or_none(stream_fragment.get('r')) or 1
2360 fragment_ctx['duration'] = int_or_none(stream_fragment.get('d'))
2361 if not fragment_ctx['duration']:
2363 next_fragment_time = int(stream_fragment[stream_fragment_index + 1].attrib['t'])
2365 next_fragment_time = duration
2366 fragment_ctx['duration'] = (next_fragment_time - fragment_ctx['time']) / fragment_repeat
2367 for _ in range(fragment_repeat):
2369 'url': re.sub(r'{start[ _]time}', compat_str(fragment_ctx['time']), track_url_pattern),
2370 'duration': fragment_ctx['duration'] / stream_timescale,
2372 fragment_ctx['time'] += fragment_ctx['duration']
2376 format_id.append(ism_id)
2378 format_id.append(stream_name)
2379 format_id.append(compat_str(tbr))
2382 'format_id': '-'.join(format_id),
2384 'manifest_url': ism_url,
2385 'ext': 'ismv' if stream_type == 'video' else 'isma',
2389 'asr': sampling_rate,
2390 'vcodec': 'none' if stream_type == 'audio' else fourcc,
2391 'acodec': 'none' if stream_type == 'video' else fourcc,
2393 'fragments': fragments,
2394 '_download_params': {
2395 'duration': duration,
2396 'timescale': stream_timescale,
2397 'width': width or 0,
2398 'height': height or 0,
2400 'codec_private_data': track.get('CodecPrivateData'),
2401 'sampling_rate': sampling_rate,
2402 'channels': int_or_none(track.get('Channels', 2)),
2403 'bits_per_sample': int_or_none(track.get('BitsPerSample', 16)),
2404 'nal_unit_length_field': int_or_none(track.get('NALUnitLengthField', 4)),
2409 def _parse_html5_media_entries(self, base_url, webpage, video_id, m3u8_id=None, m3u8_entry_protocol='m3u8', mpd_id=None, preference=None):
2410 def absolute_url(item_url):
2411 return urljoin(base_url, item_url)
2413 def parse_content_type(content_type):
2414 if not content_type:
2416 ctr = re.search(r'(?P<mimetype>[^/]+/[^;]+)(?:;\s*codecs="?(?P<codecs>[^"]+))?', content_type)
2418 mimetype, codecs = ctr.groups()
2419 f = parse_codecs(codecs)
2420 f['ext'] = mimetype2ext(mimetype)
2424 def _media_formats(src, cur_media_type, type_info={}):
2425 full_url = absolute_url(src)
2426 ext = type_info.get('ext') or determine_ext(full_url)
2428 is_plain_url = False
2429 formats = self._extract_m3u8_formats(
2430 full_url, video_id, ext='mp4',
2431 entry_protocol=m3u8_entry_protocol, m3u8_id=m3u8_id,
2432 preference=preference, fatal=False)
2434 is_plain_url = False
2435 formats = self._extract_mpd_formats(
2436 full_url, video_id, mpd_id=mpd_id, fatal=False)
2441 'vcodec': 'none' if cur_media_type == 'audio' else None,
2443 return is_plain_url, formats
2446 # amp-video and amp-audio are very similar to their HTML5 counterparts
2447 # so we wll include them right here (see
2448 # https://www.ampproject.org/docs/reference/components/amp-video)
2449 media_tags = [(media_tag, media_type, '')
2450 for media_tag, media_type
2451 in re.findall(r'(?s)(<(?:amp-)?(video|audio)[^>]*/>)', webpage)]
2452 media_tags.extend(re.findall(
2453 # We only allow video|audio followed by a whitespace or '>'.
2454 # Allowing more characters may end up in significant slow down (see
2455 # https://github.com/rg3/youtube-dl/issues/11979, example URL:
2456 # http://www.porntrex.com/maps/videositemap.xml).
2457 r'(?s)(<(?P<tag>(?:amp-)?(?:video|audio))(?:\s+[^>]*)?>)(.*?)</(?P=tag)>', webpage))
2458 for media_tag, media_type, media_content in media_tags:
2463 media_attributes = extract_attributes(media_tag)
2464 src = media_attributes.get('src')
2466 _, formats = _media_formats(src, media_type)
2467 media_info['formats'].extend(formats)
2468 media_info['thumbnail'] = absolute_url(media_attributes.get('poster'))
2470 for source_tag in re.findall(r'<source[^>]+>', media_content):
2471 source_attributes = extract_attributes(source_tag)
2472 src = source_attributes.get('src')
2475 f = parse_content_type(source_attributes.get('type'))
2476 is_plain_url, formats = _media_formats(src, media_type, f)
2478 # res attribute is not standard but seen several times
2481 'height': int_or_none(source_attributes.get('res')),
2482 'format_id': source_attributes.get('label'),
2484 f.update(formats[0])
2485 media_info['formats'].append(f)
2487 media_info['formats'].extend(formats)
2488 for track_tag in re.findall(r'<track[^>]+>', media_content):
2489 track_attributes = extract_attributes(track_tag)
2490 kind = track_attributes.get('kind')
2491 if not kind or kind in ('subtitles', 'captions'):
2492 src = track_attributes.get('src')
2495 lang = track_attributes.get('srclang') or track_attributes.get('lang') or track_attributes.get('label')
2496 media_info['subtitles'].setdefault(lang, []).append({
2497 'url': absolute_url(src),
2499 for f in media_info['formats']:
2500 f.setdefault('http_headers', {})['Referer'] = base_url
2501 if media_info['formats'] or media_info['subtitles']:
2502 entries.append(media_info)
2505 def _extract_akamai_formats(self, manifest_url, video_id, hosts={}):
2507 hdcore_sign = 'hdcore=3.7.0'
2508 f4m_url = re.sub(r'(https?://[^/]+)/i/', r'\1/z/', manifest_url).replace('/master.m3u8', '/manifest.f4m')
2509 hds_host = hosts.get('hds')
2511 f4m_url = re.sub(r'(https?://)[^/]+', r'\1' + hds_host, f4m_url)
2512 if 'hdcore=' not in f4m_url:
2513 f4m_url += ('&' if '?' in f4m_url else '?') + hdcore_sign
2514 f4m_formats = self._extract_f4m_formats(
2515 f4m_url, video_id, f4m_id='hds', fatal=False)
2516 for entry in f4m_formats:
2517 entry.update({'extra_param_to_segment_url': hdcore_sign})
2518 formats.extend(f4m_formats)
2519 m3u8_url = re.sub(r'(https?://[^/]+)/z/', r'\1/i/', manifest_url).replace('/manifest.f4m', '/master.m3u8')
2520 hls_host = hosts.get('hls')
2522 m3u8_url = re.sub(r'(https?://)[^/]+', r'\1' + hls_host, m3u8_url)
2523 formats.extend(self._extract_m3u8_formats(
2524 m3u8_url, video_id, 'mp4', 'm3u8_native',
2525 m3u8_id='hls', fatal=False))
2528 def _extract_wowza_formats(self, url, video_id, m3u8_entry_protocol='m3u8_native', skip_protocols=[]):
2529 query = compat_urlparse.urlparse(url).query
2530 url = re.sub(r'/(?:manifest|playlist|jwplayer)\.(?:m3u8|f4m|mpd|smil)', '', url)
2532 r'(?:(?:http|rtmp|rtsp)(?P<s>s)?:)?(?P<url>//[^?]+)', url)
2533 url_base = mobj.group('url')
2534 http_base_url = '%s%s:%s' % ('http', mobj.group('s') or '', url_base)
2537 def manifest_url(manifest):
2538 m_url = '%s/%s' % (http_base_url, manifest)
2540 m_url += '?%s' % query
2543 if 'm3u8' not in skip_protocols:
2544 formats.extend(self._extract_m3u8_formats(
2545 manifest_url('playlist.m3u8'), video_id, 'mp4',
2546 m3u8_entry_protocol, m3u8_id='hls', fatal=False))
2547 if 'f4m' not in skip_protocols:
2548 formats.extend(self._extract_f4m_formats(
2549 manifest_url('manifest.f4m'),
2550 video_id, f4m_id='hds', fatal=False))
2551 if 'dash' not in skip_protocols:
2552 formats.extend(self._extract_mpd_formats(
2553 manifest_url('manifest.mpd'),
2554 video_id, mpd_id='dash', fatal=False))
2555 if re.search(r'(?:/smil:|\.smil)', url_base):
2556 if 'smil' not in skip_protocols:
2557 rtmp_formats = self._extract_smil_formats(
2558 manifest_url('jwplayer.smil'),
2559 video_id, fatal=False)
2560 for rtmp_format in rtmp_formats:
2561 rtsp_format = rtmp_format.copy()
2562 rtsp_format['url'] = '%s/%s' % (rtmp_format['url'], rtmp_format['play_path'])
2563 del rtsp_format['play_path']
2564 del rtsp_format['ext']
2565 rtsp_format.update({
2566 'url': rtsp_format['url'].replace('rtmp://', 'rtsp://'),
2567 'format_id': rtmp_format['format_id'].replace('rtmp', 'rtsp'),
2570 formats.extend([rtmp_format, rtsp_format])
2572 for protocol in ('rtmp', 'rtsp'):
2573 if protocol not in skip_protocols:
2575 'url': '%s:%s' % (protocol, url_base),
2576 'format_id': protocol,
2577 'protocol': protocol,
2581 def _find_jwplayer_data(self, webpage, video_id=None, transform_source=js_to_json):
2583 r'(?s)jwplayer\((?P<quote>[\'"])[^\'" ]+(?P=quote)\)(?!</script>).*?\.setup\s*\((?P<options>[^)]+)\)',
2587 jwplayer_data = self._parse_json(mobj.group('options'),
2589 transform_source=transform_source)
2590 except ExtractorError:
2593 if isinstance(jwplayer_data, dict):
2594 return jwplayer_data
2596 def _extract_jwplayer_data(self, webpage, video_id, *args, **kwargs):
2597 jwplayer_data = self._find_jwplayer_data(
2598 webpage, video_id, transform_source=js_to_json)
2599 return self._parse_jwplayer_data(
2600 jwplayer_data, video_id, *args, **kwargs)
2602 def _parse_jwplayer_data(self, jwplayer_data, video_id=None, require_title=True,
2603 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
2604 # JWPlayer backward compatibility: flattened playlists
2605 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/api/config.js#L81-L96
2606 if 'playlist' not in jwplayer_data:
2607 jwplayer_data = {'playlist': [jwplayer_data]}
2611 # JWPlayer backward compatibility: single playlist item
2612 # https://github.com/jwplayer/jwplayer/blob/v7.7.0/src/js/playlist/playlist.js#L10
2613 if not isinstance(jwplayer_data['playlist'], list):
2614 jwplayer_data['playlist'] = [jwplayer_data['playlist']]
2616 for video_data in jwplayer_data['playlist']:
2617 # JWPlayer backward compatibility: flattened sources
2618 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/playlist/item.js#L29-L35
2619 if 'sources' not in video_data:
2620 video_data['sources'] = [video_data]
2622 this_video_id = video_id or video_data['mediaid']
2624 formats = self._parse_jwplayer_formats(
2625 video_data['sources'], video_id=this_video_id, m3u8_id=m3u8_id,
2626 mpd_id=mpd_id, rtmp_params=rtmp_params, base_url=base_url)
2629 tracks = video_data.get('tracks')
2630 if tracks and isinstance(tracks, list):
2631 for track in tracks:
2632 if not isinstance(track, dict):
2634 track_kind = track.get('kind')
2635 if not track_kind or not isinstance(track_kind, compat_str):
2637 if track_kind.lower() not in ('captions', 'subtitles'):
2639 track_url = urljoin(base_url, track.get('file'))
2642 subtitles.setdefault(track.get('label') or 'en', []).append({
2643 'url': self._proto_relative_url(track_url)
2647 'id': this_video_id,
2648 'title': unescapeHTML(video_data['title'] if require_title else video_data.get('title')),
2649 'description': video_data.get('description'),
2650 'thumbnail': urljoin(base_url, self._proto_relative_url(video_data.get('image'))),
2651 'timestamp': int_or_none(video_data.get('pubdate')),
2652 'duration': float_or_none(jwplayer_data.get('duration') or video_data.get('duration')),
2653 'subtitles': subtitles,
2655 # https://github.com/jwplayer/jwplayer/blob/master/src/js/utils/validator.js#L32
2656 if len(formats) == 1 and re.search(r'^(?:http|//).*(?:youtube\.com|youtu\.be)/.+', formats[0]['url']):
2658 '_type': 'url_transparent',
2659 'url': formats[0]['url'],
2662 self._sort_formats(formats)
2663 entry['formats'] = formats
2664 entries.append(entry)
2665 if len(entries) == 1:
2668 return self.playlist_result(entries)
2670 def _parse_jwplayer_formats(self, jwplayer_sources_data, video_id=None,
2671 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
2674 for source in jwplayer_sources_data:
2675 if not isinstance(source, dict):
2677 source_url = urljoin(
2678 base_url, self._proto_relative_url(source.get('file')))
2679 if not source_url or source_url in urls:
2681 urls.append(source_url)
2682 source_type = source.get('type') or ''
2683 ext = mimetype2ext(source_type) or determine_ext(source_url)
2684 if source_type == 'hls' or ext == 'm3u8':
2685 formats.extend(self._extract_m3u8_formats(
2686 source_url, video_id, 'mp4', entry_protocol='m3u8_native',
2687 m3u8_id=m3u8_id, fatal=False))
2688 elif source_type == 'dash' or ext == 'mpd':
2689 formats.extend(self._extract_mpd_formats(
2690 source_url, video_id, mpd_id=mpd_id, fatal=False))
2692 formats.extend(self._extract_smil_formats(
2693 source_url, video_id, fatal=False))
2694 # https://github.com/jwplayer/jwplayer/blob/master/src/js/providers/default.js#L67
2695 elif source_type.startswith('audio') or ext in (
2696 'oga', 'aac', 'mp3', 'mpeg', 'vorbis'):
2703 height = int_or_none(source.get('height'))
2705 # Often no height is provided but there is a label in
2706 # format like "1080p", "720p SD", or 1080.
2707 height = int_or_none(self._search_regex(
2708 r'^(\d{3,4})[pP]?(?:\b|$)', compat_str(source.get('label') or ''),
2709 'height', default=None))
2712 'width': int_or_none(source.get('width')),
2714 'tbr': int_or_none(source.get('bitrate')),
2717 if source_url.startswith('rtmp'):
2718 a_format['ext'] = 'flv'
2719 # See com/longtailvideo/jwplayer/media/RTMPMediaProvider.as
2720 # of jwplayer.flash.swf
2721 rtmp_url_parts = re.split(
2722 r'((?:mp4|mp3|flv):)', source_url, 1)
2723 if len(rtmp_url_parts) == 3:
2724 rtmp_url, prefix, play_path = rtmp_url_parts
2727 'play_path': prefix + play_path,
2730 a_format.update(rtmp_params)
2731 formats.append(a_format)
2734 def _live_title(self, name):
2735 """ Generate the title for a live video """
2736 now = datetime.datetime.now()
2737 now_str = now.strftime('%Y-%m-%d %H:%M')
2738 return name + ' ' + now_str
2740 def _int(self, v, name, fatal=False, **kwargs):
2741 res = int_or_none(v, **kwargs)
2742 if 'get_attr' in kwargs:
2743 print(getattr(v, kwargs['get_attr']))
2745 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
2747 raise ExtractorError(msg)
2749 self._downloader.report_warning(msg)
2752 def _float(self, v, name, fatal=False, **kwargs):
2753 res = float_or_none(v, **kwargs)
2755 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
2757 raise ExtractorError(msg)
2759 self._downloader.report_warning(msg)
2762 def _set_cookie(self, domain, name, value, expire_time=None, port=None,
2763 path='/', secure=False, discard=False, rest={}, **kwargs):
2764 cookie = compat_cookiejar.Cookie(
2765 0, name, value, port, port is not None, domain, True,
2766 domain.startswith('.'), path, True, secure, expire_time,
2767 discard, None, None, rest)
2768 self._downloader.cookiejar.set_cookie(cookie)
2770 def _get_cookies(self, url):
2771 """ Return a compat_cookies.SimpleCookie with the cookies for the url """
2772 req = sanitized_Request(url)
2773 self._downloader.cookiejar.add_cookie_header(req)
2774 return compat_cookies.SimpleCookie(req.get_header('Cookie'))
2776 def get_testcases(self, include_onlymatching=False):
2777 t = getattr(self, '_TEST', None)
2779 assert not hasattr(self, '_TESTS'), \
2780 '%s has _TEST and _TESTS' % type(self).__name__
2783 tests = getattr(self, '_TESTS', [])
2785 if not include_onlymatching and t.get('only_matching', False):
2787 t['name'] = type(self).__name__[:-len('IE')]
2790 def is_suitable(self, age_limit):
2791 """ Test whether the extractor is generally suitable for the given
2792 age limit (i.e. pornographic sites are not, all others usually are) """
2794 any_restricted = False
2795 for tc in self.get_testcases(include_onlymatching=False):
2796 if tc.get('playlist', []):
2797 tc = tc['playlist'][0]
2798 is_restricted = age_restricted(
2799 tc.get('info_dict', {}).get('age_limit'), age_limit)
2800 if not is_restricted:
2802 any_restricted = any_restricted or is_restricted
2803 return not any_restricted
2805 def extract_subtitles(self, *args, **kwargs):
2806 if (self._downloader.params.get('writesubtitles', False) or
2807 self._downloader.params.get('listsubtitles')):
2808 return self._get_subtitles(*args, **kwargs)
2811 def _get_subtitles(self, *args, **kwargs):
2812 raise NotImplementedError('This method must be implemented by subclasses')
2815 def _merge_subtitle_items(subtitle_list1, subtitle_list2):
2816 """ Merge subtitle items for one language. Items with duplicated URLs
2817 will be dropped. """
2818 list1_urls = set([item['url'] for item in subtitle_list1])
2819 ret = list(subtitle_list1)
2820 ret.extend([item for item in subtitle_list2 if item['url'] not in list1_urls])
2824 def _merge_subtitles(cls, subtitle_dict1, subtitle_dict2):
2825 """ Merge two subtitle dictionaries, language by language. """
2826 ret = dict(subtitle_dict1)
2827 for lang in subtitle_dict2:
2828 ret[lang] = cls._merge_subtitle_items(subtitle_dict1.get(lang, []), subtitle_dict2[lang])
2831 def extract_automatic_captions(self, *args, **kwargs):
2832 if (self._downloader.params.get('writeautomaticsub', False) or
2833 self._downloader.params.get('listsubtitles')):
2834 return self._get_automatic_captions(*args, **kwargs)
2837 def _get_automatic_captions(self, *args, **kwargs):
2838 raise NotImplementedError('This method must be implemented by subclasses')
2840 def mark_watched(self, *args, **kwargs):
2841 if (self._downloader.params.get('mark_watched', False) and
2842 (self._get_login_info()[0] is not None or
2843 self._downloader.params.get('cookiefile') is not None)):
2844 self._mark_watched(*args, **kwargs)
2846 def _mark_watched(self, *args, **kwargs):
2847 raise NotImplementedError('This method must be implemented by subclasses')
2849 def geo_verification_headers(self):
2851 geo_verification_proxy = self._downloader.params.get('geo_verification_proxy')
2852 if geo_verification_proxy:
2853 headers['Ytdl-request-proxy'] = geo_verification_proxy
2856 def _generic_id(self, url):
2857 return compat_urllib_parse_unquote(os.path.splitext(url.rstrip('/').split('/')[-1])[0])
2859 def _generic_title(self, url):
2860 return compat_urllib_parse_unquote(os.path.splitext(url_basename(url))[0])
2863 class SearchInfoExtractor(InfoExtractor):
2865 Base class for paged search queries extractors.
2866 They accept URLs in the format _SEARCH_KEY(|all|[0-9]):{query}
2867 Instances should define _SEARCH_KEY and _MAX_RESULTS.
2871 def _make_valid_url(cls):
2872 return r'%s(?P<prefix>|[1-9][0-9]*|all):(?P<query>[\s\S]+)' % cls._SEARCH_KEY
2875 def suitable(cls, url):
2876 return re.match(cls._make_valid_url(), url) is not None
2878 def _real_extract(self, query):
2879 mobj = re.match(self._make_valid_url(), query)
2881 raise ExtractorError('Invalid search query "%s"' % query)
2883 prefix = mobj.group('prefix')
2884 query = mobj.group('query')
2886 return self._get_n_results(query, 1)
2887 elif prefix == 'all':
2888 return self._get_n_results(query, self._MAX_RESULTS)
2892 raise ExtractorError('invalid download number %s for query "%s"' % (n, query))
2893 elif n > self._MAX_RESULTS:
2894 self._downloader.report_warning('%s returns max %i results (you requested %i)' % (self._SEARCH_KEY, self._MAX_RESULTS, n))
2895 n = self._MAX_RESULTS
2896 return self._get_n_results(query, n)
2898 def _get_n_results(self, query, n):
2899 """Get a specified number of results for a query"""
2900 raise NotImplementedError('This method must be implemented by subclasses')
2903 def SEARCH_KEY(self):
2904 return self._SEARCH_KEY