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 Language code, e.g. "de" or "en-US".
112 * language_preference Is this in the language mentioned in
114 10 if it's what the URL is about,
115 -1 for default (don't know),
116 -10 otherwise, other values reserved for now.
117 * quality Order number of the video quality of this
118 format, irrespective of the file format.
119 -1 for default (order by other properties),
120 -2 or smaller for less than default.
121 * source_preference Order number for this video source
122 (quality takes higher priority)
123 -1 for default (order by other properties),
124 -2 or smaller for less than default.
125 * http_headers A dictionary of additional HTTP headers
126 to add to the request.
127 * stretched_ratio If given and not 1, indicates that the
128 video's pixels are not square.
129 width : height ratio as float.
130 * no_resume The server does not support resuming the
131 (HTTP or RTMP) download. Boolean.
133 url: Final video URL.
134 ext: Video filename extension.
135 format: The video format, defaults to ext (used for --get-format)
136 player_url: SWF Player URL (used for rtmpdump).
138 The following fields are optional:
140 alt_title: A secondary title of the video.
141 display_id An alternative identifier for the video, not necessarily
142 unique, but available before title. Typically, id is
143 something like "4234987", title "Dancing naked mole rats",
144 and display_id "dancing-naked-mole-rats"
145 thumbnails: A list of dictionaries, with the following entries:
146 * "id" (optional, string) - Thumbnail format ID
148 * "preference" (optional, int) - quality of the image
149 * "width" (optional, int)
150 * "height" (optional, int)
151 * "resolution" (optional, string "{width}x{height"},
153 thumbnail: Full URL to a video thumbnail image.
154 description: Full video description.
155 uploader: Full name of the video uploader.
156 creator: The main artist who created the video.
157 release_date: The date (YYYYMMDD) when the video was released.
158 timestamp: UNIX timestamp of the moment the video became available.
159 upload_date: Video upload date (YYYYMMDD).
160 If not explicitly set, calculated from timestamp.
161 uploader_id: Nickname or id of the video uploader.
162 location: Physical location where the video was filmed.
163 subtitles: The available subtitles as a dictionary in the format
164 {language: subformats}. "subformats" is a list sorted from
165 lower to higher preference, each element is a dictionary
166 with the "ext" entry and one of:
167 * "data": The subtitles file contents
168 * "url": A URL pointing to the subtitles file
169 "ext" will be calculated from URL if missing
170 automatic_captions: Like 'subtitles', used by the YoutubeIE for
171 automatically generated captions
172 duration: Length of the video in seconds, as an integer or float.
173 view_count: How many users have watched the video on the platform.
174 like_count: Number of positive ratings of the video
175 dislike_count: Number of negative ratings of the video
176 repost_count: Number of reposts of the video
177 average_rating: Average rating give by users, the scale used depends on the webpage
178 comment_count: Number of comments on the video
179 comments: A list of comments, each with one or more of the following
180 properties (all but one of text or html optional):
181 * "author" - human-readable name of the comment author
182 * "author_id" - user ID of the comment author
184 * "html" - Comment as HTML
185 * "text" - Plain text of the comment
186 * "timestamp" - UNIX timestamp of comment
187 * "parent" - ID of the comment this one is replying to.
188 Set to "root" to indicate that this is a
189 comment to the original video.
190 age_limit: Age restriction for the video, as an integer (years)
191 webpage_url: The URL to the video webpage, if given to youtube-dl it
192 should allow to get the same result again. (It will be set
193 by YoutubeDL if it's missing)
194 categories: A list of categories that the video falls in, for example
196 tags: A list of tags assigned to the video, e.g. ["sweden", "pop music"]
197 is_live: True, False, or None (=unknown). Whether this video is a
198 live stream that goes on instead of a fixed-length video.
199 start_time: Time in seconds where the reproduction should start, as
200 specified in the URL.
201 end_time: Time in seconds where the reproduction should end, as
202 specified in the URL.
204 The following fields should only be used when the video belongs to some logical
207 chapter: Name or title of the chapter the video belongs to.
208 chapter_number: Number of the chapter the video belongs to, as an integer.
209 chapter_id: Id of the chapter the video belongs to, as a unicode string.
211 The following fields should only be used when the video is an episode of some
214 series: Title of the series or programme the video episode belongs to.
215 season: Title of the season the video episode belongs to.
216 season_number: Number of the season the video episode belongs to, as an integer.
217 season_id: Id of the season the video episode belongs to, as a unicode string.
218 episode: Title of the video episode. Unlike mandatory video title field,
219 this field should denote the exact title of the video episode
220 without any kind of decoration.
221 episode_number: Number of the video episode within a season, as an integer.
222 episode_id: Id of the video episode, as a unicode string.
224 Unless mentioned otherwise, the fields should be Unicode strings.
226 Unless mentioned otherwise, None is equivalent to absence of information.
229 _type "playlist" indicates multiple videos.
230 There must be a key "entries", which is a list, an iterable, or a PagedList
231 object, each element of which is a valid dictionary by this specification.
233 Additionally, playlists can have "title", "description" and "id" attributes
234 with the same semantics as videos (see above).
237 _type "multi_video" indicates that there are multiple videos that
238 form a single show, for examples multiple acts of an opera or TV episode.
239 It must have an entries key like a playlist and contain all the keys
240 required for a video at the same time.
243 _type "url" indicates that the video must be extracted from another
244 location, possibly by a different extractor. Its only required key is:
245 "url" - the next URL to extract.
246 The key "ie_key" can be set to the class name (minus the trailing "IE",
247 e.g. "Youtube") if the extractor class is known in advance.
248 Additionally, the dictionary may have any properties of the resolved entity
249 known in advance, for example "title" if the title of the referred video is
253 _type "url_transparent" entities have the same specification as "url", but
254 indicate that the given additional information is more precise than the one
255 associated with the resolved URL.
256 This is useful when a site employs a video service that hosts the video and
257 its technical metadata, but that video service does not embed a useful
258 title, description etc.
261 Subclasses of this one should re-define the _real_initialize() and
262 _real_extract() methods and define a _VALID_URL regexp.
263 Probably, they should also be added to the list of extractors.
265 Finally, the _WORKING attribute should be set to False for broken IEs
266 in order to warn the users and skip the tests.
273 def __init__(self, downloader=None):
274 """Constructor. Receives an optional downloader."""
276 self.set_downloader(downloader)
279 def suitable(cls, url):
280 """Receives a URL and returns True if suitable for this IE."""
282 # This does not use has/getattr intentionally - we want to know whether
283 # we have cached the regexp for *this* class, whereas getattr would also
284 # match the superclass
285 if '_VALID_URL_RE' not in cls.__dict__:
286 cls._VALID_URL_RE = re.compile(cls._VALID_URL)
287 return cls._VALID_URL_RE.match(url) is not None
290 def _match_id(cls, url):
291 if '_VALID_URL_RE' not in cls.__dict__:
292 cls._VALID_URL_RE = re.compile(cls._VALID_URL)
293 m = cls._VALID_URL_RE.match(url)
299 """Getter method for _WORKING."""
302 def initialize(self):
303 """Initializes an instance (authentication, etc)."""
305 self._real_initialize()
308 def extract(self, url):
309 """Extracts URL information and returns it in list of dicts."""
312 return self._real_extract(url)
313 except ExtractorError:
315 except compat_http_client.IncompleteRead as e:
316 raise ExtractorError('A network error has occured.', cause=e, expected=True)
317 except (KeyError, StopIteration) as e:
318 raise ExtractorError('An extractor error has occured.', cause=e)
320 def set_downloader(self, downloader):
321 """Sets the downloader for this IE."""
322 self._downloader = downloader
324 def _real_initialize(self):
325 """Real initialization process. Redefine in subclasses."""
328 def _real_extract(self, url):
329 """Real extraction process. Redefine in subclasses."""
334 """A string for getting the InfoExtractor with get_info_extractor"""
335 return compat_str(cls.__name__[:-2])
339 return compat_str(type(self).__name__[:-2])
341 def _request_webpage(self, url_or_request, video_id, note=None, errnote=None, fatal=True):
342 """ Returns the response handle """
344 self.report_download_webpage(video_id)
345 elif note is not False:
347 self.to_screen('%s' % (note,))
349 self.to_screen('%s: %s' % (video_id, note))
351 return self._downloader.urlopen(url_or_request)
352 except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err:
356 errnote = 'Unable to download webpage'
358 errmsg = '%s: %s' % (errnote, error_to_compat_str(err))
360 raise ExtractorError(errmsg, sys.exc_info()[2], cause=err)
362 self._downloader.report_warning(errmsg)
365 def _download_webpage_handle(self, url_or_request, video_id, note=None, errnote=None, fatal=True, encoding=None):
366 """ Returns a tuple (page content as string, URL handle) """
367 # Strip hashes from the URL (#1038)
368 if isinstance(url_or_request, (compat_str, str)):
369 url_or_request = url_or_request.partition('#')[0]
371 urlh = self._request_webpage(url_or_request, video_id, note, errnote, fatal)
375 content = self._webpage_read_content(urlh, url_or_request, video_id, note, errnote, fatal, encoding=encoding)
376 return (content, urlh)
379 def _guess_encoding_from_content(content_type, webpage_bytes):
380 m = re.match(r'[a-zA-Z0-9_.-]+/[a-zA-Z0-9_.-]+\s*;\s*charset=(.+)', content_type)
382 encoding = m.group(1)
384 m = re.search(br'<meta[^>]+charset=[\'"]?([^\'")]+)[ /\'">]',
385 webpage_bytes[:1024])
387 encoding = m.group(1).decode('ascii')
388 elif webpage_bytes.startswith(b'\xff\xfe'):
395 def _webpage_read_content(self, urlh, url_or_request, video_id, note=None, errnote=None, fatal=True, prefix=None, encoding=None):
396 content_type = urlh.headers.get('Content-Type', '')
397 webpage_bytes = urlh.read()
398 if prefix is not None:
399 webpage_bytes = prefix + webpage_bytes
401 encoding = self._guess_encoding_from_content(content_type, webpage_bytes)
402 if self._downloader.params.get('dump_intermediate_pages', False):
404 url = url_or_request.get_full_url()
405 except AttributeError:
407 self.to_screen('Dumping request to ' + url)
408 dump = base64.b64encode(webpage_bytes).decode('ascii')
409 self._downloader.to_screen(dump)
410 if self._downloader.params.get('write_pages', False):
412 url = url_or_request.get_full_url()
413 except AttributeError:
415 basen = '%s_%s' % (video_id, url)
417 h = '___' + hashlib.md5(basen.encode('utf-8')).hexdigest()
418 basen = basen[:240 - len(h)] + h
419 raw_filename = basen + '.dump'
420 filename = sanitize_filename(raw_filename, restricted=True)
421 self.to_screen('Saving request to ' + filename)
422 # Working around MAX_PATH limitation on Windows (see
423 # http://msdn.microsoft.com/en-us/library/windows/desktop/aa365247(v=vs.85).aspx)
425 absfilepath = os.path.abspath(filename)
426 if len(absfilepath) > 259:
427 filename = '\\\\?\\' + absfilepath
428 with open(filename, 'wb') as outf:
429 outf.write(webpage_bytes)
432 content = webpage_bytes.decode(encoding, 'replace')
434 content = webpage_bytes.decode('utf-8', 'replace')
436 if ('<title>Access to this site is blocked</title>' in content and
437 'Websense' in content[:512]):
438 msg = 'Access to this webpage has been blocked by Websense filtering software in your network.'
439 blocked_iframe = self._html_search_regex(
440 r'<iframe src="([^"]+)"', content,
441 'Websense information URL', default=None)
443 msg += ' Visit %s for more details' % blocked_iframe
444 raise ExtractorError(msg, expected=True)
445 if '<title>The URL you requested has been blocked</title>' in content[:512]:
447 'Access to this webpage has been blocked by Indian censorship. '
448 'Use a VPN or proxy server (with --proxy) to route around it.')
449 block_msg = self._html_search_regex(
450 r'</h1><p>(.*?)</p>',
451 content, 'block message', default=None)
453 msg += ' (Message: "%s")' % block_msg.replace('\n', ' ')
454 raise ExtractorError(msg, expected=True)
458 def _download_webpage(self, url_or_request, video_id, note=None, errnote=None, fatal=True, tries=1, timeout=5, encoding=None):
459 """ Returns the data of the page as a string """
462 while success is False:
464 res = self._download_webpage_handle(url_or_request, video_id, note, errnote, fatal, encoding=encoding)
466 except compat_http_client.IncompleteRead as e:
468 if try_count >= tries:
470 self._sleep(timeout, video_id)
477 def _download_xml(self, url_or_request, video_id,
478 note='Downloading XML', errnote='Unable to download XML',
479 transform_source=None, fatal=True, encoding=None):
480 """Return the xml as an xml.etree.ElementTree.Element"""
481 xml_string = self._download_webpage(
482 url_or_request, video_id, note, errnote, fatal=fatal, encoding=encoding)
483 if xml_string is False:
486 xml_string = transform_source(xml_string)
487 return compat_etree_fromstring(xml_string.encode('utf-8'))
489 def _download_json(self, url_or_request, video_id,
490 note='Downloading JSON metadata',
491 errnote='Unable to download JSON metadata',
492 transform_source=None,
493 fatal=True, encoding=None):
494 json_string = self._download_webpage(
495 url_or_request, video_id, note, errnote, fatal=fatal,
497 if (not fatal) and json_string is False:
499 return self._parse_json(
500 json_string, video_id, transform_source=transform_source, fatal=fatal)
502 def _parse_json(self, json_string, video_id, transform_source=None, fatal=True):
504 json_string = transform_source(json_string)
506 return json.loads(json_string)
507 except ValueError as ve:
508 errmsg = '%s: Failed to parse JSON ' % video_id
510 raise ExtractorError(errmsg, cause=ve)
512 self.report_warning(errmsg + str(ve))
514 def report_warning(self, msg, video_id=None):
515 idstr = '' if video_id is None else '%s: ' % video_id
516 self._downloader.report_warning(
517 '[%s] %s%s' % (self.IE_NAME, idstr, msg))
519 def to_screen(self, msg):
520 """Print msg to screen, prefixing it with '[ie_name]'"""
521 self._downloader.to_screen('[%s] %s' % (self.IE_NAME, msg))
523 def report_extraction(self, id_or_name):
524 """Report information extraction."""
525 self.to_screen('%s: Extracting information' % id_or_name)
527 def report_download_webpage(self, video_id):
528 """Report webpage download."""
529 self.to_screen('%s: Downloading webpage' % video_id)
531 def report_age_confirmation(self):
532 """Report attempt to confirm age."""
533 self.to_screen('Confirming age')
535 def report_login(self):
536 """Report attempt to log in."""
537 self.to_screen('Logging in')
540 def raise_login_required(msg='This video is only available for registered users'):
541 raise ExtractorError(
542 '%s. Use --username and --password or --netrc to provide account credentials.' % msg,
546 def raise_geo_restricted(msg='This video is not available from your location due to geo restriction'):
547 raise ExtractorError(
548 '%s. You might want to use --proxy to workaround.' % msg,
551 # Methods for following #608
553 def url_result(url, ie=None, video_id=None, video_title=None):
554 """Returns a URL that points to a page that should be processed"""
555 # TODO: ie should be the class used for getting the info
556 video_info = {'_type': 'url',
559 if video_id is not None:
560 video_info['id'] = video_id
561 if video_title is not None:
562 video_info['title'] = video_title
566 def playlist_result(entries, playlist_id=None, playlist_title=None, playlist_description=None):
567 """Returns a playlist"""
568 video_info = {'_type': 'playlist',
571 video_info['id'] = playlist_id
573 video_info['title'] = playlist_title
574 if playlist_description:
575 video_info['description'] = playlist_description
578 def _search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
580 Perform a regex search on the given string, using a single or a list of
581 patterns returning the first matching group.
582 In case of failure return a default value or raise a WARNING or a
583 RegexNotFoundError, depending on fatal, specifying the field name.
585 if isinstance(pattern, (str, compat_str, compiled_regex_type)):
586 mobj = re.search(pattern, string, flags)
589 mobj = re.search(p, string, flags)
593 if not self._downloader.params.get('no_color') and os.name != 'nt' and sys.stderr.isatty():
594 _name = '\033[0;34m%s\033[0m' % name
600 # return the first matching group
601 return next(g for g in mobj.groups() if g is not None)
603 return mobj.group(group)
604 elif default is not NO_DEFAULT:
607 raise RegexNotFoundError('Unable to extract %s' % _name)
609 self._downloader.report_warning('unable to extract %s' % _name + bug_reports_message())
612 def _html_search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
614 Like _search_regex, but strips HTML tags and unescapes entities.
616 res = self._search_regex(pattern, string, name, default, fatal, flags, group)
618 return clean_html(res).strip()
622 def _get_login_info(self):
624 Get the login info as (username, password)
625 It will look in the netrc file using the _NETRC_MACHINE value
626 If there's no info available, return (None, None)
628 if self._downloader is None:
633 downloader_params = self._downloader.params
635 # Attempt to use provided username and password or .netrc data
636 if downloader_params.get('username', None) is not None:
637 username = downloader_params['username']
638 password = downloader_params['password']
639 elif downloader_params.get('usenetrc', False):
641 info = netrc.netrc().authenticators(self._NETRC_MACHINE)
646 raise netrc.NetrcParseError('No authenticators for %s' % self._NETRC_MACHINE)
647 except (IOError, netrc.NetrcParseError) as err:
648 self._downloader.report_warning('parsing .netrc: %s' % error_to_compat_str(err))
650 return (username, password)
652 def _get_tfa_info(self, note='two-factor verification code'):
654 Get the two-factor authentication info
655 TODO - asking the user will be required for sms/phone verify
656 currently just uses the command line option
657 If there's no info available, return None
659 if self._downloader is None:
661 downloader_params = self._downloader.params
663 if downloader_params.get('twofactor', None) is not None:
664 return downloader_params['twofactor']
666 return compat_getpass('Type %s and press [Return]: ' % note)
668 # Helper functions for extracting OpenGraph info
670 def _og_regexes(prop):
671 content_re = r'content=(?:"([^"]+?)"|\'([^\']+?)\'|\s*([^\s"\'=<>`]+?))'
672 property_re = (r'(?:name|property)=(?:\'og:%(prop)s\'|"og:%(prop)s"|\s*og:%(prop)s\b)'
673 % {'prop': re.escape(prop)})
674 template = r'<meta[^>]+?%s[^>]+?%s'
676 template % (property_re, content_re),
677 template % (content_re, property_re),
681 def _meta_regex(prop):
682 return r'''(?isx)<meta
683 (?=[^>]+(?:itemprop|name|property|id|http-equiv)=(["\']?)%s\1)
684 [^>]+?content=(["\'])(?P<content>.*?)\2''' % re.escape(prop)
686 def _og_search_property(self, prop, html, name=None, **kargs):
688 name = 'OpenGraph %s' % prop
689 escaped = self._search_regex(self._og_regexes(prop), html, name, flags=re.DOTALL, **kargs)
692 return unescapeHTML(escaped)
694 def _og_search_thumbnail(self, html, **kargs):
695 return self._og_search_property('image', html, 'thumbnail URL', fatal=False, **kargs)
697 def _og_search_description(self, html, **kargs):
698 return self._og_search_property('description', html, fatal=False, **kargs)
700 def _og_search_title(self, html, **kargs):
701 return self._og_search_property('title', html, **kargs)
703 def _og_search_video_url(self, html, name='video url', secure=True, **kargs):
704 regexes = self._og_regexes('video') + self._og_regexes('video:url')
706 regexes = self._og_regexes('video:secure_url') + regexes
707 return self._html_search_regex(regexes, html, name, **kargs)
709 def _og_search_url(self, html, **kargs):
710 return self._og_search_property('url', html, **kargs)
712 def _html_search_meta(self, name, html, display_name=None, fatal=False, **kwargs):
713 if display_name is None:
715 return self._html_search_regex(
716 self._meta_regex(name),
717 html, display_name, fatal=fatal, group='content', **kwargs)
719 def _dc_search_uploader(self, html):
720 return self._html_search_meta('dc.creator', html, 'uploader')
722 def _rta_search(self, html):
723 # See http://www.rtalabel.org/index.php?content=howtofaq#single
724 if re.search(r'(?ix)<meta\s+name="rating"\s+'
725 r' content="RTA-5042-1996-1400-1577-RTA"',
730 def _media_rating_search(self, html):
731 # See http://www.tjg-designs.com/WP/metadata-code-examples-adding-metadata-to-your-web-pages/
732 rating = self._html_search_meta('rating', html)
744 return RATING_TABLE.get(rating.lower(), None)
746 def _family_friendly_search(self, html):
747 # See http://schema.org/VideoObject
748 family_friendly = self._html_search_meta('isFamilyFriendly', html)
750 if not family_friendly:
759 return RATING_TABLE.get(family_friendly.lower(), None)
761 def _twitter_search_player(self, html):
762 return self._html_search_meta('twitter:player', html,
763 'twitter card player')
766 def _hidden_inputs(html):
767 html = re.sub(r'<!--(?:(?!<!--).)*-->', '', html)
769 for input in re.findall(r'(?i)<input([^>]+)>', html):
770 if not re.search(r'type=(["\'])(?:hidden|submit)\1', input):
772 name = re.search(r'name=(["\'])(?P<value>.+?)\1', input)
775 value = re.search(r'value=(["\'])(?P<value>.*?)\1', input)
778 hidden_inputs[name.group('value')] = value.group('value')
781 def _form_hidden_inputs(self, form_id, html):
782 form = self._search_regex(
783 r'(?is)<form[^>]+?id=(["\'])%s\1[^>]*>(?P<form>.+?)</form>' % form_id,
784 html, '%s form' % form_id, group='form')
785 return self._hidden_inputs(form)
787 def _sort_formats(self, formats, field_preference=None):
789 raise ExtractorError('No video formats found')
792 # TODO remove the following workaround
793 from ..utils import determine_ext
794 if not f.get('ext') and 'url' in f:
795 f['ext'] = determine_ext(f['url'])
797 if isinstance(field_preference, (list, tuple)):
798 return tuple(f.get(field) if f.get(field) is not None else -1 for field in field_preference)
800 preference = f.get('preference')
801 if preference is None:
803 if f.get('ext') in ['f4f', 'f4m']: # Not yet supported
806 proto_preference = 0 if determine_protocol(f) in ['http', 'https'] else -0.1
808 if f.get('vcodec') == 'none': # audio only
809 if self._downloader.params.get('prefer_free_formats'):
810 ORDER = ['aac', 'mp3', 'm4a', 'webm', 'ogg', 'opus']
812 ORDER = ['webm', 'opus', 'ogg', 'mp3', 'aac', 'm4a']
815 audio_ext_preference = ORDER.index(f['ext'])
817 audio_ext_preference = -1
819 if self._downloader.params.get('prefer_free_formats'):
820 ORDER = ['flv', 'mp4', 'webm']
822 ORDER = ['webm', 'flv', 'mp4']
824 ext_preference = ORDER.index(f['ext'])
827 audio_ext_preference = 0
831 f.get('language_preference') if f.get('language_preference') is not None else -1,
832 f.get('quality') if f.get('quality') is not None else -1,
833 f.get('tbr') if f.get('tbr') is not None else -1,
834 f.get('filesize') if f.get('filesize') is not None else -1,
835 f.get('vbr') if f.get('vbr') is not None else -1,
836 f.get('height') if f.get('height') is not None else -1,
837 f.get('width') if f.get('width') is not None else -1,
840 f.get('abr') if f.get('abr') is not None else -1,
841 audio_ext_preference,
842 f.get('fps') if f.get('fps') is not None else -1,
843 f.get('filesize_approx') if f.get('filesize_approx') is not None else -1,
844 f.get('source_preference') if f.get('source_preference') is not None else -1,
845 f.get('format_id') if f.get('format_id') is not None else '',
847 formats.sort(key=_formats_key)
849 def _check_formats(self, formats, video_id):
852 lambda f: self._is_valid_url(
854 item='%s video format' % f.get('format_id') if f.get('format_id') else 'video'),
857 def _is_valid_url(self, url, video_id, item='video'):
858 url = self._proto_relative_url(url, scheme='http:')
859 # For now assume non HTTP(S) URLs always valid
860 if not (url.startswith('http://') or url.startswith('https://')):
863 self._request_webpage(url, video_id, 'Checking %s URL' % item)
865 except ExtractorError as e:
866 if isinstance(e.cause, compat_urllib_error.URLError):
868 '%s: %s URL is invalid, skipping' % (video_id, item))
872 def http_scheme(self):
873 """ Either "http:" or "https:", depending on the user's preferences """
876 if self._downloader.params.get('prefer_insecure', False)
879 def _proto_relative_url(self, url, scheme=None):
882 if url.startswith('//'):
884 scheme = self.http_scheme()
889 def _sleep(self, timeout, video_id, msg_template=None):
890 if msg_template is None:
891 msg_template = '%(video_id)s: Waiting for %(timeout)s seconds'
892 msg = msg_template % {'video_id': video_id, 'timeout': timeout}
896 def _extract_f4m_formats(self, manifest_url, video_id, preference=None, f4m_id=None,
897 transform_source=lambda s: fix_xml_ampersands(s).strip(),
899 manifest = self._download_xml(
900 manifest_url, video_id, 'Downloading f4m manifest',
901 'Unable to download f4m manifest',
902 # Some manifests may be malformed, e.g. prosiebensat1 generated manifests
903 # (see https://github.com/rg3/youtube-dl/issues/6215#issuecomment-121704244)
904 transform_source=transform_source,
907 if manifest is False:
911 manifest_version = '1.0'
912 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/1.0}media')
914 manifest_version = '2.0'
915 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/2.0}media')
916 base_url = xpath_text(
917 manifest, ['{http://ns.adobe.com/f4m/1.0}baseURL', '{http://ns.adobe.com/f4m/2.0}baseURL'],
918 'base URL', default=None)
920 base_url = base_url.strip()
921 for i, media_el in enumerate(media_nodes):
922 if manifest_version == '2.0':
923 media_url = media_el.attrib.get('href') or media_el.attrib.get('url')
927 media_url if media_url.startswith('http://') or media_url.startswith('https://')
928 else ((base_url or '/'.join(manifest_url.split('/')[:-1])) + '/' + media_url))
929 # If media_url is itself a f4m manifest do the recursive extraction
930 # since bitrates in parent manifest (this one) and media_url manifest
931 # may differ leading to inability to resolve the format by requested
932 # bitrate in f4m downloader
933 if determine_ext(manifest_url) == 'f4m':
934 formats.extend(self._extract_f4m_formats(
935 manifest_url, video_id, preference, f4m_id, fatal=fatal))
937 tbr = int_or_none(media_el.attrib.get('bitrate'))
939 'format_id': '-'.join(filter(None, [f4m_id, compat_str(i if tbr is None else tbr)])),
943 'width': int_or_none(media_el.attrib.get('width')),
944 'height': int_or_none(media_el.attrib.get('height')),
945 'preference': preference,
947 self._sort_formats(formats)
951 def _extract_m3u8_formats(self, m3u8_url, video_id, ext=None,
952 entry_protocol='m3u8', preference=None,
953 m3u8_id=None, note=None, errnote=None,
957 'format_id': '-'.join(filter(None, [m3u8_id, 'meta'])),
961 'preference': preference - 1 if preference else -1,
962 'resolution': 'multiple',
963 'format_note': 'Quality selection URL',
966 format_url = lambda u: (
968 if re.match(r'^https?://', u)
969 else compat_urlparse.urljoin(m3u8_url, u))
971 res = self._download_webpage_handle(
973 note=note or 'Downloading m3u8 information',
974 errnote=errnote or 'Failed to download m3u8 information',
979 m3u8_url = urlh.geturl()
983 r'(?P<key>[a-zA-Z_-]+)=(?P<val>"[^"]+"|[^",]+)(?:,|$)')
984 for line in m3u8_doc.splitlines():
985 if line.startswith('#EXT-X-STREAM-INF:'):
987 for m in kv_rex.finditer(line):
989 if v.startswith('"'):
991 last_info[m.group('key')] = v
992 elif line.startswith('#EXT-X-MEDIA:'):
994 for m in kv_rex.finditer(line):
996 if v.startswith('"'):
998 last_media[m.group('key')] = v
999 elif line.startswith('#') or not line.strip():
1002 if last_info is None:
1003 formats.append({'url': format_url(line)})
1005 tbr = int_or_none(last_info.get('BANDWIDTH'), scale=1000)
1008 format_id.append(m3u8_id)
1009 last_media_name = last_media.get('NAME') if last_media and last_media.get('TYPE') != 'SUBTITLES' else None
1010 format_id.append(last_media_name if last_media_name else '%d' % (tbr if tbr else len(formats)))
1012 'format_id': '-'.join(format_id),
1013 'url': format_url(line.strip()),
1016 'protocol': entry_protocol,
1017 'preference': preference,
1019 codecs = last_info.get('CODECS')
1021 # TODO: looks like video codec is not always necessarily goes first
1022 va_codecs = codecs.split(',')
1024 f['vcodec'] = va_codecs[0].partition('.')[0]
1025 if len(va_codecs) > 1 and va_codecs[1]:
1026 f['acodec'] = va_codecs[1].partition('.')[0]
1027 resolution = last_info.get('RESOLUTION')
1029 width_str, height_str = resolution.split('x')
1030 f['width'] = int(width_str)
1031 f['height'] = int(height_str)
1032 if last_media is not None:
1033 f['m3u8_media'] = last_media
1037 self._sort_formats(formats)
1041 def _xpath_ns(path, namespace=None):
1045 for c in path.split('/'):
1046 if not c or c == '.':
1049 out.append('{%s}%s' % (namespace, c))
1050 return '/'.join(out)
1052 def _extract_smil_formats(self, smil_url, video_id, fatal=True, f4m_params=None):
1053 smil = self._download_smil(smil_url, video_id, fatal=fatal)
1059 namespace = self._parse_smil_namespace(smil)
1061 return self._parse_smil_formats(
1062 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
1064 def _extract_smil_info(self, smil_url, video_id, fatal=True, f4m_params=None):
1065 smil = self._download_smil(smil_url, video_id, fatal=fatal)
1068 return self._parse_smil(smil, smil_url, video_id, f4m_params=f4m_params)
1070 def _download_smil(self, smil_url, video_id, fatal=True):
1071 return self._download_xml(
1072 smil_url, video_id, 'Downloading SMIL file',
1073 'Unable to download SMIL file', fatal=fatal)
1075 def _parse_smil(self, smil, smil_url, video_id, f4m_params=None):
1076 namespace = self._parse_smil_namespace(smil)
1078 formats = self._parse_smil_formats(
1079 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
1080 subtitles = self._parse_smil_subtitles(smil, namespace=namespace)
1082 video_id = os.path.splitext(url_basename(smil_url))[0]
1086 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
1087 name = meta.attrib.get('name')
1088 content = meta.attrib.get('content')
1089 if not name or not content:
1091 if not title and name == 'title':
1093 elif not description and name in ('description', 'abstract'):
1094 description = content
1095 elif not upload_date and name == 'date':
1096 upload_date = unified_strdate(content)
1099 'id': image.get('type'),
1100 'url': image.get('src'),
1101 'width': int_or_none(image.get('width')),
1102 'height': int_or_none(image.get('height')),
1103 } for image in smil.findall(self._xpath_ns('.//image', namespace)) if image.get('src')]
1107 'title': title or video_id,
1108 'description': description,
1109 'upload_date': upload_date,
1110 'thumbnails': thumbnails,
1112 'subtitles': subtitles,
1115 def _parse_smil_namespace(self, smil):
1116 return self._search_regex(
1117 r'(?i)^{([^}]+)?}smil$', smil.tag, 'namespace', default=None)
1119 def _parse_smil_formats(self, smil, smil_url, video_id, namespace=None, f4m_params=None, transform_rtmp_url=None):
1121 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
1122 b = meta.get('base') or meta.get('httpBase')
1131 videos = smil.findall(self._xpath_ns('.//video', namespace))
1132 for video in videos:
1133 src = video.get('src')
1137 bitrate = float_or_none(video.get('system-bitrate') or video.get('systemBitrate'), 1000)
1138 filesize = int_or_none(video.get('size') or video.get('fileSize'))
1139 width = int_or_none(video.get('width'))
1140 height = int_or_none(video.get('height'))
1141 proto = video.get('proto')
1142 ext = video.get('ext')
1143 src_ext = determine_ext(src)
1144 streamer = video.get('streamer') or base
1146 if proto == 'rtmp' or streamer.startswith('rtmp'):
1152 'format_id': 'rtmp-%d' % (rtmp_count if bitrate is None else bitrate),
1154 'filesize': filesize,
1158 if transform_rtmp_url:
1159 streamer, src = transform_rtmp_url(streamer, src)
1160 formats[-1].update({
1166 src_url = src if src.startswith('http') else compat_urlparse.urljoin(base, src)
1168 if proto == 'm3u8' or src_ext == 'm3u8':
1169 formats.extend(self._extract_m3u8_formats(
1170 src_url, video_id, ext or 'mp4', m3u8_id='hls', fatal=False))
1173 if src_ext == 'f4m':
1178 'plugin': 'flowplayer-3.2.0.1',
1180 f4m_url += '&' if '?' in f4m_url else '?'
1181 f4m_url += compat_urllib_parse.urlencode(f4m_params)
1182 formats.extend(self._extract_f4m_formats(f4m_url, video_id, f4m_id='hds', fatal=False))
1185 if src_url.startswith('http') and self._is_valid_url(src, video_id):
1189 'ext': ext or src_ext or 'flv',
1190 'format_id': 'http-%d' % (bitrate or http_count),
1192 'filesize': filesize,
1198 self._sort_formats(formats)
1202 def _parse_smil_subtitles(self, smil, namespace=None, subtitles_lang='en'):
1204 for num, textstream in enumerate(smil.findall(self._xpath_ns('.//textstream', namespace))):
1205 src = textstream.get('src')
1208 ext = textstream.get('ext') or determine_ext(src)
1210 type_ = textstream.get('type')
1214 'application/smptett+xml': 'tt',
1216 if type_ in SUBTITLES_TYPES:
1217 ext = SUBTITLES_TYPES[type_]
1218 lang = textstream.get('systemLanguage') or textstream.get('systemLanguageName') or textstream.get('lang') or subtitles_lang
1219 subtitles.setdefault(lang, []).append({
1225 def _extract_xspf_playlist(self, playlist_url, playlist_id, fatal=True):
1226 xspf = self._download_xml(
1227 playlist_url, playlist_id, 'Downloading xpsf playlist',
1228 'Unable to download xspf manifest', fatal=fatal)
1231 return self._parse_xspf(xspf, playlist_id)
1233 def _parse_xspf(self, playlist, playlist_id):
1235 'xspf': 'http://xspf.org/ns/0/',
1236 's1': 'http://static.streamone.nl/player/ns/0',
1240 for track in playlist.findall(xpath_with_ns('./xspf:trackList/xspf:track', NS_MAP)):
1242 track, xpath_with_ns('./xspf:title', NS_MAP), 'title', default=playlist_id)
1243 description = xpath_text(
1244 track, xpath_with_ns('./xspf:annotation', NS_MAP), 'description')
1245 thumbnail = xpath_text(
1246 track, xpath_with_ns('./xspf:image', NS_MAP), 'thumbnail')
1247 duration = float_or_none(
1248 xpath_text(track, xpath_with_ns('./xspf:duration', NS_MAP), 'duration'), 1000)
1251 'url': location.text,
1252 'format_id': location.get(xpath_with_ns('s1:label', NS_MAP)),
1253 'width': int_or_none(location.get(xpath_with_ns('s1:width', NS_MAP))),
1254 'height': int_or_none(location.get(xpath_with_ns('s1:height', NS_MAP))),
1255 } for location in track.findall(xpath_with_ns('./xspf:location', NS_MAP))]
1256 self._sort_formats(formats)
1261 'description': description,
1262 'thumbnail': thumbnail,
1263 'duration': duration,
1268 def _live_title(self, name):
1269 """ Generate the title for a live video """
1270 now = datetime.datetime.now()
1271 now_str = now.strftime("%Y-%m-%d %H:%M")
1272 return name + ' ' + now_str
1274 def _int(self, v, name, fatal=False, **kwargs):
1275 res = int_or_none(v, **kwargs)
1276 if 'get_attr' in kwargs:
1277 print(getattr(v, kwargs['get_attr']))
1279 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
1281 raise ExtractorError(msg)
1283 self._downloader.report_warning(msg)
1286 def _float(self, v, name, fatal=False, **kwargs):
1287 res = float_or_none(v, **kwargs)
1289 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
1291 raise ExtractorError(msg)
1293 self._downloader.report_warning(msg)
1296 def _set_cookie(self, domain, name, value, expire_time=None):
1297 cookie = compat_cookiejar.Cookie(
1298 0, name, value, None, None, domain, None,
1299 None, '/', True, False, expire_time, '', None, None, None)
1300 self._downloader.cookiejar.set_cookie(cookie)
1302 def _get_cookies(self, url):
1303 """ Return a compat_cookies.SimpleCookie with the cookies for the url """
1304 req = sanitized_Request(url)
1305 self._downloader.cookiejar.add_cookie_header(req)
1306 return compat_cookies.SimpleCookie(req.get_header('Cookie'))
1308 def get_testcases(self, include_onlymatching=False):
1309 t = getattr(self, '_TEST', None)
1311 assert not hasattr(self, '_TESTS'), \
1312 '%s has _TEST and _TESTS' % type(self).__name__
1315 tests = getattr(self, '_TESTS', [])
1317 if not include_onlymatching and t.get('only_matching', False):
1319 t['name'] = type(self).__name__[:-len('IE')]
1322 def is_suitable(self, age_limit):
1323 """ Test whether the extractor is generally suitable for the given
1324 age limit (i.e. pornographic sites are not, all others usually are) """
1326 any_restricted = False
1327 for tc in self.get_testcases(include_onlymatching=False):
1328 if 'playlist' in tc:
1329 tc = tc['playlist'][0]
1330 is_restricted = age_restricted(
1331 tc.get('info_dict', {}).get('age_limit'), age_limit)
1332 if not is_restricted:
1334 any_restricted = any_restricted or is_restricted
1335 return not any_restricted
1337 def extract_subtitles(self, *args, **kwargs):
1338 if (self._downloader.params.get('writesubtitles', False) or
1339 self._downloader.params.get('listsubtitles')):
1340 return self._get_subtitles(*args, **kwargs)
1343 def _get_subtitles(self, *args, **kwargs):
1344 raise NotImplementedError("This method must be implemented by subclasses")
1347 def _merge_subtitle_items(subtitle_list1, subtitle_list2):
1348 """ Merge subtitle items for one language. Items with duplicated URLs
1349 will be dropped. """
1350 list1_urls = set([item['url'] for item in subtitle_list1])
1351 ret = list(subtitle_list1)
1352 ret.extend([item for item in subtitle_list2 if item['url'] not in list1_urls])
1356 def _merge_subtitles(cls, subtitle_dict1, subtitle_dict2):
1357 """ Merge two subtitle dictionaries, language by language. """
1358 ret = dict(subtitle_dict1)
1359 for lang in subtitle_dict2:
1360 ret[lang] = cls._merge_subtitle_items(subtitle_dict1.get(lang, []), subtitle_dict2[lang])
1363 def extract_automatic_captions(self, *args, **kwargs):
1364 if (self._downloader.params.get('writeautomaticsub', False) or
1365 self._downloader.params.get('listsubtitles')):
1366 return self._get_automatic_captions(*args, **kwargs)
1369 def _get_automatic_captions(self, *args, **kwargs):
1370 raise NotImplementedError("This method must be implemented by subclasses")
1373 class SearchInfoExtractor(InfoExtractor):
1375 Base class for paged search queries extractors.
1376 They accept URLs in the format _SEARCH_KEY(|all|[0-9]):{query}
1377 Instances should define _SEARCH_KEY and _MAX_RESULTS.
1381 def _make_valid_url(cls):
1382 return r'%s(?P<prefix>|[1-9][0-9]*|all):(?P<query>[\s\S]+)' % cls._SEARCH_KEY
1385 def suitable(cls, url):
1386 return re.match(cls._make_valid_url(), url) is not None
1388 def _real_extract(self, query):
1389 mobj = re.match(self._make_valid_url(), query)
1391 raise ExtractorError('Invalid search query "%s"' % query)
1393 prefix = mobj.group('prefix')
1394 query = mobj.group('query')
1396 return self._get_n_results(query, 1)
1397 elif prefix == 'all':
1398 return self._get_n_results(query, self._MAX_RESULTS)
1402 raise ExtractorError('invalid download number %s for query "%s"' % (n, query))
1403 elif n > self._MAX_RESULTS:
1404 self._downloader.report_warning('%s returns max %i results (you requested %i)' % (self._SEARCH_KEY, self._MAX_RESULTS, n))
1405 n = self._MAX_RESULTS
1406 return self._get_n_results(query, n)
1408 def _get_n_results(self, query, n):
1409 """Get a specified number of results for a query"""
1410 raise NotImplementedError("This method must be implemented by subclasses")
1413 def SEARCH_KEY(self):
1414 return self._SEARCH_KEY