1 from __future__ import unicode_literals
15 from ..compat import (
18 compat_etree_fromstring,
24 compat_urllib_parse_urlencode,
25 compat_urllib_request,
28 from ..downloader.f4m import remove_encrypted_media
58 class InfoExtractor(object):
59 """Information Extractor class.
61 Information extractors are the classes that, given a URL, extract
62 information about the video (or videos) the URL refers to. This
63 information includes the real video URL, the video title, author and
64 others. The information is stored in a dictionary which is then
65 passed to the YoutubeDL. The YoutubeDL processes this
66 information possibly downloading the video to the file system, among
67 other possible outcomes.
69 The type field determines the type of the result.
70 By far the most common value (and the default if _type is missing) is
71 "video", which indicates a single video.
73 For a video, the dictionaries must include the following fields:
76 title: Video title, unescaped.
78 Additionally, it must contain either a formats entry or a url one:
80 formats: A list of dictionaries for each format available, ordered
81 from worst to best quality.
84 * url Mandatory. The URL of the video file
85 * ext Will be calculated from URL if missing
86 * format A human-readable description of the format
87 ("mp4 container with h264/opus").
88 Calculated from the format_id, width, height.
89 and format_note fields if missing.
90 * format_id A short description of the format
91 ("mp4_h264_opus" or "19").
92 Technically optional, but strongly recommended.
93 * format_note Additional info about the format
94 ("3D" or "DASH video")
95 * width Width of the video, if known
96 * height Height of the video, if known
97 * resolution Textual description of width and height
98 * tbr Average bitrate of audio and video in KBit/s
99 * abr Average audio bitrate in KBit/s
100 * acodec Name of the audio codec in use
101 * asr Audio sampling rate in Hertz
102 * vbr Average video bitrate in KBit/s
104 * vcodec Name of the video codec in use
105 * container Name of the container format
106 * filesize The number of bytes, if known in advance
107 * filesize_approx An estimate for the number of bytes
108 * player_url SWF Player URL (used for rtmpdump).
109 * protocol The protocol that will be used for the actual
110 download, lower-case.
111 "http", "https", "rtsp", "rtmp", "rtmpe",
112 "m3u8", "m3u8_native" or "http_dash_segments".
113 * preference Order number of this format. If this field is
114 present and not None, the formats get sorted
115 by this field, regardless of all other values.
116 -1 for default (order by other properties),
117 -2 or smaller for less than default.
118 < -1000 to hide the format (if there is
119 another one which is strictly better)
120 * language Language code, e.g. "de" or "en-US".
121 * language_preference Is this in the language mentioned in
123 10 if it's what the URL is about,
124 -1 for default (don't know),
125 -10 otherwise, other values reserved for now.
126 * quality Order number of the video quality of this
127 format, irrespective of the file format.
128 -1 for default (order by other properties),
129 -2 or smaller for less than default.
130 * source_preference Order number for this video source
131 (quality takes higher priority)
132 -1 for default (order by other properties),
133 -2 or smaller for less than default.
134 * http_headers A dictionary of additional HTTP headers
135 to add to the request.
136 * stretched_ratio If given and not 1, indicates that the
137 video's pixels are not square.
138 width : height ratio as float.
139 * no_resume The server does not support resuming the
140 (HTTP or RTMP) download. Boolean.
142 url: Final video URL.
143 ext: Video filename extension.
144 format: The video format, defaults to ext (used for --get-format)
145 player_url: SWF Player URL (used for rtmpdump).
147 The following fields are optional:
149 alt_title: A secondary title of the video.
150 display_id An alternative identifier for the video, not necessarily
151 unique, but available before title. Typically, id is
152 something like "4234987", title "Dancing naked mole rats",
153 and display_id "dancing-naked-mole-rats"
154 thumbnails: A list of dictionaries, with the following entries:
155 * "id" (optional, string) - Thumbnail format ID
157 * "preference" (optional, int) - quality of the image
158 * "width" (optional, int)
159 * "height" (optional, int)
160 * "resolution" (optional, string "{width}x{height"},
162 thumbnail: Full URL to a video thumbnail image.
163 description: Full video description.
164 uploader: Full name of the video uploader.
165 license: License name the video is licensed under.
166 creator: The main artist who created the video.
167 release_date: The date (YYYYMMDD) when the video was released.
168 timestamp: UNIX timestamp of the moment the video became available.
169 upload_date: Video upload date (YYYYMMDD).
170 If not explicitly set, calculated from timestamp.
171 uploader_id: Nickname or id of the video uploader.
172 uploader_url: Full URL to a personal webpage of the video uploader.
173 location: Physical location where the video was filmed.
174 subtitles: The available subtitles as a dictionary in the format
175 {language: subformats}. "subformats" is a list sorted from
176 lower to higher preference, each element is a dictionary
177 with the "ext" entry and one of:
178 * "data": The subtitles file contents
179 * "url": A URL pointing to the subtitles file
180 "ext" will be calculated from URL if missing
181 automatic_captions: Like 'subtitles', used by the YoutubeIE for
182 automatically generated captions
183 duration: Length of the video in seconds, as an integer or float.
184 view_count: How many users have watched the video on the platform.
185 like_count: Number of positive ratings of the video
186 dislike_count: Number of negative ratings of the video
187 repost_count: Number of reposts of the video
188 average_rating: Average rating give by users, the scale used depends on the webpage
189 comment_count: Number of comments on the video
190 comments: A list of comments, each with one or more of the following
191 properties (all but one of text or html optional):
192 * "author" - human-readable name of the comment author
193 * "author_id" - user ID of the comment author
195 * "html" - Comment as HTML
196 * "text" - Plain text of the comment
197 * "timestamp" - UNIX timestamp of comment
198 * "parent" - ID of the comment this one is replying to.
199 Set to "root" to indicate that this is a
200 comment to the original video.
201 age_limit: Age restriction for the video, as an integer (years)
202 webpage_url: The URL to the video webpage, if given to youtube-dl it
203 should allow to get the same result again. (It will be set
204 by YoutubeDL if it's missing)
205 categories: A list of categories that the video falls in, for example
207 tags: A list of tags assigned to the video, e.g. ["sweden", "pop music"]
208 is_live: True, False, or None (=unknown). Whether this video is a
209 live stream that goes on instead of a fixed-length video.
210 start_time: Time in seconds where the reproduction should start, as
211 specified in the URL.
212 end_time: Time in seconds where the reproduction should end, as
213 specified in the URL.
215 The following fields should only be used when the video belongs to some logical
218 chapter: Name or title of the chapter the video belongs to.
219 chapter_number: Number of the chapter the video belongs to, as an integer.
220 chapter_id: Id of the chapter the video belongs to, as a unicode string.
222 The following fields should only be used when the video is an episode of some
225 series: Title of the series or programme the video episode belongs to.
226 season: Title of the season the video episode belongs to.
227 season_number: Number of the season the video episode belongs to, as an integer.
228 season_id: Id of the season the video episode belongs to, as a unicode string.
229 episode: Title of the video episode. Unlike mandatory video title field,
230 this field should denote the exact title of the video episode
231 without any kind of decoration.
232 episode_number: Number of the video episode within a season, as an integer.
233 episode_id: Id of the video episode, as a unicode string.
235 Unless mentioned otherwise, the fields should be Unicode strings.
237 Unless mentioned otherwise, None is equivalent to absence of information.
240 _type "playlist" indicates multiple videos.
241 There must be a key "entries", which is a list, an iterable, or a PagedList
242 object, each element of which is a valid dictionary by this specification.
244 Additionally, playlists can have "title", "description" and "id" attributes
245 with the same semantics as videos (see above).
248 _type "multi_video" indicates that there are multiple videos that
249 form a single show, for examples multiple acts of an opera or TV episode.
250 It must have an entries key like a playlist and contain all the keys
251 required for a video at the same time.
254 _type "url" indicates that the video must be extracted from another
255 location, possibly by a different extractor. Its only required key is:
256 "url" - the next URL to extract.
257 The key "ie_key" can be set to the class name (minus the trailing "IE",
258 e.g. "Youtube") if the extractor class is known in advance.
259 Additionally, the dictionary may have any properties of the resolved entity
260 known in advance, for example "title" if the title of the referred video is
264 _type "url_transparent" entities have the same specification as "url", but
265 indicate that the given additional information is more precise than the one
266 associated with the resolved URL.
267 This is useful when a site employs a video service that hosts the video and
268 its technical metadata, but that video service does not embed a useful
269 title, description etc.
272 Subclasses of this one should re-define the _real_initialize() and
273 _real_extract() methods and define a _VALID_URL regexp.
274 Probably, they should also be added to the list of extractors.
276 Finally, the _WORKING attribute should be set to False for broken IEs
277 in order to warn the users and skip the tests.
284 def __init__(self, downloader=None):
285 """Constructor. Receives an optional downloader."""
287 self.set_downloader(downloader)
290 def suitable(cls, url):
291 """Receives a URL and returns True if suitable for this IE."""
293 # This does not use has/getattr intentionally - we want to know whether
294 # we have cached the regexp for *this* class, whereas getattr would also
295 # match the superclass
296 if '_VALID_URL_RE' not in cls.__dict__:
297 cls._VALID_URL_RE = re.compile(cls._VALID_URL)
298 return cls._VALID_URL_RE.match(url) is not None
301 def _match_id(cls, url):
302 if '_VALID_URL_RE' not in cls.__dict__:
303 cls._VALID_URL_RE = re.compile(cls._VALID_URL)
304 m = cls._VALID_URL_RE.match(url)
310 """Getter method for _WORKING."""
313 def initialize(self):
314 """Initializes an instance (authentication, etc)."""
316 self._real_initialize()
319 def extract(self, url):
320 """Extracts URL information and returns it in list of dicts."""
323 return self._real_extract(url)
324 except ExtractorError:
326 except compat_http_client.IncompleteRead as e:
327 raise ExtractorError('A network error has occurred.', cause=e, expected=True)
328 except (KeyError, StopIteration) as e:
329 raise ExtractorError('An extractor error has occurred.', cause=e)
331 def set_downloader(self, downloader):
332 """Sets the downloader for this IE."""
333 self._downloader = downloader
335 def _real_initialize(self):
336 """Real initialization process. Redefine in subclasses."""
339 def _real_extract(self, url):
340 """Real extraction process. Redefine in subclasses."""
345 """A string for getting the InfoExtractor with get_info_extractor"""
346 return compat_str(cls.__name__[:-2])
350 return compat_str(type(self).__name__[:-2])
352 def _request_webpage(self, url_or_request, video_id, note=None, errnote=None, fatal=True, data=None, headers={}, query={}):
353 """ Returns the response handle """
355 self.report_download_webpage(video_id)
356 elif note is not False:
358 self.to_screen('%s' % (note,))
360 self.to_screen('%s: %s' % (video_id, note))
361 # data, headers and query params will be ignored for `Request` objects
362 if isinstance(url_or_request, compat_urllib_request.Request):
363 url_or_request = update_Request(
364 url_or_request, data=data, headers=headers, query=query)
367 url_or_request = update_url_query(url_or_request, query)
369 url_or_request = sanitized_Request(url_or_request, data, headers)
371 return self._downloader.urlopen(url_or_request)
372 except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err:
376 errnote = 'Unable to download webpage'
378 errmsg = '%s: %s' % (errnote, error_to_compat_str(err))
380 raise ExtractorError(errmsg, sys.exc_info()[2], cause=err)
382 self._downloader.report_warning(errmsg)
385 def _download_webpage_handle(self, url_or_request, video_id, note=None, errnote=None, fatal=True, encoding=None, data=None, headers={}, query={}):
386 """ Returns a tuple (page content as string, URL handle) """
387 # Strip hashes from the URL (#1038)
388 if isinstance(url_or_request, (compat_str, str)):
389 url_or_request = url_or_request.partition('#')[0]
391 urlh = self._request_webpage(url_or_request, video_id, note, errnote, fatal, data=data, headers=headers, query=query)
395 content = self._webpage_read_content(urlh, url_or_request, video_id, note, errnote, fatal, encoding=encoding)
396 return (content, urlh)
399 def _guess_encoding_from_content(content_type, webpage_bytes):
400 m = re.match(r'[a-zA-Z0-9_.-]+/[a-zA-Z0-9_.-]+\s*;\s*charset=(.+)', content_type)
402 encoding = m.group(1)
404 m = re.search(br'<meta[^>]+charset=[\'"]?([^\'")]+)[ /\'">]',
405 webpage_bytes[:1024])
407 encoding = m.group(1).decode('ascii')
408 elif webpage_bytes.startswith(b'\xff\xfe'):
415 def _webpage_read_content(self, urlh, url_or_request, video_id, note=None, errnote=None, fatal=True, prefix=None, encoding=None):
416 content_type = urlh.headers.get('Content-Type', '')
417 webpage_bytes = urlh.read()
418 if prefix is not None:
419 webpage_bytes = prefix + webpage_bytes
421 encoding = self._guess_encoding_from_content(content_type, webpage_bytes)
422 if self._downloader.params.get('dump_intermediate_pages', False):
424 url = url_or_request.get_full_url()
425 except AttributeError:
427 self.to_screen('Dumping request to ' + url)
428 dump = base64.b64encode(webpage_bytes).decode('ascii')
429 self._downloader.to_screen(dump)
430 if self._downloader.params.get('write_pages', False):
432 url = url_or_request.get_full_url()
433 except AttributeError:
435 basen = '%s_%s' % (video_id, url)
437 h = '___' + hashlib.md5(basen.encode('utf-8')).hexdigest()
438 basen = basen[:240 - len(h)] + h
439 raw_filename = basen + '.dump'
440 filename = sanitize_filename(raw_filename, restricted=True)
441 self.to_screen('Saving request to ' + filename)
442 # Working around MAX_PATH limitation on Windows (see
443 # http://msdn.microsoft.com/en-us/library/windows/desktop/aa365247(v=vs.85).aspx)
444 if compat_os_name == 'nt':
445 absfilepath = os.path.abspath(filename)
446 if len(absfilepath) > 259:
447 filename = '\\\\?\\' + absfilepath
448 with open(filename, 'wb') as outf:
449 outf.write(webpage_bytes)
452 content = webpage_bytes.decode(encoding, 'replace')
454 content = webpage_bytes.decode('utf-8', 'replace')
456 if ('<title>Access to this site is blocked</title>' in content and
457 'Websense' in content[:512]):
458 msg = 'Access to this webpage has been blocked by Websense filtering software in your network.'
459 blocked_iframe = self._html_search_regex(
460 r'<iframe src="([^"]+)"', content,
461 'Websense information URL', default=None)
463 msg += ' Visit %s for more details' % blocked_iframe
464 raise ExtractorError(msg, expected=True)
465 if '<title>The URL you requested has been blocked</title>' in content[:512]:
467 'Access to this webpage has been blocked by Indian censorship. '
468 'Use a VPN or proxy server (with --proxy) to route around it.')
469 block_msg = self._html_search_regex(
470 r'</h1><p>(.*?)</p>',
471 content, 'block message', default=None)
473 msg += ' (Message: "%s")' % block_msg.replace('\n', ' ')
474 raise ExtractorError(msg, expected=True)
478 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={}):
479 """ Returns the data of the page as a string """
482 while success is False:
484 res = self._download_webpage_handle(url_or_request, video_id, note, errnote, fatal, encoding=encoding, data=data, headers=headers, query=query)
486 except compat_http_client.IncompleteRead as e:
488 if try_count >= tries:
490 self._sleep(timeout, video_id)
497 def _download_xml(self, url_or_request, video_id,
498 note='Downloading XML', errnote='Unable to download XML',
499 transform_source=None, fatal=True, encoding=None, data=None, headers={}, query={}):
500 """Return the xml as an xml.etree.ElementTree.Element"""
501 xml_string = self._download_webpage(
502 url_or_request, video_id, note, errnote, fatal=fatal, encoding=encoding, data=data, headers=headers, query=query)
503 if xml_string is False:
506 xml_string = transform_source(xml_string)
507 return compat_etree_fromstring(xml_string.encode('utf-8'))
509 def _download_json(self, url_or_request, video_id,
510 note='Downloading JSON metadata',
511 errnote='Unable to download JSON metadata',
512 transform_source=None,
513 fatal=True, encoding=None, data=None, headers={}, query={}):
514 json_string = self._download_webpage(
515 url_or_request, video_id, note, errnote, fatal=fatal,
516 encoding=encoding, data=data, headers=headers, query=query)
517 if (not fatal) and json_string is False:
519 return self._parse_json(
520 json_string, video_id, transform_source=transform_source, fatal=fatal)
522 def _parse_json(self, json_string, video_id, transform_source=None, fatal=True):
524 json_string = transform_source(json_string)
526 return json.loads(json_string)
527 except ValueError as ve:
528 errmsg = '%s: Failed to parse JSON ' % video_id
530 raise ExtractorError(errmsg, cause=ve)
532 self.report_warning(errmsg + str(ve))
534 def report_warning(self, msg, video_id=None):
535 idstr = '' if video_id is None else '%s: ' % video_id
536 self._downloader.report_warning(
537 '[%s] %s%s' % (self.IE_NAME, idstr, msg))
539 def to_screen(self, msg):
540 """Print msg to screen, prefixing it with '[ie_name]'"""
541 self._downloader.to_screen('[%s] %s' % (self.IE_NAME, msg))
543 def report_extraction(self, id_or_name):
544 """Report information extraction."""
545 self.to_screen('%s: Extracting information' % id_or_name)
547 def report_download_webpage(self, video_id):
548 """Report webpage download."""
549 self.to_screen('%s: Downloading webpage' % video_id)
551 def report_age_confirmation(self):
552 """Report attempt to confirm age."""
553 self.to_screen('Confirming age')
555 def report_login(self):
556 """Report attempt to log in."""
557 self.to_screen('Logging in')
560 def raise_login_required(msg='This video is only available for registered users'):
561 raise ExtractorError(
562 '%s. Use --username and --password or --netrc to provide account credentials.' % msg,
566 def raise_geo_restricted(msg='This video is not available from your location due to geo restriction'):
567 raise ExtractorError(
568 '%s. You might want to use --proxy to workaround.' % msg,
571 # Methods for following #608
573 def url_result(url, ie=None, video_id=None, video_title=None):
574 """Returns a URL that points to a page that should be processed"""
575 # TODO: ie should be the class used for getting the info
576 video_info = {'_type': 'url',
579 if video_id is not None:
580 video_info['id'] = video_id
581 if video_title is not None:
582 video_info['title'] = video_title
586 def playlist_result(entries, playlist_id=None, playlist_title=None, playlist_description=None):
587 """Returns a playlist"""
588 video_info = {'_type': 'playlist',
591 video_info['id'] = playlist_id
593 video_info['title'] = playlist_title
594 if playlist_description:
595 video_info['description'] = playlist_description
598 def _search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
600 Perform a regex search on the given string, using a single or a list of
601 patterns returning the first matching group.
602 In case of failure return a default value or raise a WARNING or a
603 RegexNotFoundError, depending on fatal, specifying the field name.
605 if isinstance(pattern, (str, compat_str, compiled_regex_type)):
606 mobj = re.search(pattern, string, flags)
609 mobj = re.search(p, string, flags)
613 if not self._downloader.params.get('no_color') and compat_os_name != 'nt' and sys.stderr.isatty():
614 _name = '\033[0;34m%s\033[0m' % name
620 # return the first matching group
621 return next(g for g in mobj.groups() if g is not None)
623 return mobj.group(group)
624 elif default is not NO_DEFAULT:
627 raise RegexNotFoundError('Unable to extract %s' % _name)
629 self._downloader.report_warning('unable to extract %s' % _name + bug_reports_message())
632 def _html_search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
634 Like _search_regex, but strips HTML tags and unescapes entities.
636 res = self._search_regex(pattern, string, name, default, fatal, flags, group)
638 return clean_html(res).strip()
642 def _get_login_info(self):
644 Get the login info as (username, password)
645 It will look in the netrc file using the _NETRC_MACHINE value
646 If there's no info available, return (None, None)
648 if self._downloader is None:
653 downloader_params = self._downloader.params
655 # Attempt to use provided username and password or .netrc data
656 if downloader_params.get('username') is not None:
657 username = downloader_params['username']
658 password = downloader_params['password']
659 elif downloader_params.get('usenetrc', False):
661 info = netrc.netrc().authenticators(self._NETRC_MACHINE)
666 raise netrc.NetrcParseError('No authenticators for %s' % self._NETRC_MACHINE)
667 except (IOError, netrc.NetrcParseError) as err:
668 self._downloader.report_warning('parsing .netrc: %s' % error_to_compat_str(err))
670 return (username, password)
672 def _get_tfa_info(self, note='two-factor verification code'):
674 Get the two-factor authentication info
675 TODO - asking the user will be required for sms/phone verify
676 currently just uses the command line option
677 If there's no info available, return None
679 if self._downloader is None:
681 downloader_params = self._downloader.params
683 if downloader_params.get('twofactor') is not None:
684 return downloader_params['twofactor']
686 return compat_getpass('Type %s and press [Return]: ' % note)
688 # Helper functions for extracting OpenGraph info
690 def _og_regexes(prop):
691 content_re = r'content=(?:"([^"]+?)"|\'([^\']+?)\'|\s*([^\s"\'=<>`]+?))'
692 property_re = (r'(?:name|property)=(?:\'og:%(prop)s\'|"og:%(prop)s"|\s*og:%(prop)s\b)'
693 % {'prop': re.escape(prop)})
694 template = r'<meta[^>]+?%s[^>]+?%s'
696 template % (property_re, content_re),
697 template % (content_re, property_re),
701 def _meta_regex(prop):
702 return r'''(?isx)<meta
703 (?=[^>]+(?:itemprop|name|property|id|http-equiv)=(["\']?)%s\1)
704 [^>]+?content=(["\'])(?P<content>.*?)\2''' % re.escape(prop)
706 def _og_search_property(self, prop, html, name=None, **kargs):
708 name = 'OpenGraph %s' % prop
709 escaped = self._search_regex(self._og_regexes(prop), html, name, flags=re.DOTALL, **kargs)
712 return unescapeHTML(escaped)
714 def _og_search_thumbnail(self, html, **kargs):
715 return self._og_search_property('image', html, 'thumbnail URL', fatal=False, **kargs)
717 def _og_search_description(self, html, **kargs):
718 return self._og_search_property('description', html, fatal=False, **kargs)
720 def _og_search_title(self, html, **kargs):
721 return self._og_search_property('title', html, **kargs)
723 def _og_search_video_url(self, html, name='video url', secure=True, **kargs):
724 regexes = self._og_regexes('video') + self._og_regexes('video:url')
726 regexes = self._og_regexes('video:secure_url') + regexes
727 return self._html_search_regex(regexes, html, name, **kargs)
729 def _og_search_url(self, html, **kargs):
730 return self._og_search_property('url', html, **kargs)
732 def _html_search_meta(self, name, html, display_name=None, fatal=False, **kwargs):
733 if display_name is None:
735 return self._html_search_regex(
736 self._meta_regex(name),
737 html, display_name, fatal=fatal, group='content', **kwargs)
739 def _dc_search_uploader(self, html):
740 return self._html_search_meta('dc.creator', html, 'uploader')
742 def _rta_search(self, html):
743 # See http://www.rtalabel.org/index.php?content=howtofaq#single
744 if re.search(r'(?ix)<meta\s+name="rating"\s+'
745 r' content="RTA-5042-1996-1400-1577-RTA"',
750 def _media_rating_search(self, html):
751 # See http://www.tjg-designs.com/WP/metadata-code-examples-adding-metadata-to-your-web-pages/
752 rating = self._html_search_meta('rating', html)
764 return RATING_TABLE.get(rating.lower())
766 def _family_friendly_search(self, html):
767 # See http://schema.org/VideoObject
768 family_friendly = self._html_search_meta('isFamilyFriendly', html)
770 if not family_friendly:
779 return RATING_TABLE.get(family_friendly.lower())
781 def _twitter_search_player(self, html):
782 return self._html_search_meta('twitter:player', html,
783 'twitter card player')
785 def _search_json_ld(self, html, video_id, **kwargs):
786 json_ld = self._search_regex(
787 r'(?s)<script[^>]+type=(["\'])application/ld\+json\1[^>]*>(?P<json_ld>.+?)</script>',
788 html, 'JSON-LD', group='json_ld', **kwargs)
791 return self._json_ld(json_ld, video_id, fatal=kwargs.get('fatal', True))
793 def _json_ld(self, json_ld, video_id, fatal=True):
794 if isinstance(json_ld, compat_str):
795 json_ld = self._parse_json(json_ld, video_id, fatal=fatal)
799 if json_ld.get('@context') == 'http://schema.org':
800 item_type = json_ld.get('@type')
801 if item_type == 'TVEpisode':
803 'episode': unescapeHTML(json_ld.get('name')),
804 'episode_number': int_or_none(json_ld.get('episodeNumber')),
805 'description': unescapeHTML(json_ld.get('description')),
807 part_of_season = json_ld.get('partOfSeason')
808 if isinstance(part_of_season, dict) and part_of_season.get('@type') == 'TVSeason':
809 info['season_number'] = int_or_none(part_of_season.get('seasonNumber'))
810 part_of_series = json_ld.get('partOfSeries')
811 if isinstance(part_of_series, dict) and part_of_series.get('@type') == 'TVSeries':
812 info['series'] = unescapeHTML(part_of_series.get('name'))
813 elif item_type == 'Article':
815 'timestamp': parse_iso8601(json_ld.get('datePublished')),
816 'title': unescapeHTML(json_ld.get('headline')),
817 'description': unescapeHTML(json_ld.get('articleBody')),
819 return dict((k, v) for k, v in info.items() if v is not None)
822 def _hidden_inputs(html):
823 html = re.sub(r'<!--(?:(?!<!--).)*-->', '', html)
825 for input in re.findall(r'(?i)<input([^>]+)>', html):
826 if not re.search(r'type=(["\'])(?:hidden|submit)\1', input):
828 name = re.search(r'name=(["\'])(?P<value>.+?)\1', input)
831 value = re.search(r'value=(["\'])(?P<value>.*?)\1', input)
834 hidden_inputs[name.group('value')] = value.group('value')
837 def _form_hidden_inputs(self, form_id, html):
838 form = self._search_regex(
839 r'(?is)<form[^>]+?id=(["\'])%s\1[^>]*>(?P<form>.+?)</form>' % form_id,
840 html, '%s form' % form_id, group='form')
841 return self._hidden_inputs(form)
843 def _sort_formats(self, formats, field_preference=None):
845 raise ExtractorError('No video formats found')
848 # Automatically determine tbr when missing based on abr and vbr (improves
849 # formats sorting in some cases)
850 if 'tbr' not in f and f.get('abr') is not None and f.get('vbr') is not None:
851 f['tbr'] = f['abr'] + f['vbr']
854 # TODO remove the following workaround
855 from ..utils import determine_ext
856 if not f.get('ext') and 'url' in f:
857 f['ext'] = determine_ext(f['url'])
859 if isinstance(field_preference, (list, tuple)):
860 return tuple(f.get(field) if f.get(field) is not None else -1 for field in field_preference)
862 preference = f.get('preference')
863 if preference is None:
865 if f.get('ext') in ['f4f', 'f4m']: # Not yet supported
868 proto_preference = 0 if determine_protocol(f) in ['http', 'https'] else -0.1
870 if f.get('vcodec') == 'none': # audio only
872 if self._downloader.params.get('prefer_free_formats'):
873 ORDER = ['aac', 'mp3', 'm4a', 'webm', 'ogg', 'opus']
875 ORDER = ['webm', 'opus', 'ogg', 'mp3', 'aac', 'm4a']
878 audio_ext_preference = ORDER.index(f['ext'])
880 audio_ext_preference = -1
882 if f.get('acodec') == 'none': # video only
884 if self._downloader.params.get('prefer_free_formats'):
885 ORDER = ['flv', 'mp4', 'webm']
887 ORDER = ['webm', 'flv', 'mp4']
889 ext_preference = ORDER.index(f['ext'])
892 audio_ext_preference = 0
896 f.get('language_preference') if f.get('language_preference') is not None else -1,
897 f.get('quality') if f.get('quality') is not None else -1,
898 f.get('tbr') if f.get('tbr') is not None else -1,
899 f.get('filesize') if f.get('filesize') is not None else -1,
900 f.get('vbr') if f.get('vbr') is not None else -1,
901 f.get('height') if f.get('height') is not None else -1,
902 f.get('width') if f.get('width') is not None else -1,
905 f.get('abr') if f.get('abr') is not None else -1,
906 audio_ext_preference,
907 f.get('fps') if f.get('fps') is not None else -1,
908 f.get('filesize_approx') if f.get('filesize_approx') is not None else -1,
909 f.get('source_preference') if f.get('source_preference') is not None else -1,
910 f.get('format_id') if f.get('format_id') is not None else '',
912 formats.sort(key=_formats_key)
914 def _check_formats(self, formats, video_id):
917 lambda f: self._is_valid_url(
919 item='%s video format' % f.get('format_id') if f.get('format_id') else 'video'),
923 def _remove_duplicate_formats(formats):
927 if f['url'] not in format_urls:
928 format_urls.add(f['url'])
929 unique_formats.append(f)
930 formats[:] = unique_formats
932 def _is_valid_url(self, url, video_id, item='video'):
933 url = self._proto_relative_url(url, scheme='http:')
934 # For now assume non HTTP(S) URLs always valid
935 if not (url.startswith('http://') or url.startswith('https://')):
938 self._request_webpage(url, video_id, 'Checking %s URL' % item)
940 except ExtractorError as e:
941 if isinstance(e.cause, compat_urllib_error.URLError):
943 '%s: %s URL is invalid, skipping' % (video_id, item))
947 def http_scheme(self):
948 """ Either "http:" or "https:", depending on the user's preferences """
951 if self._downloader.params.get('prefer_insecure', False)
954 def _proto_relative_url(self, url, scheme=None):
957 if url.startswith('//'):
959 scheme = self.http_scheme()
964 def _sleep(self, timeout, video_id, msg_template=None):
965 if msg_template is None:
966 msg_template = '%(video_id)s: Waiting for %(timeout)s seconds'
967 msg = msg_template % {'video_id': video_id, 'timeout': timeout}
971 def _extract_f4m_formats(self, manifest_url, video_id, preference=None, f4m_id=None,
972 transform_source=lambda s: fix_xml_ampersands(s).strip(),
974 manifest = self._download_xml(
975 manifest_url, video_id, 'Downloading f4m manifest',
976 'Unable to download f4m manifest',
977 # Some manifests may be malformed, e.g. prosiebensat1 generated manifests
978 # (see https://github.com/rg3/youtube-dl/issues/6215#issuecomment-121704244)
979 transform_source=transform_source,
982 if manifest is False:
985 return self._parse_f4m_formats(
986 manifest, manifest_url, video_id, preference=preference, f4m_id=f4m_id,
987 transform_source=transform_source, fatal=fatal)
989 def _parse_f4m_formats(self, manifest, manifest_url, video_id, preference=None, f4m_id=None,
990 transform_source=lambda s: fix_xml_ampersands(s).strip(),
993 manifest_version = '1.0'
994 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/1.0}media')
996 manifest_version = '2.0'
997 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/2.0}media')
998 # Remove unsupported DRM protected media from final formats
999 # rendition (see https://github.com/rg3/youtube-dl/issues/8573).
1000 media_nodes = remove_encrypted_media(media_nodes)
1003 base_url = xpath_text(
1004 manifest, ['{http://ns.adobe.com/f4m/1.0}baseURL', '{http://ns.adobe.com/f4m/2.0}baseURL'],
1005 'base URL', default=None)
1007 base_url = base_url.strip()
1008 for i, media_el in enumerate(media_nodes):
1009 if manifest_version == '2.0':
1010 media_url = media_el.attrib.get('href') or media_el.attrib.get('url')
1014 media_url if media_url.startswith('http://') or media_url.startswith('https://')
1015 else ((base_url or '/'.join(manifest_url.split('/')[:-1])) + '/' + media_url))
1016 # If media_url is itself a f4m manifest do the recursive extraction
1017 # since bitrates in parent manifest (this one) and media_url manifest
1018 # may differ leading to inability to resolve the format by requested
1019 # bitrate in f4m downloader
1020 if determine_ext(manifest_url) == 'f4m':
1021 formats.extend(self._extract_f4m_formats(
1022 manifest_url, video_id, preference=preference, f4m_id=f4m_id,
1023 transform_source=transform_source, fatal=fatal))
1025 tbr = int_or_none(media_el.attrib.get('bitrate'))
1027 'format_id': '-'.join(filter(None, [f4m_id, compat_str(i if tbr is None else tbr)])),
1028 'url': manifest_url,
1031 'width': int_or_none(media_el.attrib.get('width')),
1032 'height': int_or_none(media_el.attrib.get('height')),
1033 'preference': preference,
1037 def _extract_m3u8_formats(self, m3u8_url, video_id, ext=None,
1038 entry_protocol='m3u8', preference=None,
1039 m3u8_id=None, note=None, errnote=None,
1043 'format_id': '-'.join(filter(None, [m3u8_id, 'meta'])),
1047 'preference': preference - 1 if preference else -1,
1048 'resolution': 'multiple',
1049 'format_note': 'Quality selection URL',
1052 format_url = lambda u: (
1054 if re.match(r'^https?://', u)
1055 else compat_urlparse.urljoin(m3u8_url, u))
1057 res = self._download_webpage_handle(
1059 note=note or 'Downloading m3u8 information',
1060 errnote=errnote or 'Failed to download m3u8 information',
1064 m3u8_doc, urlh = res
1065 m3u8_url = urlh.geturl()
1067 # We should try extracting formats only from master playlists [1], i.e.
1068 # playlists that describe available qualities. On the other hand media
1069 # playlists [2] should be returned as is since they contain just the media
1070 # without qualities renditions.
1071 # Fortunately, master playlist can be easily distinguished from media
1072 # playlist based on particular tags availability. As of [1, 2] master
1073 # playlist tags MUST NOT appear in a media playist and vice versa.
1074 # As of [3] #EXT-X-TARGETDURATION tag is REQUIRED for every media playlist
1075 # and MUST NOT appear in master playlist thus we can clearly detect media
1076 # playlist with this criterion.
1077 # 1. https://tools.ietf.org/html/draft-pantos-http-live-streaming-17#section-4.3.4
1078 # 2. https://tools.ietf.org/html/draft-pantos-http-live-streaming-17#section-4.3.3
1079 # 3. https://tools.ietf.org/html/draft-pantos-http-live-streaming-17#section-4.3.3.1
1080 if '#EXT-X-TARGETDURATION' in m3u8_doc: # media playlist, return as is
1083 'format_id': m3u8_id,
1085 'protocol': entry_protocol,
1086 'preference': preference,
1090 kv_rex = re.compile(
1091 r'(?P<key>[a-zA-Z_-]+)=(?P<val>"[^"]+"|[^",]+)(?:,|$)')
1092 for line in m3u8_doc.splitlines():
1093 if line.startswith('#EXT-X-STREAM-INF:'):
1095 for m in kv_rex.finditer(line):
1097 if v.startswith('"'):
1099 last_info[m.group('key')] = v
1100 elif line.startswith('#EXT-X-MEDIA:'):
1102 for m in kv_rex.finditer(line):
1104 if v.startswith('"'):
1106 last_media[m.group('key')] = v
1107 elif line.startswith('#') or not line.strip():
1110 if last_info is None:
1111 formats.append({'url': format_url(line)})
1113 tbr = int_or_none(last_info.get('BANDWIDTH'), scale=1000)
1116 format_id.append(m3u8_id)
1117 last_media_name = last_media.get('NAME') if last_media and last_media.get('TYPE') != 'SUBTITLES' else None
1118 format_id.append(last_media_name if last_media_name else '%d' % (tbr if tbr else len(formats)))
1120 'format_id': '-'.join(format_id),
1121 'url': format_url(line.strip()),
1124 'protocol': entry_protocol,
1125 'preference': preference,
1127 resolution = last_info.get('RESOLUTION')
1129 width_str, height_str = resolution.split('x')
1130 f['width'] = int(width_str)
1131 f['height'] = int(height_str)
1132 codecs = last_info.get('CODECS')
1134 vcodec, acodec = [None] * 2
1135 va_codecs = codecs.split(',')
1136 if len(va_codecs) == 1:
1137 # Audio only entries usually come with single codec and
1138 # no resolution. For more robustness we also check it to
1140 if not resolution and va_codecs[0].startswith('mp4a'):
1141 vcodec, acodec = 'none', va_codecs[0]
1143 vcodec = va_codecs[0]
1145 vcodec, acodec = va_codecs[:2]
1150 if last_media is not None:
1151 f['m3u8_media'] = last_media
1158 def _xpath_ns(path, namespace=None):
1162 for c in path.split('/'):
1163 if not c or c == '.':
1166 out.append('{%s}%s' % (namespace, c))
1167 return '/'.join(out)
1169 def _extract_smil_formats(self, smil_url, video_id, fatal=True, f4m_params=None, transform_source=None):
1170 smil = self._download_smil(smil_url, video_id, fatal=fatal, transform_source=transform_source)
1176 namespace = self._parse_smil_namespace(smil)
1178 return self._parse_smil_formats(
1179 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
1181 def _extract_smil_info(self, smil_url, video_id, fatal=True, f4m_params=None):
1182 smil = self._download_smil(smil_url, video_id, fatal=fatal)
1185 return self._parse_smil(smil, smil_url, video_id, f4m_params=f4m_params)
1187 def _download_smil(self, smil_url, video_id, fatal=True, transform_source=None):
1188 return self._download_xml(
1189 smil_url, video_id, 'Downloading SMIL file',
1190 'Unable to download SMIL file', fatal=fatal, transform_source=transform_source)
1192 def _parse_smil(self, smil, smil_url, video_id, f4m_params=None):
1193 namespace = self._parse_smil_namespace(smil)
1195 formats = self._parse_smil_formats(
1196 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
1197 subtitles = self._parse_smil_subtitles(smil, namespace=namespace)
1199 video_id = os.path.splitext(url_basename(smil_url))[0]
1203 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
1204 name = meta.attrib.get('name')
1205 content = meta.attrib.get('content')
1206 if not name or not content:
1208 if not title and name == 'title':
1210 elif not description and name in ('description', 'abstract'):
1211 description = content
1212 elif not upload_date and name == 'date':
1213 upload_date = unified_strdate(content)
1216 'id': image.get('type'),
1217 'url': image.get('src'),
1218 'width': int_or_none(image.get('width')),
1219 'height': int_or_none(image.get('height')),
1220 } for image in smil.findall(self._xpath_ns('.//image', namespace)) if image.get('src')]
1224 'title': title or video_id,
1225 'description': description,
1226 'upload_date': upload_date,
1227 'thumbnails': thumbnails,
1229 'subtitles': subtitles,
1232 def _parse_smil_namespace(self, smil):
1233 return self._search_regex(
1234 r'(?i)^{([^}]+)?}smil$', smil.tag, 'namespace', default=None)
1236 def _parse_smil_formats(self, smil, smil_url, video_id, namespace=None, f4m_params=None, transform_rtmp_url=None):
1238 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
1239 b = meta.get('base') or meta.get('httpBase')
1250 videos = smil.findall(self._xpath_ns('.//video', namespace))
1251 for video in videos:
1252 src = video.get('src')
1253 if not src or src in srcs:
1257 bitrate = float_or_none(video.get('system-bitrate') or video.get('systemBitrate'), 1000)
1258 filesize = int_or_none(video.get('size') or video.get('fileSize'))
1259 width = int_or_none(video.get('width'))
1260 height = int_or_none(video.get('height'))
1261 proto = video.get('proto')
1262 ext = video.get('ext')
1263 src_ext = determine_ext(src)
1264 streamer = video.get('streamer') or base
1266 if proto == 'rtmp' or streamer.startswith('rtmp'):
1272 'format_id': 'rtmp-%d' % (rtmp_count if bitrate is None else bitrate),
1274 'filesize': filesize,
1278 if transform_rtmp_url:
1279 streamer, src = transform_rtmp_url(streamer, src)
1280 formats[-1].update({
1286 src_url = src if src.startswith('http') else compat_urlparse.urljoin(base, src)
1287 src_url = src_url.strip()
1289 if proto == 'm3u8' or src_ext == 'm3u8':
1290 m3u8_formats = self._extract_m3u8_formats(
1291 src_url, video_id, ext or 'mp4', m3u8_id='hls', fatal=False)
1292 if len(m3u8_formats) == 1:
1294 m3u8_formats[0].update({
1295 'format_id': 'hls-%d' % (m3u8_count if bitrate is None else bitrate),
1300 formats.extend(m3u8_formats)
1303 if src_ext == 'f4m':
1308 'plugin': 'flowplayer-3.2.0.1',
1310 f4m_url += '&' if '?' in f4m_url else '?'
1311 f4m_url += compat_urllib_parse_urlencode(f4m_params)
1312 formats.extend(self._extract_f4m_formats(f4m_url, video_id, f4m_id='hds', fatal=False))
1315 if src_url.startswith('http') and self._is_valid_url(src, video_id):
1319 'ext': ext or src_ext or 'flv',
1320 'format_id': 'http-%d' % (bitrate or http_count),
1322 'filesize': filesize,
1330 def _parse_smil_subtitles(self, smil, namespace=None, subtitles_lang='en'):
1333 for num, textstream in enumerate(smil.findall(self._xpath_ns('.//textstream', namespace))):
1334 src = textstream.get('src')
1335 if not src or src in urls:
1338 ext = textstream.get('ext') or determine_ext(src) or mimetype2ext(textstream.get('type'))
1339 lang = textstream.get('systemLanguage') or textstream.get('systemLanguageName') or textstream.get('lang') or subtitles_lang
1340 subtitles.setdefault(lang, []).append({
1346 def _extract_xspf_playlist(self, playlist_url, playlist_id, fatal=True):
1347 xspf = self._download_xml(
1348 playlist_url, playlist_id, 'Downloading xpsf playlist',
1349 'Unable to download xspf manifest', fatal=fatal)
1352 return self._parse_xspf(xspf, playlist_id)
1354 def _parse_xspf(self, playlist, playlist_id):
1356 'xspf': 'http://xspf.org/ns/0/',
1357 's1': 'http://static.streamone.nl/player/ns/0',
1361 for track in playlist.findall(xpath_with_ns('./xspf:trackList/xspf:track', NS_MAP)):
1363 track, xpath_with_ns('./xspf:title', NS_MAP), 'title', default=playlist_id)
1364 description = xpath_text(
1365 track, xpath_with_ns('./xspf:annotation', NS_MAP), 'description')
1366 thumbnail = xpath_text(
1367 track, xpath_with_ns('./xspf:image', NS_MAP), 'thumbnail')
1368 duration = float_or_none(
1369 xpath_text(track, xpath_with_ns('./xspf:duration', NS_MAP), 'duration'), 1000)
1372 'url': location.text,
1373 'format_id': location.get(xpath_with_ns('s1:label', NS_MAP)),
1374 'width': int_or_none(location.get(xpath_with_ns('s1:width', NS_MAP))),
1375 'height': int_or_none(location.get(xpath_with_ns('s1:height', NS_MAP))),
1376 } for location in track.findall(xpath_with_ns('./xspf:location', NS_MAP))]
1377 self._sort_formats(formats)
1382 'description': description,
1383 'thumbnail': thumbnail,
1384 'duration': duration,
1389 def _extract_mpd_formats(self, mpd_url, video_id, mpd_id=None, note=None, errnote=None, fatal=True, formats_dict={}):
1390 res = self._download_webpage_handle(
1392 note=note or 'Downloading MPD manifest',
1393 errnote=errnote or 'Failed to download MPD manifest',
1398 mpd_base_url = re.match(r'https?://.+/', urlh.geturl()).group()
1400 return self._parse_mpd_formats(
1401 compat_etree_fromstring(mpd.encode('utf-8')), mpd_id, mpd_base_url, formats_dict=formats_dict)
1403 def _parse_mpd_formats(self, mpd_doc, mpd_id=None, mpd_base_url='', formats_dict={}):
1404 if mpd_doc.get('type') == 'dynamic':
1407 namespace = self._search_regex(r'(?i)^{([^}]+)?}MPD$', mpd_doc.tag, 'namespace', default=None)
1410 return self._xpath_ns(path, namespace)
1412 def is_drm_protected(element):
1413 return element.find(_add_ns('ContentProtection')) is not None
1415 def extract_multisegment_info(element, ms_parent_info):
1416 ms_info = ms_parent_info.copy()
1417 segment_list = element.find(_add_ns('SegmentList'))
1418 if segment_list is not None:
1419 segment_urls_e = segment_list.findall(_add_ns('SegmentURL'))
1421 ms_info['segment_urls'] = [segment.attrib['media'] for segment in segment_urls_e]
1422 initialization = segment_list.find(_add_ns('Initialization'))
1423 if initialization is not None:
1424 ms_info['initialization_url'] = initialization.attrib['sourceURL']
1426 segment_template = element.find(_add_ns('SegmentTemplate'))
1427 if segment_template is not None:
1428 start_number = segment_template.get('startNumber')
1430 ms_info['start_number'] = int(start_number)
1431 segment_timeline = segment_template.find(_add_ns('SegmentTimeline'))
1432 if segment_timeline is not None:
1433 s_e = segment_timeline.findall(_add_ns('S'))
1435 ms_info['total_number'] = 0
1437 ms_info['total_number'] += 1 + int(s.get('r', '0'))
1439 timescale = segment_template.get('timescale')
1441 ms_info['timescale'] = int(timescale)
1442 segment_duration = segment_template.get('duration')
1443 if segment_duration:
1444 ms_info['segment_duration'] = int(segment_duration)
1445 media_template = segment_template.get('media')
1447 ms_info['media_template'] = media_template
1448 initialization = segment_template.get('initialization')
1450 ms_info['initialization_url'] = initialization
1452 initialization = segment_template.find(_add_ns('Initialization'))
1453 if initialization is not None:
1454 ms_info['initialization_url'] = initialization.attrib['sourceURL']
1457 mpd_duration = parse_duration(mpd_doc.get('mediaPresentationDuration'))
1459 for period in mpd_doc.findall(_add_ns('Period')):
1460 period_duration = parse_duration(period.get('duration')) or mpd_duration
1461 period_ms_info = extract_multisegment_info(period, {
1465 for adaptation_set in period.findall(_add_ns('AdaptationSet')):
1466 if is_drm_protected(adaptation_set):
1468 adaption_set_ms_info = extract_multisegment_info(adaptation_set, period_ms_info)
1469 for representation in adaptation_set.findall(_add_ns('Representation')):
1470 if is_drm_protected(representation):
1472 representation_attrib = adaptation_set.attrib.copy()
1473 representation_attrib.update(representation.attrib)
1474 # According to page 41 of ISO/IEC 29001-1:2014, @mimeType is mandatory
1475 mime_type = representation_attrib['mimeType']
1476 content_type = mime_type.split('/')[0]
1477 if content_type == 'text':
1478 # TODO implement WebVTT downloading
1480 elif content_type == 'video' or content_type == 'audio':
1482 for element in (representation, adaptation_set, period, mpd_doc):
1483 base_url_e = element.find(_add_ns('BaseURL'))
1484 if base_url_e is not None:
1485 base_url = base_url_e.text + base_url
1486 if re.match(r'^https?://', base_url):
1488 if mpd_base_url and not re.match(r'^https?://', base_url):
1489 if not mpd_base_url.endswith('/') and not base_url.startswith('/'):
1491 base_url = mpd_base_url + base_url
1492 representation_id = representation_attrib.get('id')
1493 lang = representation_attrib.get('lang')
1494 url_el = representation.find(_add_ns('BaseURL'))
1495 filesize = int_or_none(url_el.attrib.get('{http://youtube.com/yt/2012/10/10}contentLength') if url_el is not None else None)
1497 'format_id': '%s-%s' % (mpd_id, representation_id) if mpd_id else representation_id,
1499 'ext': mimetype2ext(mime_type),
1500 'width': int_or_none(representation_attrib.get('width')),
1501 'height': int_or_none(representation_attrib.get('height')),
1502 'tbr': int_or_none(representation_attrib.get('bandwidth'), 1000),
1503 'asr': int_or_none(representation_attrib.get('audioSamplingRate')),
1504 'fps': int_or_none(representation_attrib.get('frameRate')),
1505 'vcodec': 'none' if content_type == 'audio' else representation_attrib.get('codecs'),
1506 'acodec': 'none' if content_type == 'video' else representation_attrib.get('codecs'),
1507 'language': lang if lang not in ('mul', 'und', 'zxx', 'mis') else None,
1508 'format_note': 'DASH %s' % content_type,
1509 'filesize': filesize,
1511 representation_ms_info = extract_multisegment_info(representation, adaption_set_ms_info)
1512 if 'segment_urls' not in representation_ms_info and 'media_template' in representation_ms_info:
1513 if 'total_number' not in representation_ms_info and 'segment_duration':
1514 segment_duration = float(representation_ms_info['segment_duration']) / float(representation_ms_info['timescale'])
1515 representation_ms_info['total_number'] = int(math.ceil(float(period_duration) / segment_duration))
1516 media_template = representation_ms_info['media_template']
1517 media_template = media_template.replace('$RepresentationID$', representation_id)
1518 media_template = re.sub(r'\$(Number|Bandwidth)(?:%(0\d+)d)?\$', r'%(\1)\2d', media_template)
1519 media_template.replace('$$', '$')
1520 representation_ms_info['segment_urls'] = [media_template % {'Number': segment_number, 'Bandwidth': representation_attrib.get('bandwidth')} for segment_number in range(representation_ms_info['start_number'], representation_ms_info['total_number'] + representation_ms_info['start_number'])]
1521 if 'segment_urls' in representation_ms_info:
1523 'segment_urls': representation_ms_info['segment_urls'],
1524 'protocol': 'http_dash_segments',
1526 if 'initialization_url' in representation_ms_info:
1527 initialization_url = representation_ms_info['initialization_url'].replace('$RepresentationID$', representation_id)
1529 'initialization_url': initialization_url,
1531 if not f.get('url'):
1532 f['url'] = initialization_url
1534 existing_format = next(
1535 fo for fo in formats
1536 if fo['format_id'] == representation_id)
1537 except StopIteration:
1538 full_info = formats_dict.get(representation_id, {}).copy()
1540 formats.append(full_info)
1542 existing_format.update(f)
1544 self.report_warning('Unknown MIME type %s in DASH manifest' % mime_type)
1547 def _live_title(self, name):
1548 """ Generate the title for a live video """
1549 now = datetime.datetime.now()
1550 now_str = now.strftime('%Y-%m-%d %H:%M')
1551 return name + ' ' + now_str
1553 def _int(self, v, name, fatal=False, **kwargs):
1554 res = int_or_none(v, **kwargs)
1555 if 'get_attr' in kwargs:
1556 print(getattr(v, kwargs['get_attr']))
1558 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
1560 raise ExtractorError(msg)
1562 self._downloader.report_warning(msg)
1565 def _float(self, v, name, fatal=False, **kwargs):
1566 res = float_or_none(v, **kwargs)
1568 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
1570 raise ExtractorError(msg)
1572 self._downloader.report_warning(msg)
1575 def _set_cookie(self, domain, name, value, expire_time=None):
1576 cookie = compat_cookiejar.Cookie(
1577 0, name, value, None, None, domain, None,
1578 None, '/', True, False, expire_time, '', None, None, None)
1579 self._downloader.cookiejar.set_cookie(cookie)
1581 def _get_cookies(self, url):
1582 """ Return a compat_cookies.SimpleCookie with the cookies for the url """
1583 req = sanitized_Request(url)
1584 self._downloader.cookiejar.add_cookie_header(req)
1585 return compat_cookies.SimpleCookie(req.get_header('Cookie'))
1587 def get_testcases(self, include_onlymatching=False):
1588 t = getattr(self, '_TEST', None)
1590 assert not hasattr(self, '_TESTS'), \
1591 '%s has _TEST and _TESTS' % type(self).__name__
1594 tests = getattr(self, '_TESTS', [])
1596 if not include_onlymatching and t.get('only_matching', False):
1598 t['name'] = type(self).__name__[:-len('IE')]
1601 def is_suitable(self, age_limit):
1602 """ Test whether the extractor is generally suitable for the given
1603 age limit (i.e. pornographic sites are not, all others usually are) """
1605 any_restricted = False
1606 for tc in self.get_testcases(include_onlymatching=False):
1607 if 'playlist' in tc:
1608 tc = tc['playlist'][0]
1609 is_restricted = age_restricted(
1610 tc.get('info_dict', {}).get('age_limit'), age_limit)
1611 if not is_restricted:
1613 any_restricted = any_restricted or is_restricted
1614 return not any_restricted
1616 def extract_subtitles(self, *args, **kwargs):
1617 if (self._downloader.params.get('writesubtitles', False) or
1618 self._downloader.params.get('listsubtitles')):
1619 return self._get_subtitles(*args, **kwargs)
1622 def _get_subtitles(self, *args, **kwargs):
1623 raise NotImplementedError('This method must be implemented by subclasses')
1626 def _merge_subtitle_items(subtitle_list1, subtitle_list2):
1627 """ Merge subtitle items for one language. Items with duplicated URLs
1628 will be dropped. """
1629 list1_urls = set([item['url'] for item in subtitle_list1])
1630 ret = list(subtitle_list1)
1631 ret.extend([item for item in subtitle_list2 if item['url'] not in list1_urls])
1635 def _merge_subtitles(cls, subtitle_dict1, subtitle_dict2):
1636 """ Merge two subtitle dictionaries, language by language. """
1637 ret = dict(subtitle_dict1)
1638 for lang in subtitle_dict2:
1639 ret[lang] = cls._merge_subtitle_items(subtitle_dict1.get(lang, []), subtitle_dict2[lang])
1642 def extract_automatic_captions(self, *args, **kwargs):
1643 if (self._downloader.params.get('writeautomaticsub', False) or
1644 self._downloader.params.get('listsubtitles')):
1645 return self._get_automatic_captions(*args, **kwargs)
1648 def _get_automatic_captions(self, *args, **kwargs):
1649 raise NotImplementedError('This method must be implemented by subclasses')
1651 def mark_watched(self, *args, **kwargs):
1652 if (self._downloader.params.get('mark_watched', False) and
1653 (self._get_login_info()[0] is not None or
1654 self._downloader.params.get('cookiefile') is not None)):
1655 self._mark_watched(*args, **kwargs)
1657 def _mark_watched(self, *args, **kwargs):
1658 raise NotImplementedError('This method must be implemented by subclasses')
1661 class SearchInfoExtractor(InfoExtractor):
1663 Base class for paged search queries extractors.
1664 They accept URLs in the format _SEARCH_KEY(|all|[0-9]):{query}
1665 Instances should define _SEARCH_KEY and _MAX_RESULTS.
1669 def _make_valid_url(cls):
1670 return r'%s(?P<prefix>|[1-9][0-9]*|all):(?P<query>[\s\S]+)' % cls._SEARCH_KEY
1673 def suitable(cls, url):
1674 return re.match(cls._make_valid_url(), url) is not None
1676 def _real_extract(self, query):
1677 mobj = re.match(self._make_valid_url(), query)
1679 raise ExtractorError('Invalid search query "%s"' % query)
1681 prefix = mobj.group('prefix')
1682 query = mobj.group('query')
1684 return self._get_n_results(query, 1)
1685 elif prefix == 'all':
1686 return self._get_n_results(query, self._MAX_RESULTS)
1690 raise ExtractorError('invalid download number %s for query "%s"' % (n, query))
1691 elif n > self._MAX_RESULTS:
1692 self._downloader.report_warning('%s returns max %i results (you requested %i)' % (self._SEARCH_KEY, self._MAX_RESULTS, n))
1693 n = self._MAX_RESULTS
1694 return self._get_n_results(query, n)
1696 def _get_n_results(self, query, n):
1697 """Get a specified number of results for a query"""
1698 raise NotImplementedError('This method must be implemented by subclasses')
1701 def SEARCH_KEY(self):
1702 return self._SEARCH_KEY