1 from __future__ import unicode_literals
13 import xml.etree.ElementTree
15 from ..compat import (
23 compat_urllib_parse_urlparse,
24 compat_urllib_request,
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 automatic_captions: Like 'subtitles', used by the YoutubeIE for
169 automatically generated captions
170 duration: Length of the video in seconds, as an integer.
171 view_count: How many users have watched the video on the platform.
172 like_count: Number of positive ratings of the video
173 dislike_count: Number of negative ratings of the video
174 average_rating: Average rating give by users, the scale used depends on the webpage
175 comment_count: Number of comments on the video
176 comments: A list of comments, each with one or more of the following
177 properties (all but one of text or html optional):
178 * "author" - human-readable name of the comment author
179 * "author_id" - user ID of the comment author
181 * "html" - Comment as HTML
182 * "text" - Plain text of the comment
183 * "timestamp" - UNIX timestamp of comment
184 * "parent" - ID of the comment this one is replying to.
185 Set to "root" to indicate that this is a
186 comment to the original video.
187 age_limit: Age restriction for the video, as an integer (years)
188 webpage_url: The URL to the video webpage, if given to youtube-dl it
189 should allow to get the same result again. (It will be set
190 by YoutubeDL if it's missing)
191 categories: A list of categories that the video falls in, for example
193 tags: A list of tags assigned to the video, e.g. ["sweden", "pop music"]
194 is_live: True, False, or None (=unknown). Whether this video is a
195 live stream that goes on instead of a fixed-length video.
196 start_time: Time in seconds where the reproduction should start, as
197 specified in the URL.
198 end_time: Time in seconds where the reproduction should end, as
199 specified in the URL.
201 Unless mentioned otherwise, the fields should be Unicode strings.
203 Unless mentioned otherwise, None is equivalent to absence of information.
206 _type "playlist" indicates multiple videos.
207 There must be a key "entries", which is a list, an iterable, or a PagedList
208 object, each element of which is a valid dictionary by this specification.
210 Additionally, playlists can have "title", "description" and "id" attributes
211 with the same semantics as videos (see above).
214 _type "multi_video" indicates that there are multiple videos that
215 form a single show, for examples multiple acts of an opera or TV episode.
216 It must have an entries key like a playlist and contain all the keys
217 required for a video at the same time.
220 _type "url" indicates that the video must be extracted from another
221 location, possibly by a different extractor. Its only required key is:
222 "url" - the next URL to extract.
223 The key "ie_key" can be set to the class name (minus the trailing "IE",
224 e.g. "Youtube") if the extractor class is known in advance.
225 Additionally, the dictionary may have any properties of the resolved entity
226 known in advance, for example "title" if the title of the referred video is
230 _type "url_transparent" entities have the same specification as "url", but
231 indicate that the given additional information is more precise than the one
232 associated with the resolved URL.
233 This is useful when a site employs a video service that hosts the video and
234 its technical metadata, but that video service does not embed a useful
235 title, description etc.
238 Subclasses of this one should re-define the _real_initialize() and
239 _real_extract() methods and define a _VALID_URL regexp.
240 Probably, they should also be added to the list of extractors.
242 Finally, the _WORKING attribute should be set to False for broken IEs
243 in order to warn the users and skip the tests.
250 def __init__(self, downloader=None):
251 """Constructor. Receives an optional downloader."""
253 self.set_downloader(downloader)
256 def suitable(cls, url):
257 """Receives a URL and returns True if suitable for this IE."""
259 # This does not use has/getattr intentionally - we want to know whether
260 # we have cached the regexp for *this* class, whereas getattr would also
261 # match the superclass
262 if '_VALID_URL_RE' not in cls.__dict__:
263 cls._VALID_URL_RE = re.compile(cls._VALID_URL)
264 return cls._VALID_URL_RE.match(url) is not None
267 def _match_id(cls, url):
268 if '_VALID_URL_RE' not in cls.__dict__:
269 cls._VALID_URL_RE = re.compile(cls._VALID_URL)
270 m = cls._VALID_URL_RE.match(url)
276 """Getter method for _WORKING."""
279 def initialize(self):
280 """Initializes an instance (authentication, etc)."""
282 self._real_initialize()
285 def extract(self, url):
286 """Extracts URL information and returns it in list of dicts."""
289 return self._real_extract(url)
290 except ExtractorError:
292 except compat_http_client.IncompleteRead as e:
293 raise ExtractorError('A network error has occured.', cause=e, expected=True)
294 except (KeyError, StopIteration) as e:
295 raise ExtractorError('An extractor error has occured.', cause=e)
297 def set_downloader(self, downloader):
298 """Sets the downloader for this IE."""
299 self._downloader = downloader
301 def _real_initialize(self):
302 """Real initialization process. Redefine in subclasses."""
305 def _real_extract(self, url):
306 """Real extraction process. Redefine in subclasses."""
311 """A string for getting the InfoExtractor with get_info_extractor"""
312 return cls.__name__[:-2]
316 return type(self).__name__[:-2]
318 def _request_webpage(self, url_or_request, video_id, note=None, errnote=None, fatal=True):
319 """ Returns the response handle """
321 self.report_download_webpage(video_id)
322 elif note is not False:
324 self.to_screen('%s' % (note,))
326 self.to_screen('%s: %s' % (video_id, note))
328 return self._downloader.urlopen(url_or_request)
329 except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err:
333 errnote = 'Unable to download webpage'
334 errmsg = '%s: %s' % (errnote, compat_str(err))
336 raise ExtractorError(errmsg, sys.exc_info()[2], cause=err)
338 self._downloader.report_warning(errmsg)
341 def _download_webpage_handle(self, url_or_request, video_id, note=None, errnote=None, fatal=True, encoding=None):
342 """ Returns a tuple (page content as string, URL handle) """
343 # Strip hashes from the URL (#1038)
344 if isinstance(url_or_request, (compat_str, str)):
345 url_or_request = url_or_request.partition('#')[0]
347 urlh = self._request_webpage(url_or_request, video_id, note, errnote, fatal)
351 content = self._webpage_read_content(urlh, url_or_request, video_id, note, errnote, fatal, encoding=encoding)
352 return (content, urlh)
355 def _guess_encoding_from_content(content_type, webpage_bytes):
356 m = re.match(r'[a-zA-Z0-9_.-]+/[a-zA-Z0-9_.-]+\s*;\s*charset=(.+)', content_type)
358 encoding = m.group(1)
360 m = re.search(br'<meta[^>]+charset=[\'"]?([^\'")]+)[ /\'">]',
361 webpage_bytes[:1024])
363 encoding = m.group(1).decode('ascii')
364 elif webpage_bytes.startswith(b'\xff\xfe'):
371 def _webpage_read_content(self, urlh, url_or_request, video_id, note=None, errnote=None, fatal=True, prefix=None, encoding=None):
372 content_type = urlh.headers.get('Content-Type', '')
373 webpage_bytes = urlh.read()
374 if prefix is not None:
375 webpage_bytes = prefix + webpage_bytes
377 encoding = self._guess_encoding_from_content(content_type, webpage_bytes)
378 if self._downloader.params.get('dump_intermediate_pages', False):
380 url = url_or_request.get_full_url()
381 except AttributeError:
383 self.to_screen('Dumping request to ' + url)
384 dump = base64.b64encode(webpage_bytes).decode('ascii')
385 self._downloader.to_screen(dump)
386 if self._downloader.params.get('write_pages', False):
388 url = url_or_request.get_full_url()
389 except AttributeError:
391 basen = '%s_%s' % (video_id, url)
393 h = '___' + hashlib.md5(basen.encode('utf-8')).hexdigest()
394 basen = basen[:240 - len(h)] + h
395 raw_filename = basen + '.dump'
396 filename = sanitize_filename(raw_filename, restricted=True)
397 self.to_screen('Saving request to ' + filename)
398 # Working around MAX_PATH limitation on Windows (see
399 # http://msdn.microsoft.com/en-us/library/windows/desktop/aa365247(v=vs.85).aspx)
401 absfilepath = os.path.abspath(filename)
402 if len(absfilepath) > 259:
403 filename = '\\\\?\\' + absfilepath
404 with open(filename, 'wb') as outf:
405 outf.write(webpage_bytes)
408 content = webpage_bytes.decode(encoding, 'replace')
410 content = webpage_bytes.decode('utf-8', 'replace')
412 if ('<title>Access to this site is blocked</title>' in content and
413 'Websense' in content[:512]):
414 msg = 'Access to this webpage has been blocked by Websense filtering software in your network.'
415 blocked_iframe = self._html_search_regex(
416 r'<iframe src="([^"]+)"', content,
417 'Websense information URL', default=None)
419 msg += ' Visit %s for more details' % blocked_iframe
420 raise ExtractorError(msg, expected=True)
421 if '<title>The URL you requested has been blocked</title>' in content[:512]:
423 'Access to this webpage has been blocked by Indian censorship. '
424 'Use a VPN or proxy server (with --proxy) to route around it.')
425 block_msg = self._html_search_regex(
426 r'</h1><p>(.*?)</p>',
427 content, 'block message', default=None)
429 msg += ' (Message: "%s")' % block_msg.replace('\n', ' ')
430 raise ExtractorError(msg, expected=True)
434 def _download_webpage(self, url_or_request, video_id, note=None, errnote=None, fatal=True, tries=1, timeout=5, encoding=None):
435 """ Returns the data of the page as a string """
438 while success is False:
440 res = self._download_webpage_handle(url_or_request, video_id, note, errnote, fatal, encoding=encoding)
442 except compat_http_client.IncompleteRead as e:
444 if try_count >= tries:
446 self._sleep(timeout, video_id)
453 def _download_xml(self, url_or_request, video_id,
454 note='Downloading XML', errnote='Unable to download XML',
455 transform_source=None, fatal=True, encoding=None):
456 """Return the xml as an xml.etree.ElementTree.Element"""
457 xml_string = self._download_webpage(
458 url_or_request, video_id, note, errnote, fatal=fatal, encoding=encoding)
459 if xml_string is False:
462 xml_string = transform_source(xml_string)
463 return xml.etree.ElementTree.fromstring(xml_string.encode('utf-8'))
465 def _download_json(self, url_or_request, video_id,
466 note='Downloading JSON metadata',
467 errnote='Unable to download JSON metadata',
468 transform_source=None,
469 fatal=True, encoding=None):
470 json_string = self._download_webpage(
471 url_or_request, video_id, note, errnote, fatal=fatal,
473 if (not fatal) and json_string is False:
475 return self._parse_json(
476 json_string, video_id, transform_source=transform_source, fatal=fatal)
478 def _parse_json(self, json_string, video_id, transform_source=None, fatal=True):
480 json_string = transform_source(json_string)
482 return json.loads(json_string)
483 except ValueError as ve:
484 errmsg = '%s: Failed to parse JSON ' % video_id
486 raise ExtractorError(errmsg, cause=ve)
488 self.report_warning(errmsg + str(ve))
490 def report_warning(self, msg, video_id=None):
491 idstr = '' if video_id is None else '%s: ' % video_id
492 self._downloader.report_warning(
493 '[%s] %s%s' % (self.IE_NAME, idstr, msg))
495 def to_screen(self, msg):
496 """Print msg to screen, prefixing it with '[ie_name]'"""
497 self._downloader.to_screen('[%s] %s' % (self.IE_NAME, msg))
499 def report_extraction(self, id_or_name):
500 """Report information extraction."""
501 self.to_screen('%s: Extracting information' % id_or_name)
503 def report_download_webpage(self, video_id):
504 """Report webpage download."""
505 self.to_screen('%s: Downloading webpage' % video_id)
507 def report_age_confirmation(self):
508 """Report attempt to confirm age."""
509 self.to_screen('Confirming age')
511 def report_login(self):
512 """Report attempt to log in."""
513 self.to_screen('Logging in')
516 def raise_login_required(msg='This video is only available for registered users'):
517 raise ExtractorError(
518 '%s. Use --username and --password or --netrc to provide account credentials.' % msg,
522 def raise_geo_restricted(msg='This video is not available from your location due to geo restriction'):
523 raise ExtractorError(
524 '%s. You might want to use --proxy to workaround.' % msg,
527 # Methods for following #608
529 def url_result(url, ie=None, video_id=None, video_title=None):
530 """Returns a URL that points to a page that should be processed"""
531 # TODO: ie should be the class used for getting the info
532 video_info = {'_type': 'url',
535 if video_id is not None:
536 video_info['id'] = video_id
537 if video_title is not None:
538 video_info['title'] = video_title
542 def playlist_result(entries, playlist_id=None, playlist_title=None, playlist_description=None):
543 """Returns a playlist"""
544 video_info = {'_type': 'playlist',
547 video_info['id'] = playlist_id
549 video_info['title'] = playlist_title
550 if playlist_description:
551 video_info['description'] = playlist_description
554 def _search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
556 Perform a regex search on the given string, using a single or a list of
557 patterns returning the first matching group.
558 In case of failure return a default value or raise a WARNING or a
559 RegexNotFoundError, depending on fatal, specifying the field name.
561 if isinstance(pattern, (str, compat_str, compiled_regex_type)):
562 mobj = re.search(pattern, string, flags)
565 mobj = re.search(p, string, flags)
569 if not self._downloader.params.get('no_color') and os.name != 'nt' and sys.stderr.isatty():
570 _name = '\033[0;34m%s\033[0m' % name
576 # return the first matching group
577 return next(g for g in mobj.groups() if g is not None)
579 return mobj.group(group)
580 elif default is not NO_DEFAULT:
583 raise RegexNotFoundError('Unable to extract %s' % _name)
585 self._downloader.report_warning('unable to extract %s' % _name + bug_reports_message())
588 def _html_search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
590 Like _search_regex, but strips HTML tags and unescapes entities.
592 res = self._search_regex(pattern, string, name, default, fatal, flags, group)
594 return clean_html(res).strip()
598 def _get_login_info(self):
600 Get the login info as (username, password)
601 It will look in the netrc file using the _NETRC_MACHINE value
602 If there's no info available, return (None, None)
604 if self._downloader is None:
609 downloader_params = self._downloader.params
611 # Attempt to use provided username and password or .netrc data
612 if downloader_params.get('username', None) is not None:
613 username = downloader_params['username']
614 password = downloader_params['password']
615 elif downloader_params.get('usenetrc', False):
617 info = netrc.netrc().authenticators(self._NETRC_MACHINE)
622 raise netrc.NetrcParseError('No authenticators for %s' % self._NETRC_MACHINE)
623 except (IOError, netrc.NetrcParseError) as err:
624 self._downloader.report_warning('parsing .netrc: %s' % compat_str(err))
626 return (username, password)
628 def _get_tfa_info(self, note='two-factor verification code'):
630 Get the two-factor authentication info
631 TODO - asking the user will be required for sms/phone verify
632 currently just uses the command line option
633 If there's no info available, return None
635 if self._downloader is None:
637 downloader_params = self._downloader.params
639 if downloader_params.get('twofactor', None) is not None:
640 return downloader_params['twofactor']
642 return compat_getpass('Type %s and press [Return]: ' % note)
644 # Helper functions for extracting OpenGraph info
646 def _og_regexes(prop):
647 content_re = r'content=(?:"([^>]+?)"|\'([^>]+?)\')'
648 property_re = r'(?:name|property)=[\'"]og:%s[\'"]' % re.escape(prop)
649 template = r'<meta[^>]+?%s[^>]+?%s'
651 template % (property_re, content_re),
652 template % (content_re, property_re),
656 def _meta_regex(prop):
657 return r'''(?isx)<meta
658 (?=[^>]+(?:itemprop|name|property|id|http-equiv)=(["\']?)%s\1)
659 [^>]+?content=(["\'])(?P<content>.*?)\2''' % re.escape(prop)
661 def _og_search_property(self, prop, html, name=None, **kargs):
663 name = 'OpenGraph %s' % prop
664 escaped = self._search_regex(self._og_regexes(prop), html, name, flags=re.DOTALL, **kargs)
667 return unescapeHTML(escaped)
669 def _og_search_thumbnail(self, html, **kargs):
670 return self._og_search_property('image', html, 'thumbnail URL', fatal=False, **kargs)
672 def _og_search_description(self, html, **kargs):
673 return self._og_search_property('description', html, fatal=False, **kargs)
675 def _og_search_title(self, html, **kargs):
676 return self._og_search_property('title', html, **kargs)
678 def _og_search_video_url(self, html, name='video url', secure=True, **kargs):
679 regexes = self._og_regexes('video') + self._og_regexes('video:url')
681 regexes = self._og_regexes('video:secure_url') + regexes
682 return self._html_search_regex(regexes, html, name, **kargs)
684 def _og_search_url(self, html, **kargs):
685 return self._og_search_property('url', html, **kargs)
687 def _html_search_meta(self, name, html, display_name=None, fatal=False, **kwargs):
688 if display_name is None:
690 return self._html_search_regex(
691 self._meta_regex(name),
692 html, display_name, fatal=fatal, group='content', **kwargs)
694 def _dc_search_uploader(self, html):
695 return self._html_search_meta('dc.creator', html, 'uploader')
697 def _rta_search(self, html):
698 # See http://www.rtalabel.org/index.php?content=howtofaq#single
699 if re.search(r'(?ix)<meta\s+name="rating"\s+'
700 r' content="RTA-5042-1996-1400-1577-RTA"',
705 def _media_rating_search(self, html):
706 # See http://www.tjg-designs.com/WP/metadata-code-examples-adding-metadata-to-your-web-pages/
707 rating = self._html_search_meta('rating', html)
719 return RATING_TABLE.get(rating.lower(), None)
721 def _family_friendly_search(self, html):
722 # See http://schema.org/VideoObject
723 family_friendly = self._html_search_meta('isFamilyFriendly', html)
725 if not family_friendly:
734 return RATING_TABLE.get(family_friendly.lower(), None)
736 def _twitter_search_player(self, html):
737 return self._html_search_meta('twitter:player', html,
738 'twitter card player')
741 def _hidden_inputs(html):
742 html = re.sub(r'<!--(?:(?!<!--).)*-->', '', html)
744 for input in re.findall(r'(?i)<input([^>]+)>', html):
745 if not re.search(r'type=(["\'])(?:hidden|submit)\1', input):
747 name = re.search(r'name=(["\'])(?P<value>.+?)\1', input)
750 value = re.search(r'value=(["\'])(?P<value>.*?)\1', input)
753 hidden_inputs[name.group('value')] = value.group('value')
756 def _form_hidden_inputs(self, form_id, html):
757 form = self._search_regex(
758 r'(?is)<form[^>]+?id=(["\'])%s\1[^>]*>(?P<form>.+?)</form>' % form_id,
759 html, '%s form' % form_id, group='form')
760 return self._hidden_inputs(form)
762 def _sort_formats(self, formats, field_preference=None):
764 raise ExtractorError('No video formats found')
767 # TODO remove the following workaround
768 from ..utils import determine_ext
769 if not f.get('ext') and 'url' in f:
770 f['ext'] = determine_ext(f['url'])
772 if isinstance(field_preference, (list, tuple)):
773 return tuple(f.get(field) if f.get(field) is not None else -1 for field in field_preference)
775 preference = f.get('preference')
776 if preference is None:
777 proto = f.get('protocol')
779 proto = compat_urllib_parse_urlparse(f.get('url', '')).scheme
781 preference = 0 if proto in ['http', 'https'] else -0.1
782 if f.get('ext') in ['f4f', 'f4m']: # Not yet supported
785 if f.get('vcodec') == 'none': # audio only
786 if self._downloader.params.get('prefer_free_formats'):
787 ORDER = ['aac', 'mp3', 'm4a', 'webm', 'ogg', 'opus']
789 ORDER = ['webm', 'opus', 'ogg', 'mp3', 'aac', 'm4a']
792 audio_ext_preference = ORDER.index(f['ext'])
794 audio_ext_preference = -1
796 if self._downloader.params.get('prefer_free_formats'):
797 ORDER = ['flv', 'mp4', 'webm']
799 ORDER = ['webm', 'flv', 'mp4']
801 ext_preference = ORDER.index(f['ext'])
804 audio_ext_preference = 0
808 f.get('language_preference') if f.get('language_preference') is not None else -1,
809 f.get('quality') if f.get('quality') is not None else -1,
810 f.get('tbr') if f.get('tbr') is not None else -1,
811 f.get('filesize') if f.get('filesize') is not None else -1,
812 f.get('vbr') if f.get('vbr') is not None else -1,
813 f.get('height') if f.get('height') is not None else -1,
814 f.get('width') if f.get('width') is not None else -1,
816 f.get('abr') if f.get('abr') is not None else -1,
817 audio_ext_preference,
818 f.get('fps') if f.get('fps') is not None else -1,
819 f.get('filesize_approx') if f.get('filesize_approx') is not None else -1,
820 f.get('source_preference') if f.get('source_preference') is not None else -1,
821 f.get('format_id') if f.get('format_id') is not None else '',
823 formats.sort(key=_formats_key)
825 def _check_formats(self, formats, video_id):
828 lambda f: self._is_valid_url(
830 item='%s video format' % f.get('format_id') if f.get('format_id') else 'video'),
833 def _is_valid_url(self, url, video_id, item='video'):
834 url = self._proto_relative_url(url, scheme='http:')
835 # For now assume non HTTP(S) URLs always valid
836 if not (url.startswith('http://') or url.startswith('https://')):
839 self._request_webpage(url, video_id, 'Checking %s URL' % item)
841 except ExtractorError as e:
842 if isinstance(e.cause, compat_HTTPError):
844 '%s: %s URL is invalid, skipping' % (video_id, item))
848 def http_scheme(self):
849 """ Either "http:" or "https:", depending on the user's preferences """
852 if self._downloader.params.get('prefer_insecure', False)
855 def _proto_relative_url(self, url, scheme=None):
858 if url.startswith('//'):
860 scheme = self.http_scheme()
865 def _sleep(self, timeout, video_id, msg_template=None):
866 if msg_template is None:
867 msg_template = '%(video_id)s: Waiting for %(timeout)s seconds'
868 msg = msg_template % {'video_id': video_id, 'timeout': timeout}
872 def _extract_f4m_formats(self, manifest_url, video_id, preference=None, f4m_id=None,
873 transform_source=lambda s: fix_xml_ampersands(s).strip(),
875 manifest = self._download_xml(
876 manifest_url, video_id, 'Downloading f4m manifest',
877 'Unable to download f4m manifest',
878 # Some manifests may be malformed, e.g. prosiebensat1 generated manifests
879 # (see https://github.com/rg3/youtube-dl/issues/6215#issuecomment-121704244)
880 transform_source=transform_source,
883 if manifest is False:
887 manifest_version = '1.0'
888 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/1.0}media')
890 manifest_version = '2.0'
891 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/2.0}media')
892 for i, media_el in enumerate(media_nodes):
893 if manifest_version == '2.0':
894 media_url = media_el.attrib.get('href') or media_el.attrib.get('url')
898 media_url if media_url.startswith('http://') or media_url.startswith('https://')
899 else ('/'.join(manifest_url.split('/')[:-1]) + '/' + media_url))
900 # If media_url is itself a f4m manifest do the recursive extraction
901 # since bitrates in parent manifest (this one) and media_url manifest
902 # may differ leading to inability to resolve the format by requested
903 # bitrate in f4m downloader
904 if determine_ext(manifest_url) == 'f4m':
905 f4m_formats = self._extract_f4m_formats(
906 manifest_url, video_id, preference, f4m_id, fatal=fatal)
908 formats.extend(f4m_formats)
910 tbr = int_or_none(media_el.attrib.get('bitrate'))
912 'format_id': '-'.join(filter(None, [f4m_id, compat_str(i if tbr is None else tbr)])),
916 'width': int_or_none(media_el.attrib.get('width')),
917 'height': int_or_none(media_el.attrib.get('height')),
918 'preference': preference,
920 self._sort_formats(formats)
924 def _extract_m3u8_formats(self, m3u8_url, video_id, ext=None,
925 entry_protocol='m3u8', preference=None,
926 m3u8_id=None, note=None, errnote=None,
930 'format_id': '-'.join(filter(None, [m3u8_id, 'meta'])),
934 'preference': preference - 1 if preference else -1,
935 'resolution': 'multiple',
936 'format_note': 'Quality selection URL',
939 format_url = lambda u: (
941 if re.match(r'^https?://', u)
942 else compat_urlparse.urljoin(m3u8_url, u))
944 m3u8_doc = self._download_webpage(
946 note=note or 'Downloading m3u8 information',
947 errnote=errnote or 'Failed to download m3u8 information',
949 if m3u8_doc is False:
954 r'(?P<key>[a-zA-Z_-]+)=(?P<val>"[^"]+"|[^",]+)(?:,|$)')
955 for line in m3u8_doc.splitlines():
956 if line.startswith('#EXT-X-STREAM-INF:'):
958 for m in kv_rex.finditer(line):
960 if v.startswith('"'):
962 last_info[m.group('key')] = v
963 elif line.startswith('#EXT-X-MEDIA:'):
965 for m in kv_rex.finditer(line):
967 if v.startswith('"'):
969 last_media[m.group('key')] = v
970 elif line.startswith('#') or not line.strip():
973 if last_info is None:
974 formats.append({'url': format_url(line)})
976 tbr = int_or_none(last_info.get('BANDWIDTH'), scale=1000)
979 format_id.append(m3u8_id)
980 last_media_name = last_media.get('NAME') if last_media and last_media.get('TYPE') != 'SUBTITLES' else None
981 format_id.append(last_media_name if last_media_name else '%d' % (tbr if tbr else len(formats)))
983 'format_id': '-'.join(format_id),
984 'url': format_url(line.strip()),
987 'protocol': entry_protocol,
988 'preference': preference,
990 codecs = last_info.get('CODECS')
992 # TODO: looks like video codec is not always necessarily goes first
993 va_codecs = codecs.split(',')
995 f['vcodec'] = va_codecs[0].partition('.')[0]
996 if len(va_codecs) > 1 and va_codecs[1]:
997 f['acodec'] = va_codecs[1].partition('.')[0]
998 resolution = last_info.get('RESOLUTION')
1000 width_str, height_str = resolution.split('x')
1001 f['width'] = int(width_str)
1002 f['height'] = int(height_str)
1003 if last_media is not None:
1004 f['m3u8_media'] = last_media
1008 self._sort_formats(formats)
1012 def _xpath_ns(path, namespace=None):
1016 for c in path.split('/'):
1017 if not c or c == '.':
1020 out.append('{%s}%s' % (namespace, c))
1021 return '/'.join(out)
1023 def _extract_smil_formats(self, smil_url, video_id, fatal=True, f4m_params=None):
1024 smil = self._download_smil(smil_url, video_id, fatal=fatal)
1030 namespace = self._parse_smil_namespace(smil)
1032 return self._parse_smil_formats(
1033 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
1035 def _extract_smil_info(self, smil_url, video_id, fatal=True, f4m_params=None):
1036 smil = self._download_smil(smil_url, video_id, fatal=fatal)
1039 return self._parse_smil(smil, smil_url, video_id, f4m_params=f4m_params)
1041 def _download_smil(self, smil_url, video_id, fatal=True):
1042 return self._download_xml(
1043 smil_url, video_id, 'Downloading SMIL file',
1044 'Unable to download SMIL file', fatal=fatal)
1046 def _parse_smil(self, smil, smil_url, video_id, f4m_params=None):
1047 namespace = self._parse_smil_namespace(smil)
1049 formats = self._parse_smil_formats(
1050 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
1051 subtitles = self._parse_smil_subtitles(smil, namespace=namespace)
1053 video_id = os.path.splitext(url_basename(smil_url))[0]
1057 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
1058 name = meta.attrib.get('name')
1059 content = meta.attrib.get('content')
1060 if not name or not content:
1062 if not title and name == 'title':
1064 elif not description and name in ('description', 'abstract'):
1065 description = content
1066 elif not upload_date and name == 'date':
1067 upload_date = unified_strdate(content)
1070 'id': image.get('type'),
1071 'url': image.get('src'),
1072 'width': int_or_none(image.get('width')),
1073 'height': int_or_none(image.get('height')),
1074 } for image in smil.findall(self._xpath_ns('.//image', namespace)) if image.get('src')]
1078 'title': title or video_id,
1079 'description': description,
1080 'upload_date': upload_date,
1081 'thumbnails': thumbnails,
1083 'subtitles': subtitles,
1086 def _parse_smil_namespace(self, smil):
1087 return self._search_regex(
1088 r'(?i)^{([^}]+)?}smil$', smil.tag, 'namespace', default=None)
1090 def _parse_smil_formats(self, smil, smil_url, video_id, namespace=None, f4m_params=None, transform_rtmp_url=None):
1092 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
1093 b = meta.get('base') or meta.get('httpBase')
1102 videos = smil.findall(self._xpath_ns('.//video', namespace))
1103 for video in videos:
1104 src = video.get('src')
1108 bitrate = float_or_none(video.get('system-bitrate') or video.get('systemBitrate'), 1000)
1109 filesize = int_or_none(video.get('size') or video.get('fileSize'))
1110 width = int_or_none(video.get('width'))
1111 height = int_or_none(video.get('height'))
1112 proto = video.get('proto')
1113 ext = video.get('ext')
1114 src_ext = determine_ext(src)
1115 streamer = video.get('streamer') or base
1117 if proto == 'rtmp' or streamer.startswith('rtmp'):
1123 'format_id': 'rtmp-%d' % (rtmp_count if bitrate is None else bitrate),
1125 'filesize': filesize,
1129 if transform_rtmp_url:
1130 streamer, src = transform_rtmp_url(streamer, src)
1131 formats[-1].update({
1137 src_url = src if src.startswith('http') else compat_urlparse.urljoin(base, src)
1139 if proto == 'm3u8' or src_ext == 'm3u8':
1140 m3u8_formats = self._extract_m3u8_formats(
1141 src_url, video_id, ext or 'mp4', m3u8_id='hls', fatal=False)
1143 formats.extend(m3u8_formats)
1146 if src_ext == 'f4m':
1151 'plugin': 'flowplayer-3.2.0.1',
1153 f4m_url += '&' if '?' in f4m_url else '?'
1154 f4m_url += compat_urllib_parse.urlencode(f4m_params)
1155 f4m_formats = self._extract_f4m_formats(f4m_url, video_id, f4m_id='hds', fatal=False)
1157 formats.extend(f4m_formats)
1160 if src_url.startswith('http') and self._is_valid_url(src, video_id):
1164 'ext': ext or src_ext or 'flv',
1165 'format_id': 'http-%d' % (bitrate or http_count),
1167 'filesize': filesize,
1173 self._sort_formats(formats)
1177 def _parse_smil_subtitles(self, smil, namespace=None, subtitles_lang='en'):
1179 for num, textstream in enumerate(smil.findall(self._xpath_ns('.//textstream', namespace))):
1180 src = textstream.get('src')
1183 ext = textstream.get('ext') or determine_ext(src)
1185 type_ = textstream.get('type')
1189 'application/smptett+xml': 'tt',
1191 if type_ in SUBTITLES_TYPES:
1192 ext = SUBTITLES_TYPES[type_]
1193 lang = textstream.get('systemLanguage') or textstream.get('systemLanguageName') or textstream.get('lang') or subtitles_lang
1194 subtitles.setdefault(lang, []).append({
1200 def _extract_xspf_playlist(self, playlist_url, playlist_id, fatal=True):
1201 xspf = self._download_xml(
1202 playlist_url, playlist_id, 'Downloading xpsf playlist',
1203 'Unable to download xspf manifest', fatal=fatal)
1206 return self._parse_xspf(xspf, playlist_id)
1208 def _parse_xspf(self, playlist, playlist_id):
1210 'xspf': 'http://xspf.org/ns/0/',
1211 's1': 'http://static.streamone.nl/player/ns/0',
1215 for track in playlist.findall(xpath_with_ns('./xspf:trackList/xspf:track', NS_MAP)):
1217 track, xpath_with_ns('./xspf:title', NS_MAP), 'title', default=playlist_id)
1218 description = xpath_text(
1219 track, xpath_with_ns('./xspf:annotation', NS_MAP), 'description')
1220 thumbnail = xpath_text(
1221 track, xpath_with_ns('./xspf:image', NS_MAP), 'thumbnail')
1222 duration = float_or_none(
1223 xpath_text(track, xpath_with_ns('./xspf:duration', NS_MAP), 'duration'), 1000)
1226 'url': location.text,
1227 'format_id': location.get(xpath_with_ns('s1:label', NS_MAP)),
1228 'width': int_or_none(location.get(xpath_with_ns('s1:width', NS_MAP))),
1229 'height': int_or_none(location.get(xpath_with_ns('s1:height', NS_MAP))),
1230 } for location in track.findall(xpath_with_ns('./xspf:location', NS_MAP))]
1231 self._sort_formats(formats)
1236 'description': description,
1237 'thumbnail': thumbnail,
1238 'duration': duration,
1243 def _live_title(self, name):
1244 """ Generate the title for a live video """
1245 now = datetime.datetime.now()
1246 now_str = now.strftime("%Y-%m-%d %H:%M")
1247 return name + ' ' + now_str
1249 def _int(self, v, name, fatal=False, **kwargs):
1250 res = int_or_none(v, **kwargs)
1251 if 'get_attr' in kwargs:
1252 print(getattr(v, kwargs['get_attr']))
1254 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
1256 raise ExtractorError(msg)
1258 self._downloader.report_warning(msg)
1261 def _float(self, v, name, fatal=False, **kwargs):
1262 res = float_or_none(v, **kwargs)
1264 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
1266 raise ExtractorError(msg)
1268 self._downloader.report_warning(msg)
1271 def _set_cookie(self, domain, name, value, expire_time=None):
1272 cookie = compat_cookiejar.Cookie(
1273 0, name, value, None, None, domain, None,
1274 None, '/', True, False, expire_time, '', None, None, None)
1275 self._downloader.cookiejar.set_cookie(cookie)
1277 def _get_cookies(self, url):
1278 """ Return a compat_cookies.SimpleCookie with the cookies for the url """
1279 req = compat_urllib_request.Request(url)
1280 self._downloader.cookiejar.add_cookie_header(req)
1281 return compat_cookies.SimpleCookie(req.get_header('Cookie'))
1283 def get_testcases(self, include_onlymatching=False):
1284 t = getattr(self, '_TEST', None)
1286 assert not hasattr(self, '_TESTS'), \
1287 '%s has _TEST and _TESTS' % type(self).__name__
1290 tests = getattr(self, '_TESTS', [])
1292 if not include_onlymatching and t.get('only_matching', False):
1294 t['name'] = type(self).__name__[:-len('IE')]
1297 def is_suitable(self, age_limit):
1298 """ Test whether the extractor is generally suitable for the given
1299 age limit (i.e. pornographic sites are not, all others usually are) """
1301 any_restricted = False
1302 for tc in self.get_testcases(include_onlymatching=False):
1303 if 'playlist' in tc:
1304 tc = tc['playlist'][0]
1305 is_restricted = age_restricted(
1306 tc.get('info_dict', {}).get('age_limit'), age_limit)
1307 if not is_restricted:
1309 any_restricted = any_restricted or is_restricted
1310 return not any_restricted
1312 def extract_subtitles(self, *args, **kwargs):
1313 if (self._downloader.params.get('writesubtitles', False) or
1314 self._downloader.params.get('listsubtitles')):
1315 return self._get_subtitles(*args, **kwargs)
1318 def _get_subtitles(self, *args, **kwargs):
1319 raise NotImplementedError("This method must be implemented by subclasses")
1322 def _merge_subtitle_items(subtitle_list1, subtitle_list2):
1323 """ Merge subtitle items for one language. Items with duplicated URLs
1324 will be dropped. """
1325 list1_urls = set([item['url'] for item in subtitle_list1])
1326 ret = list(subtitle_list1)
1327 ret.extend([item for item in subtitle_list2 if item['url'] not in list1_urls])
1331 def _merge_subtitles(cls, subtitle_dict1, subtitle_dict2):
1332 """ Merge two subtitle dictionaries, language by language. """
1333 ret = dict(subtitle_dict1)
1334 for lang in subtitle_dict2:
1335 ret[lang] = cls._merge_subtitle_items(subtitle_dict1.get(lang, []), subtitle_dict2[lang])
1338 def extract_automatic_captions(self, *args, **kwargs):
1339 if (self._downloader.params.get('writeautomaticsub', False) or
1340 self._downloader.params.get('listsubtitles')):
1341 return self._get_automatic_captions(*args, **kwargs)
1344 def _get_automatic_captions(self, *args, **kwargs):
1345 raise NotImplementedError("This method must be implemented by subclasses")
1348 class SearchInfoExtractor(InfoExtractor):
1350 Base class for paged search queries extractors.
1351 They accept URLs in the format _SEARCH_KEY(|all|[0-9]):{query}
1352 Instances should define _SEARCH_KEY and _MAX_RESULTS.
1356 def _make_valid_url(cls):
1357 return r'%s(?P<prefix>|[1-9][0-9]*|all):(?P<query>[\s\S]+)' % cls._SEARCH_KEY
1360 def suitable(cls, url):
1361 return re.match(cls._make_valid_url(), url) is not None
1363 def _real_extract(self, query):
1364 mobj = re.match(self._make_valid_url(), query)
1366 raise ExtractorError('Invalid search query "%s"' % query)
1368 prefix = mobj.group('prefix')
1369 query = mobj.group('query')
1371 return self._get_n_results(query, 1)
1372 elif prefix == 'all':
1373 return self._get_n_results(query, self._MAX_RESULTS)
1377 raise ExtractorError('invalid download number %s for query "%s"' % (n, query))
1378 elif n > self._MAX_RESULTS:
1379 self._downloader.report_warning('%s returns max %i results (you requested %i)' % (self._SEARCH_KEY, self._MAX_RESULTS, n))
1380 n = self._MAX_RESULTS
1381 return self._get_n_results(query, n)
1383 def _get_n_results(self, query, n):
1384 """Get a specified number of results for a query"""
1385 raise NotImplementedError("This method must be implemented by subclasses")
1388 def SEARCH_KEY(self):
1389 return self._SEARCH_KEY