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 Unless mentioned otherwise, the fields should be Unicode strings.
206 Unless mentioned otherwise, None is equivalent to absence of information.
209 _type "playlist" indicates multiple videos.
210 There must be a key "entries", which is a list, an iterable, or a PagedList
211 object, each element of which is a valid dictionary by this specification.
213 Additionally, playlists can have "title", "description" and "id" attributes
214 with the same semantics as videos (see above).
217 _type "multi_video" indicates that there are multiple videos that
218 form a single show, for examples multiple acts of an opera or TV episode.
219 It must have an entries key like a playlist and contain all the keys
220 required for a video at the same time.
223 _type "url" indicates that the video must be extracted from another
224 location, possibly by a different extractor. Its only required key is:
225 "url" - the next URL to extract.
226 The key "ie_key" can be set to the class name (minus the trailing "IE",
227 e.g. "Youtube") if the extractor class is known in advance.
228 Additionally, the dictionary may have any properties of the resolved entity
229 known in advance, for example "title" if the title of the referred video is
233 _type "url_transparent" entities have the same specification as "url", but
234 indicate that the given additional information is more precise than the one
235 associated with the resolved URL.
236 This is useful when a site employs a video service that hosts the video and
237 its technical metadata, but that video service does not embed a useful
238 title, description etc.
241 Subclasses of this one should re-define the _real_initialize() and
242 _real_extract() methods and define a _VALID_URL regexp.
243 Probably, they should also be added to the list of extractors.
245 Finally, the _WORKING attribute should be set to False for broken IEs
246 in order to warn the users and skip the tests.
253 def __init__(self, downloader=None):
254 """Constructor. Receives an optional downloader."""
256 self.set_downloader(downloader)
259 def suitable(cls, url):
260 """Receives a URL and returns True if suitable for this IE."""
262 # This does not use has/getattr intentionally - we want to know whether
263 # we have cached the regexp for *this* class, whereas getattr would also
264 # match the superclass
265 if '_VALID_URL_RE' not in cls.__dict__:
266 cls._VALID_URL_RE = re.compile(cls._VALID_URL)
267 return cls._VALID_URL_RE.match(url) is not None
270 def _match_id(cls, url):
271 if '_VALID_URL_RE' not in cls.__dict__:
272 cls._VALID_URL_RE = re.compile(cls._VALID_URL)
273 m = cls._VALID_URL_RE.match(url)
279 """Getter method for _WORKING."""
282 def initialize(self):
283 """Initializes an instance (authentication, etc)."""
285 self._real_initialize()
288 def extract(self, url):
289 """Extracts URL information and returns it in list of dicts."""
292 return self._real_extract(url)
293 except ExtractorError:
295 except compat_http_client.IncompleteRead as e:
296 raise ExtractorError('A network error has occured.', cause=e, expected=True)
297 except (KeyError, StopIteration) as e:
298 raise ExtractorError('An extractor error has occured.', cause=e)
300 def set_downloader(self, downloader):
301 """Sets the downloader for this IE."""
302 self._downloader = downloader
304 def _real_initialize(self):
305 """Real initialization process. Redefine in subclasses."""
308 def _real_extract(self, url):
309 """Real extraction process. Redefine in subclasses."""
314 """A string for getting the InfoExtractor with get_info_extractor"""
315 return compat_str(cls.__name__[:-2])
319 return compat_str(type(self).__name__[:-2])
321 def _request_webpage(self, url_or_request, video_id, note=None, errnote=None, fatal=True):
322 """ Returns the response handle """
324 self.report_download_webpage(video_id)
325 elif note is not False:
327 self.to_screen('%s' % (note,))
329 self.to_screen('%s: %s' % (video_id, note))
331 return self._downloader.urlopen(url_or_request)
332 except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err:
336 errnote = 'Unable to download webpage'
338 errmsg = '%s: %s' % (errnote, error_to_compat_str(err))
340 raise ExtractorError(errmsg, sys.exc_info()[2], cause=err)
342 self._downloader.report_warning(errmsg)
345 def _download_webpage_handle(self, url_or_request, video_id, note=None, errnote=None, fatal=True, encoding=None):
346 """ Returns a tuple (page content as string, URL handle) """
347 # Strip hashes from the URL (#1038)
348 if isinstance(url_or_request, (compat_str, str)):
349 url_or_request = url_or_request.partition('#')[0]
351 urlh = self._request_webpage(url_or_request, video_id, note, errnote, fatal)
355 content = self._webpage_read_content(urlh, url_or_request, video_id, note, errnote, fatal, encoding=encoding)
356 return (content, urlh)
359 def _guess_encoding_from_content(content_type, webpage_bytes):
360 m = re.match(r'[a-zA-Z0-9_.-]+/[a-zA-Z0-9_.-]+\s*;\s*charset=(.+)', content_type)
362 encoding = m.group(1)
364 m = re.search(br'<meta[^>]+charset=[\'"]?([^\'")]+)[ /\'">]',
365 webpage_bytes[:1024])
367 encoding = m.group(1).decode('ascii')
368 elif webpage_bytes.startswith(b'\xff\xfe'):
375 def _webpage_read_content(self, urlh, url_or_request, video_id, note=None, errnote=None, fatal=True, prefix=None, encoding=None):
376 content_type = urlh.headers.get('Content-Type', '')
377 webpage_bytes = urlh.read()
378 if prefix is not None:
379 webpage_bytes = prefix + webpage_bytes
381 encoding = self._guess_encoding_from_content(content_type, webpage_bytes)
382 if self._downloader.params.get('dump_intermediate_pages', False):
384 url = url_or_request.get_full_url()
385 except AttributeError:
387 self.to_screen('Dumping request to ' + url)
388 dump = base64.b64encode(webpage_bytes).decode('ascii')
389 self._downloader.to_screen(dump)
390 if self._downloader.params.get('write_pages', False):
392 url = url_or_request.get_full_url()
393 except AttributeError:
395 basen = '%s_%s' % (video_id, url)
397 h = '___' + hashlib.md5(basen.encode('utf-8')).hexdigest()
398 basen = basen[:240 - len(h)] + h
399 raw_filename = basen + '.dump'
400 filename = sanitize_filename(raw_filename, restricted=True)
401 self.to_screen('Saving request to ' + filename)
402 # Working around MAX_PATH limitation on Windows (see
403 # http://msdn.microsoft.com/en-us/library/windows/desktop/aa365247(v=vs.85).aspx)
405 absfilepath = os.path.abspath(filename)
406 if len(absfilepath) > 259:
407 filename = '\\\\?\\' + absfilepath
408 with open(filename, 'wb') as outf:
409 outf.write(webpage_bytes)
412 content = webpage_bytes.decode(encoding, 'replace')
414 content = webpage_bytes.decode('utf-8', 'replace')
416 if ('<title>Access to this site is blocked</title>' in content and
417 'Websense' in content[:512]):
418 msg = 'Access to this webpage has been blocked by Websense filtering software in your network.'
419 blocked_iframe = self._html_search_regex(
420 r'<iframe src="([^"]+)"', content,
421 'Websense information URL', default=None)
423 msg += ' Visit %s for more details' % blocked_iframe
424 raise ExtractorError(msg, expected=True)
425 if '<title>The URL you requested has been blocked</title>' in content[:512]:
427 'Access to this webpage has been blocked by Indian censorship. '
428 'Use a VPN or proxy server (with --proxy) to route around it.')
429 block_msg = self._html_search_regex(
430 r'</h1><p>(.*?)</p>',
431 content, 'block message', default=None)
433 msg += ' (Message: "%s")' % block_msg.replace('\n', ' ')
434 raise ExtractorError(msg, expected=True)
438 def _download_webpage(self, url_or_request, video_id, note=None, errnote=None, fatal=True, tries=1, timeout=5, encoding=None):
439 """ Returns the data of the page as a string """
442 while success is False:
444 res = self._download_webpage_handle(url_or_request, video_id, note, errnote, fatal, encoding=encoding)
446 except compat_http_client.IncompleteRead as e:
448 if try_count >= tries:
450 self._sleep(timeout, video_id)
457 def _download_xml(self, url_or_request, video_id,
458 note='Downloading XML', errnote='Unable to download XML',
459 transform_source=None, fatal=True, encoding=None):
460 """Return the xml as an xml.etree.ElementTree.Element"""
461 xml_string = self._download_webpage(
462 url_or_request, video_id, note, errnote, fatal=fatal, encoding=encoding)
463 if xml_string is False:
466 xml_string = transform_source(xml_string)
467 return compat_etree_fromstring(xml_string.encode('utf-8'))
469 def _download_json(self, url_or_request, video_id,
470 note='Downloading JSON metadata',
471 errnote='Unable to download JSON metadata',
472 transform_source=None,
473 fatal=True, encoding=None):
474 json_string = self._download_webpage(
475 url_or_request, video_id, note, errnote, fatal=fatal,
477 if (not fatal) and json_string is False:
479 return self._parse_json(
480 json_string, video_id, transform_source=transform_source, fatal=fatal)
482 def _parse_json(self, json_string, video_id, transform_source=None, fatal=True):
484 json_string = transform_source(json_string)
486 return json.loads(json_string)
487 except ValueError as ve:
488 errmsg = '%s: Failed to parse JSON ' % video_id
490 raise ExtractorError(errmsg, cause=ve)
492 self.report_warning(errmsg + str(ve))
494 def report_warning(self, msg, video_id=None):
495 idstr = '' if video_id is None else '%s: ' % video_id
496 self._downloader.report_warning(
497 '[%s] %s%s' % (self.IE_NAME, idstr, msg))
499 def to_screen(self, msg):
500 """Print msg to screen, prefixing it with '[ie_name]'"""
501 self._downloader.to_screen('[%s] %s' % (self.IE_NAME, msg))
503 def report_extraction(self, id_or_name):
504 """Report information extraction."""
505 self.to_screen('%s: Extracting information' % id_or_name)
507 def report_download_webpage(self, video_id):
508 """Report webpage download."""
509 self.to_screen('%s: Downloading webpage' % video_id)
511 def report_age_confirmation(self):
512 """Report attempt to confirm age."""
513 self.to_screen('Confirming age')
515 def report_login(self):
516 """Report attempt to log in."""
517 self.to_screen('Logging in')
520 def raise_login_required(msg='This video is only available for registered users'):
521 raise ExtractorError(
522 '%s. Use --username and --password or --netrc to provide account credentials.' % msg,
526 def raise_geo_restricted(msg='This video is not available from your location due to geo restriction'):
527 raise ExtractorError(
528 '%s. You might want to use --proxy to workaround.' % msg,
531 # Methods for following #608
533 def url_result(url, ie=None, video_id=None, video_title=None):
534 """Returns a URL that points to a page that should be processed"""
535 # TODO: ie should be the class used for getting the info
536 video_info = {'_type': 'url',
539 if video_id is not None:
540 video_info['id'] = video_id
541 if video_title is not None:
542 video_info['title'] = video_title
546 def playlist_result(entries, playlist_id=None, playlist_title=None, playlist_description=None):
547 """Returns a playlist"""
548 video_info = {'_type': 'playlist',
551 video_info['id'] = playlist_id
553 video_info['title'] = playlist_title
554 if playlist_description:
555 video_info['description'] = playlist_description
558 def _search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
560 Perform a regex search on the given string, using a single or a list of
561 patterns returning the first matching group.
562 In case of failure return a default value or raise a WARNING or a
563 RegexNotFoundError, depending on fatal, specifying the field name.
565 if isinstance(pattern, (str, compat_str, compiled_regex_type)):
566 mobj = re.search(pattern, string, flags)
569 mobj = re.search(p, string, flags)
573 if not self._downloader.params.get('no_color') and os.name != 'nt' and sys.stderr.isatty():
574 _name = '\033[0;34m%s\033[0m' % name
580 # return the first matching group
581 return next(g for g in mobj.groups() if g is not None)
583 return mobj.group(group)
584 elif default is not NO_DEFAULT:
587 raise RegexNotFoundError('Unable to extract %s' % _name)
589 self._downloader.report_warning('unable to extract %s' % _name + bug_reports_message())
592 def _html_search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
594 Like _search_regex, but strips HTML tags and unescapes entities.
596 res = self._search_regex(pattern, string, name, default, fatal, flags, group)
598 return clean_html(res).strip()
602 def _get_login_info(self):
604 Get the login info as (username, password)
605 It will look in the netrc file using the _NETRC_MACHINE value
606 If there's no info available, return (None, None)
608 if self._downloader is None:
613 downloader_params = self._downloader.params
615 # Attempt to use provided username and password or .netrc data
616 if downloader_params.get('username', None) is not None:
617 username = downloader_params['username']
618 password = downloader_params['password']
619 elif downloader_params.get('usenetrc', False):
621 info = netrc.netrc().authenticators(self._NETRC_MACHINE)
626 raise netrc.NetrcParseError('No authenticators for %s' % self._NETRC_MACHINE)
627 except (IOError, netrc.NetrcParseError) as err:
628 self._downloader.report_warning('parsing .netrc: %s' % error_to_compat_str(err))
630 return (username, password)
632 def _get_tfa_info(self, note='two-factor verification code'):
634 Get the two-factor authentication info
635 TODO - asking the user will be required for sms/phone verify
636 currently just uses the command line option
637 If there's no info available, return None
639 if self._downloader is None:
641 downloader_params = self._downloader.params
643 if downloader_params.get('twofactor', None) is not None:
644 return downloader_params['twofactor']
646 return compat_getpass('Type %s and press [Return]: ' % note)
648 # Helper functions for extracting OpenGraph info
650 def _og_regexes(prop):
651 content_re = r'content=(?:"([^"]+?)"|\'([^\']+?)\'|\s*([^\s"\'=<>`]+?))'
652 property_re = (r'(?:name|property)=(?:\'og:%(prop)s\'|"og:%(prop)s"|\s*og:%(prop)s\b)'
653 % {'prop': re.escape(prop)})
654 template = r'<meta[^>]+?%s[^>]+?%s'
656 template % (property_re, content_re),
657 template % (content_re, property_re),
661 def _meta_regex(prop):
662 return r'''(?isx)<meta
663 (?=[^>]+(?:itemprop|name|property|id|http-equiv)=(["\']?)%s\1)
664 [^>]+?content=(["\'])(?P<content>.*?)\2''' % re.escape(prop)
666 def _og_search_property(self, prop, html, name=None, **kargs):
668 name = 'OpenGraph %s' % prop
669 escaped = self._search_regex(self._og_regexes(prop), html, name, flags=re.DOTALL, **kargs)
672 return unescapeHTML(escaped)
674 def _og_search_thumbnail(self, html, **kargs):
675 return self._og_search_property('image', html, 'thumbnail URL', fatal=False, **kargs)
677 def _og_search_description(self, html, **kargs):
678 return self._og_search_property('description', html, fatal=False, **kargs)
680 def _og_search_title(self, html, **kargs):
681 return self._og_search_property('title', html, **kargs)
683 def _og_search_video_url(self, html, name='video url', secure=True, **kargs):
684 regexes = self._og_regexes('video') + self._og_regexes('video:url')
686 regexes = self._og_regexes('video:secure_url') + regexes
687 return self._html_search_regex(regexes, html, name, **kargs)
689 def _og_search_url(self, html, **kargs):
690 return self._og_search_property('url', html, **kargs)
692 def _html_search_meta(self, name, html, display_name=None, fatal=False, **kwargs):
693 if display_name is None:
695 return self._html_search_regex(
696 self._meta_regex(name),
697 html, display_name, fatal=fatal, group='content', **kwargs)
699 def _dc_search_uploader(self, html):
700 return self._html_search_meta('dc.creator', html, 'uploader')
702 def _rta_search(self, html):
703 # See http://www.rtalabel.org/index.php?content=howtofaq#single
704 if re.search(r'(?ix)<meta\s+name="rating"\s+'
705 r' content="RTA-5042-1996-1400-1577-RTA"',
710 def _media_rating_search(self, html):
711 # See http://www.tjg-designs.com/WP/metadata-code-examples-adding-metadata-to-your-web-pages/
712 rating = self._html_search_meta('rating', html)
724 return RATING_TABLE.get(rating.lower(), None)
726 def _family_friendly_search(self, html):
727 # See http://schema.org/VideoObject
728 family_friendly = self._html_search_meta('isFamilyFriendly', html)
730 if not family_friendly:
739 return RATING_TABLE.get(family_friendly.lower(), None)
741 def _twitter_search_player(self, html):
742 return self._html_search_meta('twitter:player', html,
743 'twitter card player')
746 def _hidden_inputs(html):
747 html = re.sub(r'<!--(?:(?!<!--).)*-->', '', html)
749 for input in re.findall(r'(?i)<input([^>]+)>', html):
750 if not re.search(r'type=(["\'])(?:hidden|submit)\1', input):
752 name = re.search(r'name=(["\'])(?P<value>.+?)\1', input)
755 value = re.search(r'value=(["\'])(?P<value>.*?)\1', input)
758 hidden_inputs[name.group('value')] = value.group('value')
761 def _form_hidden_inputs(self, form_id, html):
762 form = self._search_regex(
763 r'(?is)<form[^>]+?id=(["\'])%s\1[^>]*>(?P<form>.+?)</form>' % form_id,
764 html, '%s form' % form_id, group='form')
765 return self._hidden_inputs(form)
767 def _sort_formats(self, formats, field_preference=None):
769 raise ExtractorError('No video formats found')
772 # TODO remove the following workaround
773 from ..utils import determine_ext
774 if not f.get('ext') and 'url' in f:
775 f['ext'] = determine_ext(f['url'])
777 if isinstance(field_preference, (list, tuple)):
778 return tuple(f.get(field) if f.get(field) is not None else -1 for field in field_preference)
780 preference = f.get('preference')
781 if preference is None:
783 if f.get('ext') in ['f4f', 'f4m']: # Not yet supported
786 proto_preference = 0 if determine_protocol(f) in ['http', 'https'] else -0.1
788 if f.get('vcodec') == 'none': # audio only
789 if self._downloader.params.get('prefer_free_formats'):
790 ORDER = ['aac', 'mp3', 'm4a', 'webm', 'ogg', 'opus']
792 ORDER = ['webm', 'opus', 'ogg', 'mp3', 'aac', 'm4a']
795 audio_ext_preference = ORDER.index(f['ext'])
797 audio_ext_preference = -1
799 if self._downloader.params.get('prefer_free_formats'):
800 ORDER = ['flv', 'mp4', 'webm']
802 ORDER = ['webm', 'flv', 'mp4']
804 ext_preference = ORDER.index(f['ext'])
807 audio_ext_preference = 0
811 f.get('language_preference') if f.get('language_preference') is not None else -1,
812 f.get('quality') if f.get('quality') is not None else -1,
813 f.get('tbr') if f.get('tbr') is not None else -1,
814 f.get('filesize') if f.get('filesize') is not None else -1,
815 f.get('vbr') if f.get('vbr') is not None else -1,
816 f.get('height') if f.get('height') is not None else -1,
817 f.get('width') if f.get('width') is not None else -1,
820 f.get('abr') if f.get('abr') is not None else -1,
821 audio_ext_preference,
822 f.get('fps') if f.get('fps') is not None else -1,
823 f.get('filesize_approx') if f.get('filesize_approx') is not None else -1,
824 f.get('source_preference') if f.get('source_preference') is not None else -1,
825 f.get('format_id') if f.get('format_id') is not None else '',
827 formats.sort(key=_formats_key)
829 def _check_formats(self, formats, video_id):
832 lambda f: self._is_valid_url(
834 item='%s video format' % f.get('format_id') if f.get('format_id') else 'video'),
837 def _is_valid_url(self, url, video_id, item='video'):
838 url = self._proto_relative_url(url, scheme='http:')
839 # For now assume non HTTP(S) URLs always valid
840 if not (url.startswith('http://') or url.startswith('https://')):
843 self._request_webpage(url, video_id, 'Checking %s URL' % item)
845 except ExtractorError as e:
846 if isinstance(e.cause, compat_urllib_error.URLError):
848 '%s: %s URL is invalid, skipping' % (video_id, item))
852 def http_scheme(self):
853 """ Either "http:" or "https:", depending on the user's preferences """
856 if self._downloader.params.get('prefer_insecure', False)
859 def _proto_relative_url(self, url, scheme=None):
862 if url.startswith('//'):
864 scheme = self.http_scheme()
869 def _sleep(self, timeout, video_id, msg_template=None):
870 if msg_template is None:
871 msg_template = '%(video_id)s: Waiting for %(timeout)s seconds'
872 msg = msg_template % {'video_id': video_id, 'timeout': timeout}
876 def _extract_f4m_formats(self, manifest_url, video_id, preference=None, f4m_id=None,
877 transform_source=lambda s: fix_xml_ampersands(s).strip(),
879 manifest = self._download_xml(
880 manifest_url, video_id, 'Downloading f4m manifest',
881 'Unable to download f4m manifest',
882 # Some manifests may be malformed, e.g. prosiebensat1 generated manifests
883 # (see https://github.com/rg3/youtube-dl/issues/6215#issuecomment-121704244)
884 transform_source=transform_source,
887 if manifest is False:
891 manifest_version = '1.0'
892 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/1.0}media')
894 manifest_version = '2.0'
895 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/2.0}media')
896 base_url = xpath_text(
897 manifest, ['{http://ns.adobe.com/f4m/1.0}baseURL', '{http://ns.adobe.com/f4m/2.0}baseURL'],
898 'base URL', default=None)
900 base_url = base_url.strip()
901 for i, media_el in enumerate(media_nodes):
902 if manifest_version == '2.0':
903 media_url = media_el.attrib.get('href') or media_el.attrib.get('url')
907 media_url if media_url.startswith('http://') or media_url.startswith('https://')
908 else ((base_url or '/'.join(manifest_url.split('/')[:-1])) + '/' + media_url))
909 # If media_url is itself a f4m manifest do the recursive extraction
910 # since bitrates in parent manifest (this one) and media_url manifest
911 # may differ leading to inability to resolve the format by requested
912 # bitrate in f4m downloader
913 if determine_ext(manifest_url) == 'f4m':
914 formats.extend(self._extract_f4m_formats(
915 manifest_url, video_id, preference, f4m_id, fatal=fatal))
917 tbr = int_or_none(media_el.attrib.get('bitrate'))
919 'format_id': '-'.join(filter(None, [f4m_id, compat_str(i if tbr is None else tbr)])),
923 'width': int_or_none(media_el.attrib.get('width')),
924 'height': int_or_none(media_el.attrib.get('height')),
925 'preference': preference,
927 self._sort_formats(formats)
931 def _extract_m3u8_formats(self, m3u8_url, video_id, ext=None,
932 entry_protocol='m3u8', preference=None,
933 m3u8_id=None, note=None, errnote=None,
937 'format_id': '-'.join(filter(None, [m3u8_id, 'meta'])),
941 'preference': preference - 1 if preference else -1,
942 'resolution': 'multiple',
943 'format_note': 'Quality selection URL',
946 format_url = lambda u: (
948 if re.match(r'^https?://', u)
949 else compat_urlparse.urljoin(m3u8_url, u))
951 res = self._download_webpage_handle(
953 note=note or 'Downloading m3u8 information',
954 errnote=errnote or 'Failed to download m3u8 information',
959 m3u8_url = urlh.geturl()
963 r'(?P<key>[a-zA-Z_-]+)=(?P<val>"[^"]+"|[^",]+)(?:,|$)')
964 for line in m3u8_doc.splitlines():
965 if line.startswith('#EXT-X-STREAM-INF:'):
967 for m in kv_rex.finditer(line):
969 if v.startswith('"'):
971 last_info[m.group('key')] = v
972 elif line.startswith('#EXT-X-MEDIA:'):
974 for m in kv_rex.finditer(line):
976 if v.startswith('"'):
978 last_media[m.group('key')] = v
979 elif line.startswith('#') or not line.strip():
982 if last_info is None:
983 formats.append({'url': format_url(line)})
985 tbr = int_or_none(last_info.get('BANDWIDTH'), scale=1000)
988 format_id.append(m3u8_id)
989 last_media_name = last_media.get('NAME') if last_media and last_media.get('TYPE') != 'SUBTITLES' else None
990 format_id.append(last_media_name if last_media_name else '%d' % (tbr if tbr else len(formats)))
992 'format_id': '-'.join(format_id),
993 'url': format_url(line.strip()),
996 'protocol': entry_protocol,
997 'preference': preference,
999 codecs = last_info.get('CODECS')
1001 # TODO: looks like video codec is not always necessarily goes first
1002 va_codecs = codecs.split(',')
1004 f['vcodec'] = va_codecs[0].partition('.')[0]
1005 if len(va_codecs) > 1 and va_codecs[1]:
1006 f['acodec'] = va_codecs[1].partition('.')[0]
1007 resolution = last_info.get('RESOLUTION')
1009 width_str, height_str = resolution.split('x')
1010 f['width'] = int(width_str)
1011 f['height'] = int(height_str)
1012 if last_media is not None:
1013 f['m3u8_media'] = last_media
1017 self._sort_formats(formats)
1021 def _xpath_ns(path, namespace=None):
1025 for c in path.split('/'):
1026 if not c or c == '.':
1029 out.append('{%s}%s' % (namespace, c))
1030 return '/'.join(out)
1032 def _extract_smil_formats(self, smil_url, video_id, fatal=True, f4m_params=None):
1033 smil = self._download_smil(smil_url, video_id, fatal=fatal)
1039 namespace = self._parse_smil_namespace(smil)
1041 return self._parse_smil_formats(
1042 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
1044 def _extract_smil_info(self, smil_url, video_id, fatal=True, f4m_params=None):
1045 smil = self._download_smil(smil_url, video_id, fatal=fatal)
1048 return self._parse_smil(smil, smil_url, video_id, f4m_params=f4m_params)
1050 def _download_smil(self, smil_url, video_id, fatal=True):
1051 return self._download_xml(
1052 smil_url, video_id, 'Downloading SMIL file',
1053 'Unable to download SMIL file', fatal=fatal)
1055 def _parse_smil(self, smil, smil_url, video_id, f4m_params=None):
1056 namespace = self._parse_smil_namespace(smil)
1058 formats = self._parse_smil_formats(
1059 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
1060 subtitles = self._parse_smil_subtitles(smil, namespace=namespace)
1062 video_id = os.path.splitext(url_basename(smil_url))[0]
1066 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
1067 name = meta.attrib.get('name')
1068 content = meta.attrib.get('content')
1069 if not name or not content:
1071 if not title and name == 'title':
1073 elif not description and name in ('description', 'abstract'):
1074 description = content
1075 elif not upload_date and name == 'date':
1076 upload_date = unified_strdate(content)
1079 'id': image.get('type'),
1080 'url': image.get('src'),
1081 'width': int_or_none(image.get('width')),
1082 'height': int_or_none(image.get('height')),
1083 } for image in smil.findall(self._xpath_ns('.//image', namespace)) if image.get('src')]
1087 'title': title or video_id,
1088 'description': description,
1089 'upload_date': upload_date,
1090 'thumbnails': thumbnails,
1092 'subtitles': subtitles,
1095 def _parse_smil_namespace(self, smil):
1096 return self._search_regex(
1097 r'(?i)^{([^}]+)?}smil$', smil.tag, 'namespace', default=None)
1099 def _parse_smil_formats(self, smil, smil_url, video_id, namespace=None, f4m_params=None, transform_rtmp_url=None):
1101 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
1102 b = meta.get('base') or meta.get('httpBase')
1111 videos = smil.findall(self._xpath_ns('.//video', namespace))
1112 for video in videos:
1113 src = video.get('src')
1117 bitrate = float_or_none(video.get('system-bitrate') or video.get('systemBitrate'), 1000)
1118 filesize = int_or_none(video.get('size') or video.get('fileSize'))
1119 width = int_or_none(video.get('width'))
1120 height = int_or_none(video.get('height'))
1121 proto = video.get('proto')
1122 ext = video.get('ext')
1123 src_ext = determine_ext(src)
1124 streamer = video.get('streamer') or base
1126 if proto == 'rtmp' or streamer.startswith('rtmp'):
1132 'format_id': 'rtmp-%d' % (rtmp_count if bitrate is None else bitrate),
1134 'filesize': filesize,
1138 if transform_rtmp_url:
1139 streamer, src = transform_rtmp_url(streamer, src)
1140 formats[-1].update({
1146 src_url = src if src.startswith('http') else compat_urlparse.urljoin(base, src)
1148 if proto == 'm3u8' or src_ext == 'm3u8':
1149 formats.extend(self._extract_m3u8_formats(
1150 src_url, video_id, ext or 'mp4', m3u8_id='hls', fatal=False))
1153 if src_ext == 'f4m':
1158 'plugin': 'flowplayer-3.2.0.1',
1160 f4m_url += '&' if '?' in f4m_url else '?'
1161 f4m_url += compat_urllib_parse.urlencode(f4m_params)
1162 formats.extend(self._extract_f4m_formats(f4m_url, video_id, f4m_id='hds', fatal=False))
1165 if src_url.startswith('http') and self._is_valid_url(src, video_id):
1169 'ext': ext or src_ext or 'flv',
1170 'format_id': 'http-%d' % (bitrate or http_count),
1172 'filesize': filesize,
1178 self._sort_formats(formats)
1182 def _parse_smil_subtitles(self, smil, namespace=None, subtitles_lang='en'):
1184 for num, textstream in enumerate(smil.findall(self._xpath_ns('.//textstream', namespace))):
1185 src = textstream.get('src')
1188 ext = textstream.get('ext') or determine_ext(src)
1190 type_ = textstream.get('type')
1194 'application/smptett+xml': 'tt',
1196 if type_ in SUBTITLES_TYPES:
1197 ext = SUBTITLES_TYPES[type_]
1198 lang = textstream.get('systemLanguage') or textstream.get('systemLanguageName') or textstream.get('lang') or subtitles_lang
1199 subtitles.setdefault(lang, []).append({
1205 def _extract_xspf_playlist(self, playlist_url, playlist_id, fatal=True):
1206 xspf = self._download_xml(
1207 playlist_url, playlist_id, 'Downloading xpsf playlist',
1208 'Unable to download xspf manifest', fatal=fatal)
1211 return self._parse_xspf(xspf, playlist_id)
1213 def _parse_xspf(self, playlist, playlist_id):
1215 'xspf': 'http://xspf.org/ns/0/',
1216 's1': 'http://static.streamone.nl/player/ns/0',
1220 for track in playlist.findall(xpath_with_ns('./xspf:trackList/xspf:track', NS_MAP)):
1222 track, xpath_with_ns('./xspf:title', NS_MAP), 'title', default=playlist_id)
1223 description = xpath_text(
1224 track, xpath_with_ns('./xspf:annotation', NS_MAP), 'description')
1225 thumbnail = xpath_text(
1226 track, xpath_with_ns('./xspf:image', NS_MAP), 'thumbnail')
1227 duration = float_or_none(
1228 xpath_text(track, xpath_with_ns('./xspf:duration', NS_MAP), 'duration'), 1000)
1231 'url': location.text,
1232 'format_id': location.get(xpath_with_ns('s1:label', NS_MAP)),
1233 'width': int_or_none(location.get(xpath_with_ns('s1:width', NS_MAP))),
1234 'height': int_or_none(location.get(xpath_with_ns('s1:height', NS_MAP))),
1235 } for location in track.findall(xpath_with_ns('./xspf:location', NS_MAP))]
1236 self._sort_formats(formats)
1241 'description': description,
1242 'thumbnail': thumbnail,
1243 'duration': duration,
1248 def _live_title(self, name):
1249 """ Generate the title for a live video """
1250 now = datetime.datetime.now()
1251 now_str = now.strftime("%Y-%m-%d %H:%M")
1252 return name + ' ' + now_str
1254 def _int(self, v, name, fatal=False, **kwargs):
1255 res = int_or_none(v, **kwargs)
1256 if 'get_attr' in kwargs:
1257 print(getattr(v, kwargs['get_attr']))
1259 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
1261 raise ExtractorError(msg)
1263 self._downloader.report_warning(msg)
1266 def _float(self, v, name, fatal=False, **kwargs):
1267 res = float_or_none(v, **kwargs)
1269 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
1271 raise ExtractorError(msg)
1273 self._downloader.report_warning(msg)
1276 def _set_cookie(self, domain, name, value, expire_time=None):
1277 cookie = compat_cookiejar.Cookie(
1278 0, name, value, None, None, domain, None,
1279 None, '/', True, False, expire_time, '', None, None, None)
1280 self._downloader.cookiejar.set_cookie(cookie)
1282 def _get_cookies(self, url):
1283 """ Return a compat_cookies.SimpleCookie with the cookies for the url """
1284 req = sanitized_Request(url)
1285 self._downloader.cookiejar.add_cookie_header(req)
1286 return compat_cookies.SimpleCookie(req.get_header('Cookie'))
1288 def get_testcases(self, include_onlymatching=False):
1289 t = getattr(self, '_TEST', None)
1291 assert not hasattr(self, '_TESTS'), \
1292 '%s has _TEST and _TESTS' % type(self).__name__
1295 tests = getattr(self, '_TESTS', [])
1297 if not include_onlymatching and t.get('only_matching', False):
1299 t['name'] = type(self).__name__[:-len('IE')]
1302 def is_suitable(self, age_limit):
1303 """ Test whether the extractor is generally suitable for the given
1304 age limit (i.e. pornographic sites are not, all others usually are) """
1306 any_restricted = False
1307 for tc in self.get_testcases(include_onlymatching=False):
1308 if 'playlist' in tc:
1309 tc = tc['playlist'][0]
1310 is_restricted = age_restricted(
1311 tc.get('info_dict', {}).get('age_limit'), age_limit)
1312 if not is_restricted:
1314 any_restricted = any_restricted or is_restricted
1315 return not any_restricted
1317 def extract_subtitles(self, *args, **kwargs):
1318 if (self._downloader.params.get('writesubtitles', False) or
1319 self._downloader.params.get('listsubtitles')):
1320 return self._get_subtitles(*args, **kwargs)
1323 def _get_subtitles(self, *args, **kwargs):
1324 raise NotImplementedError("This method must be implemented by subclasses")
1327 def _merge_subtitle_items(subtitle_list1, subtitle_list2):
1328 """ Merge subtitle items for one language. Items with duplicated URLs
1329 will be dropped. """
1330 list1_urls = set([item['url'] for item in subtitle_list1])
1331 ret = list(subtitle_list1)
1332 ret.extend([item for item in subtitle_list2 if item['url'] not in list1_urls])
1336 def _merge_subtitles(cls, subtitle_dict1, subtitle_dict2):
1337 """ Merge two subtitle dictionaries, language by language. """
1338 ret = dict(subtitle_dict1)
1339 for lang in subtitle_dict2:
1340 ret[lang] = cls._merge_subtitle_items(subtitle_dict1.get(lang, []), subtitle_dict2[lang])
1343 def extract_automatic_captions(self, *args, **kwargs):
1344 if (self._downloader.params.get('writeautomaticsub', False) or
1345 self._downloader.params.get('listsubtitles')):
1346 return self._get_automatic_captions(*args, **kwargs)
1349 def _get_automatic_captions(self, *args, **kwargs):
1350 raise NotImplementedError("This method must be implemented by subclasses")
1353 class SearchInfoExtractor(InfoExtractor):
1355 Base class for paged search queries extractors.
1356 They accept URLs in the format _SEARCH_KEY(|all|[0-9]):{query}
1357 Instances should define _SEARCH_KEY and _MAX_RESULTS.
1361 def _make_valid_url(cls):
1362 return r'%s(?P<prefix>|[1-9][0-9]*|all):(?P<query>[\s\S]+)' % cls._SEARCH_KEY
1365 def suitable(cls, url):
1366 return re.match(cls._make_valid_url(), url) is not None
1368 def _real_extract(self, query):
1369 mobj = re.match(self._make_valid_url(), query)
1371 raise ExtractorError('Invalid search query "%s"' % query)
1373 prefix = mobj.group('prefix')
1374 query = mobj.group('query')
1376 return self._get_n_results(query, 1)
1377 elif prefix == 'all':
1378 return self._get_n_results(query, self._MAX_RESULTS)
1382 raise ExtractorError('invalid download number %s for query "%s"' % (n, query))
1383 elif n > self._MAX_RESULTS:
1384 self._downloader.report_warning('%s returns max %i results (you requested %i)' % (self._SEARCH_KEY, self._MAX_RESULTS, n))
1385 n = self._MAX_RESULTS
1386 return self._get_n_results(query, n)
1388 def _get_n_results(self, query, n):
1389 """Get a specified number of results for a query"""
1390 raise NotImplementedError("This method must be implemented by subclasses")
1393 def SEARCH_KEY(self):
1394 return self._SEARCH_KEY