1 from __future__ import unicode_literals
16 from ..compat import (
19 compat_etree_fromstring,
25 compat_urllib_parse_unquote,
26 compat_urllib_parse_urlencode,
27 compat_urllib_request,
30 from ..downloader.f4m import remove_encrypted_media
63 parse_m3u8_attributes,
70 class InfoExtractor(object):
71 """Information Extractor class.
73 Information extractors are the classes that, given a URL, extract
74 information about the video (or videos) the URL refers to. This
75 information includes the real video URL, the video title, author and
76 others. The information is stored in a dictionary which is then
77 passed to the YoutubeDL. The YoutubeDL processes this
78 information possibly downloading the video to the file system, among
79 other possible outcomes.
81 The type field determines the type of the result.
82 By far the most common value (and the default if _type is missing) is
83 "video", which indicates a single video.
85 For a video, the dictionaries must include the following fields:
88 title: Video title, unescaped.
90 Additionally, it must contain either a formats entry or a url one:
92 formats: A list of dictionaries for each format available, ordered
93 from worst to best quality.
96 * url Mandatory. The URL of the video file
98 The URL of the manifest file in case of
99 fragmented media (DASH, hls, hds)
100 * ext Will be calculated from URL if missing
101 * format A human-readable description of the format
102 ("mp4 container with h264/opus").
103 Calculated from the format_id, width, height.
104 and format_note fields if missing.
105 * format_id A short description of the format
106 ("mp4_h264_opus" or "19").
107 Technically optional, but strongly recommended.
108 * format_note Additional info about the format
109 ("3D" or "DASH video")
110 * width Width of the video, if known
111 * height Height of the video, if known
112 * resolution Textual description of width and height
113 * tbr Average bitrate of audio and video in KBit/s
114 * abr Average audio bitrate in KBit/s
115 * acodec Name of the audio codec in use
116 * asr Audio sampling rate in Hertz
117 * vbr Average video bitrate in KBit/s
119 * vcodec Name of the video codec in use
120 * container Name of the container format
121 * filesize The number of bytes, if known in advance
122 * filesize_approx An estimate for the number of bytes
123 * player_url SWF Player URL (used for rtmpdump).
124 * protocol The protocol that will be used for the actual
125 download, lower-case.
126 "http", "https", "rtsp", "rtmp", "rtmpe",
127 "m3u8", "m3u8_native" or "http_dash_segments".
129 Base URL for fragments. Each fragment's path
130 value (if present) will be relative to
132 * fragments A list of fragments of a fragmented media.
133 Each fragment entry must contain either an url
134 or a path. If an url is present it should be
135 considered by a client. Otherwise both path and
136 fragment_base_url must be present. Here is
137 the list of all potential fields:
138 * "url" - fragment's URL
139 * "path" - fragment's path relative to
141 * "duration" (optional, int or float)
142 * "filesize" (optional, int)
143 * preference Order number of this format. If this field is
144 present and not None, the formats get sorted
145 by this field, regardless of all other values.
146 -1 for default (order by other properties),
147 -2 or smaller for less than default.
148 < -1000 to hide the format (if there is
149 another one which is strictly better)
150 * language Language code, e.g. "de" or "en-US".
151 * language_preference Is this in the language mentioned in
153 10 if it's what the URL is about,
154 -1 for default (don't know),
155 -10 otherwise, other values reserved for now.
156 * quality Order number of the video quality of this
157 format, irrespective of the file format.
158 -1 for default (order by other properties),
159 -2 or smaller for less than default.
160 * source_preference Order number for this video source
161 (quality takes higher priority)
162 -1 for default (order by other properties),
163 -2 or smaller for less than default.
164 * http_headers A dictionary of additional HTTP headers
165 to add to the request.
166 * stretched_ratio If given and not 1, indicates that the
167 video's pixels are not square.
168 width : height ratio as float.
169 * no_resume The server does not support resuming the
170 (HTTP or RTMP) download. Boolean.
172 url: Final video URL.
173 ext: Video filename extension.
174 format: The video format, defaults to ext (used for --get-format)
175 player_url: SWF Player URL (used for rtmpdump).
177 The following fields are optional:
179 alt_title: A secondary title of the video.
180 display_id An alternative identifier for the video, not necessarily
181 unique, but available before title. Typically, id is
182 something like "4234987", title "Dancing naked mole rats",
183 and display_id "dancing-naked-mole-rats"
184 thumbnails: A list of dictionaries, with the following entries:
185 * "id" (optional, string) - Thumbnail format ID
187 * "preference" (optional, int) - quality of the image
188 * "width" (optional, int)
189 * "height" (optional, int)
190 * "resolution" (optional, string "{width}x{height"},
192 * "filesize" (optional, int)
193 thumbnail: Full URL to a video thumbnail image.
194 description: Full video description.
195 uploader: Full name of the video uploader.
196 license: License name the video is licensed under.
197 creator: The creator of the video.
198 release_date: The date (YYYYMMDD) when the video was released.
199 timestamp: UNIX timestamp of the moment the video became available.
200 upload_date: Video upload date (YYYYMMDD).
201 If not explicitly set, calculated from timestamp.
202 uploader_id: Nickname or id of the video uploader.
203 uploader_url: Full URL to a personal webpage of the video uploader.
204 location: Physical location where the video was filmed.
205 subtitles: The available subtitles as a dictionary in the format
206 {tag: subformats}. "tag" is usually a language code, and
207 "subformats" is a list sorted from lower to higher
208 preference, each element is a dictionary with the "ext"
210 * "data": The subtitles file contents
211 * "url": A URL pointing to the subtitles file
212 "ext" will be calculated from URL if missing
213 automatic_captions: Like 'subtitles', used by the YoutubeIE for
214 automatically generated captions
215 duration: Length of the video in seconds, as an integer or float.
216 view_count: How many users have watched the video on the platform.
217 like_count: Number of positive ratings of the video
218 dislike_count: Number of negative ratings of the video
219 repost_count: Number of reposts of the video
220 average_rating: Average rating give by users, the scale used depends on the webpage
221 comment_count: Number of comments on the video
222 comments: A list of comments, each with one or more of the following
223 properties (all but one of text or html optional):
224 * "author" - human-readable name of the comment author
225 * "author_id" - user ID of the comment author
227 * "html" - Comment as HTML
228 * "text" - Plain text of the comment
229 * "timestamp" - UNIX timestamp of comment
230 * "parent" - ID of the comment this one is replying to.
231 Set to "root" to indicate that this is a
232 comment to the original video.
233 age_limit: Age restriction for the video, as an integer (years)
234 webpage_url: The URL to the video webpage, if given to youtube-dl it
235 should allow to get the same result again. (It will be set
236 by YoutubeDL if it's missing)
237 categories: A list of categories that the video falls in, for example
239 tags: A list of tags assigned to the video, e.g. ["sweden", "pop music"]
240 is_live: True, False, or None (=unknown). Whether this video is a
241 live stream that goes on instead of a fixed-length video.
242 start_time: Time in seconds where the reproduction should start, as
243 specified in the URL.
244 end_time: Time in seconds where the reproduction should end, as
245 specified in the URL.
247 The following fields should only be used when the video belongs to some logical
250 chapter: Name or title of the chapter the video belongs to.
251 chapter_number: Number of the chapter the video belongs to, as an integer.
252 chapter_id: Id of the chapter the video belongs to, as a unicode string.
254 The following fields should only be used when the video is an episode of some
255 series, programme or podcast:
257 series: Title of the series or programme the video episode belongs to.
258 season: Title of the season the video episode belongs to.
259 season_number: Number of the season the video episode belongs to, as an integer.
260 season_id: Id of the season the video episode belongs to, as a unicode string.
261 episode: Title of the video episode. Unlike mandatory video title field,
262 this field should denote the exact title of the video episode
263 without any kind of decoration.
264 episode_number: Number of the video episode within a season, as an integer.
265 episode_id: Id of the video episode, as a unicode string.
267 The following fields should only be used when the media is a track or a part of
270 track: Title of the track.
271 track_number: Number of the track within an album or a disc, as an integer.
272 track_id: Id of the track (useful in case of custom indexing, e.g. 6.iii),
274 artist: Artist(s) of the track.
275 genre: Genre(s) of the track.
276 album: Title of the album the track belongs to.
277 album_type: Type of the album (e.g. "Demo", "Full-length", "Split", "Compilation", etc).
278 album_artist: List of all artists appeared on the album (e.g.
279 "Ash Borer / Fell Voices" or "Various Artists", useful for splits
281 disc_number: Number of the disc or other physical medium the track belongs to,
283 release_year: Year (YYYY) when the album was released.
285 Unless mentioned otherwise, the fields should be Unicode strings.
287 Unless mentioned otherwise, None is equivalent to absence of information.
290 _type "playlist" indicates multiple videos.
291 There must be a key "entries", which is a list, an iterable, or a PagedList
292 object, each element of which is a valid dictionary by this specification.
294 Additionally, playlists can have "title", "description" and "id" attributes
295 with the same semantics as videos (see above).
298 _type "multi_video" indicates that there are multiple videos that
299 form a single show, for examples multiple acts of an opera or TV episode.
300 It must have an entries key like a playlist and contain all the keys
301 required for a video at the same time.
304 _type "url" indicates that the video must be extracted from another
305 location, possibly by a different extractor. Its only required key is:
306 "url" - the next URL to extract.
307 The key "ie_key" can be set to the class name (minus the trailing "IE",
308 e.g. "Youtube") if the extractor class is known in advance.
309 Additionally, the dictionary may have any properties of the resolved entity
310 known in advance, for example "title" if the title of the referred video is
314 _type "url_transparent" entities have the same specification as "url", but
315 indicate that the given additional information is more precise than the one
316 associated with the resolved URL.
317 This is useful when a site employs a video service that hosts the video and
318 its technical metadata, but that video service does not embed a useful
319 title, description etc.
322 Subclasses of this one should re-define the _real_initialize() and
323 _real_extract() methods and define a _VALID_URL regexp.
324 Probably, they should also be added to the list of extractors.
326 _GEO_BYPASS attribute may be set to False in order to disable
327 geo restriction bypass mechanisms for a particular extractor.
328 Though it won't disable explicit geo restriction bypass based on
329 country code provided with geo_bypass_country. (experimental)
331 _GEO_COUNTRIES attribute may contain a list of presumably geo unrestricted
332 countries for this extractor. One of these countries will be used by
333 geo restriction bypass mechanism right away in order to bypass
334 geo restriction, of course, if the mechanism is not disabled. (experimental)
336 NB: both these geo attributes are experimental and may change in future
337 or be completely removed.
339 Finally, the _WORKING attribute should be set to False for broken IEs
340 in order to warn the users and skip the tests.
345 _x_forwarded_for_ip = None
347 _GEO_COUNTRIES = None
350 def __init__(self, downloader=None):
351 """Constructor. Receives an optional downloader."""
353 self._x_forwarded_for_ip = None
354 self.set_downloader(downloader)
357 def suitable(cls, url):
358 """Receives a URL and returns True if suitable for this IE."""
360 # This does not use has/getattr intentionally - we want to know whether
361 # we have cached the regexp for *this* class, whereas getattr would also
362 # match the superclass
363 if '_VALID_URL_RE' not in cls.__dict__:
364 cls._VALID_URL_RE = re.compile(cls._VALID_URL)
365 return cls._VALID_URL_RE.match(url) is not None
368 def _match_id(cls, url):
369 if '_VALID_URL_RE' not in cls.__dict__:
370 cls._VALID_URL_RE = re.compile(cls._VALID_URL)
371 m = cls._VALID_URL_RE.match(url)
377 """Getter method for _WORKING."""
380 def initialize(self):
381 """Initializes an instance (authentication, etc)."""
382 self.__initialize_geo_bypass()
384 self._real_initialize()
387 def __initialize_geo_bypass(self):
388 if not self._x_forwarded_for_ip:
389 country_code = self._downloader.params.get('geo_bypass_country', None)
390 # If there is no explicit country for geo bypass specified and
391 # the extractor is known to be geo restricted let's fake IP
392 # as X-Forwarded-For right away.
393 if (not country_code and
395 self._downloader.params.get('geo_bypass', True) and
396 self._GEO_COUNTRIES):
397 country_code = random.choice(self._GEO_COUNTRIES)
399 self._x_forwarded_for_ip = GeoUtils.random_ipv4(country_code)
400 if self._downloader.params.get('verbose', False):
401 self._downloader.to_stdout(
402 '[debug] Using fake %s IP as X-Forwarded-For.' % self._x_forwarded_for_ip)
404 def extract(self, url):
405 """Extracts URL information and returns it in list of dicts."""
410 ie_result = self._real_extract(url)
411 if self._x_forwarded_for_ip:
412 ie_result['__x_forwarded_for_ip'] = self._x_forwarded_for_ip
414 except GeoRestrictedError as e:
415 if self.__maybe_fake_ip_and_retry(e.countries):
418 except ExtractorError:
420 except compat_http_client.IncompleteRead as e:
421 raise ExtractorError('A network error has occurred.', cause=e, expected=True)
422 except (KeyError, StopIteration) as e:
423 raise ExtractorError('An extractor error has occurred.', cause=e)
425 def __maybe_fake_ip_and_retry(self, countries):
426 if (not self._downloader.params.get('geo_bypass_country', None) and
428 self._downloader.params.get('geo_bypass', True) and
429 not self._x_forwarded_for_ip and
431 self._x_forwarded_for_ip = GeoUtils.random_ipv4(random.choice(countries))
432 if self._x_forwarded_for_ip:
434 'Video is geo restricted. Retrying extraction with fake %s IP as X-Forwarded-For.' % self._x_forwarded_for_ip)
438 def set_downloader(self, downloader):
439 """Sets the downloader for this IE."""
440 self._downloader = downloader
442 def _real_initialize(self):
443 """Real initialization process. Redefine in subclasses."""
446 def _real_extract(self, url):
447 """Real extraction process. Redefine in subclasses."""
452 """A string for getting the InfoExtractor with get_info_extractor"""
453 return compat_str(cls.__name__[:-2])
457 return compat_str(type(self).__name__[:-2])
459 def _request_webpage(self, url_or_request, video_id, note=None, errnote=None, fatal=True, data=None, headers={}, query={}):
460 """ Returns the response handle """
462 self.report_download_webpage(video_id)
463 elif note is not False:
465 self.to_screen('%s' % (note,))
467 self.to_screen('%s: %s' % (video_id, note))
468 if isinstance(url_or_request, compat_urllib_request.Request):
469 url_or_request = update_Request(
470 url_or_request, data=data, headers=headers, query=query)
473 url_or_request = update_url_query(url_or_request, query)
474 if data is not None or headers:
475 url_or_request = sanitized_Request(url_or_request, data, headers)
477 return self._downloader.urlopen(url_or_request)
478 except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err:
482 errnote = 'Unable to download webpage'
484 errmsg = '%s: %s' % (errnote, error_to_compat_str(err))
486 raise ExtractorError(errmsg, sys.exc_info()[2], cause=err)
488 self._downloader.report_warning(errmsg)
491 def _download_webpage_handle(self, url_or_request, video_id, note=None, errnote=None, fatal=True, encoding=None, data=None, headers={}, query={}):
492 """ Returns a tuple (page content as string, URL handle) """
493 # Strip hashes from the URL (#1038)
494 if isinstance(url_or_request, (compat_str, str)):
495 url_or_request = url_or_request.partition('#')[0]
497 # Some sites check X-Forwarded-For HTTP header in order to figure out
498 # the origin of the client behind proxy. This allows bypassing geo
499 # restriction by faking this header's value to IP that belongs to some
500 # geo unrestricted country. We will do so once we encounter any
501 # geo restriction error.
502 if self._x_forwarded_for_ip:
503 if 'X-Forwarded-For' not in headers:
504 headers['X-Forwarded-For'] = self._x_forwarded_for_ip
506 urlh = self._request_webpage(url_or_request, video_id, note, errnote, fatal, data=data, headers=headers, query=query)
510 content = self._webpage_read_content(urlh, url_or_request, video_id, note, errnote, fatal, encoding=encoding)
511 return (content, urlh)
514 def _guess_encoding_from_content(content_type, webpage_bytes):
515 m = re.match(r'[a-zA-Z0-9_.-]+/[a-zA-Z0-9_.-]+\s*;\s*charset=(.+)', content_type)
517 encoding = m.group(1)
519 m = re.search(br'<meta[^>]+charset=[\'"]?([^\'")]+)[ /\'">]',
520 webpage_bytes[:1024])
522 encoding = m.group(1).decode('ascii')
523 elif webpage_bytes.startswith(b'\xff\xfe'):
530 def _webpage_read_content(self, urlh, url_or_request, video_id, note=None, errnote=None, fatal=True, prefix=None, encoding=None):
531 content_type = urlh.headers.get('Content-Type', '')
532 webpage_bytes = urlh.read()
533 if prefix is not None:
534 webpage_bytes = prefix + webpage_bytes
536 encoding = self._guess_encoding_from_content(content_type, webpage_bytes)
537 if self._downloader.params.get('dump_intermediate_pages', False):
539 url = url_or_request.get_full_url()
540 except AttributeError:
542 self.to_screen('Dumping request to ' + url)
543 dump = base64.b64encode(webpage_bytes).decode('ascii')
544 self._downloader.to_screen(dump)
545 if self._downloader.params.get('write_pages', False):
547 url = url_or_request.get_full_url()
548 except AttributeError:
550 basen = '%s_%s' % (video_id, url)
552 h = '___' + hashlib.md5(basen.encode('utf-8')).hexdigest()
553 basen = basen[:240 - len(h)] + h
554 raw_filename = basen + '.dump'
555 filename = sanitize_filename(raw_filename, restricted=True)
556 self.to_screen('Saving request to ' + filename)
557 # Working around MAX_PATH limitation on Windows (see
558 # http://msdn.microsoft.com/en-us/library/windows/desktop/aa365247(v=vs.85).aspx)
559 if compat_os_name == 'nt':
560 absfilepath = os.path.abspath(filename)
561 if len(absfilepath) > 259:
562 filename = '\\\\?\\' + absfilepath
563 with open(filename, 'wb') as outf:
564 outf.write(webpage_bytes)
567 content = webpage_bytes.decode(encoding, 'replace')
569 content = webpage_bytes.decode('utf-8', 'replace')
571 if ('<title>Access to this site is blocked</title>' in content and
572 'Websense' in content[:512]):
573 msg = 'Access to this webpage has been blocked by Websense filtering software in your network.'
574 blocked_iframe = self._html_search_regex(
575 r'<iframe src="([^"]+)"', content,
576 'Websense information URL', default=None)
578 msg += ' Visit %s for more details' % blocked_iframe
579 raise ExtractorError(msg, expected=True)
580 if '<title>The URL you requested has been blocked</title>' in content[:512]:
582 'Access to this webpage has been blocked by Indian censorship. '
583 'Use a VPN or proxy server (with --proxy) to route around it.')
584 block_msg = self._html_search_regex(
585 r'</h1><p>(.*?)</p>',
586 content, 'block message', default=None)
588 msg += ' (Message: "%s")' % block_msg.replace('\n', ' ')
589 raise ExtractorError(msg, expected=True)
593 def _download_webpage(self, url_or_request, video_id, note=None, errnote=None, fatal=True, tries=1, timeout=5, encoding=None, data=None, headers={}, query={}):
594 """ Returns the data of the page as a string """
597 while success is False:
599 res = self._download_webpage_handle(url_or_request, video_id, note, errnote, fatal, encoding=encoding, data=data, headers=headers, query=query)
601 except compat_http_client.IncompleteRead as e:
603 if try_count >= tries:
605 self._sleep(timeout, video_id)
612 def _download_xml(self, url_or_request, video_id,
613 note='Downloading XML', errnote='Unable to download XML',
614 transform_source=None, fatal=True, encoding=None, data=None, headers={}, query={}):
615 """Return the xml as an xml.etree.ElementTree.Element"""
616 xml_string = self._download_webpage(
617 url_or_request, video_id, note, errnote, fatal=fatal, encoding=encoding, data=data, headers=headers, query=query)
618 if xml_string is False:
621 xml_string = transform_source(xml_string)
622 return compat_etree_fromstring(xml_string.encode('utf-8'))
624 def _download_json(self, url_or_request, video_id,
625 note='Downloading JSON metadata',
626 errnote='Unable to download JSON metadata',
627 transform_source=None,
628 fatal=True, encoding=None, data=None, headers={}, query={}):
629 json_string = self._download_webpage(
630 url_or_request, video_id, note, errnote, fatal=fatal,
631 encoding=encoding, data=data, headers=headers, query=query)
632 if (not fatal) and json_string is False:
634 return self._parse_json(
635 json_string, video_id, transform_source=transform_source, fatal=fatal)
637 def _parse_json(self, json_string, video_id, transform_source=None, fatal=True):
639 json_string = transform_source(json_string)
641 return json.loads(json_string)
642 except ValueError as ve:
643 errmsg = '%s: Failed to parse JSON ' % video_id
645 raise ExtractorError(errmsg, cause=ve)
647 self.report_warning(errmsg + str(ve))
649 def report_warning(self, msg, video_id=None):
650 idstr = '' if video_id is None else '%s: ' % video_id
651 self._downloader.report_warning(
652 '[%s] %s%s' % (self.IE_NAME, idstr, msg))
654 def to_screen(self, msg):
655 """Print msg to screen, prefixing it with '[ie_name]'"""
656 self._downloader.to_screen('[%s] %s' % (self.IE_NAME, msg))
658 def report_extraction(self, id_or_name):
659 """Report information extraction."""
660 self.to_screen('%s: Extracting information' % id_or_name)
662 def report_download_webpage(self, video_id):
663 """Report webpage download."""
664 self.to_screen('%s: Downloading webpage' % video_id)
666 def report_age_confirmation(self):
667 """Report attempt to confirm age."""
668 self.to_screen('Confirming age')
670 def report_login(self):
671 """Report attempt to log in."""
672 self.to_screen('Logging in')
675 def raise_login_required(msg='This video is only available for registered users'):
676 raise ExtractorError(
677 '%s. Use --username and --password or --netrc to provide account credentials.' % msg,
681 def raise_geo_restricted(msg='This video is not available from your location due to geo restriction', countries=None):
682 raise GeoRestrictedError(msg, countries=countries)
684 # Methods for following #608
686 def url_result(url, ie=None, video_id=None, video_title=None):
687 """Returns a URL that points to a page that should be processed"""
688 # TODO: ie should be the class used for getting the info
689 video_info = {'_type': 'url',
692 if video_id is not None:
693 video_info['id'] = video_id
694 if video_title is not None:
695 video_info['title'] = video_title
699 def playlist_result(entries, playlist_id=None, playlist_title=None, playlist_description=None):
700 """Returns a playlist"""
701 video_info = {'_type': 'playlist',
704 video_info['id'] = playlist_id
706 video_info['title'] = playlist_title
707 if playlist_description:
708 video_info['description'] = playlist_description
711 def _search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
713 Perform a regex search on the given string, using a single or a list of
714 patterns returning the first matching group.
715 In case of failure return a default value or raise a WARNING or a
716 RegexNotFoundError, depending on fatal, specifying the field name.
718 if isinstance(pattern, (str, compat_str, compiled_regex_type)):
719 mobj = re.search(pattern, string, flags)
722 mobj = re.search(p, string, flags)
726 if not self._downloader.params.get('no_color') and compat_os_name != 'nt' and sys.stderr.isatty():
727 _name = '\033[0;34m%s\033[0m' % name
733 # return the first matching group
734 return next(g for g in mobj.groups() if g is not None)
736 return mobj.group(group)
737 elif default is not NO_DEFAULT:
740 raise RegexNotFoundError('Unable to extract %s' % _name)
742 self._downloader.report_warning('unable to extract %s' % _name + bug_reports_message())
745 def _html_search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
747 Like _search_regex, but strips HTML tags and unescapes entities.
749 res = self._search_regex(pattern, string, name, default, fatal, flags, group)
751 return clean_html(res).strip()
755 def _get_netrc_login_info(self, netrc_machine=None):
758 netrc_machine = netrc_machine or self._NETRC_MACHINE
760 if self._downloader.params.get('usenetrc', False):
762 info = netrc.netrc().authenticators(netrc_machine)
767 raise netrc.NetrcParseError(
768 'No authenticators for %s' % netrc_machine)
769 except (IOError, netrc.NetrcParseError) as err:
770 self._downloader.report_warning(
771 'parsing .netrc: %s' % error_to_compat_str(err))
773 return username, password
775 def _get_login_info(self, username_option='username', password_option='password', netrc_machine=None):
777 Get the login info as (username, password)
778 First look for the manually specified credentials using username_option
779 and password_option as keys in params dictionary. If no such credentials
780 available look in the netrc file using the netrc_machine or _NETRC_MACHINE
782 If there's no info available, return (None, None)
784 if self._downloader is None:
787 downloader_params = self._downloader.params
789 # Attempt to use provided username and password or .netrc data
790 if downloader_params.get(username_option) is not None:
791 username = downloader_params[username_option]
792 password = downloader_params[password_option]
794 username, password = self._get_netrc_login_info(netrc_machine)
796 return username, password
798 def _get_tfa_info(self, note='two-factor verification code'):
800 Get the two-factor authentication info
801 TODO - asking the user will be required for sms/phone verify
802 currently just uses the command line option
803 If there's no info available, return None
805 if self._downloader is None:
807 downloader_params = self._downloader.params
809 if downloader_params.get('twofactor') is not None:
810 return downloader_params['twofactor']
812 return compat_getpass('Type %s and press [Return]: ' % note)
814 # Helper functions for extracting OpenGraph info
816 def _og_regexes(prop):
817 content_re = r'content=(?:"([^"]+?)"|\'([^\']+?)\'|\s*([^\s"\'=<>`]+?))'
818 property_re = (r'(?:name|property)=(?:\'og:%(prop)s\'|"og:%(prop)s"|\s*og:%(prop)s\b)'
819 % {'prop': re.escape(prop)})
820 template = r'<meta[^>]+?%s[^>]+?%s'
822 template % (property_re, content_re),
823 template % (content_re, property_re),
827 def _meta_regex(prop):
828 return r'''(?isx)<meta
829 (?=[^>]+(?:itemprop|name|property|id|http-equiv)=(["\']?)%s\1)
830 [^>]+?content=(["\'])(?P<content>.*?)\2''' % re.escape(prop)
832 def _og_search_property(self, prop, html, name=None, **kargs):
833 if not isinstance(prop, (list, tuple)):
836 name = 'OpenGraph %s' % prop[0]
839 og_regexes.extend(self._og_regexes(p))
840 escaped = self._search_regex(og_regexes, html, name, flags=re.DOTALL, **kargs)
843 return unescapeHTML(escaped)
845 def _og_search_thumbnail(self, html, **kargs):
846 return self._og_search_property('image', html, 'thumbnail URL', fatal=False, **kargs)
848 def _og_search_description(self, html, **kargs):
849 return self._og_search_property('description', html, fatal=False, **kargs)
851 def _og_search_title(self, html, **kargs):
852 return self._og_search_property('title', html, **kargs)
854 def _og_search_video_url(self, html, name='video url', secure=True, **kargs):
855 regexes = self._og_regexes('video') + self._og_regexes('video:url')
857 regexes = self._og_regexes('video:secure_url') + regexes
858 return self._html_search_regex(regexes, html, name, **kargs)
860 def _og_search_url(self, html, **kargs):
861 return self._og_search_property('url', html, **kargs)
863 def _html_search_meta(self, name, html, display_name=None, fatal=False, **kwargs):
864 if not isinstance(name, (list, tuple)):
866 if display_name is None:
867 display_name = name[0]
868 return self._html_search_regex(
869 [self._meta_regex(n) for n in name],
870 html, display_name, fatal=fatal, group='content', **kwargs)
872 def _dc_search_uploader(self, html):
873 return self._html_search_meta('dc.creator', html, 'uploader')
875 def _rta_search(self, html):
876 # See http://www.rtalabel.org/index.php?content=howtofaq#single
877 if re.search(r'(?ix)<meta\s+name="rating"\s+'
878 r' content="RTA-5042-1996-1400-1577-RTA"',
883 def _media_rating_search(self, html):
884 # See http://www.tjg-designs.com/WP/metadata-code-examples-adding-metadata-to-your-web-pages/
885 rating = self._html_search_meta('rating', html)
897 return RATING_TABLE.get(rating.lower())
899 def _family_friendly_search(self, html):
900 # See http://schema.org/VideoObject
901 family_friendly = self._html_search_meta('isFamilyFriendly', html)
903 if not family_friendly:
912 return RATING_TABLE.get(family_friendly.lower())
914 def _twitter_search_player(self, html):
915 return self._html_search_meta('twitter:player', html,
916 'twitter card player')
918 def _search_json_ld(self, html, video_id, expected_type=None, **kwargs):
919 json_ld = self._search_regex(
920 r'(?s)<script[^>]+type=(["\'])application/ld\+json\1[^>]*>(?P<json_ld>.+?)</script>',
921 html, 'JSON-LD', group='json_ld', **kwargs)
922 default = kwargs.get('default', NO_DEFAULT)
924 return default if default is not NO_DEFAULT else {}
925 # JSON-LD may be malformed and thus `fatal` should be respected.
926 # At the same time `default` may be passed that assumes `fatal=False`
927 # for _search_regex. Let's simulate the same behavior here as well.
928 fatal = kwargs.get('fatal', True) if default == NO_DEFAULT else False
929 return self._json_ld(json_ld, video_id, fatal=fatal, expected_type=expected_type)
931 def _json_ld(self, json_ld, video_id, fatal=True, expected_type=None):
932 if isinstance(json_ld, compat_str):
933 json_ld = self._parse_json(json_ld, video_id, fatal=fatal)
937 if not isinstance(json_ld, (list, tuple, dict)):
939 if isinstance(json_ld, dict):
942 if e.get('@context') == 'http://schema.org':
943 item_type = e.get('@type')
944 if expected_type is not None and expected_type != item_type:
946 if item_type == 'TVEpisode':
948 'episode': unescapeHTML(e.get('name')),
949 'episode_number': int_or_none(e.get('episodeNumber')),
950 'description': unescapeHTML(e.get('description')),
952 part_of_season = e.get('partOfSeason')
953 if isinstance(part_of_season, dict) and part_of_season.get('@type') == 'TVSeason':
954 info['season_number'] = int_or_none(part_of_season.get('seasonNumber'))
955 part_of_series = e.get('partOfSeries') or e.get('partOfTVSeries')
956 if isinstance(part_of_series, dict) and part_of_series.get('@type') == 'TVSeries':
957 info['series'] = unescapeHTML(part_of_series.get('name'))
958 elif item_type == 'Article':
960 'timestamp': parse_iso8601(e.get('datePublished')),
961 'title': unescapeHTML(e.get('headline')),
962 'description': unescapeHTML(e.get('articleBody')),
964 elif item_type == 'VideoObject':
966 'url': e.get('contentUrl'),
967 'title': unescapeHTML(e.get('name')),
968 'description': unescapeHTML(e.get('description')),
969 'thumbnail': e.get('thumbnailUrl') or e.get('thumbnailURL'),
970 'duration': parse_duration(e.get('duration')),
971 'timestamp': unified_timestamp(e.get('uploadDate')),
972 'filesize': float_or_none(e.get('contentSize')),
973 'tbr': int_or_none(e.get('bitrate')),
974 'width': int_or_none(e.get('width')),
975 'height': int_or_none(e.get('height')),
978 return dict((k, v) for k, v in info.items() if v is not None)
981 def _hidden_inputs(html):
982 html = re.sub(r'<!--(?:(?!<!--).)*-->', '', html)
984 for input in re.findall(r'(?i)(<input[^>]+>)', html):
985 attrs = extract_attributes(input)
988 if attrs.get('type') not in ('hidden', 'submit'):
990 name = attrs.get('name') or attrs.get('id')
991 value = attrs.get('value')
992 if name and value is not None:
993 hidden_inputs[name] = value
996 def _form_hidden_inputs(self, form_id, html):
997 form = self._search_regex(
998 r'(?is)<form[^>]+?id=(["\'])%s\1[^>]*>(?P<form>.+?)</form>' % form_id,
999 html, '%s form' % form_id, group='form')
1000 return self._hidden_inputs(form)
1002 def _sort_formats(self, formats, field_preference=None):
1004 raise ExtractorError('No video formats found')
1007 # Automatically determine tbr when missing based on abr and vbr (improves
1008 # formats sorting in some cases)
1009 if 'tbr' not in f and f.get('abr') is not None and f.get('vbr') is not None:
1010 f['tbr'] = f['abr'] + f['vbr']
1012 def _formats_key(f):
1013 # TODO remove the following workaround
1014 from ..utils import determine_ext
1015 if not f.get('ext') and 'url' in f:
1016 f['ext'] = determine_ext(f['url'])
1018 if isinstance(field_preference, (list, tuple)):
1021 if f.get(field) is not None
1022 else ('' if field == 'format_id' else -1)
1023 for field in field_preference)
1025 preference = f.get('preference')
1026 if preference is None:
1028 if f.get('ext') in ['f4f', 'f4m']: # Not yet supported
1031 protocol = f.get('protocol') or determine_protocol(f)
1032 proto_preference = 0 if protocol in ['http', 'https'] else (-0.5 if protocol == 'rtsp' else -0.1)
1034 if f.get('vcodec') == 'none': # audio only
1036 if self._downloader.params.get('prefer_free_formats'):
1037 ORDER = ['aac', 'mp3', 'm4a', 'webm', 'ogg', 'opus']
1039 ORDER = ['webm', 'opus', 'ogg', 'mp3', 'aac', 'm4a']
1042 audio_ext_preference = ORDER.index(f['ext'])
1044 audio_ext_preference = -1
1046 if f.get('acodec') == 'none': # video only
1048 if self._downloader.params.get('prefer_free_formats'):
1049 ORDER = ['flv', 'mp4', 'webm']
1051 ORDER = ['webm', 'flv', 'mp4']
1053 ext_preference = ORDER.index(f['ext'])
1056 audio_ext_preference = 0
1060 f.get('language_preference') if f.get('language_preference') is not None else -1,
1061 f.get('quality') if f.get('quality') is not None else -1,
1062 f.get('tbr') if f.get('tbr') is not None else -1,
1063 f.get('filesize') if f.get('filesize') is not None else -1,
1064 f.get('vbr') if f.get('vbr') is not None else -1,
1065 f.get('height') if f.get('height') is not None else -1,
1066 f.get('width') if f.get('width') is not None else -1,
1069 f.get('abr') if f.get('abr') is not None else -1,
1070 audio_ext_preference,
1071 f.get('fps') if f.get('fps') is not None else -1,
1072 f.get('filesize_approx') if f.get('filesize_approx') is not None else -1,
1073 f.get('source_preference') if f.get('source_preference') is not None else -1,
1074 f.get('format_id') if f.get('format_id') is not None else '',
1076 formats.sort(key=_formats_key)
1078 def _check_formats(self, formats, video_id):
1080 formats[:] = filter(
1081 lambda f: self._is_valid_url(
1083 item='%s video format' % f.get('format_id') if f.get('format_id') else 'video'),
1087 def _remove_duplicate_formats(formats):
1091 if f['url'] not in format_urls:
1092 format_urls.add(f['url'])
1093 unique_formats.append(f)
1094 formats[:] = unique_formats
1096 def _is_valid_url(self, url, video_id, item='video', headers={}):
1097 url = self._proto_relative_url(url, scheme='http:')
1098 # For now assume non HTTP(S) URLs always valid
1099 if not (url.startswith('http://') or url.startswith('https://')):
1102 self._request_webpage(url, video_id, 'Checking %s URL' % item, headers=headers)
1104 except ExtractorError as e:
1105 if isinstance(e.cause, compat_urllib_error.URLError):
1107 '%s: %s URL is invalid, skipping' % (video_id, item))
1111 def http_scheme(self):
1112 """ Either "http:" or "https:", depending on the user's preferences """
1115 if self._downloader.params.get('prefer_insecure', False)
1118 def _proto_relative_url(self, url, scheme=None):
1121 if url.startswith('//'):
1123 scheme = self.http_scheme()
1128 def _sleep(self, timeout, video_id, msg_template=None):
1129 if msg_template is None:
1130 msg_template = '%(video_id)s: Waiting for %(timeout)s seconds'
1131 msg = msg_template % {'video_id': video_id, 'timeout': timeout}
1135 def _extract_f4m_formats(self, manifest_url, video_id, preference=None, f4m_id=None,
1136 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1137 fatal=True, m3u8_id=None):
1138 manifest = self._download_xml(
1139 manifest_url, video_id, 'Downloading f4m manifest',
1140 'Unable to download f4m manifest',
1141 # Some manifests may be malformed, e.g. prosiebensat1 generated manifests
1142 # (see https://github.com/rg3/youtube-dl/issues/6215#issuecomment-121704244)
1143 transform_source=transform_source,
1146 if manifest is False:
1149 return self._parse_f4m_formats(
1150 manifest, manifest_url, video_id, preference=preference, f4m_id=f4m_id,
1151 transform_source=transform_source, fatal=fatal, m3u8_id=m3u8_id)
1153 def _parse_f4m_formats(self, manifest, manifest_url, video_id, preference=None, f4m_id=None,
1154 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1155 fatal=True, m3u8_id=None):
1156 # currently youtube-dl cannot decode the playerVerificationChallenge as Akamai uses Adobe Alchemy
1157 akamai_pv = manifest.find('{http://ns.adobe.com/f4m/1.0}pv-2.0')
1158 if akamai_pv is not None and ';' in akamai_pv.text:
1159 playerVerificationChallenge = akamai_pv.text.split(';')[0]
1160 if playerVerificationChallenge.strip() != '':
1164 manifest_version = '1.0'
1165 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/1.0}media')
1167 manifest_version = '2.0'
1168 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/2.0}media')
1169 # Remove unsupported DRM protected media from final formats
1170 # rendition (see https://github.com/rg3/youtube-dl/issues/8573).
1171 media_nodes = remove_encrypted_media(media_nodes)
1174 base_url = xpath_text(
1175 manifest, ['{http://ns.adobe.com/f4m/1.0}baseURL', '{http://ns.adobe.com/f4m/2.0}baseURL'],
1176 'base URL', default=None)
1178 base_url = base_url.strip()
1180 bootstrap_info = xpath_element(
1181 manifest, ['{http://ns.adobe.com/f4m/1.0}bootstrapInfo', '{http://ns.adobe.com/f4m/2.0}bootstrapInfo'],
1182 'bootstrap info', default=None)
1185 mime_type = xpath_text(
1186 manifest, ['{http://ns.adobe.com/f4m/1.0}mimeType', '{http://ns.adobe.com/f4m/2.0}mimeType'],
1187 'base URL', default=None)
1188 if mime_type and mime_type.startswith('audio/'):
1191 for i, media_el in enumerate(media_nodes):
1192 tbr = int_or_none(media_el.attrib.get('bitrate'))
1193 width = int_or_none(media_el.attrib.get('width'))
1194 height = int_or_none(media_el.attrib.get('height'))
1195 format_id = '-'.join(filter(None, [f4m_id, compat_str(i if tbr is None else tbr)]))
1196 # If <bootstrapInfo> is present, the specified f4m is a
1197 # stream-level manifest, and only set-level manifests may refer to
1198 # external resources. See section 11.4 and section 4 of F4M spec
1199 if bootstrap_info is None:
1201 # @href is introduced in 2.0, see section 11.6 of F4M spec
1202 if manifest_version == '2.0':
1203 media_url = media_el.attrib.get('href')
1204 if media_url is None:
1205 media_url = media_el.attrib.get('url')
1209 media_url if media_url.startswith('http://') or media_url.startswith('https://')
1210 else ((base_url or '/'.join(manifest_url.split('/')[:-1])) + '/' + media_url))
1211 # If media_url is itself a f4m manifest do the recursive extraction
1212 # since bitrates in parent manifest (this one) and media_url manifest
1213 # may differ leading to inability to resolve the format by requested
1214 # bitrate in f4m downloader
1215 ext = determine_ext(manifest_url)
1217 f4m_formats = self._extract_f4m_formats(
1218 manifest_url, video_id, preference=preference, f4m_id=f4m_id,
1219 transform_source=transform_source, fatal=fatal)
1220 # Sometimes stream-level manifest contains single media entry that
1221 # does not contain any quality metadata (e.g. http://matchtv.ru/#live-player).
1222 # At the same time parent's media entry in set-level manifest may
1223 # contain it. We will copy it from parent in such cases.
1224 if len(f4m_formats) == 1:
1227 'tbr': f.get('tbr') or tbr,
1228 'width': f.get('width') or width,
1229 'height': f.get('height') or height,
1230 'format_id': f.get('format_id') if not tbr else format_id,
1233 formats.extend(f4m_formats)
1236 formats.extend(self._extract_m3u8_formats(
1237 manifest_url, video_id, 'mp4', preference=preference,
1238 m3u8_id=m3u8_id, fatal=fatal))
1241 'format_id': format_id,
1242 'url': manifest_url,
1243 'manifest_url': manifest_url,
1244 'ext': 'flv' if bootstrap_info is not None else None,
1249 'preference': preference,
1253 def _m3u8_meta_format(self, m3u8_url, ext=None, preference=None, m3u8_id=None):
1255 'format_id': '-'.join(filter(None, [m3u8_id, 'meta'])),
1259 'preference': preference - 100 if preference else -100,
1260 'resolution': 'multiple',
1261 'format_note': 'Quality selection URL',
1264 def _extract_m3u8_formats(self, m3u8_url, video_id, ext=None,
1265 entry_protocol='m3u8', preference=None,
1266 m3u8_id=None, note=None, errnote=None,
1267 fatal=True, live=False):
1269 res = self._download_webpage_handle(
1271 note=note or 'Downloading m3u8 information',
1272 errnote=errnote or 'Failed to download m3u8 information',
1276 m3u8_doc, urlh = res
1277 m3u8_url = urlh.geturl()
1279 if '#EXT-X-FAXS-CM:' in m3u8_doc: # Adobe Flash Access
1282 formats = [self._m3u8_meta_format(m3u8_url, ext, preference, m3u8_id)]
1284 format_url = lambda u: (
1286 if re.match(r'^https?://', u)
1287 else compat_urlparse.urljoin(m3u8_url, u))
1289 # We should try extracting formats only from master playlists [1], i.e.
1290 # playlists that describe available qualities. On the other hand media
1291 # playlists [2] should be returned as is since they contain just the media
1292 # without qualities renditions.
1293 # Fortunately, master playlist can be easily distinguished from media
1294 # playlist based on particular tags availability. As of [1, 2] master
1295 # playlist tags MUST NOT appear in a media playist and vice versa.
1296 # As of [3] #EXT-X-TARGETDURATION tag is REQUIRED for every media playlist
1297 # and MUST NOT appear in master playlist thus we can clearly detect media
1298 # playlist with this criterion.
1299 # 1. https://tools.ietf.org/html/draft-pantos-http-live-streaming-17#section-4.3.4
1300 # 2. https://tools.ietf.org/html/draft-pantos-http-live-streaming-17#section-4.3.3
1301 # 3. https://tools.ietf.org/html/draft-pantos-http-live-streaming-17#section-4.3.3.1
1302 if '#EXT-X-TARGETDURATION' in m3u8_doc: # media playlist, return as is
1305 'format_id': m3u8_id,
1307 'protocol': entry_protocol,
1308 'preference': preference,
1310 audio_in_video_stream = {}
1313 for line in m3u8_doc.splitlines():
1314 if line.startswith('#EXT-X-STREAM-INF:'):
1315 last_info = parse_m3u8_attributes(line)
1316 elif line.startswith('#EXT-X-MEDIA:'):
1317 media = parse_m3u8_attributes(line)
1318 media_type = media.get('TYPE')
1319 if media_type in ('VIDEO', 'AUDIO'):
1320 group_id = media.get('GROUP-ID')
1321 media_url = media.get('URI')
1324 for v in (group_id, media.get('NAME')):
1328 'format_id': '-'.join(format_id),
1329 'url': format_url(media_url),
1330 'language': media.get('LANGUAGE'),
1332 'protocol': entry_protocol,
1333 'preference': preference,
1335 if media_type == 'AUDIO':
1336 f['vcodec'] = 'none'
1337 if group_id and not audio_in_video_stream.get(group_id):
1338 audio_in_video_stream[group_id] = False
1341 # When there is no URI in EXT-X-MEDIA let this tag's
1342 # data be used by regular URI lines below
1344 if media_type == 'AUDIO' and group_id:
1345 audio_in_video_stream[group_id] = True
1346 elif line.startswith('#') or not line.strip():
1349 tbr = int_or_none(last_info.get('AVERAGE-BANDWIDTH') or last_info.get('BANDWIDTH'), scale=1000)
1352 format_id.append(m3u8_id)
1353 # Despite specification does not mention NAME attribute for
1354 # EXT-X-STREAM-INF it still sometimes may be present
1355 stream_name = last_info.get('NAME') or last_media.get('NAME')
1356 # Bandwidth of live streams may differ over time thus making
1357 # format_id unpredictable. So it's better to keep provided
1360 format_id.append(stream_name if stream_name else '%d' % (tbr if tbr else len(formats)))
1361 manifest_url = format_url(line.strip())
1363 'format_id': '-'.join(format_id),
1364 'url': manifest_url,
1365 'manifest_url': manifest_url,
1368 'fps': float_or_none(last_info.get('FRAME-RATE')),
1369 'protocol': entry_protocol,
1370 'preference': preference,
1372 resolution = last_info.get('RESOLUTION')
1374 mobj = re.search(r'(?P<width>\d+)[xX](?P<height>\d+)', resolution)
1376 f['width'] = int(mobj.group('width'))
1377 f['height'] = int(mobj.group('height'))
1378 # Unified Streaming Platform
1380 r'audio.*?(?:%3D|=)(\d+)(?:-video.*?(?:%3D|=)(\d+))?', f['url'])
1382 abr, vbr = mobj.groups()
1383 abr, vbr = float_or_none(abr, 1000), float_or_none(vbr, 1000)
1388 f.update(parse_codecs(last_info.get('CODECS')))
1389 if audio_in_video_stream.get(last_info.get('AUDIO')) is False and f['vcodec'] != 'none':
1390 # TODO: update acodec for audio only formats with the same GROUP-ID
1391 f['acodec'] = 'none'
1398 def _xpath_ns(path, namespace=None):
1402 for c in path.split('/'):
1403 if not c or c == '.':
1406 out.append('{%s}%s' % (namespace, c))
1407 return '/'.join(out)
1409 def _extract_smil_formats(self, smil_url, video_id, fatal=True, f4m_params=None, transform_source=None):
1410 smil = self._download_smil(smil_url, video_id, fatal=fatal, transform_source=transform_source)
1416 namespace = self._parse_smil_namespace(smil)
1418 return self._parse_smil_formats(
1419 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
1421 def _extract_smil_info(self, smil_url, video_id, fatal=True, f4m_params=None):
1422 smil = self._download_smil(smil_url, video_id, fatal=fatal)
1425 return self._parse_smil(smil, smil_url, video_id, f4m_params=f4m_params)
1427 def _download_smil(self, smil_url, video_id, fatal=True, transform_source=None):
1428 return self._download_xml(
1429 smil_url, video_id, 'Downloading SMIL file',
1430 'Unable to download SMIL file', fatal=fatal, transform_source=transform_source)
1432 def _parse_smil(self, smil, smil_url, video_id, f4m_params=None):
1433 namespace = self._parse_smil_namespace(smil)
1435 formats = self._parse_smil_formats(
1436 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
1437 subtitles = self._parse_smil_subtitles(smil, namespace=namespace)
1439 video_id = os.path.splitext(url_basename(smil_url))[0]
1443 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
1444 name = meta.attrib.get('name')
1445 content = meta.attrib.get('content')
1446 if not name or not content:
1448 if not title and name == 'title':
1450 elif not description and name in ('description', 'abstract'):
1451 description = content
1452 elif not upload_date and name == 'date':
1453 upload_date = unified_strdate(content)
1456 'id': image.get('type'),
1457 'url': image.get('src'),
1458 'width': int_or_none(image.get('width')),
1459 'height': int_or_none(image.get('height')),
1460 } for image in smil.findall(self._xpath_ns('.//image', namespace)) if image.get('src')]
1464 'title': title or video_id,
1465 'description': description,
1466 'upload_date': upload_date,
1467 'thumbnails': thumbnails,
1469 'subtitles': subtitles,
1472 def _parse_smil_namespace(self, smil):
1473 return self._search_regex(
1474 r'(?i)^{([^}]+)?}smil$', smil.tag, 'namespace', default=None)
1476 def _parse_smil_formats(self, smil, smil_url, video_id, namespace=None, f4m_params=None, transform_rtmp_url=None):
1478 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
1479 b = meta.get('base') or meta.get('httpBase')
1490 media = smil.findall(self._xpath_ns('.//video', namespace)) + smil.findall(self._xpath_ns('.//audio', namespace))
1491 for medium in media:
1492 src = medium.get('src')
1493 if not src or src in srcs:
1497 bitrate = float_or_none(medium.get('system-bitrate') or medium.get('systemBitrate'), 1000)
1498 filesize = int_or_none(medium.get('size') or medium.get('fileSize'))
1499 width = int_or_none(medium.get('width'))
1500 height = int_or_none(medium.get('height'))
1501 proto = medium.get('proto')
1502 ext = medium.get('ext')
1503 src_ext = determine_ext(src)
1504 streamer = medium.get('streamer') or base
1506 if proto == 'rtmp' or streamer.startswith('rtmp'):
1512 'format_id': 'rtmp-%d' % (rtmp_count if bitrate is None else bitrate),
1514 'filesize': filesize,
1518 if transform_rtmp_url:
1519 streamer, src = transform_rtmp_url(streamer, src)
1520 formats[-1].update({
1526 src_url = src if src.startswith('http') else compat_urlparse.urljoin(base, src)
1527 src_url = src_url.strip()
1529 if proto == 'm3u8' or src_ext == 'm3u8':
1530 m3u8_formats = self._extract_m3u8_formats(
1531 src_url, video_id, ext or 'mp4', m3u8_id='hls', fatal=False)
1532 if len(m3u8_formats) == 1:
1534 m3u8_formats[0].update({
1535 'format_id': 'hls-%d' % (m3u8_count if bitrate is None else bitrate),
1540 formats.extend(m3u8_formats)
1543 if src_ext == 'f4m':
1548 'plugin': 'flowplayer-3.2.0.1',
1550 f4m_url += '&' if '?' in f4m_url else '?'
1551 f4m_url += compat_urllib_parse_urlencode(f4m_params)
1552 formats.extend(self._extract_f4m_formats(f4m_url, video_id, f4m_id='hds', fatal=False))
1555 if src_url.startswith('http') and self._is_valid_url(src, video_id):
1559 'ext': ext or src_ext or 'flv',
1560 'format_id': 'http-%d' % (bitrate or http_count),
1562 'filesize': filesize,
1570 def _parse_smil_subtitles(self, smil, namespace=None, subtitles_lang='en'):
1573 for num, textstream in enumerate(smil.findall(self._xpath_ns('.//textstream', namespace))):
1574 src = textstream.get('src')
1575 if not src or src in urls:
1578 ext = textstream.get('ext') or mimetype2ext(textstream.get('type')) or determine_ext(src)
1579 lang = textstream.get('systemLanguage') or textstream.get('systemLanguageName') or textstream.get('lang') or subtitles_lang
1580 subtitles.setdefault(lang, []).append({
1586 def _extract_xspf_playlist(self, playlist_url, playlist_id, fatal=True):
1587 xspf = self._download_xml(
1588 playlist_url, playlist_id, 'Downloading xpsf playlist',
1589 'Unable to download xspf manifest', fatal=fatal)
1592 return self._parse_xspf(xspf, playlist_id)
1594 def _parse_xspf(self, playlist, playlist_id):
1596 'xspf': 'http://xspf.org/ns/0/',
1597 's1': 'http://static.streamone.nl/player/ns/0',
1601 for track in playlist.findall(xpath_with_ns('./xspf:trackList/xspf:track', NS_MAP)):
1603 track, xpath_with_ns('./xspf:title', NS_MAP), 'title', default=playlist_id)
1604 description = xpath_text(
1605 track, xpath_with_ns('./xspf:annotation', NS_MAP), 'description')
1606 thumbnail = xpath_text(
1607 track, xpath_with_ns('./xspf:image', NS_MAP), 'thumbnail')
1608 duration = float_or_none(
1609 xpath_text(track, xpath_with_ns('./xspf:duration', NS_MAP), 'duration'), 1000)
1612 'url': location.text,
1613 'format_id': location.get(xpath_with_ns('s1:label', NS_MAP)),
1614 'width': int_or_none(location.get(xpath_with_ns('s1:width', NS_MAP))),
1615 'height': int_or_none(location.get(xpath_with_ns('s1:height', NS_MAP))),
1616 } for location in track.findall(xpath_with_ns('./xspf:location', NS_MAP))]
1617 self._sort_formats(formats)
1622 'description': description,
1623 'thumbnail': thumbnail,
1624 'duration': duration,
1629 def _extract_mpd_formats(self, mpd_url, video_id, mpd_id=None, note=None, errnote=None, fatal=True, formats_dict={}):
1630 res = self._download_webpage_handle(
1632 note=note or 'Downloading MPD manifest',
1633 errnote=errnote or 'Failed to download MPD manifest',
1638 mpd_base_url = base_url(urlh.geturl())
1640 return self._parse_mpd_formats(
1641 compat_etree_fromstring(mpd.encode('utf-8')), mpd_id, mpd_base_url,
1642 formats_dict=formats_dict, mpd_url=mpd_url)
1644 def _parse_mpd_formats(self, mpd_doc, mpd_id=None, mpd_base_url='', formats_dict={}, mpd_url=None):
1646 Parse formats from MPD manifest.
1648 1. MPEG-DASH Standard, ISO/IEC 23009-1:2014(E),
1649 http://standards.iso.org/ittf/PubliclyAvailableStandards/c065274_ISO_IEC_23009-1_2014.zip
1650 2. https://en.wikipedia.org/wiki/Dynamic_Adaptive_Streaming_over_HTTP
1652 if mpd_doc.get('type') == 'dynamic':
1655 namespace = self._search_regex(r'(?i)^{([^}]+)?}MPD$', mpd_doc.tag, 'namespace', default=None)
1658 return self._xpath_ns(path, namespace)
1660 def is_drm_protected(element):
1661 return element.find(_add_ns('ContentProtection')) is not None
1663 def extract_multisegment_info(element, ms_parent_info):
1664 ms_info = ms_parent_info.copy()
1666 # As per [1, 5.3.9.2.2] SegmentList and SegmentTemplate share some
1667 # common attributes and elements. We will only extract relevant
1669 def extract_common(source):
1670 segment_timeline = source.find(_add_ns('SegmentTimeline'))
1671 if segment_timeline is not None:
1672 s_e = segment_timeline.findall(_add_ns('S'))
1674 ms_info['total_number'] = 0
1677 r = int(s.get('r', 0))
1678 ms_info['total_number'] += 1 + r
1679 ms_info['s'].append({
1680 't': int(s.get('t', 0)),
1681 # @d is mandatory (see [1, 5.3.9.6.2, Table 17, page 60])
1682 'd': int(s.attrib['d']),
1685 start_number = source.get('startNumber')
1687 ms_info['start_number'] = int(start_number)
1688 timescale = source.get('timescale')
1690 ms_info['timescale'] = int(timescale)
1691 segment_duration = source.get('duration')
1692 if segment_duration:
1693 ms_info['segment_duration'] = int(segment_duration)
1695 def extract_Initialization(source):
1696 initialization = source.find(_add_ns('Initialization'))
1697 if initialization is not None:
1698 ms_info['initialization_url'] = initialization.attrib['sourceURL']
1700 segment_list = element.find(_add_ns('SegmentList'))
1701 if segment_list is not None:
1702 extract_common(segment_list)
1703 extract_Initialization(segment_list)
1704 segment_urls_e = segment_list.findall(_add_ns('SegmentURL'))
1706 ms_info['segment_urls'] = [segment.attrib['media'] for segment in segment_urls_e]
1708 segment_template = element.find(_add_ns('SegmentTemplate'))
1709 if segment_template is not None:
1710 extract_common(segment_template)
1711 media = segment_template.get('media')
1713 ms_info['media'] = media
1714 initialization = segment_template.get('initialization')
1716 ms_info['initialization'] = initialization
1718 extract_Initialization(segment_template)
1721 mpd_duration = parse_duration(mpd_doc.get('mediaPresentationDuration'))
1723 for period in mpd_doc.findall(_add_ns('Period')):
1724 period_duration = parse_duration(period.get('duration')) or mpd_duration
1725 period_ms_info = extract_multisegment_info(period, {
1729 for adaptation_set in period.findall(_add_ns('AdaptationSet')):
1730 if is_drm_protected(adaptation_set):
1732 adaption_set_ms_info = extract_multisegment_info(adaptation_set, period_ms_info)
1733 for representation in adaptation_set.findall(_add_ns('Representation')):
1734 if is_drm_protected(representation):
1736 representation_attrib = adaptation_set.attrib.copy()
1737 representation_attrib.update(representation.attrib)
1738 # According to [1, 5.3.7.2, Table 9, page 41], @mimeType is mandatory
1739 mime_type = representation_attrib['mimeType']
1740 content_type = mime_type.split('/')[0]
1741 if content_type == 'text':
1742 # TODO implement WebVTT downloading
1744 elif content_type == 'video' or content_type == 'audio':
1746 for element in (representation, adaptation_set, period, mpd_doc):
1747 base_url_e = element.find(_add_ns('BaseURL'))
1748 if base_url_e is not None:
1749 base_url = base_url_e.text + base_url
1750 if re.match(r'^https?://', base_url):
1752 if mpd_base_url and not re.match(r'^https?://', base_url):
1753 if not mpd_base_url.endswith('/') and not base_url.startswith('/'):
1755 base_url = mpd_base_url + base_url
1756 representation_id = representation_attrib.get('id')
1757 lang = representation_attrib.get('lang')
1758 url_el = representation.find(_add_ns('BaseURL'))
1759 filesize = int_or_none(url_el.attrib.get('{http://youtube.com/yt/2012/10/10}contentLength') if url_el is not None else None)
1760 bandwidth = int_or_none(representation_attrib.get('bandwidth'))
1762 'format_id': '%s-%s' % (mpd_id, representation_id) if mpd_id else representation_id,
1764 'manifest_url': mpd_url,
1765 'ext': mimetype2ext(mime_type),
1766 'width': int_or_none(representation_attrib.get('width')),
1767 'height': int_or_none(representation_attrib.get('height')),
1768 'tbr': int_or_none(bandwidth, 1000),
1769 'asr': int_or_none(representation_attrib.get('audioSamplingRate')),
1770 'fps': int_or_none(representation_attrib.get('frameRate')),
1771 'language': lang if lang not in ('mul', 'und', 'zxx', 'mis') else None,
1772 'format_note': 'DASH %s' % content_type,
1773 'filesize': filesize,
1775 f.update(parse_codecs(representation_attrib.get('codecs')))
1776 representation_ms_info = extract_multisegment_info(representation, adaption_set_ms_info)
1778 def prepare_template(template_name, identifiers):
1779 t = representation_ms_info[template_name]
1780 t = t.replace('$RepresentationID$', representation_id)
1781 t = re.sub(r'\$(%s)\$' % '|'.join(identifiers), r'%(\1)d', t)
1782 t = re.sub(r'\$(%s)%%([^$]+)\$' % '|'.join(identifiers), r'%(\1)\2', t)
1783 t.replace('$$', '$')
1786 # @initialization is a regular template like @media one
1787 # so it should be handled just the same way (see
1788 # https://github.com/rg3/youtube-dl/issues/11605)
1789 if 'initialization' in representation_ms_info:
1790 initialization_template = prepare_template(
1792 # As per [1, 5.3.9.4.2, Table 15, page 54] $Number$ and
1793 # $Time$ shall not be included for @initialization thus
1794 # only $Bandwidth$ remains
1796 representation_ms_info['initialization_url'] = initialization_template % {
1797 'Bandwidth': bandwidth,
1800 if 'segment_urls' not in representation_ms_info and 'media' in representation_ms_info:
1802 media_template = prepare_template('media', ('Number', 'Bandwidth', 'Time'))
1804 # As per [1, 5.3.9.4.4, Table 16, page 55] $Number$ and $Time$
1805 # can't be used at the same time
1806 if '%(Number' in media_template and 's' not in representation_ms_info:
1807 segment_duration = None
1808 if 'total_number' not in representation_ms_info and 'segment_duration':
1809 segment_duration = float_or_none(representation_ms_info['segment_duration'], representation_ms_info['timescale'])
1810 representation_ms_info['total_number'] = int(math.ceil(float(period_duration) / segment_duration))
1811 representation_ms_info['fragments'] = [{
1812 'url': media_template % {
1813 'Number': segment_number,
1814 'Bandwidth': bandwidth,
1816 'duration': segment_duration,
1817 } for segment_number in range(
1818 representation_ms_info['start_number'],
1819 representation_ms_info['total_number'] + representation_ms_info['start_number'])]
1821 # $Number*$ or $Time$ in media template with S list available
1822 # Example $Number*$: http://www.svtplay.se/klipp/9023742/stopptid-om-bjorn-borg
1823 # Example $Time$: https://play.arkena.com/embed/avp/v2/player/media/b41dda37-d8e7-4d3f-b1b5-9a9db578bdfe/1/129411
1824 representation_ms_info['fragments'] = []
1827 segment_number = representation_ms_info['start_number']
1829 def add_segment_url():
1830 segment_url = media_template % {
1831 'Time': segment_time,
1832 'Bandwidth': bandwidth,
1833 'Number': segment_number,
1835 representation_ms_info['fragments'].append({
1837 'duration': float_or_none(segment_d, representation_ms_info['timescale']),
1840 for num, s in enumerate(representation_ms_info['s']):
1841 segment_time = s.get('t') or segment_time
1845 for r in range(s.get('r', 0)):
1846 segment_time += segment_d
1849 segment_time += segment_d
1850 elif 'segment_urls' in representation_ms_info and 's' in representation_ms_info:
1852 # Example: https://www.youtube.com/watch?v=iXZV5uAYMJI
1853 # or any YouTube dashsegments video
1856 timescale = representation_ms_info['timescale']
1857 for s in representation_ms_info['s']:
1858 duration = float_or_none(s['d'], timescale)
1859 for r in range(s.get('r', 0) + 1):
1861 'url': representation_ms_info['segment_urls'][segment_index],
1862 'duration': duration,
1865 representation_ms_info['fragments'] = fragments
1866 # NB: MPD manifest may contain direct URLs to unfragmented media.
1867 # No fragments key is present in this case.
1868 if 'fragments' in representation_ms_info:
1871 'protocol': 'http_dash_segments',
1873 if 'initialization_url' in representation_ms_info:
1874 initialization_url = representation_ms_info['initialization_url']
1875 if not f.get('url'):
1876 f['url'] = initialization_url
1877 f['fragments'].append({'url': initialization_url})
1878 f['fragments'].extend(representation_ms_info['fragments'])
1879 for fragment in f['fragments']:
1880 fragment['url'] = urljoin(base_url, fragment['url'])
1882 existing_format = next(
1883 fo for fo in formats
1884 if fo['format_id'] == representation_id)
1885 except StopIteration:
1886 full_info = formats_dict.get(representation_id, {}).copy()
1888 formats.append(full_info)
1890 existing_format.update(f)
1892 self.report_warning('Unknown MIME type %s in DASH manifest' % mime_type)
1895 def _extract_ism_formats(self, ism_url, video_id, ism_id=None, note=None, errnote=None, fatal=True):
1896 res = self._download_webpage_handle(
1898 note=note or 'Downloading ISM manifest',
1899 errnote=errnote or 'Failed to download ISM manifest',
1905 return self._parse_ism_formats(
1906 compat_etree_fromstring(ism.encode('utf-8')), urlh.geturl(), ism_id)
1908 def _parse_ism_formats(self, ism_doc, ism_url, ism_id=None):
1909 if ism_doc.get('IsLive') == 'TRUE' or ism_doc.find('Protection') is not None:
1912 duration = int(ism_doc.attrib['Duration'])
1913 timescale = int_or_none(ism_doc.get('TimeScale')) or 10000000
1916 for stream in ism_doc.findall('StreamIndex'):
1917 stream_type = stream.get('Type')
1918 if stream_type not in ('video', 'audio'):
1920 url_pattern = stream.attrib['Url']
1921 stream_timescale = int_or_none(stream.get('TimeScale')) or timescale
1922 stream_name = stream.get('Name')
1923 for track in stream.findall('QualityLevel'):
1924 fourcc = track.get('FourCC')
1925 # TODO: add support for WVC1 and WMAP
1926 if fourcc not in ('H264', 'AVC1', 'AACL'):
1927 self.report_warning('%s is not a supported codec' % fourcc)
1929 tbr = int(track.attrib['Bitrate']) // 1000
1930 width = int_or_none(track.get('MaxWidth'))
1931 height = int_or_none(track.get('MaxHeight'))
1932 sampling_rate = int_or_none(track.get('SamplingRate'))
1934 track_url_pattern = re.sub(r'{[Bb]itrate}', track.attrib['Bitrate'], url_pattern)
1935 track_url_pattern = compat_urlparse.urljoin(ism_url, track_url_pattern)
1941 stream_fragments = stream.findall('c')
1942 for stream_fragment_index, stream_fragment in enumerate(stream_fragments):
1943 fragment_ctx['time'] = int_or_none(stream_fragment.get('t')) or fragment_ctx['time']
1944 fragment_repeat = int_or_none(stream_fragment.get('r')) or 1
1945 fragment_ctx['duration'] = int_or_none(stream_fragment.get('d'))
1946 if not fragment_ctx['duration']:
1948 next_fragment_time = int(stream_fragment[stream_fragment_index + 1].attrib['t'])
1950 next_fragment_time = duration
1951 fragment_ctx['duration'] = (next_fragment_time - fragment_ctx['time']) / fragment_repeat
1952 for _ in range(fragment_repeat):
1954 'url': re.sub(r'{start[ _]time}', compat_str(fragment_ctx['time']), track_url_pattern),
1955 'duration': fragment_ctx['duration'] / stream_timescale,
1957 fragment_ctx['time'] += fragment_ctx['duration']
1961 format_id.append(ism_id)
1963 format_id.append(stream_name)
1964 format_id.append(compat_str(tbr))
1967 'format_id': '-'.join(format_id),
1969 'manifest_url': ism_url,
1970 'ext': 'ismv' if stream_type == 'video' else 'isma',
1974 'asr': sampling_rate,
1975 'vcodec': 'none' if stream_type == 'audio' else fourcc,
1976 'acodec': 'none' if stream_type == 'video' else fourcc,
1978 'fragments': fragments,
1979 '_download_params': {
1980 'duration': duration,
1981 'timescale': stream_timescale,
1982 'width': width or 0,
1983 'height': height or 0,
1985 'codec_private_data': track.get('CodecPrivateData'),
1986 'sampling_rate': sampling_rate,
1987 'channels': int_or_none(track.get('Channels', 2)),
1988 'bits_per_sample': int_or_none(track.get('BitsPerSample', 16)),
1989 'nal_unit_length_field': int_or_none(track.get('NALUnitLengthField', 4)),
1994 def _parse_html5_media_entries(self, base_url, webpage, video_id, m3u8_id=None, m3u8_entry_protocol='m3u8', mpd_id=None):
1995 def absolute_url(video_url):
1996 return compat_urlparse.urljoin(base_url, video_url)
1998 def parse_content_type(content_type):
1999 if not content_type:
2001 ctr = re.search(r'(?P<mimetype>[^/]+/[^;]+)(?:;\s*codecs="?(?P<codecs>[^"]+))?', content_type)
2003 mimetype, codecs = ctr.groups()
2004 f = parse_codecs(codecs)
2005 f['ext'] = mimetype2ext(mimetype)
2009 def _media_formats(src, cur_media_type):
2010 full_url = absolute_url(src)
2011 ext = determine_ext(full_url)
2013 is_plain_url = False
2014 formats = self._extract_m3u8_formats(
2015 full_url, video_id, ext='mp4',
2016 entry_protocol=m3u8_entry_protocol, m3u8_id=m3u8_id)
2018 is_plain_url = False
2019 formats = self._extract_mpd_formats(
2020 full_url, video_id, mpd_id=mpd_id)
2025 'vcodec': 'none' if cur_media_type == 'audio' else None,
2027 return is_plain_url, formats
2030 media_tags = [(media_tag, media_type, '')
2031 for media_tag, media_type
2032 in re.findall(r'(?s)(<(video|audio)[^>]*/>)', webpage)]
2033 media_tags.extend(re.findall(
2034 # We only allow video|audio followed by a whitespace or '>'.
2035 # Allowing more characters may end up in significant slow down (see
2036 # https://github.com/rg3/youtube-dl/issues/11979, example URL:
2037 # http://www.porntrex.com/maps/videositemap.xml).
2038 r'(?s)(<(?P<tag>video|audio)(?:\s+[^>]*)?>)(.*?)</(?P=tag)>', webpage))
2039 for media_tag, media_type, media_content in media_tags:
2044 media_attributes = extract_attributes(media_tag)
2045 src = media_attributes.get('src')
2047 _, formats = _media_formats(src, media_type)
2048 media_info['formats'].extend(formats)
2049 media_info['thumbnail'] = media_attributes.get('poster')
2051 for source_tag in re.findall(r'<source[^>]+>', media_content):
2052 source_attributes = extract_attributes(source_tag)
2053 src = source_attributes.get('src')
2056 is_plain_url, formats = _media_formats(src, media_type)
2058 f = parse_content_type(source_attributes.get('type'))
2059 f.update(formats[0])
2060 media_info['formats'].append(f)
2062 media_info['formats'].extend(formats)
2063 for track_tag in re.findall(r'<track[^>]+>', media_content):
2064 track_attributes = extract_attributes(track_tag)
2065 kind = track_attributes.get('kind')
2066 if not kind or kind in ('subtitles', 'captions'):
2067 src = track_attributes.get('src')
2070 lang = track_attributes.get('srclang') or track_attributes.get('lang') or track_attributes.get('label')
2071 media_info['subtitles'].setdefault(lang, []).append({
2072 'url': absolute_url(src),
2074 if media_info['formats'] or media_info['subtitles']:
2075 entries.append(media_info)
2078 def _extract_akamai_formats(self, manifest_url, video_id, hosts={}):
2080 hdcore_sign = 'hdcore=3.7.0'
2081 f4m_url = re.sub(r'(https?://[^/+])/i/', r'\1/z/', manifest_url).replace('/master.m3u8', '/manifest.f4m')
2082 hds_host = hosts.get('hds')
2084 f4m_url = re.sub(r'(https?://)[^/]+', r'\1' + hds_host, f4m_url)
2085 if 'hdcore=' not in f4m_url:
2086 f4m_url += ('&' if '?' in f4m_url else '?') + hdcore_sign
2087 f4m_formats = self._extract_f4m_formats(
2088 f4m_url, video_id, f4m_id='hds', fatal=False)
2089 for entry in f4m_formats:
2090 entry.update({'extra_param_to_segment_url': hdcore_sign})
2091 formats.extend(f4m_formats)
2092 m3u8_url = re.sub(r'(https?://[^/]+)/z/', r'\1/i/', manifest_url).replace('/manifest.f4m', '/master.m3u8')
2093 hls_host = hosts.get('hls')
2095 m3u8_url = re.sub(r'(https?://)[^/]+', r'\1' + hls_host, m3u8_url)
2096 formats.extend(self._extract_m3u8_formats(
2097 m3u8_url, video_id, 'mp4', 'm3u8_native',
2098 m3u8_id='hls', fatal=False))
2101 def _extract_wowza_formats(self, url, video_id, m3u8_entry_protocol='m3u8_native', skip_protocols=[]):
2102 url = re.sub(r'/(?:manifest|playlist|jwplayer)\.(?:m3u8|f4m|mpd|smil)', '', url)
2103 url_base = self._search_regex(r'(?:https?|rtmp|rtsp)(://[^?]+)', url, 'format url')
2104 http_base_url = 'http' + url_base
2106 if 'm3u8' not in skip_protocols:
2107 formats.extend(self._extract_m3u8_formats(
2108 http_base_url + '/playlist.m3u8', video_id, 'mp4',
2109 m3u8_entry_protocol, m3u8_id='hls', fatal=False))
2110 if 'f4m' not in skip_protocols:
2111 formats.extend(self._extract_f4m_formats(
2112 http_base_url + '/manifest.f4m',
2113 video_id, f4m_id='hds', fatal=False))
2114 if 'dash' not in skip_protocols:
2115 formats.extend(self._extract_mpd_formats(
2116 http_base_url + '/manifest.mpd',
2117 video_id, mpd_id='dash', fatal=False))
2118 if re.search(r'(?:/smil:|\.smil)', url_base):
2119 if 'smil' not in skip_protocols:
2120 rtmp_formats = self._extract_smil_formats(
2121 http_base_url + '/jwplayer.smil',
2122 video_id, fatal=False)
2123 for rtmp_format in rtmp_formats:
2124 rtsp_format = rtmp_format.copy()
2125 rtsp_format['url'] = '%s/%s' % (rtmp_format['url'], rtmp_format['play_path'])
2126 del rtsp_format['play_path']
2127 del rtsp_format['ext']
2128 rtsp_format.update({
2129 'url': rtsp_format['url'].replace('rtmp://', 'rtsp://'),
2130 'format_id': rtmp_format['format_id'].replace('rtmp', 'rtsp'),
2133 formats.extend([rtmp_format, rtsp_format])
2135 for protocol in ('rtmp', 'rtsp'):
2136 if protocol not in skip_protocols:
2138 'url': protocol + url_base,
2139 'format_id': protocol,
2140 'protocol': protocol,
2145 def _find_jwplayer_data(webpage):
2147 r'jwplayer\((?P<quote>[\'"])[^\'" ]+(?P=quote)\)\.setup\s*\((?P<options>[^)]+)\)',
2150 return mobj.group('options')
2152 def _extract_jwplayer_data(self, webpage, video_id, *args, **kwargs):
2153 jwplayer_data = self._parse_json(
2154 self._find_jwplayer_data(webpage), video_id,
2155 transform_source=js_to_json)
2156 return self._parse_jwplayer_data(
2157 jwplayer_data, video_id, *args, **kwargs)
2159 def _parse_jwplayer_data(self, jwplayer_data, video_id=None, require_title=True,
2160 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
2161 # JWPlayer backward compatibility: flattened playlists
2162 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/api/config.js#L81-L96
2163 if 'playlist' not in jwplayer_data:
2164 jwplayer_data = {'playlist': [jwplayer_data]}
2168 # JWPlayer backward compatibility: single playlist item
2169 # https://github.com/jwplayer/jwplayer/blob/v7.7.0/src/js/playlist/playlist.js#L10
2170 if not isinstance(jwplayer_data['playlist'], list):
2171 jwplayer_data['playlist'] = [jwplayer_data['playlist']]
2173 for video_data in jwplayer_data['playlist']:
2174 # JWPlayer backward compatibility: flattened sources
2175 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/playlist/item.js#L29-L35
2176 if 'sources' not in video_data:
2177 video_data['sources'] = [video_data]
2179 this_video_id = video_id or video_data['mediaid']
2182 for source in video_data['sources']:
2183 source_url = self._proto_relative_url(source['file'])
2185 source_url = compat_urlparse.urljoin(base_url, source_url)
2186 source_type = source.get('type') or ''
2187 ext = mimetype2ext(source_type) or determine_ext(source_url)
2188 if source_type == 'hls' or ext == 'm3u8':
2189 formats.extend(self._extract_m3u8_formats(
2190 source_url, this_video_id, 'mp4', 'm3u8_native', m3u8_id=m3u8_id, fatal=False))
2192 formats.extend(self._extract_mpd_formats(
2193 source_url, this_video_id, mpd_id=mpd_id, fatal=False))
2194 # https://github.com/jwplayer/jwplayer/blob/master/src/js/providers/default.js#L67
2195 elif source_type.startswith('audio') or ext in ('oga', 'aac', 'mp3', 'mpeg', 'vorbis'):
2202 height = int_or_none(source.get('height'))
2204 # Often no height is provided but there is a label in
2205 # format like 1080p.
2206 height = int_or_none(self._search_regex(
2207 r'^(\d{3,})[pP]$', source.get('label') or '',
2208 'height', default=None))
2211 'width': int_or_none(source.get('width')),
2215 if source_url.startswith('rtmp'):
2216 a_format['ext'] = 'flv'
2218 # See com/longtailvideo/jwplayer/media/RTMPMediaProvider.as
2219 # of jwplayer.flash.swf
2220 rtmp_url_parts = re.split(
2221 r'((?:mp4|mp3|flv):)', source_url, 1)
2222 if len(rtmp_url_parts) == 3:
2223 rtmp_url, prefix, play_path = rtmp_url_parts
2226 'play_path': prefix + play_path,
2229 a_format.update(rtmp_params)
2230 formats.append(a_format)
2231 self._sort_formats(formats)
2234 tracks = video_data.get('tracks')
2235 if tracks and isinstance(tracks, list):
2236 for track in tracks:
2237 if track.get('kind') != 'captions':
2239 track_url = urljoin(base_url, track.get('file'))
2242 subtitles.setdefault(track.get('label') or 'en', []).append({
2243 'url': self._proto_relative_url(track_url)
2247 'id': this_video_id,
2248 'title': video_data['title'] if require_title else video_data.get('title'),
2249 'description': video_data.get('description'),
2250 'thumbnail': self._proto_relative_url(video_data.get('image')),
2251 'timestamp': int_or_none(video_data.get('pubdate')),
2252 'duration': float_or_none(jwplayer_data.get('duration') or video_data.get('duration')),
2253 'subtitles': subtitles,
2256 if len(entries) == 1:
2259 return self.playlist_result(entries)
2261 def _live_title(self, name):
2262 """ Generate the title for a live video """
2263 now = datetime.datetime.now()
2264 now_str = now.strftime('%Y-%m-%d %H:%M')
2265 return name + ' ' + now_str
2267 def _int(self, v, name, fatal=False, **kwargs):
2268 res = int_or_none(v, **kwargs)
2269 if 'get_attr' in kwargs:
2270 print(getattr(v, kwargs['get_attr']))
2272 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
2274 raise ExtractorError(msg)
2276 self._downloader.report_warning(msg)
2279 def _float(self, v, name, fatal=False, **kwargs):
2280 res = float_or_none(v, **kwargs)
2282 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
2284 raise ExtractorError(msg)
2286 self._downloader.report_warning(msg)
2289 def _set_cookie(self, domain, name, value, expire_time=None):
2290 cookie = compat_cookiejar.Cookie(
2291 0, name, value, None, None, domain, None,
2292 None, '/', True, False, expire_time, '', None, None, None)
2293 self._downloader.cookiejar.set_cookie(cookie)
2295 def _get_cookies(self, url):
2296 """ Return a compat_cookies.SimpleCookie with the cookies for the url """
2297 req = sanitized_Request(url)
2298 self._downloader.cookiejar.add_cookie_header(req)
2299 return compat_cookies.SimpleCookie(req.get_header('Cookie'))
2301 def get_testcases(self, include_onlymatching=False):
2302 t = getattr(self, '_TEST', None)
2304 assert not hasattr(self, '_TESTS'), \
2305 '%s has _TEST and _TESTS' % type(self).__name__
2308 tests = getattr(self, '_TESTS', [])
2310 if not include_onlymatching and t.get('only_matching', False):
2312 t['name'] = type(self).__name__[:-len('IE')]
2315 def is_suitable(self, age_limit):
2316 """ Test whether the extractor is generally suitable for the given
2317 age limit (i.e. pornographic sites are not, all others usually are) """
2319 any_restricted = False
2320 for tc in self.get_testcases(include_onlymatching=False):
2321 if tc.get('playlist', []):
2322 tc = tc['playlist'][0]
2323 is_restricted = age_restricted(
2324 tc.get('info_dict', {}).get('age_limit'), age_limit)
2325 if not is_restricted:
2327 any_restricted = any_restricted or is_restricted
2328 return not any_restricted
2330 def extract_subtitles(self, *args, **kwargs):
2331 if (self._downloader.params.get('writesubtitles', False) or
2332 self._downloader.params.get('listsubtitles')):
2333 return self._get_subtitles(*args, **kwargs)
2336 def _get_subtitles(self, *args, **kwargs):
2337 raise NotImplementedError('This method must be implemented by subclasses')
2340 def _merge_subtitle_items(subtitle_list1, subtitle_list2):
2341 """ Merge subtitle items for one language. Items with duplicated URLs
2342 will be dropped. """
2343 list1_urls = set([item['url'] for item in subtitle_list1])
2344 ret = list(subtitle_list1)
2345 ret.extend([item for item in subtitle_list2 if item['url'] not in list1_urls])
2349 def _merge_subtitles(cls, subtitle_dict1, subtitle_dict2):
2350 """ Merge two subtitle dictionaries, language by language. """
2351 ret = dict(subtitle_dict1)
2352 for lang in subtitle_dict2:
2353 ret[lang] = cls._merge_subtitle_items(subtitle_dict1.get(lang, []), subtitle_dict2[lang])
2356 def extract_automatic_captions(self, *args, **kwargs):
2357 if (self._downloader.params.get('writeautomaticsub', False) or
2358 self._downloader.params.get('listsubtitles')):
2359 return self._get_automatic_captions(*args, **kwargs)
2362 def _get_automatic_captions(self, *args, **kwargs):
2363 raise NotImplementedError('This method must be implemented by subclasses')
2365 def mark_watched(self, *args, **kwargs):
2366 if (self._downloader.params.get('mark_watched', False) and
2367 (self._get_login_info()[0] is not None or
2368 self._downloader.params.get('cookiefile') is not None)):
2369 self._mark_watched(*args, **kwargs)
2371 def _mark_watched(self, *args, **kwargs):
2372 raise NotImplementedError('This method must be implemented by subclasses')
2374 def geo_verification_headers(self):
2376 geo_verification_proxy = self._downloader.params.get('geo_verification_proxy')
2377 if geo_verification_proxy:
2378 headers['Ytdl-request-proxy'] = geo_verification_proxy
2381 def _generic_id(self, url):
2382 return compat_urllib_parse_unquote(os.path.splitext(url.rstrip('/').split('/')[-1])[0])
2384 def _generic_title(self, url):
2385 return compat_urllib_parse_unquote(os.path.splitext(url_basename(url))[0])
2388 class SearchInfoExtractor(InfoExtractor):
2390 Base class for paged search queries extractors.
2391 They accept URLs in the format _SEARCH_KEY(|all|[0-9]):{query}
2392 Instances should define _SEARCH_KEY and _MAX_RESULTS.
2396 def _make_valid_url(cls):
2397 return r'%s(?P<prefix>|[1-9][0-9]*|all):(?P<query>[\s\S]+)' % cls._SEARCH_KEY
2400 def suitable(cls, url):
2401 return re.match(cls._make_valid_url(), url) is not None
2403 def _real_extract(self, query):
2404 mobj = re.match(self._make_valid_url(), query)
2406 raise ExtractorError('Invalid search query "%s"' % query)
2408 prefix = mobj.group('prefix')
2409 query = mobj.group('query')
2411 return self._get_n_results(query, 1)
2412 elif prefix == 'all':
2413 return self._get_n_results(query, self._MAX_RESULTS)
2417 raise ExtractorError('invalid download number %s for query "%s"' % (n, query))
2418 elif n > self._MAX_RESULTS:
2419 self._downloader.report_warning('%s returns max %i results (you requested %i)' % (self._SEARCH_KEY, self._MAX_RESULTS, n))
2420 n = self._MAX_RESULTS
2421 return self._get_n_results(query, n)
2423 def _get_n_results(self, query, n):
2424 """Get a specified number of results for a query"""
2425 raise NotImplementedError('This method must be implemented by subclasses')
2428 def SEARCH_KEY(self):
2429 return self._SEARCH_KEY