1 from __future__ import unicode_literals
14 from ..compat import (
23 compat_etree_fromstring,
49 class InfoExtractor(object):
50 """Information Extractor class.
52 Information extractors are the classes that, given a URL, extract
53 information about the video (or videos) the URL refers to. This
54 information includes the real video URL, the video title, author and
55 others. The information is stored in a dictionary which is then
56 passed to the YoutubeDL. The YoutubeDL processes this
57 information possibly downloading the video to the file system, among
58 other possible outcomes.
60 The type field determines the type of the result.
61 By far the most common value (and the default if _type is missing) is
62 "video", which indicates a single video.
64 For a video, the dictionaries must include the following fields:
67 title: Video title, unescaped.
69 Additionally, it must contain either a formats entry or a url one:
71 formats: A list of dictionaries for each format available, ordered
72 from worst to best quality.
75 * url Mandatory. The URL of the video file
76 * ext Will be calculated from URL if missing
77 * format A human-readable description of the format
78 ("mp4 container with h264/opus").
79 Calculated from the format_id, width, height.
80 and format_note fields if missing.
81 * format_id A short description of the format
82 ("mp4_h264_opus" or "19").
83 Technically optional, but strongly recommended.
84 * format_note Additional info about the format
85 ("3D" or "DASH video")
86 * width Width of the video, if known
87 * height Height of the video, if known
88 * resolution Textual description of width and height
89 * tbr Average bitrate of audio and video in KBit/s
90 * abr Average audio bitrate in KBit/s
91 * acodec Name of the audio codec in use
92 * asr Audio sampling rate in Hertz
93 * vbr Average video bitrate in KBit/s
95 * vcodec Name of the video codec in use
96 * container Name of the container format
97 * filesize The number of bytes, if known in advance
98 * filesize_approx An estimate for the number of bytes
99 * player_url SWF Player URL (used for rtmpdump).
100 * protocol The protocol that will be used for the actual
101 download, lower-case.
102 "http", "https", "rtsp", "rtmp", "rtmpe",
103 "m3u8", or "m3u8_native".
104 * preference Order number of this format. If this field is
105 present and not None, the formats get sorted
106 by this field, regardless of all other values.
107 -1 for default (order by other properties),
108 -2 or smaller for less than default.
109 < -1000 to hide the format (if there is
110 another one which is strictly better)
111 * language_preference Is this in the correct requested
113 10 if it's what the URL is about,
114 -1 for default (don't know),
115 -10 otherwise, other values reserved for now.
116 * quality Order number of the video quality of this
117 format, irrespective of the file format.
118 -1 for default (order by other properties),
119 -2 or smaller for less than default.
120 * source_preference Order number for this video source
121 (quality takes higher priority)
122 -1 for default (order by other properties),
123 -2 or smaller for less than default.
124 * http_headers A dictionary of additional HTTP headers
125 to add to the request.
126 * stretched_ratio If given and not 1, indicates that the
127 video's pixels are not square.
128 width : height ratio as float.
129 * no_resume The server does not support resuming the
130 (HTTP or RTMP) download. Boolean.
132 url: Final video URL.
133 ext: Video filename extension.
134 format: The video format, defaults to ext (used for --get-format)
135 player_url: SWF Player URL (used for rtmpdump).
137 The following fields are optional:
139 alt_title: A secondary title of the video.
140 display_id An alternative identifier for the video, not necessarily
141 unique, but available before title. Typically, id is
142 something like "4234987", title "Dancing naked mole rats",
143 and display_id "dancing-naked-mole-rats"
144 thumbnails: A list of dictionaries, with the following entries:
145 * "id" (optional, string) - Thumbnail format ID
147 * "preference" (optional, int) - quality of the image
148 * "width" (optional, int)
149 * "height" (optional, int)
150 * "resolution" (optional, string "{width}x{height"},
152 thumbnail: Full URL to a video thumbnail image.
153 description: Full video description.
154 uploader: Full name of the video uploader.
155 creator: The main artist who created the video.
156 release_date: The date (YYYYMMDD) when the video was released.
157 timestamp: UNIX timestamp of the moment the video became available.
158 upload_date: Video upload date (YYYYMMDD).
159 If not explicitly set, calculated from timestamp.
160 uploader_id: Nickname or id of the video uploader.
161 location: Physical location where the video was filmed.
162 subtitles: The available subtitles as a dictionary in the format
163 {language: subformats}. "subformats" is a list sorted from
164 lower to higher preference, each element is a dictionary
165 with the "ext" entry and one of:
166 * "data": The subtitles file contents
167 * "url": A URL pointing to the subtitles file
168 "ext" will be calculated from URL if missing
169 automatic_captions: Like 'subtitles', used by the YoutubeIE for
170 automatically generated captions
171 duration: Length of the video in seconds, as an integer or float.
172 view_count: How many users have watched the video on the platform.
173 like_count: Number of positive ratings of the video
174 dislike_count: Number of negative ratings of the video
175 repost_count: Number of reposts of the video
176 average_rating: Average rating give by users, the scale used depends on the webpage
177 comment_count: Number of comments on the video
178 comments: A list of comments, each with one or more of the following
179 properties (all but one of text or html optional):
180 * "author" - human-readable name of the comment author
181 * "author_id" - user ID of the comment author
183 * "html" - Comment as HTML
184 * "text" - Plain text of the comment
185 * "timestamp" - UNIX timestamp of comment
186 * "parent" - ID of the comment this one is replying to.
187 Set to "root" to indicate that this is a
188 comment to the original video.
189 age_limit: Age restriction for the video, as an integer (years)
190 webpage_url: The URL to the video webpage, if given to youtube-dl it
191 should allow to get the same result again. (It will be set
192 by YoutubeDL if it's missing)
193 categories: A list of categories that the video falls in, for example
195 tags: A list of tags assigned to the video, e.g. ["sweden", "pop music"]
196 is_live: True, False, or None (=unknown). Whether this video is a
197 live stream that goes on instead of a fixed-length video.
198 start_time: Time in seconds where the reproduction should start, as
199 specified in the URL.
200 end_time: Time in seconds where the reproduction should end, as
201 specified in the URL.
203 The following fields should only be used when the video belongs to some logical
206 chapter: Name or title of the chapter the video belongs to.
207 chapter_number: Number of the chapter the video belongs to, as an integer.
208 chapter_id: Id of the chapter the video belongs to, as a unicode string.
210 The following fields should only be used when the video is an episode of some
213 series: Title of the series or programme the video episode belongs to.
214 season: Title of the season the video episode belongs to.
215 season_number: Number of the season the video episode belongs to, as an integer.
216 season_id: Id of the season the video episode belongs to, as a unicode string.
217 episode: Title of the video episode. Unlike mandatory video title field,
218 this field should denote the exact title of the video episode
219 without any kind of decoration.
220 episode_number: Number of the video episode within a season, as an integer.
221 episode_id: Id of the video episode, as a unicode string.
223 Unless mentioned otherwise, the fields should be Unicode strings.
225 Unless mentioned otherwise, None is equivalent to absence of information.
228 _type "playlist" indicates multiple videos.
229 There must be a key "entries", which is a list, an iterable, or a PagedList
230 object, each element of which is a valid dictionary by this specification.
232 Additionally, playlists can have "title", "description" and "id" attributes
233 with the same semantics as videos (see above).
236 _type "multi_video" indicates that there are multiple videos that
237 form a single show, for examples multiple acts of an opera or TV episode.
238 It must have an entries key like a playlist and contain all the keys
239 required for a video at the same time.
242 _type "url" indicates that the video must be extracted from another
243 location, possibly by a different extractor. Its only required key is:
244 "url" - the next URL to extract.
245 The key "ie_key" can be set to the class name (minus the trailing "IE",
246 e.g. "Youtube") if the extractor class is known in advance.
247 Additionally, the dictionary may have any properties of the resolved entity
248 known in advance, for example "title" if the title of the referred video is
252 _type "url_transparent" entities have the same specification as "url", but
253 indicate that the given additional information is more precise than the one
254 associated with the resolved URL.
255 This is useful when a site employs a video service that hosts the video and
256 its technical metadata, but that video service does not embed a useful
257 title, description etc.
260 Subclasses of this one should re-define the _real_initialize() and
261 _real_extract() methods and define a _VALID_URL regexp.
262 Probably, they should also be added to the list of extractors.
264 Finally, the _WORKING attribute should be set to False for broken IEs
265 in order to warn the users and skip the tests.
272 def __init__(self, downloader=None):
273 """Constructor. Receives an optional downloader."""
275 self.set_downloader(downloader)
278 def suitable(cls, url):
279 """Receives a URL and returns True if suitable for this IE."""
281 # This does not use has/getattr intentionally - we want to know whether
282 # we have cached the regexp for *this* class, whereas getattr would also
283 # match the superclass
284 if '_VALID_URL_RE' not in cls.__dict__:
285 cls._VALID_URL_RE = re.compile(cls._VALID_URL)
286 return cls._VALID_URL_RE.match(url) is not None
289 def _match_id(cls, url):
290 if '_VALID_URL_RE' not in cls.__dict__:
291 cls._VALID_URL_RE = re.compile(cls._VALID_URL)
292 m = cls._VALID_URL_RE.match(url)
298 """Getter method for _WORKING."""
301 def initialize(self):
302 """Initializes an instance (authentication, etc)."""
304 self._real_initialize()
307 def extract(self, url):
308 """Extracts URL information and returns it in list of dicts."""
311 return self._real_extract(url)
312 except ExtractorError:
314 except compat_http_client.IncompleteRead as e:
315 raise ExtractorError('A network error has occured.', cause=e, expected=True)
316 except (KeyError, StopIteration) as e:
317 raise ExtractorError('An extractor error has occured.', cause=e)
319 def set_downloader(self, downloader):
320 """Sets the downloader for this IE."""
321 self._downloader = downloader
323 def _real_initialize(self):
324 """Real initialization process. Redefine in subclasses."""
327 def _real_extract(self, url):
328 """Real extraction process. Redefine in subclasses."""
333 """A string for getting the InfoExtractor with get_info_extractor"""
334 return compat_str(cls.__name__[:-2])
338 return compat_str(type(self).__name__[:-2])
340 def _request_webpage(self, url_or_request, video_id, note=None, errnote=None, fatal=True):
341 """ Returns the response handle """
343 self.report_download_webpage(video_id)
344 elif note is not False:
346 self.to_screen('%s' % (note,))
348 self.to_screen('%s: %s' % (video_id, note))
350 return self._downloader.urlopen(url_or_request)
351 except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err:
355 errnote = 'Unable to download webpage'
357 errmsg = '%s: %s' % (errnote, error_to_compat_str(err))
359 raise ExtractorError(errmsg, sys.exc_info()[2], cause=err)
361 self._downloader.report_warning(errmsg)
364 def _download_webpage_handle(self, url_or_request, video_id, note=None, errnote=None, fatal=True, encoding=None):
365 """ Returns a tuple (page content as string, URL handle) """
366 # Strip hashes from the URL (#1038)
367 if isinstance(url_or_request, (compat_str, str)):
368 url_or_request = url_or_request.partition('#')[0]
370 urlh = self._request_webpage(url_or_request, video_id, note, errnote, fatal)
374 content = self._webpage_read_content(urlh, url_or_request, video_id, note, errnote, fatal, encoding=encoding)
375 return (content, urlh)
378 def _guess_encoding_from_content(content_type, webpage_bytes):
379 m = re.match(r'[a-zA-Z0-9_.-]+/[a-zA-Z0-9_.-]+\s*;\s*charset=(.+)', content_type)
381 encoding = m.group(1)
383 m = re.search(br'<meta[^>]+charset=[\'"]?([^\'")]+)[ /\'">]',
384 webpage_bytes[:1024])
386 encoding = m.group(1).decode('ascii')
387 elif webpage_bytes.startswith(b'\xff\xfe'):
394 def _webpage_read_content(self, urlh, url_or_request, video_id, note=None, errnote=None, fatal=True, prefix=None, encoding=None):
395 content_type = urlh.headers.get('Content-Type', '')
396 webpage_bytes = urlh.read()
397 if prefix is not None:
398 webpage_bytes = prefix + webpage_bytes
400 encoding = self._guess_encoding_from_content(content_type, webpage_bytes)
401 if self._downloader.params.get('dump_intermediate_pages', False):
403 url = url_or_request.get_full_url()
404 except AttributeError:
406 self.to_screen('Dumping request to ' + url)
407 dump = base64.b64encode(webpage_bytes).decode('ascii')
408 self._downloader.to_screen(dump)
409 if self._downloader.params.get('write_pages', False):
411 url = url_or_request.get_full_url()
412 except AttributeError:
414 basen = '%s_%s' % (video_id, url)
416 h = '___' + hashlib.md5(basen.encode('utf-8')).hexdigest()
417 basen = basen[:240 - len(h)] + h
418 raw_filename = basen + '.dump'
419 filename = sanitize_filename(raw_filename, restricted=True)
420 self.to_screen('Saving request to ' + filename)
421 # Working around MAX_PATH limitation on Windows (see
422 # http://msdn.microsoft.com/en-us/library/windows/desktop/aa365247(v=vs.85).aspx)
424 absfilepath = os.path.abspath(filename)
425 if len(absfilepath) > 259:
426 filename = '\\\\?\\' + absfilepath
427 with open(filename, 'wb') as outf:
428 outf.write(webpage_bytes)
431 content = webpage_bytes.decode(encoding, 'replace')
433 content = webpage_bytes.decode('utf-8', 'replace')
435 if ('<title>Access to this site is blocked</title>' in content and
436 'Websense' in content[:512]):
437 msg = 'Access to this webpage has been blocked by Websense filtering software in your network.'
438 blocked_iframe = self._html_search_regex(
439 r'<iframe src="([^"]+)"', content,
440 'Websense information URL', default=None)
442 msg += ' Visit %s for more details' % blocked_iframe
443 raise ExtractorError(msg, expected=True)
444 if '<title>The URL you requested has been blocked</title>' in content[:512]:
446 'Access to this webpage has been blocked by Indian censorship. '
447 'Use a VPN or proxy server (with --proxy) to route around it.')
448 block_msg = self._html_search_regex(
449 r'</h1><p>(.*?)</p>',
450 content, 'block message', default=None)
452 msg += ' (Message: "%s")' % block_msg.replace('\n', ' ')
453 raise ExtractorError(msg, expected=True)
457 def _download_webpage(self, url_or_request, video_id, note=None, errnote=None, fatal=True, tries=1, timeout=5, encoding=None):
458 """ Returns the data of the page as a string """
461 while success is False:
463 res = self._download_webpage_handle(url_or_request, video_id, note, errnote, fatal, encoding=encoding)
465 except compat_http_client.IncompleteRead as e:
467 if try_count >= tries:
469 self._sleep(timeout, video_id)
476 def _download_xml(self, url_or_request, video_id,
477 note='Downloading XML', errnote='Unable to download XML',
478 transform_source=None, fatal=True, encoding=None):
479 """Return the xml as an xml.etree.ElementTree.Element"""
480 xml_string = self._download_webpage(
481 url_or_request, video_id, note, errnote, fatal=fatal, encoding=encoding)
482 if xml_string is False:
485 xml_string = transform_source(xml_string)
486 return compat_etree_fromstring(xml_string.encode('utf-8'))
488 def _download_json(self, url_or_request, video_id,
489 note='Downloading JSON metadata',
490 errnote='Unable to download JSON metadata',
491 transform_source=None,
492 fatal=True, encoding=None):
493 json_string = self._download_webpage(
494 url_or_request, video_id, note, errnote, fatal=fatal,
496 if (not fatal) and json_string is False:
498 return self._parse_json(
499 json_string, video_id, transform_source=transform_source, fatal=fatal)
501 def _parse_json(self, json_string, video_id, transform_source=None, fatal=True):
503 json_string = transform_source(json_string)
505 return json.loads(json_string)
506 except ValueError as ve:
507 errmsg = '%s: Failed to parse JSON ' % video_id
509 raise ExtractorError(errmsg, cause=ve)
511 self.report_warning(errmsg + str(ve))
513 def report_warning(self, msg, video_id=None):
514 idstr = '' if video_id is None else '%s: ' % video_id
515 self._downloader.report_warning(
516 '[%s] %s%s' % (self.IE_NAME, idstr, msg))
518 def to_screen(self, msg):
519 """Print msg to screen, prefixing it with '[ie_name]'"""
520 self._downloader.to_screen('[%s] %s' % (self.IE_NAME, msg))
522 def report_extraction(self, id_or_name):
523 """Report information extraction."""
524 self.to_screen('%s: Extracting information' % id_or_name)
526 def report_download_webpage(self, video_id):
527 """Report webpage download."""
528 self.to_screen('%s: Downloading webpage' % video_id)
530 def report_age_confirmation(self):
531 """Report attempt to confirm age."""
532 self.to_screen('Confirming age')
534 def report_login(self):
535 """Report attempt to log in."""
536 self.to_screen('Logging in')
539 def raise_login_required(msg='This video is only available for registered users'):
540 raise ExtractorError(
541 '%s. Use --username and --password or --netrc to provide account credentials.' % msg,
545 def raise_geo_restricted(msg='This video is not available from your location due to geo restriction'):
546 raise ExtractorError(
547 '%s. You might want to use --proxy to workaround.' % msg,
550 # Methods for following #608
552 def url_result(url, ie=None, video_id=None, video_title=None):
553 """Returns a URL that points to a page that should be processed"""
554 # TODO: ie should be the class used for getting the info
555 video_info = {'_type': 'url',
558 if video_id is not None:
559 video_info['id'] = video_id
560 if video_title is not None:
561 video_info['title'] = video_title
565 def playlist_result(entries, playlist_id=None, playlist_title=None, playlist_description=None):
566 """Returns a playlist"""
567 video_info = {'_type': 'playlist',
570 video_info['id'] = playlist_id
572 video_info['title'] = playlist_title
573 if playlist_description:
574 video_info['description'] = playlist_description
577 def _search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
579 Perform a regex search on the given string, using a single or a list of
580 patterns returning the first matching group.
581 In case of failure return a default value or raise a WARNING or a
582 RegexNotFoundError, depending on fatal, specifying the field name.
584 if isinstance(pattern, (str, compat_str, compiled_regex_type)):
585 mobj = re.search(pattern, string, flags)
588 mobj = re.search(p, string, flags)
592 if not self._downloader.params.get('no_color') and os.name != 'nt' and sys.stderr.isatty():
593 _name = '\033[0;34m%s\033[0m' % name
599 # return the first matching group
600 return next(g for g in mobj.groups() if g is not None)
602 return mobj.group(group)
603 elif default is not NO_DEFAULT:
606 raise RegexNotFoundError('Unable to extract %s' % _name)
608 self._downloader.report_warning('unable to extract %s' % _name + bug_reports_message())
611 def _html_search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
613 Like _search_regex, but strips HTML tags and unescapes entities.
615 res = self._search_regex(pattern, string, name, default, fatal, flags, group)
617 return clean_html(res).strip()
621 def _get_login_info(self):
623 Get the login info as (username, password)
624 It will look in the netrc file using the _NETRC_MACHINE value
625 If there's no info available, return (None, None)
627 if self._downloader is None:
632 downloader_params = self._downloader.params
634 # Attempt to use provided username and password or .netrc data
635 if downloader_params.get('username', None) is not None:
636 username = downloader_params['username']
637 password = downloader_params['password']
638 elif downloader_params.get('usenetrc', False):
640 info = netrc.netrc().authenticators(self._NETRC_MACHINE)
645 raise netrc.NetrcParseError('No authenticators for %s' % self._NETRC_MACHINE)
646 except (IOError, netrc.NetrcParseError) as err:
647 self._downloader.report_warning('parsing .netrc: %s' % error_to_compat_str(err))
649 return (username, password)
651 def _get_tfa_info(self, note='two-factor verification code'):
653 Get the two-factor authentication info
654 TODO - asking the user will be required for sms/phone verify
655 currently just uses the command line option
656 If there's no info available, return None
658 if self._downloader is None:
660 downloader_params = self._downloader.params
662 if downloader_params.get('twofactor', None) is not None:
663 return downloader_params['twofactor']
665 return compat_getpass('Type %s and press [Return]: ' % note)
667 # Helper functions for extracting OpenGraph info
669 def _og_regexes(prop):
670 content_re = r'content=(?:"([^"]+?)"|\'([^\']+?)\'|\s*([^\s"\'=<>`]+?))'
671 property_re = (r'(?:name|property)=(?:\'og:%(prop)s\'|"og:%(prop)s"|\s*og:%(prop)s\b)'
672 % {'prop': re.escape(prop)})
673 template = r'<meta[^>]+?%s[^>]+?%s'
675 template % (property_re, content_re),
676 template % (content_re, property_re),
680 def _meta_regex(prop):
681 return r'''(?isx)<meta
682 (?=[^>]+(?:itemprop|name|property|id|http-equiv)=(["\']?)%s\1)
683 [^>]+?content=(["\'])(?P<content>.*?)\2''' % re.escape(prop)
685 def _og_search_property(self, prop, html, name=None, **kargs):
687 name = 'OpenGraph %s' % prop
688 escaped = self._search_regex(self._og_regexes(prop), html, name, flags=re.DOTALL, **kargs)
691 return unescapeHTML(escaped)
693 def _og_search_thumbnail(self, html, **kargs):
694 return self._og_search_property('image', html, 'thumbnail URL', fatal=False, **kargs)
696 def _og_search_description(self, html, **kargs):
697 return self._og_search_property('description', html, fatal=False, **kargs)
699 def _og_search_title(self, html, **kargs):
700 return self._og_search_property('title', html, **kargs)
702 def _og_search_video_url(self, html, name='video url', secure=True, **kargs):
703 regexes = self._og_regexes('video') + self._og_regexes('video:url')
705 regexes = self._og_regexes('video:secure_url') + regexes
706 return self._html_search_regex(regexes, html, name, **kargs)
708 def _og_search_url(self, html, **kargs):
709 return self._og_search_property('url', html, **kargs)
711 def _html_search_meta(self, name, html, display_name=None, fatal=False, **kwargs):
712 if display_name is None:
714 return self._html_search_regex(
715 self._meta_regex(name),
716 html, display_name, fatal=fatal, group='content', **kwargs)
718 def _dc_search_uploader(self, html):
719 return self._html_search_meta('dc.creator', html, 'uploader')
721 def _rta_search(self, html):
722 # See http://www.rtalabel.org/index.php?content=howtofaq#single
723 if re.search(r'(?ix)<meta\s+name="rating"\s+'
724 r' content="RTA-5042-1996-1400-1577-RTA"',
729 def _media_rating_search(self, html):
730 # See http://www.tjg-designs.com/WP/metadata-code-examples-adding-metadata-to-your-web-pages/
731 rating = self._html_search_meta('rating', html)
743 return RATING_TABLE.get(rating.lower(), None)
745 def _family_friendly_search(self, html):
746 # See http://schema.org/VideoObject
747 family_friendly = self._html_search_meta('isFamilyFriendly', html)
749 if not family_friendly:
758 return RATING_TABLE.get(family_friendly.lower(), None)
760 def _twitter_search_player(self, html):
761 return self._html_search_meta('twitter:player', html,
762 'twitter card player')
765 def _hidden_inputs(html):
766 html = re.sub(r'<!--(?:(?!<!--).)*-->', '', html)
768 for input in re.findall(r'(?i)<input([^>]+)>', html):
769 if not re.search(r'type=(["\'])(?:hidden|submit)\1', input):
771 name = re.search(r'name=(["\'])(?P<value>.+?)\1', input)
774 value = re.search(r'value=(["\'])(?P<value>.*?)\1', input)
777 hidden_inputs[name.group('value')] = value.group('value')
780 def _form_hidden_inputs(self, form_id, html):
781 form = self._search_regex(
782 r'(?is)<form[^>]+?id=(["\'])%s\1[^>]*>(?P<form>.+?)</form>' % form_id,
783 html, '%s form' % form_id, group='form')
784 return self._hidden_inputs(form)
786 def _sort_formats(self, formats, field_preference=None):
788 raise ExtractorError('No video formats found')
791 # TODO remove the following workaround
792 from ..utils import determine_ext
793 if not f.get('ext') and 'url' in f:
794 f['ext'] = determine_ext(f['url'])
796 if isinstance(field_preference, (list, tuple)):
797 return tuple(f.get(field) if f.get(field) is not None else -1 for field in field_preference)
799 preference = f.get('preference')
800 if preference is None:
802 if f.get('ext') in ['f4f', 'f4m']: # Not yet supported
805 proto_preference = 0 if determine_protocol(f) in ['http', 'https'] else -0.1
807 if f.get('vcodec') == 'none': # audio only
808 if self._downloader.params.get('prefer_free_formats'):
809 ORDER = ['aac', 'mp3', 'm4a', 'webm', 'ogg', 'opus']
811 ORDER = ['webm', 'opus', 'ogg', 'mp3', 'aac', 'm4a']
814 audio_ext_preference = ORDER.index(f['ext'])
816 audio_ext_preference = -1
818 if self._downloader.params.get('prefer_free_formats'):
819 ORDER = ['flv', 'mp4', 'webm']
821 ORDER = ['webm', 'flv', 'mp4']
823 ext_preference = ORDER.index(f['ext'])
826 audio_ext_preference = 0
830 f.get('language_preference') if f.get('language_preference') is not None else -1,
831 f.get('quality') if f.get('quality') is not None else -1,
832 f.get('tbr') if f.get('tbr') is not None else -1,
833 f.get('filesize') if f.get('filesize') is not None else -1,
834 f.get('vbr') if f.get('vbr') is not None else -1,
835 f.get('height') if f.get('height') is not None else -1,
836 f.get('width') if f.get('width') is not None else -1,
839 f.get('abr') if f.get('abr') is not None else -1,
840 audio_ext_preference,
841 f.get('fps') if f.get('fps') is not None else -1,
842 f.get('filesize_approx') if f.get('filesize_approx') is not None else -1,
843 f.get('source_preference') if f.get('source_preference') is not None else -1,
844 f.get('format_id') if f.get('format_id') is not None else '',
846 formats.sort(key=_formats_key)
848 def _check_formats(self, formats, video_id):
851 lambda f: self._is_valid_url(
853 item='%s video format' % f.get('format_id') if f.get('format_id') else 'video'),
856 def _is_valid_url(self, url, video_id, item='video'):
857 url = self._proto_relative_url(url, scheme='http:')
858 # For now assume non HTTP(S) URLs always valid
859 if not (url.startswith('http://') or url.startswith('https://')):
862 self._request_webpage(url, video_id, 'Checking %s URL' % item)
864 except ExtractorError as e:
865 if isinstance(e.cause, compat_urllib_error.URLError):
867 '%s: %s URL is invalid, skipping' % (video_id, item))
871 def http_scheme(self):
872 """ Either "http:" or "https:", depending on the user's preferences """
875 if self._downloader.params.get('prefer_insecure', False)
878 def _proto_relative_url(self, url, scheme=None):
881 if url.startswith('//'):
883 scheme = self.http_scheme()
888 def _sleep(self, timeout, video_id, msg_template=None):
889 if msg_template is None:
890 msg_template = '%(video_id)s: Waiting for %(timeout)s seconds'
891 msg = msg_template % {'video_id': video_id, 'timeout': timeout}
895 def _extract_f4m_formats(self, manifest_url, video_id, preference=None, f4m_id=None,
896 transform_source=lambda s: fix_xml_ampersands(s).strip(),
898 manifest = self._download_xml(
899 manifest_url, video_id, 'Downloading f4m manifest',
900 'Unable to download f4m manifest',
901 # Some manifests may be malformed, e.g. prosiebensat1 generated manifests
902 # (see https://github.com/rg3/youtube-dl/issues/6215#issuecomment-121704244)
903 transform_source=transform_source,
906 if manifest is False:
910 manifest_version = '1.0'
911 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/1.0}media')
913 manifest_version = '2.0'
914 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/2.0}media')
915 base_url = xpath_text(
916 manifest, ['{http://ns.adobe.com/f4m/1.0}baseURL', '{http://ns.adobe.com/f4m/2.0}baseURL'],
917 'base URL', default=None)
919 base_url = base_url.strip()
920 for i, media_el in enumerate(media_nodes):
921 if manifest_version == '2.0':
922 media_url = media_el.attrib.get('href') or media_el.attrib.get('url')
926 media_url if media_url.startswith('http://') or media_url.startswith('https://')
927 else ((base_url or '/'.join(manifest_url.split('/')[:-1])) + '/' + media_url))
928 # If media_url is itself a f4m manifest do the recursive extraction
929 # since bitrates in parent manifest (this one) and media_url manifest
930 # may differ leading to inability to resolve the format by requested
931 # bitrate in f4m downloader
932 if determine_ext(manifest_url) == 'f4m':
933 formats.extend(self._extract_f4m_formats(
934 manifest_url, video_id, preference, f4m_id, fatal=fatal))
936 tbr = int_or_none(media_el.attrib.get('bitrate'))
938 'format_id': '-'.join(filter(None, [f4m_id, compat_str(i if tbr is None else tbr)])),
942 'width': int_or_none(media_el.attrib.get('width')),
943 'height': int_or_none(media_el.attrib.get('height')),
944 'preference': preference,
946 self._sort_formats(formats)
950 def _extract_m3u8_formats(self, m3u8_url, video_id, ext=None,
951 entry_protocol='m3u8', preference=None,
952 m3u8_id=None, note=None, errnote=None,
956 'format_id': '-'.join(filter(None, [m3u8_id, 'meta'])),
960 'preference': preference - 1 if preference else -1,
961 'resolution': 'multiple',
962 'format_note': 'Quality selection URL',
965 format_url = lambda u: (
967 if re.match(r'^https?://', u)
968 else compat_urlparse.urljoin(m3u8_url, u))
970 res = self._download_webpage_handle(
972 note=note or 'Downloading m3u8 information',
973 errnote=errnote or 'Failed to download m3u8 information',
978 m3u8_url = urlh.geturl()
982 r'(?P<key>[a-zA-Z_-]+)=(?P<val>"[^"]+"|[^",]+)(?:,|$)')
983 for line in m3u8_doc.splitlines():
984 if line.startswith('#EXT-X-STREAM-INF:'):
986 for m in kv_rex.finditer(line):
988 if v.startswith('"'):
990 last_info[m.group('key')] = v
991 elif line.startswith('#EXT-X-MEDIA:'):
993 for m in kv_rex.finditer(line):
995 if v.startswith('"'):
997 last_media[m.group('key')] = v
998 elif line.startswith('#') or not line.strip():
1001 if last_info is None:
1002 formats.append({'url': format_url(line)})
1004 tbr = int_or_none(last_info.get('BANDWIDTH'), scale=1000)
1007 format_id.append(m3u8_id)
1008 last_media_name = last_media.get('NAME') if last_media and last_media.get('TYPE') != 'SUBTITLES' else None
1009 format_id.append(last_media_name if last_media_name else '%d' % (tbr if tbr else len(formats)))
1011 'format_id': '-'.join(format_id),
1012 'url': format_url(line.strip()),
1015 'protocol': entry_protocol,
1016 'preference': preference,
1018 codecs = last_info.get('CODECS')
1020 # TODO: looks like video codec is not always necessarily goes first
1021 va_codecs = codecs.split(',')
1023 f['vcodec'] = va_codecs[0].partition('.')[0]
1024 if len(va_codecs) > 1 and va_codecs[1]:
1025 f['acodec'] = va_codecs[1].partition('.')[0]
1026 resolution = last_info.get('RESOLUTION')
1028 width_str, height_str = resolution.split('x')
1029 f['width'] = int(width_str)
1030 f['height'] = int(height_str)
1031 if last_media is not None:
1032 f['m3u8_media'] = last_media
1036 self._sort_formats(formats)
1040 def _xpath_ns(path, namespace=None):
1044 for c in path.split('/'):
1045 if not c or c == '.':
1048 out.append('{%s}%s' % (namespace, c))
1049 return '/'.join(out)
1051 def _extract_smil_formats(self, smil_url, video_id, fatal=True, f4m_params=None):
1052 smil = self._download_smil(smil_url, video_id, fatal=fatal)
1058 namespace = self._parse_smil_namespace(smil)
1060 return self._parse_smil_formats(
1061 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
1063 def _extract_smil_info(self, smil_url, video_id, fatal=True, f4m_params=None):
1064 smil = self._download_smil(smil_url, video_id, fatal=fatal)
1067 return self._parse_smil(smil, smil_url, video_id, f4m_params=f4m_params)
1069 def _download_smil(self, smil_url, video_id, fatal=True):
1070 return self._download_xml(
1071 smil_url, video_id, 'Downloading SMIL file',
1072 'Unable to download SMIL file', fatal=fatal)
1074 def _parse_smil(self, smil, smil_url, video_id, f4m_params=None):
1075 namespace = self._parse_smil_namespace(smil)
1077 formats = self._parse_smil_formats(
1078 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
1079 subtitles = self._parse_smil_subtitles(smil, namespace=namespace)
1081 video_id = os.path.splitext(url_basename(smil_url))[0]
1085 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
1086 name = meta.attrib.get('name')
1087 content = meta.attrib.get('content')
1088 if not name or not content:
1090 if not title and name == 'title':
1092 elif not description and name in ('description', 'abstract'):
1093 description = content
1094 elif not upload_date and name == 'date':
1095 upload_date = unified_strdate(content)
1098 'id': image.get('type'),
1099 'url': image.get('src'),
1100 'width': int_or_none(image.get('width')),
1101 'height': int_or_none(image.get('height')),
1102 } for image in smil.findall(self._xpath_ns('.//image', namespace)) if image.get('src')]
1106 'title': title or video_id,
1107 'description': description,
1108 'upload_date': upload_date,
1109 'thumbnails': thumbnails,
1111 'subtitles': subtitles,
1114 def _parse_smil_namespace(self, smil):
1115 return self._search_regex(
1116 r'(?i)^{([^}]+)?}smil$', smil.tag, 'namespace', default=None)
1118 def _parse_smil_formats(self, smil, smil_url, video_id, namespace=None, f4m_params=None, transform_rtmp_url=None):
1120 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
1121 b = meta.get('base') or meta.get('httpBase')
1130 videos = smil.findall(self._xpath_ns('.//video', namespace))
1131 for video in videos:
1132 src = video.get('src')
1136 bitrate = float_or_none(video.get('system-bitrate') or video.get('systemBitrate'), 1000)
1137 filesize = int_or_none(video.get('size') or video.get('fileSize'))
1138 width = int_or_none(video.get('width'))
1139 height = int_or_none(video.get('height'))
1140 proto = video.get('proto')
1141 ext = video.get('ext')
1142 src_ext = determine_ext(src)
1143 streamer = video.get('streamer') or base
1145 if proto == 'rtmp' or streamer.startswith('rtmp'):
1151 'format_id': 'rtmp-%d' % (rtmp_count if bitrate is None else bitrate),
1153 'filesize': filesize,
1157 if transform_rtmp_url:
1158 streamer, src = transform_rtmp_url(streamer, src)
1159 formats[-1].update({
1165 src_url = src if src.startswith('http') else compat_urlparse.urljoin(base, src)
1167 if proto == 'm3u8' or src_ext == 'm3u8':
1168 formats.extend(self._extract_m3u8_formats(
1169 src_url, video_id, ext or 'mp4', m3u8_id='hls', fatal=False))
1172 if src_ext == 'f4m':
1177 'plugin': 'flowplayer-3.2.0.1',
1179 f4m_url += '&' if '?' in f4m_url else '?'
1180 f4m_url += compat_urllib_parse.urlencode(f4m_params)
1181 formats.extend(self._extract_f4m_formats(f4m_url, video_id, f4m_id='hds', fatal=False))
1184 if src_url.startswith('http') and self._is_valid_url(src, video_id):
1188 'ext': ext or src_ext or 'flv',
1189 'format_id': 'http-%d' % (bitrate or http_count),
1191 'filesize': filesize,
1197 self._sort_formats(formats)
1201 def _parse_smil_subtitles(self, smil, namespace=None, subtitles_lang='en'):
1203 for num, textstream in enumerate(smil.findall(self._xpath_ns('.//textstream', namespace))):
1204 src = textstream.get('src')
1207 ext = textstream.get('ext') or determine_ext(src)
1209 type_ = textstream.get('type')
1213 'application/smptett+xml': 'tt',
1215 if type_ in SUBTITLES_TYPES:
1216 ext = SUBTITLES_TYPES[type_]
1217 lang = textstream.get('systemLanguage') or textstream.get('systemLanguageName') or textstream.get('lang') or subtitles_lang
1218 subtitles.setdefault(lang, []).append({
1224 def _extract_xspf_playlist(self, playlist_url, playlist_id, fatal=True):
1225 xspf = self._download_xml(
1226 playlist_url, playlist_id, 'Downloading xpsf playlist',
1227 'Unable to download xspf manifest', fatal=fatal)
1230 return self._parse_xspf(xspf, playlist_id)
1232 def _parse_xspf(self, playlist, playlist_id):
1234 'xspf': 'http://xspf.org/ns/0/',
1235 's1': 'http://static.streamone.nl/player/ns/0',
1239 for track in playlist.findall(xpath_with_ns('./xspf:trackList/xspf:track', NS_MAP)):
1241 track, xpath_with_ns('./xspf:title', NS_MAP), 'title', default=playlist_id)
1242 description = xpath_text(
1243 track, xpath_with_ns('./xspf:annotation', NS_MAP), 'description')
1244 thumbnail = xpath_text(
1245 track, xpath_with_ns('./xspf:image', NS_MAP), 'thumbnail')
1246 duration = float_or_none(
1247 xpath_text(track, xpath_with_ns('./xspf:duration', NS_MAP), 'duration'), 1000)
1250 'url': location.text,
1251 'format_id': location.get(xpath_with_ns('s1:label', NS_MAP)),
1252 'width': int_or_none(location.get(xpath_with_ns('s1:width', NS_MAP))),
1253 'height': int_or_none(location.get(xpath_with_ns('s1:height', NS_MAP))),
1254 } for location in track.findall(xpath_with_ns('./xspf:location', NS_MAP))]
1255 self._sort_formats(formats)
1260 'description': description,
1261 'thumbnail': thumbnail,
1262 'duration': duration,
1267 def _live_title(self, name):
1268 """ Generate the title for a live video """
1269 now = datetime.datetime.now()
1270 now_str = now.strftime("%Y-%m-%d %H:%M")
1271 return name + ' ' + now_str
1273 def _int(self, v, name, fatal=False, **kwargs):
1274 res = int_or_none(v, **kwargs)
1275 if 'get_attr' in kwargs:
1276 print(getattr(v, kwargs['get_attr']))
1278 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
1280 raise ExtractorError(msg)
1282 self._downloader.report_warning(msg)
1285 def _float(self, v, name, fatal=False, **kwargs):
1286 res = float_or_none(v, **kwargs)
1288 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
1290 raise ExtractorError(msg)
1292 self._downloader.report_warning(msg)
1295 def _set_cookie(self, domain, name, value, expire_time=None):
1296 cookie = compat_cookiejar.Cookie(
1297 0, name, value, None, None, domain, None,
1298 None, '/', True, False, expire_time, '', None, None, None)
1299 self._downloader.cookiejar.set_cookie(cookie)
1301 def _get_cookies(self, url):
1302 """ Return a compat_cookies.SimpleCookie with the cookies for the url """
1303 req = sanitized_Request(url)
1304 self._downloader.cookiejar.add_cookie_header(req)
1305 return compat_cookies.SimpleCookie(req.get_header('Cookie'))
1307 def get_testcases(self, include_onlymatching=False):
1308 t = getattr(self, '_TEST', None)
1310 assert not hasattr(self, '_TESTS'), \
1311 '%s has _TEST and _TESTS' % type(self).__name__
1314 tests = getattr(self, '_TESTS', [])
1316 if not include_onlymatching and t.get('only_matching', False):
1318 t['name'] = type(self).__name__[:-len('IE')]
1321 def is_suitable(self, age_limit):
1322 """ Test whether the extractor is generally suitable for the given
1323 age limit (i.e. pornographic sites are not, all others usually are) """
1325 any_restricted = False
1326 for tc in self.get_testcases(include_onlymatching=False):
1327 if 'playlist' in tc:
1328 tc = tc['playlist'][0]
1329 is_restricted = age_restricted(
1330 tc.get('info_dict', {}).get('age_limit'), age_limit)
1331 if not is_restricted:
1333 any_restricted = any_restricted or is_restricted
1334 return not any_restricted
1336 def extract_subtitles(self, *args, **kwargs):
1337 if (self._downloader.params.get('writesubtitles', False) or
1338 self._downloader.params.get('listsubtitles')):
1339 return self._get_subtitles(*args, **kwargs)
1342 def _get_subtitles(self, *args, **kwargs):
1343 raise NotImplementedError("This method must be implemented by subclasses")
1346 def _merge_subtitle_items(subtitle_list1, subtitle_list2):
1347 """ Merge subtitle items for one language. Items with duplicated URLs
1348 will be dropped. """
1349 list1_urls = set([item['url'] for item in subtitle_list1])
1350 ret = list(subtitle_list1)
1351 ret.extend([item for item in subtitle_list2 if item['url'] not in list1_urls])
1355 def _merge_subtitles(cls, subtitle_dict1, subtitle_dict2):
1356 """ Merge two subtitle dictionaries, language by language. """
1357 ret = dict(subtitle_dict1)
1358 for lang in subtitle_dict2:
1359 ret[lang] = cls._merge_subtitle_items(subtitle_dict1.get(lang, []), subtitle_dict2[lang])
1362 def extract_automatic_captions(self, *args, **kwargs):
1363 if (self._downloader.params.get('writeautomaticsub', False) or
1364 self._downloader.params.get('listsubtitles')):
1365 return self._get_automatic_captions(*args, **kwargs)
1368 def _get_automatic_captions(self, *args, **kwargs):
1369 raise NotImplementedError("This method must be implemented by subclasses")
1372 class SearchInfoExtractor(InfoExtractor):
1374 Base class for paged search queries extractors.
1375 They accept URLs in the format _SEARCH_KEY(|all|[0-9]):{query}
1376 Instances should define _SEARCH_KEY and _MAX_RESULTS.
1380 def _make_valid_url(cls):
1381 return r'%s(?P<prefix>|[1-9][0-9]*|all):(?P<query>[\s\S]+)' % cls._SEARCH_KEY
1384 def suitable(cls, url):
1385 return re.match(cls._make_valid_url(), url) is not None
1387 def _real_extract(self, query):
1388 mobj = re.match(self._make_valid_url(), query)
1390 raise ExtractorError('Invalid search query "%s"' % query)
1392 prefix = mobj.group('prefix')
1393 query = mobj.group('query')
1395 return self._get_n_results(query, 1)
1396 elif prefix == 'all':
1397 return self._get_n_results(query, self._MAX_RESULTS)
1401 raise ExtractorError('invalid download number %s for query "%s"' % (n, query))
1402 elif n > self._MAX_RESULTS:
1403 self._downloader.report_warning('%s returns max %i results (you requested %i)' % (self._SEARCH_KEY, self._MAX_RESULTS, n))
1404 n = self._MAX_RESULTS
1405 return self._get_n_results(query, n)
1407 def _get_n_results(self, query, n):
1408 """Get a specified number of results for a query"""
1409 raise NotImplementedError("This method must be implemented by subclasses")
1412 def SEARCH_KEY(self):
1413 return self._SEARCH_KEY