1 from __future__ import unicode_literals
14 from ..compat import (
21 compat_urllib_parse_urlparse,
24 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 Unless mentioned otherwise, the fields should be Unicode strings.
205 Unless mentioned otherwise, None is equivalent to absence of information.
208 _type "playlist" indicates multiple videos.
209 There must be a key "entries", which is a list, an iterable, or a PagedList
210 object, each element of which is a valid dictionary by this specification.
212 Additionally, playlists can have "title", "description" and "id" attributes
213 with the same semantics as videos (see above).
216 _type "multi_video" indicates that there are multiple videos that
217 form a single show, for examples multiple acts of an opera or TV episode.
218 It must have an entries key like a playlist and contain all the keys
219 required for a video at the same time.
222 _type "url" indicates that the video must be extracted from another
223 location, possibly by a different extractor. Its only required key is:
224 "url" - the next URL to extract.
225 The key "ie_key" can be set to the class name (minus the trailing "IE",
226 e.g. "Youtube") if the extractor class is known in advance.
227 Additionally, the dictionary may have any properties of the resolved entity
228 known in advance, for example "title" if the title of the referred video is
232 _type "url_transparent" entities have the same specification as "url", but
233 indicate that the given additional information is more precise than the one
234 associated with the resolved URL.
235 This is useful when a site employs a video service that hosts the video and
236 its technical metadata, but that video service does not embed a useful
237 title, description etc.
240 Subclasses of this one should re-define the _real_initialize() and
241 _real_extract() methods and define a _VALID_URL regexp.
242 Probably, they should also be added to the list of extractors.
244 Finally, the _WORKING attribute should be set to False for broken IEs
245 in order to warn the users and skip the tests.
252 def __init__(self, downloader=None):
253 """Constructor. Receives an optional downloader."""
255 self.set_downloader(downloader)
258 def suitable(cls, url):
259 """Receives a URL and returns True if suitable for this IE."""
261 # This does not use has/getattr intentionally - we want to know whether
262 # we have cached the regexp for *this* class, whereas getattr would also
263 # match the superclass
264 if '_VALID_URL_RE' not in cls.__dict__:
265 cls._VALID_URL_RE = re.compile(cls._VALID_URL)
266 return cls._VALID_URL_RE.match(url) is not None
269 def _match_id(cls, url):
270 if '_VALID_URL_RE' not in cls.__dict__:
271 cls._VALID_URL_RE = re.compile(cls._VALID_URL)
272 m = cls._VALID_URL_RE.match(url)
278 """Getter method for _WORKING."""
281 def initialize(self):
282 """Initializes an instance (authentication, etc)."""
284 self._real_initialize()
287 def extract(self, url):
288 """Extracts URL information and returns it in list of dicts."""
291 return self._real_extract(url)
292 except ExtractorError:
294 except compat_http_client.IncompleteRead as e:
295 raise ExtractorError('A network error has occured.', cause=e, expected=True)
296 except (KeyError, StopIteration) as e:
297 raise ExtractorError('An extractor error has occured.', cause=e)
299 def set_downloader(self, downloader):
300 """Sets the downloader for this IE."""
301 self._downloader = downloader
303 def _real_initialize(self):
304 """Real initialization process. Redefine in subclasses."""
307 def _real_extract(self, url):
308 """Real extraction process. Redefine in subclasses."""
313 """A string for getting the InfoExtractor with get_info_extractor"""
314 return compat_str(cls.__name__[:-2])
318 return compat_str(type(self).__name__[:-2])
320 def _request_webpage(self, url_or_request, video_id, note=None, errnote=None, fatal=True):
321 """ Returns the response handle """
323 self.report_download_webpage(video_id)
324 elif note is not False:
326 self.to_screen('%s' % (note,))
328 self.to_screen('%s: %s' % (video_id, note))
330 return self._downloader.urlopen(url_or_request)
331 except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err:
335 errnote = 'Unable to download webpage'
337 errmsg = '%s: %s' % (errnote, error_to_compat_str(err))
339 raise ExtractorError(errmsg, sys.exc_info()[2], cause=err)
341 self._downloader.report_warning(errmsg)
344 def _download_webpage_handle(self, url_or_request, video_id, note=None, errnote=None, fatal=True, encoding=None):
345 """ Returns a tuple (page content as string, URL handle) """
346 # Strip hashes from the URL (#1038)
347 if isinstance(url_or_request, (compat_str, str)):
348 url_or_request = url_or_request.partition('#')[0]
350 urlh = self._request_webpage(url_or_request, video_id, note, errnote, fatal)
354 content = self._webpage_read_content(urlh, url_or_request, video_id, note, errnote, fatal, encoding=encoding)
355 return (content, urlh)
358 def _guess_encoding_from_content(content_type, webpage_bytes):
359 m = re.match(r'[a-zA-Z0-9_.-]+/[a-zA-Z0-9_.-]+\s*;\s*charset=(.+)', content_type)
361 encoding = m.group(1)
363 m = re.search(br'<meta[^>]+charset=[\'"]?([^\'")]+)[ /\'">]',
364 webpage_bytes[:1024])
366 encoding = m.group(1).decode('ascii')
367 elif webpage_bytes.startswith(b'\xff\xfe'):
374 def _webpage_read_content(self, urlh, url_or_request, video_id, note=None, errnote=None, fatal=True, prefix=None, encoding=None):
375 content_type = urlh.headers.get('Content-Type', '')
376 webpage_bytes = urlh.read()
377 if prefix is not None:
378 webpage_bytes = prefix + webpage_bytes
380 encoding = self._guess_encoding_from_content(content_type, webpage_bytes)
381 if self._downloader.params.get('dump_intermediate_pages', False):
383 url = url_or_request.get_full_url()
384 except AttributeError:
386 self.to_screen('Dumping request to ' + url)
387 dump = base64.b64encode(webpage_bytes).decode('ascii')
388 self._downloader.to_screen(dump)
389 if self._downloader.params.get('write_pages', False):
391 url = url_or_request.get_full_url()
392 except AttributeError:
394 basen = '%s_%s' % (video_id, url)
396 h = '___' + hashlib.md5(basen.encode('utf-8')).hexdigest()
397 basen = basen[:240 - len(h)] + h
398 raw_filename = basen + '.dump'
399 filename = sanitize_filename(raw_filename, restricted=True)
400 self.to_screen('Saving request to ' + filename)
401 # Working around MAX_PATH limitation on Windows (see
402 # http://msdn.microsoft.com/en-us/library/windows/desktop/aa365247(v=vs.85).aspx)
404 absfilepath = os.path.abspath(filename)
405 if len(absfilepath) > 259:
406 filename = '\\\\?\\' + absfilepath
407 with open(filename, 'wb') as outf:
408 outf.write(webpage_bytes)
411 content = webpage_bytes.decode(encoding, 'replace')
413 content = webpage_bytes.decode('utf-8', 'replace')
415 if ('<title>Access to this site is blocked</title>' in content and
416 'Websense' in content[:512]):
417 msg = 'Access to this webpage has been blocked by Websense filtering software in your network.'
418 blocked_iframe = self._html_search_regex(
419 r'<iframe src="([^"]+)"', content,
420 'Websense information URL', default=None)
422 msg += ' Visit %s for more details' % blocked_iframe
423 raise ExtractorError(msg, expected=True)
424 if '<title>The URL you requested has been blocked</title>' in content[:512]:
426 'Access to this webpage has been blocked by Indian censorship. '
427 'Use a VPN or proxy server (with --proxy) to route around it.')
428 block_msg = self._html_search_regex(
429 r'</h1><p>(.*?)</p>',
430 content, 'block message', default=None)
432 msg += ' (Message: "%s")' % block_msg.replace('\n', ' ')
433 raise ExtractorError(msg, expected=True)
437 def _download_webpage(self, url_or_request, video_id, note=None, errnote=None, fatal=True, tries=1, timeout=5, encoding=None):
438 """ Returns the data of the page as a string """
441 while success is False:
443 res = self._download_webpage_handle(url_or_request, video_id, note, errnote, fatal, encoding=encoding)
445 except compat_http_client.IncompleteRead as e:
447 if try_count >= tries:
449 self._sleep(timeout, video_id)
456 def _download_xml(self, url_or_request, video_id,
457 note='Downloading XML', errnote='Unable to download XML',
458 transform_source=None, fatal=True, encoding=None):
459 """Return the xml as an xml.etree.ElementTree.Element"""
460 xml_string = self._download_webpage(
461 url_or_request, video_id, note, errnote, fatal=fatal, encoding=encoding)
462 if xml_string is False:
465 xml_string = transform_source(xml_string)
466 return compat_etree_fromstring(xml_string.encode('utf-8'))
468 def _download_json(self, url_or_request, video_id,
469 note='Downloading JSON metadata',
470 errnote='Unable to download JSON metadata',
471 transform_source=None,
472 fatal=True, encoding=None):
473 json_string = self._download_webpage(
474 url_or_request, video_id, note, errnote, fatal=fatal,
476 if (not fatal) and json_string is False:
478 return self._parse_json(
479 json_string, video_id, transform_source=transform_source, fatal=fatal)
481 def _parse_json(self, json_string, video_id, transform_source=None, fatal=True):
483 json_string = transform_source(json_string)
485 return json.loads(json_string)
486 except ValueError as ve:
487 errmsg = '%s: Failed to parse JSON ' % video_id
489 raise ExtractorError(errmsg, cause=ve)
491 self.report_warning(errmsg + str(ve))
493 def report_warning(self, msg, video_id=None):
494 idstr = '' if video_id is None else '%s: ' % video_id
495 self._downloader.report_warning(
496 '[%s] %s%s' % (self.IE_NAME, idstr, msg))
498 def to_screen(self, msg):
499 """Print msg to screen, prefixing it with '[ie_name]'"""
500 self._downloader.to_screen('[%s] %s' % (self.IE_NAME, msg))
502 def report_extraction(self, id_or_name):
503 """Report information extraction."""
504 self.to_screen('%s: Extracting information' % id_or_name)
506 def report_download_webpage(self, video_id):
507 """Report webpage download."""
508 self.to_screen('%s: Downloading webpage' % video_id)
510 def report_age_confirmation(self):
511 """Report attempt to confirm age."""
512 self.to_screen('Confirming age')
514 def report_login(self):
515 """Report attempt to log in."""
516 self.to_screen('Logging in')
519 def raise_login_required(msg='This video is only available for registered users'):
520 raise ExtractorError(
521 '%s. Use --username and --password or --netrc to provide account credentials.' % msg,
525 def raise_geo_restricted(msg='This video is not available from your location due to geo restriction'):
526 raise ExtractorError(
527 '%s. You might want to use --proxy to workaround.' % msg,
530 # Methods for following #608
532 def url_result(url, ie=None, video_id=None, video_title=None):
533 """Returns a URL that points to a page that should be processed"""
534 # TODO: ie should be the class used for getting the info
535 video_info = {'_type': 'url',
538 if video_id is not None:
539 video_info['id'] = video_id
540 if video_title is not None:
541 video_info['title'] = video_title
545 def playlist_result(entries, playlist_id=None, playlist_title=None, playlist_description=None):
546 """Returns a playlist"""
547 video_info = {'_type': 'playlist',
550 video_info['id'] = playlist_id
552 video_info['title'] = playlist_title
553 if playlist_description:
554 video_info['description'] = playlist_description
557 def _search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
559 Perform a regex search on the given string, using a single or a list of
560 patterns returning the first matching group.
561 In case of failure return a default value or raise a WARNING or a
562 RegexNotFoundError, depending on fatal, specifying the field name.
564 if isinstance(pattern, (str, compat_str, compiled_regex_type)):
565 mobj = re.search(pattern, string, flags)
568 mobj = re.search(p, string, flags)
572 if not self._downloader.params.get('no_color') and os.name != 'nt' and sys.stderr.isatty():
573 _name = '\033[0;34m%s\033[0m' % name
579 # return the first matching group
580 return next(g for g in mobj.groups() if g is not None)
582 return mobj.group(group)
583 elif default is not NO_DEFAULT:
586 raise RegexNotFoundError('Unable to extract %s' % _name)
588 self._downloader.report_warning('unable to extract %s' % _name + bug_reports_message())
591 def _html_search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
593 Like _search_regex, but strips HTML tags and unescapes entities.
595 res = self._search_regex(pattern, string, name, default, fatal, flags, group)
597 return clean_html(res).strip()
601 def _get_login_info(self):
603 Get the login info as (username, password)
604 It will look in the netrc file using the _NETRC_MACHINE value
605 If there's no info available, return (None, None)
607 if self._downloader is None:
612 downloader_params = self._downloader.params
614 # Attempt to use provided username and password or .netrc data
615 if downloader_params.get('username', None) is not None:
616 username = downloader_params['username']
617 password = downloader_params['password']
618 elif downloader_params.get('usenetrc', False):
620 info = netrc.netrc().authenticators(self._NETRC_MACHINE)
625 raise netrc.NetrcParseError('No authenticators for %s' % self._NETRC_MACHINE)
626 except (IOError, netrc.NetrcParseError) as err:
627 self._downloader.report_warning('parsing .netrc: %s' % error_to_compat_str(err))
629 return (username, password)
631 def _get_tfa_info(self, note='two-factor verification code'):
633 Get the two-factor authentication info
634 TODO - asking the user will be required for sms/phone verify
635 currently just uses the command line option
636 If there's no info available, return None
638 if self._downloader is None:
640 downloader_params = self._downloader.params
642 if downloader_params.get('twofactor', None) is not None:
643 return downloader_params['twofactor']
645 return compat_getpass('Type %s and press [Return]: ' % note)
647 # Helper functions for extracting OpenGraph info
649 def _og_regexes(prop):
650 content_re = r'content=(?:"([^"]+?)"|\'([^\']+?)\'|\s*([^\s"\'=<>`]+?))'
651 property_re = (r'(?:name|property)=(?:\'og:%(prop)s\'|"og:%(prop)s"|\s*og:%(prop)s\b)'
652 % {'prop': re.escape(prop)})
653 template = r'<meta[^>]+?%s[^>]+?%s'
655 template % (property_re, content_re),
656 template % (content_re, property_re),
660 def _meta_regex(prop):
661 return r'''(?isx)<meta
662 (?=[^>]+(?:itemprop|name|property|id|http-equiv)=(["\']?)%s\1)
663 [^>]+?content=(["\'])(?P<content>.*?)\2''' % re.escape(prop)
665 def _og_search_property(self, prop, html, name=None, **kargs):
667 name = 'OpenGraph %s' % prop
668 escaped = self._search_regex(self._og_regexes(prop), html, name, flags=re.DOTALL, **kargs)
671 return unescapeHTML(escaped)
673 def _og_search_thumbnail(self, html, **kargs):
674 return self._og_search_property('image', html, 'thumbnail URL', fatal=False, **kargs)
676 def _og_search_description(self, html, **kargs):
677 return self._og_search_property('description', html, fatal=False, **kargs)
679 def _og_search_title(self, html, **kargs):
680 return self._og_search_property('title', html, **kargs)
682 def _og_search_video_url(self, html, name='video url', secure=True, **kargs):
683 regexes = self._og_regexes('video') + self._og_regexes('video:url')
685 regexes = self._og_regexes('video:secure_url') + regexes
686 return self._html_search_regex(regexes, html, name, **kargs)
688 def _og_search_url(self, html, **kargs):
689 return self._og_search_property('url', html, **kargs)
691 def _html_search_meta(self, name, html, display_name=None, fatal=False, **kwargs):
692 if display_name is None:
694 return self._html_search_regex(
695 self._meta_regex(name),
696 html, display_name, fatal=fatal, group='content', **kwargs)
698 def _dc_search_uploader(self, html):
699 return self._html_search_meta('dc.creator', html, 'uploader')
701 def _rta_search(self, html):
702 # See http://www.rtalabel.org/index.php?content=howtofaq#single
703 if re.search(r'(?ix)<meta\s+name="rating"\s+'
704 r' content="RTA-5042-1996-1400-1577-RTA"',
709 def _media_rating_search(self, html):
710 # See http://www.tjg-designs.com/WP/metadata-code-examples-adding-metadata-to-your-web-pages/
711 rating = self._html_search_meta('rating', html)
723 return RATING_TABLE.get(rating.lower(), None)
725 def _family_friendly_search(self, html):
726 # See http://schema.org/VideoObject
727 family_friendly = self._html_search_meta('isFamilyFriendly', html)
729 if not family_friendly:
738 return RATING_TABLE.get(family_friendly.lower(), None)
740 def _twitter_search_player(self, html):
741 return self._html_search_meta('twitter:player', html,
742 'twitter card player')
745 def _hidden_inputs(html):
746 html = re.sub(r'<!--(?:(?!<!--).)*-->', '', html)
748 for input in re.findall(r'(?i)<input([^>]+)>', html):
749 if not re.search(r'type=(["\'])(?:hidden|submit)\1', input):
751 name = re.search(r'name=(["\'])(?P<value>.+?)\1', input)
754 value = re.search(r'value=(["\'])(?P<value>.*?)\1', input)
757 hidden_inputs[name.group('value')] = value.group('value')
760 def _form_hidden_inputs(self, form_id, html):
761 form = self._search_regex(
762 r'(?is)<form[^>]+?id=(["\'])%s\1[^>]*>(?P<form>.+?)</form>' % form_id,
763 html, '%s form' % form_id, group='form')
764 return self._hidden_inputs(form)
766 def _sort_formats(self, formats, field_preference=None):
768 raise ExtractorError('No video formats found')
771 # TODO remove the following workaround
772 from ..utils import determine_ext
773 if not f.get('ext') and 'url' in f:
774 f['ext'] = determine_ext(f['url'])
776 if isinstance(field_preference, (list, tuple)):
777 return tuple(f.get(field) if f.get(field) is not None else -1 for field in field_preference)
779 preference = f.get('preference')
780 if preference is None:
781 proto = f.get('protocol')
783 proto = compat_urllib_parse_urlparse(f.get('url', '')).scheme
785 preference = 0 if proto in ['http', 'https'] else -0.1
786 if f.get('ext') in ['f4f', 'f4m']: # Not yet supported
789 if f.get('vcodec') == 'none': # audio only
790 if self._downloader.params.get('prefer_free_formats'):
791 ORDER = ['aac', 'mp3', 'm4a', 'webm', 'ogg', 'opus']
793 ORDER = ['webm', 'opus', 'ogg', 'mp3', 'aac', 'm4a']
796 audio_ext_preference = ORDER.index(f['ext'])
798 audio_ext_preference = -1
800 if self._downloader.params.get('prefer_free_formats'):
801 ORDER = ['flv', 'mp4', 'webm']
803 ORDER = ['webm', 'flv', 'mp4']
805 ext_preference = ORDER.index(f['ext'])
808 audio_ext_preference = 0
812 f.get('language_preference') if f.get('language_preference') is not None else -1,
813 f.get('quality') if f.get('quality') is not None else -1,
814 f.get('tbr') if f.get('tbr') is not None else -1,
815 f.get('filesize') if f.get('filesize') is not None else -1,
816 f.get('vbr') if f.get('vbr') is not None else -1,
817 f.get('height') if f.get('height') is not None else -1,
818 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 f4m_formats = self._extract_f4m_formats(
915 manifest_url, video_id, preference, f4m_id, fatal=fatal)
917 formats.extend(f4m_formats)
919 tbr = int_or_none(media_el.attrib.get('bitrate'))
921 'format_id': '-'.join(filter(None, [f4m_id, compat_str(i if tbr is None else tbr)])),
925 'width': int_or_none(media_el.attrib.get('width')),
926 'height': int_or_none(media_el.attrib.get('height')),
927 'preference': preference,
929 self._sort_formats(formats)
933 def _extract_m3u8_formats(self, m3u8_url, video_id, ext=None,
934 entry_protocol='m3u8', preference=None,
935 m3u8_id=None, note=None, errnote=None,
939 'format_id': '-'.join(filter(None, [m3u8_id, 'meta'])),
943 'preference': preference - 1 if preference else -1,
944 'resolution': 'multiple',
945 'format_note': 'Quality selection URL',
948 format_url = lambda u: (
950 if re.match(r'^https?://', u)
951 else compat_urlparse.urljoin(m3u8_url, u))
953 res = self._download_webpage_handle(
955 note=note or 'Downloading m3u8 information',
956 errnote=errnote or 'Failed to download m3u8 information',
961 m3u8_url = urlh.geturl()
965 r'(?P<key>[a-zA-Z_-]+)=(?P<val>"[^"]+"|[^",]+)(?:,|$)')
966 for line in m3u8_doc.splitlines():
967 if line.startswith('#EXT-X-STREAM-INF:'):
969 for m in kv_rex.finditer(line):
971 if v.startswith('"'):
973 last_info[m.group('key')] = v
974 elif line.startswith('#EXT-X-MEDIA:'):
976 for m in kv_rex.finditer(line):
978 if v.startswith('"'):
980 last_media[m.group('key')] = v
981 elif line.startswith('#') or not line.strip():
984 if last_info is None:
985 formats.append({'url': format_url(line)})
987 tbr = int_or_none(last_info.get('BANDWIDTH'), scale=1000)
990 format_id.append(m3u8_id)
991 last_media_name = last_media.get('NAME') if last_media and last_media.get('TYPE') != 'SUBTITLES' else None
992 format_id.append(last_media_name if last_media_name else '%d' % (tbr if tbr else len(formats)))
994 'format_id': '-'.join(format_id),
995 'url': format_url(line.strip()),
998 'protocol': entry_protocol,
999 'preference': preference,
1001 codecs = last_info.get('CODECS')
1003 # TODO: looks like video codec is not always necessarily goes first
1004 va_codecs = codecs.split(',')
1006 f['vcodec'] = va_codecs[0].partition('.')[0]
1007 if len(va_codecs) > 1 and va_codecs[1]:
1008 f['acodec'] = va_codecs[1].partition('.')[0]
1009 resolution = last_info.get('RESOLUTION')
1011 width_str, height_str = resolution.split('x')
1012 f['width'] = int(width_str)
1013 f['height'] = int(height_str)
1014 if last_media is not None:
1015 f['m3u8_media'] = last_media
1019 self._sort_formats(formats)
1023 def _xpath_ns(path, namespace=None):
1027 for c in path.split('/'):
1028 if not c or c == '.':
1031 out.append('{%s}%s' % (namespace, c))
1032 return '/'.join(out)
1034 def _extract_smil_formats(self, smil_url, video_id, fatal=True, f4m_params=None):
1035 smil = self._download_smil(smil_url, video_id, fatal=fatal)
1041 namespace = self._parse_smil_namespace(smil)
1043 return self._parse_smil_formats(
1044 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
1046 def _extract_smil_info(self, smil_url, video_id, fatal=True, f4m_params=None):
1047 smil = self._download_smil(smil_url, video_id, fatal=fatal)
1050 return self._parse_smil(smil, smil_url, video_id, f4m_params=f4m_params)
1052 def _download_smil(self, smil_url, video_id, fatal=True):
1053 return self._download_xml(
1054 smil_url, video_id, 'Downloading SMIL file',
1055 'Unable to download SMIL file', fatal=fatal)
1057 def _parse_smil(self, smil, smil_url, video_id, f4m_params=None):
1058 namespace = self._parse_smil_namespace(smil)
1060 formats = self._parse_smil_formats(
1061 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
1062 subtitles = self._parse_smil_subtitles(smil, namespace=namespace)
1064 video_id = os.path.splitext(url_basename(smil_url))[0]
1068 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
1069 name = meta.attrib.get('name')
1070 content = meta.attrib.get('content')
1071 if not name or not content:
1073 if not title and name == 'title':
1075 elif not description and name in ('description', 'abstract'):
1076 description = content
1077 elif not upload_date and name == 'date':
1078 upload_date = unified_strdate(content)
1081 'id': image.get('type'),
1082 'url': image.get('src'),
1083 'width': int_or_none(image.get('width')),
1084 'height': int_or_none(image.get('height')),
1085 } for image in smil.findall(self._xpath_ns('.//image', namespace)) if image.get('src')]
1089 'title': title or video_id,
1090 'description': description,
1091 'upload_date': upload_date,
1092 'thumbnails': thumbnails,
1094 'subtitles': subtitles,
1097 def _parse_smil_namespace(self, smil):
1098 return self._search_regex(
1099 r'(?i)^{([^}]+)?}smil$', smil.tag, 'namespace', default=None)
1101 def _parse_smil_formats(self, smil, smil_url, video_id, namespace=None, f4m_params=None, transform_rtmp_url=None):
1103 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
1104 b = meta.get('base') or meta.get('httpBase')
1113 videos = smil.findall(self._xpath_ns('.//video', namespace))
1114 for video in videos:
1115 src = video.get('src')
1119 bitrate = float_or_none(video.get('system-bitrate') or video.get('systemBitrate'), 1000)
1120 filesize = int_or_none(video.get('size') or video.get('fileSize'))
1121 width = int_or_none(video.get('width'))
1122 height = int_or_none(video.get('height'))
1123 proto = video.get('proto')
1124 ext = video.get('ext')
1125 src_ext = determine_ext(src)
1126 streamer = video.get('streamer') or base
1128 if proto == 'rtmp' or streamer.startswith('rtmp'):
1134 'format_id': 'rtmp-%d' % (rtmp_count if bitrate is None else bitrate),
1136 'filesize': filesize,
1140 if transform_rtmp_url:
1141 streamer, src = transform_rtmp_url(streamer, src)
1142 formats[-1].update({
1148 src_url = src if src.startswith('http') else compat_urlparse.urljoin(base, src)
1150 if proto == 'm3u8' or src_ext == 'm3u8':
1151 m3u8_formats = self._extract_m3u8_formats(
1152 src_url, video_id, ext or 'mp4', m3u8_id='hls', fatal=False)
1154 formats.extend(m3u8_formats)
1157 if src_ext == 'f4m':
1162 'plugin': 'flowplayer-3.2.0.1',
1164 f4m_url += '&' if '?' in f4m_url else '?'
1165 f4m_url += compat_urllib_parse.urlencode(f4m_params)
1166 f4m_formats = self._extract_f4m_formats(f4m_url, video_id, f4m_id='hds', fatal=False)
1168 formats.extend(f4m_formats)
1171 if src_url.startswith('http') and self._is_valid_url(src, video_id):
1175 'ext': ext or src_ext or 'flv',
1176 'format_id': 'http-%d' % (bitrate or http_count),
1178 'filesize': filesize,
1184 self._sort_formats(formats)
1188 def _parse_smil_subtitles(self, smil, namespace=None, subtitles_lang='en'):
1190 for num, textstream in enumerate(smil.findall(self._xpath_ns('.//textstream', namespace))):
1191 src = textstream.get('src')
1194 ext = textstream.get('ext') or determine_ext(src)
1196 type_ = textstream.get('type')
1200 'application/smptett+xml': 'tt',
1202 if type_ in SUBTITLES_TYPES:
1203 ext = SUBTITLES_TYPES[type_]
1204 lang = textstream.get('systemLanguage') or textstream.get('systemLanguageName') or textstream.get('lang') or subtitles_lang
1205 subtitles.setdefault(lang, []).append({
1211 def _extract_xspf_playlist(self, playlist_url, playlist_id, fatal=True):
1212 xspf = self._download_xml(
1213 playlist_url, playlist_id, 'Downloading xpsf playlist',
1214 'Unable to download xspf manifest', fatal=fatal)
1217 return self._parse_xspf(xspf, playlist_id)
1219 def _parse_xspf(self, playlist, playlist_id):
1221 'xspf': 'http://xspf.org/ns/0/',
1222 's1': 'http://static.streamone.nl/player/ns/0',
1226 for track in playlist.findall(xpath_with_ns('./xspf:trackList/xspf:track', NS_MAP)):
1228 track, xpath_with_ns('./xspf:title', NS_MAP), 'title', default=playlist_id)
1229 description = xpath_text(
1230 track, xpath_with_ns('./xspf:annotation', NS_MAP), 'description')
1231 thumbnail = xpath_text(
1232 track, xpath_with_ns('./xspf:image', NS_MAP), 'thumbnail')
1233 duration = float_or_none(
1234 xpath_text(track, xpath_with_ns('./xspf:duration', NS_MAP), 'duration'), 1000)
1237 'url': location.text,
1238 'format_id': location.get(xpath_with_ns('s1:label', NS_MAP)),
1239 'width': int_or_none(location.get(xpath_with_ns('s1:width', NS_MAP))),
1240 'height': int_or_none(location.get(xpath_with_ns('s1:height', NS_MAP))),
1241 } for location in track.findall(xpath_with_ns('./xspf:location', NS_MAP))]
1242 self._sort_formats(formats)
1247 'description': description,
1248 'thumbnail': thumbnail,
1249 'duration': duration,
1254 def _live_title(self, name):
1255 """ Generate the title for a live video """
1256 now = datetime.datetime.now()
1257 now_str = now.strftime("%Y-%m-%d %H:%M")
1258 return name + ' ' + now_str
1260 def _int(self, v, name, fatal=False, **kwargs):
1261 res = int_or_none(v, **kwargs)
1262 if 'get_attr' in kwargs:
1263 print(getattr(v, kwargs['get_attr']))
1265 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
1267 raise ExtractorError(msg)
1269 self._downloader.report_warning(msg)
1272 def _float(self, v, name, fatal=False, **kwargs):
1273 res = float_or_none(v, **kwargs)
1275 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
1277 raise ExtractorError(msg)
1279 self._downloader.report_warning(msg)
1282 def _set_cookie(self, domain, name, value, expire_time=None):
1283 cookie = compat_cookiejar.Cookie(
1284 0, name, value, None, None, domain, None,
1285 None, '/', True, False, expire_time, '', None, None, None)
1286 self._downloader.cookiejar.set_cookie(cookie)
1288 def _get_cookies(self, url):
1289 """ Return a compat_cookies.SimpleCookie with the cookies for the url """
1290 req = sanitized_Request(url)
1291 self._downloader.cookiejar.add_cookie_header(req)
1292 return compat_cookies.SimpleCookie(req.get_header('Cookie'))
1294 def get_testcases(self, include_onlymatching=False):
1295 t = getattr(self, '_TEST', None)
1297 assert not hasattr(self, '_TESTS'), \
1298 '%s has _TEST and _TESTS' % type(self).__name__
1301 tests = getattr(self, '_TESTS', [])
1303 if not include_onlymatching and t.get('only_matching', False):
1305 t['name'] = type(self).__name__[:-len('IE')]
1308 def is_suitable(self, age_limit):
1309 """ Test whether the extractor is generally suitable for the given
1310 age limit (i.e. pornographic sites are not, all others usually are) """
1312 any_restricted = False
1313 for tc in self.get_testcases(include_onlymatching=False):
1314 if 'playlist' in tc:
1315 tc = tc['playlist'][0]
1316 is_restricted = age_restricted(
1317 tc.get('info_dict', {}).get('age_limit'), age_limit)
1318 if not is_restricted:
1320 any_restricted = any_restricted or is_restricted
1321 return not any_restricted
1323 def extract_subtitles(self, *args, **kwargs):
1324 if (self._downloader.params.get('writesubtitles', False) or
1325 self._downloader.params.get('listsubtitles')):
1326 return self._get_subtitles(*args, **kwargs)
1329 def _get_subtitles(self, *args, **kwargs):
1330 raise NotImplementedError("This method must be implemented by subclasses")
1333 def _merge_subtitle_items(subtitle_list1, subtitle_list2):
1334 """ Merge subtitle items for one language. Items with duplicated URLs
1335 will be dropped. """
1336 list1_urls = set([item['url'] for item in subtitle_list1])
1337 ret = list(subtitle_list1)
1338 ret.extend([item for item in subtitle_list2 if item['url'] not in list1_urls])
1342 def _merge_subtitles(cls, subtitle_dict1, subtitle_dict2):
1343 """ Merge two subtitle dictionaries, language by language. """
1344 ret = dict(subtitle_dict1)
1345 for lang in subtitle_dict2:
1346 ret[lang] = cls._merge_subtitle_items(subtitle_dict1.get(lang, []), subtitle_dict2[lang])
1349 def extract_automatic_captions(self, *args, **kwargs):
1350 if (self._downloader.params.get('writeautomaticsub', False) or
1351 self._downloader.params.get('listsubtitles')):
1352 return self._get_automatic_captions(*args, **kwargs)
1355 def _get_automatic_captions(self, *args, **kwargs):
1356 raise NotImplementedError("This method must be implemented by subclasses")
1359 class SearchInfoExtractor(InfoExtractor):
1361 Base class for paged search queries extractors.
1362 They accept URLs in the format _SEARCH_KEY(|all|[0-9]):{query}
1363 Instances should define _SEARCH_KEY and _MAX_RESULTS.
1367 def _make_valid_url(cls):
1368 return r'%s(?P<prefix>|[1-9][0-9]*|all):(?P<query>[\s\S]+)' % cls._SEARCH_KEY
1371 def suitable(cls, url):
1372 return re.match(cls._make_valid_url(), url) is not None
1374 def _real_extract(self, query):
1375 mobj = re.match(self._make_valid_url(), query)
1377 raise ExtractorError('Invalid search query "%s"' % query)
1379 prefix = mobj.group('prefix')
1380 query = mobj.group('query')
1382 return self._get_n_results(query, 1)
1383 elif prefix == 'all':
1384 return self._get_n_results(query, self._MAX_RESULTS)
1388 raise ExtractorError('invalid download number %s for query "%s"' % (n, query))
1389 elif n > self._MAX_RESULTS:
1390 self._downloader.report_warning('%s returns max %i results (you requested %i)' % (self._SEARCH_KEY, self._MAX_RESULTS, n))
1391 n = self._MAX_RESULTS
1392 return self._get_n_results(query, n)
1394 def _get_n_results(self, query, n):
1395 """Get a specified number of results for a query"""
1396 raise NotImplementedError("This method must be implemented by subclasses")
1399 def SEARCH_KEY(self):
1400 return self._SEARCH_KEY