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 "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.
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 cls.__name__[:-2]
318 return 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'
336 errmsg = '%s: %s' % (errnote, compat_str(err))
338 raise ExtractorError(errmsg, sys.exc_info()[2], cause=err)
340 self._downloader.report_warning(errmsg)
343 def _download_webpage_handle(self, url_or_request, video_id, note=None, errnote=None, fatal=True, encoding=None):
344 """ Returns a tuple (page content as string, URL handle) """
345 # Strip hashes from the URL (#1038)
346 if isinstance(url_or_request, (compat_str, str)):
347 url_or_request = url_or_request.partition('#')[0]
349 urlh = self._request_webpage(url_or_request, video_id, note, errnote, fatal)
353 content = self._webpage_read_content(urlh, url_or_request, video_id, note, errnote, fatal, encoding=encoding)
354 return (content, urlh)
357 def _guess_encoding_from_content(content_type, webpage_bytes):
358 m = re.match(r'[a-zA-Z0-9_.-]+/[a-zA-Z0-9_.-]+\s*;\s*charset=(.+)', content_type)
360 encoding = m.group(1)
362 m = re.search(br'<meta[^>]+charset=[\'"]?([^\'")]+)[ /\'">]',
363 webpage_bytes[:1024])
365 encoding = m.group(1).decode('ascii')
366 elif webpage_bytes.startswith(b'\xff\xfe'):
373 def _webpage_read_content(self, urlh, url_or_request, video_id, note=None, errnote=None, fatal=True, prefix=None, encoding=None):
374 content_type = urlh.headers.get('Content-Type', '')
375 webpage_bytes = urlh.read()
376 if prefix is not None:
377 webpage_bytes = prefix + webpage_bytes
379 encoding = self._guess_encoding_from_content(content_type, webpage_bytes)
380 if self._downloader.params.get('dump_intermediate_pages', False):
382 url = url_or_request.get_full_url()
383 except AttributeError:
385 self.to_screen('Dumping request to ' + url)
386 dump = base64.b64encode(webpage_bytes).decode('ascii')
387 self._downloader.to_screen(dump)
388 if self._downloader.params.get('write_pages', False):
390 url = url_or_request.get_full_url()
391 except AttributeError:
393 basen = '%s_%s' % (video_id, url)
395 h = '___' + hashlib.md5(basen.encode('utf-8')).hexdigest()
396 basen = basen[:240 - len(h)] + h
397 raw_filename = basen + '.dump'
398 filename = sanitize_filename(raw_filename, restricted=True)
399 self.to_screen('Saving request to ' + filename)
400 # Working around MAX_PATH limitation on Windows (see
401 # http://msdn.microsoft.com/en-us/library/windows/desktop/aa365247(v=vs.85).aspx)
403 absfilepath = os.path.abspath(filename)
404 if len(absfilepath) > 259:
405 filename = '\\\\?\\' + absfilepath
406 with open(filename, 'wb') as outf:
407 outf.write(webpage_bytes)
410 content = webpage_bytes.decode(encoding, 'replace')
412 content = webpage_bytes.decode('utf-8', 'replace')
414 if ('<title>Access to this site is blocked</title>' in content and
415 'Websense' in content[:512]):
416 msg = 'Access to this webpage has been blocked by Websense filtering software in your network.'
417 blocked_iframe = self._html_search_regex(
418 r'<iframe src="([^"]+)"', content,
419 'Websense information URL', default=None)
421 msg += ' Visit %s for more details' % blocked_iframe
422 raise ExtractorError(msg, expected=True)
423 if '<title>The URL you requested has been blocked</title>' in content[:512]:
425 'Access to this webpage has been blocked by Indian censorship. '
426 'Use a VPN or proxy server (with --proxy) to route around it.')
427 block_msg = self._html_search_regex(
428 r'</h1><p>(.*?)</p>',
429 content, 'block message', default=None)
431 msg += ' (Message: "%s")' % block_msg.replace('\n', ' ')
432 raise ExtractorError(msg, expected=True)
436 def _download_webpage(self, url_or_request, video_id, note=None, errnote=None, fatal=True, tries=1, timeout=5, encoding=None):
437 """ Returns the data of the page as a string """
440 while success is False:
442 res = self._download_webpage_handle(url_or_request, video_id, note, errnote, fatal, encoding=encoding)
444 except compat_http_client.IncompleteRead as e:
446 if try_count >= tries:
448 self._sleep(timeout, video_id)
455 def _download_xml(self, url_or_request, video_id,
456 note='Downloading XML', errnote='Unable to download XML',
457 transform_source=None, fatal=True, encoding=None):
458 """Return the xml as an xml.etree.ElementTree.Element"""
459 xml_string = self._download_webpage(
460 url_or_request, video_id, note, errnote, fatal=fatal, encoding=encoding)
461 if xml_string is False:
464 xml_string = transform_source(xml_string)
465 return xml.etree.ElementTree.fromstring(xml_string.encode('utf-8'))
467 def _download_json(self, url_or_request, video_id,
468 note='Downloading JSON metadata',
469 errnote='Unable to download JSON metadata',
470 transform_source=None,
471 fatal=True, encoding=None):
472 json_string = self._download_webpage(
473 url_or_request, video_id, note, errnote, fatal=fatal,
475 if (not fatal) and json_string is False:
477 return self._parse_json(
478 json_string, video_id, transform_source=transform_source, fatal=fatal)
480 def _parse_json(self, json_string, video_id, transform_source=None, fatal=True):
482 json_string = transform_source(json_string)
484 return json.loads(json_string)
485 except ValueError as ve:
486 errmsg = '%s: Failed to parse JSON ' % video_id
488 raise ExtractorError(errmsg, cause=ve)
490 self.report_warning(errmsg + str(ve))
492 def report_warning(self, msg, video_id=None):
493 idstr = '' if video_id is None else '%s: ' % video_id
494 self._downloader.report_warning(
495 '[%s] %s%s' % (self.IE_NAME, idstr, msg))
497 def to_screen(self, msg):
498 """Print msg to screen, prefixing it with '[ie_name]'"""
499 self._downloader.to_screen('[%s] %s' % (self.IE_NAME, msg))
501 def report_extraction(self, id_or_name):
502 """Report information extraction."""
503 self.to_screen('%s: Extracting information' % id_or_name)
505 def report_download_webpage(self, video_id):
506 """Report webpage download."""
507 self.to_screen('%s: Downloading webpage' % video_id)
509 def report_age_confirmation(self):
510 """Report attempt to confirm age."""
511 self.to_screen('Confirming age')
513 def report_login(self):
514 """Report attempt to log in."""
515 self.to_screen('Logging in')
518 def raise_login_required(msg='This video is only available for registered users'):
519 raise ExtractorError(
520 '%s. Use --username and --password or --netrc to provide account credentials.' % msg,
524 def raise_geo_restricted(msg='This video is not available from your location due to geo restriction'):
525 raise ExtractorError(
526 '%s. You might want to use --proxy to workaround.' % msg,
529 # Methods for following #608
531 def url_result(url, ie=None, video_id=None, video_title=None):
532 """Returns a URL that points to a page that should be processed"""
533 # TODO: ie should be the class used for getting the info
534 video_info = {'_type': 'url',
537 if video_id is not None:
538 video_info['id'] = video_id
539 if video_title is not None:
540 video_info['title'] = video_title
544 def playlist_result(entries, playlist_id=None, playlist_title=None, playlist_description=None):
545 """Returns a playlist"""
546 video_info = {'_type': 'playlist',
549 video_info['id'] = playlist_id
551 video_info['title'] = playlist_title
552 if playlist_description:
553 video_info['description'] = playlist_description
556 def _search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
558 Perform a regex search on the given string, using a single or a list of
559 patterns returning the first matching group.
560 In case of failure return a default value or raise a WARNING or a
561 RegexNotFoundError, depending on fatal, specifying the field name.
563 if isinstance(pattern, (str, compat_str, compiled_regex_type)):
564 mobj = re.search(pattern, string, flags)
567 mobj = re.search(p, string, flags)
571 if not self._downloader.params.get('no_color') and os.name != 'nt' and sys.stderr.isatty():
572 _name = '\033[0;34m%s\033[0m' % name
578 # return the first matching group
579 return next(g for g in mobj.groups() if g is not None)
581 return mobj.group(group)
582 elif default is not NO_DEFAULT:
585 raise RegexNotFoundError('Unable to extract %s' % _name)
587 self._downloader.report_warning('unable to extract %s' % _name + bug_reports_message())
590 def _html_search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
592 Like _search_regex, but strips HTML tags and unescapes entities.
594 res = self._search_regex(pattern, string, name, default, fatal, flags, group)
596 return clean_html(res).strip()
600 def _get_login_info(self):
602 Get the login info as (username, password)
603 It will look in the netrc file using the _NETRC_MACHINE value
604 If there's no info available, return (None, None)
606 if self._downloader is None:
611 downloader_params = self._downloader.params
613 # Attempt to use provided username and password or .netrc data
614 if downloader_params.get('username', None) is not None:
615 username = downloader_params['username']
616 password = downloader_params['password']
617 elif downloader_params.get('usenetrc', False):
619 info = netrc.netrc().authenticators(self._NETRC_MACHINE)
624 raise netrc.NetrcParseError('No authenticators for %s' % self._NETRC_MACHINE)
625 except (IOError, netrc.NetrcParseError) as err:
626 self._downloader.report_warning('parsing .netrc: %s' % compat_str(err))
628 return (username, password)
630 def _get_tfa_info(self, note='two-factor verification code'):
632 Get the two-factor authentication info
633 TODO - asking the user will be required for sms/phone verify
634 currently just uses the command line option
635 If there's no info available, return None
637 if self._downloader is None:
639 downloader_params = self._downloader.params
641 if downloader_params.get('twofactor', None) is not None:
642 return downloader_params['twofactor']
644 return compat_getpass('Type %s and press [Return]: ' % note)
646 # Helper functions for extracting OpenGraph info
648 def _og_regexes(prop):
649 content_re = r'content=(?:"([^"]+?)"|\'([^\']+?)\'|\s*([^\s"\'=<>`]+?))'
650 property_re = (r'(?:name|property)=(?:\'og:%(prop)s\'|"og:%(prop)s"|\s*og:%(prop)s\b)'
651 % {'prop': re.escape(prop)})
652 template = r'<meta[^>]+?%s[^>]+?%s'
654 template % (property_re, content_re),
655 template % (content_re, property_re),
659 def _meta_regex(prop):
660 return r'''(?isx)<meta
661 (?=[^>]+(?:itemprop|name|property|id|http-equiv)=(["\']?)%s\1)
662 [^>]+?content=(["\'])(?P<content>.*?)\2''' % re.escape(prop)
664 def _og_search_property(self, prop, html, name=None, **kargs):
666 name = 'OpenGraph %s' % prop
667 escaped = self._search_regex(self._og_regexes(prop), html, name, flags=re.DOTALL, **kargs)
670 return unescapeHTML(escaped)
672 def _og_search_thumbnail(self, html, **kargs):
673 return self._og_search_property('image', html, 'thumbnail URL', fatal=False, **kargs)
675 def _og_search_description(self, html, **kargs):
676 return self._og_search_property('description', html, fatal=False, **kargs)
678 def _og_search_title(self, html, **kargs):
679 return self._og_search_property('title', html, **kargs)
681 def _og_search_video_url(self, html, name='video url', secure=True, **kargs):
682 regexes = self._og_regexes('video') + self._og_regexes('video:url')
684 regexes = self._og_regexes('video:secure_url') + regexes
685 return self._html_search_regex(regexes, html, name, **kargs)
687 def _og_search_url(self, html, **kargs):
688 return self._og_search_property('url', html, **kargs)
690 def _html_search_meta(self, name, html, display_name=None, fatal=False, **kwargs):
691 if display_name is None:
693 return self._html_search_regex(
694 self._meta_regex(name),
695 html, display_name, fatal=fatal, group='content', **kwargs)
697 def _dc_search_uploader(self, html):
698 return self._html_search_meta('dc.creator', html, 'uploader')
700 def _rta_search(self, html):
701 # See http://www.rtalabel.org/index.php?content=howtofaq#single
702 if re.search(r'(?ix)<meta\s+name="rating"\s+'
703 r' content="RTA-5042-1996-1400-1577-RTA"',
708 def _media_rating_search(self, html):
709 # See http://www.tjg-designs.com/WP/metadata-code-examples-adding-metadata-to-your-web-pages/
710 rating = self._html_search_meta('rating', html)
722 return RATING_TABLE.get(rating.lower(), None)
724 def _family_friendly_search(self, html):
725 # See http://schema.org/VideoObject
726 family_friendly = self._html_search_meta('isFamilyFriendly', html)
728 if not family_friendly:
737 return RATING_TABLE.get(family_friendly.lower(), None)
739 def _twitter_search_player(self, html):
740 return self._html_search_meta('twitter:player', html,
741 'twitter card player')
744 def _hidden_inputs(html):
745 html = re.sub(r'<!--(?:(?!<!--).)*-->', '', html)
747 for input in re.findall(r'(?i)<input([^>]+)>', html):
748 if not re.search(r'type=(["\'])(?:hidden|submit)\1', input):
750 name = re.search(r'name=(["\'])(?P<value>.+?)\1', input)
753 value = re.search(r'value=(["\'])(?P<value>.*?)\1', input)
756 hidden_inputs[name.group('value')] = value.group('value')
759 def _form_hidden_inputs(self, form_id, html):
760 form = self._search_regex(
761 r'(?is)<form[^>]+?id=(["\'])%s\1[^>]*>(?P<form>.+?)</form>' % form_id,
762 html, '%s form' % form_id, group='form')
763 return self._hidden_inputs(form)
765 def _sort_formats(self, formats, field_preference=None):
767 raise ExtractorError('No video formats found')
770 # TODO remove the following workaround
771 from ..utils import determine_ext
772 if not f.get('ext') and 'url' in f:
773 f['ext'] = determine_ext(f['url'])
775 if isinstance(field_preference, (list, tuple)):
776 return tuple(f.get(field) if f.get(field) is not None else -1 for field in field_preference)
778 preference = f.get('preference')
779 if preference is None:
780 proto = f.get('protocol')
782 proto = compat_urllib_parse_urlparse(f.get('url', '')).scheme
784 preference = 0 if proto in ['http', 'https'] else -0.1
785 if f.get('ext') in ['f4f', 'f4m']: # Not yet supported
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,
819 f.get('abr') if f.get('abr') is not None else -1,
820 audio_ext_preference,
821 f.get('fps') if f.get('fps') is not None else -1,
822 f.get('filesize_approx') if f.get('filesize_approx') is not None else -1,
823 f.get('source_preference') if f.get('source_preference') is not None else -1,
824 f.get('format_id') if f.get('format_id') is not None else '',
826 formats.sort(key=_formats_key)
828 def _check_formats(self, formats, video_id):
831 lambda f: self._is_valid_url(
833 item='%s video format' % f.get('format_id') if f.get('format_id') else 'video'),
836 def _is_valid_url(self, url, video_id, item='video'):
837 url = self._proto_relative_url(url, scheme='http:')
838 # For now assume non HTTP(S) URLs always valid
839 if not (url.startswith('http://') or url.startswith('https://')):
842 self._request_webpage(url, video_id, 'Checking %s URL' % item)
844 except ExtractorError as e:
845 if isinstance(e.cause, compat_HTTPError):
847 '%s: %s URL is invalid, skipping' % (video_id, item))
851 def http_scheme(self):
852 """ Either "http:" or "https:", depending on the user's preferences """
855 if self._downloader.params.get('prefer_insecure', False)
858 def _proto_relative_url(self, url, scheme=None):
861 if url.startswith('//'):
863 scheme = self.http_scheme()
868 def _sleep(self, timeout, video_id, msg_template=None):
869 if msg_template is None:
870 msg_template = '%(video_id)s: Waiting for %(timeout)s seconds'
871 msg = msg_template % {'video_id': video_id, 'timeout': timeout}
875 def _extract_f4m_formats(self, manifest_url, video_id, preference=None, f4m_id=None,
876 transform_source=lambda s: fix_xml_ampersands(s).strip(),
878 manifest = self._download_xml(
879 manifest_url, video_id, 'Downloading f4m manifest',
880 'Unable to download f4m manifest',
881 # Some manifests may be malformed, e.g. prosiebensat1 generated manifests
882 # (see https://github.com/rg3/youtube-dl/issues/6215#issuecomment-121704244)
883 transform_source=transform_source,
886 if manifest is False:
890 manifest_version = '1.0'
891 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/1.0}media')
893 manifest_version = '2.0'
894 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/2.0}media')
895 for i, media_el in enumerate(media_nodes):
896 if manifest_version == '2.0':
897 media_url = media_el.attrib.get('href') or media_el.attrib.get('url')
901 media_url if media_url.startswith('http://') or media_url.startswith('https://')
902 else ('/'.join(manifest_url.split('/')[:-1]) + '/' + media_url))
903 # If media_url is itself a f4m manifest do the recursive extraction
904 # since bitrates in parent manifest (this one) and media_url manifest
905 # may differ leading to inability to resolve the format by requested
906 # bitrate in f4m downloader
907 if determine_ext(manifest_url) == 'f4m':
908 f4m_formats = self._extract_f4m_formats(
909 manifest_url, video_id, preference, f4m_id, fatal=fatal)
911 formats.extend(f4m_formats)
913 tbr = int_or_none(media_el.attrib.get('bitrate'))
915 'format_id': '-'.join(filter(None, [f4m_id, compat_str(i if tbr is None else tbr)])),
919 'width': int_or_none(media_el.attrib.get('width')),
920 'height': int_or_none(media_el.attrib.get('height')),
921 'preference': preference,
923 self._sort_formats(formats)
927 def _extract_m3u8_formats(self, m3u8_url, video_id, ext=None,
928 entry_protocol='m3u8', preference=None,
929 m3u8_id=None, note=None, errnote=None,
933 'format_id': '-'.join(filter(None, [m3u8_id, 'meta'])),
937 'preference': preference - 1 if preference else -1,
938 'resolution': 'multiple',
939 'format_note': 'Quality selection URL',
942 format_url = lambda u: (
944 if re.match(r'^https?://', u)
945 else compat_urlparse.urljoin(m3u8_url, u))
947 m3u8_doc = self._download_webpage(
949 note=note or 'Downloading m3u8 information',
950 errnote=errnote or 'Failed to download m3u8 information',
952 if m3u8_doc is False:
957 r'(?P<key>[a-zA-Z_-]+)=(?P<val>"[^"]+"|[^",]+)(?:,|$)')
958 for line in m3u8_doc.splitlines():
959 if line.startswith('#EXT-X-STREAM-INF:'):
961 for m in kv_rex.finditer(line):
963 if v.startswith('"'):
965 last_info[m.group('key')] = v
966 elif line.startswith('#EXT-X-MEDIA:'):
968 for m in kv_rex.finditer(line):
970 if v.startswith('"'):
972 last_media[m.group('key')] = v
973 elif line.startswith('#') or not line.strip():
976 if last_info is None:
977 formats.append({'url': format_url(line)})
979 tbr = int_or_none(last_info.get('BANDWIDTH'), scale=1000)
982 format_id.append(m3u8_id)
983 last_media_name = last_media.get('NAME') if last_media and last_media.get('TYPE') != 'SUBTITLES' else None
984 format_id.append(last_media_name if last_media_name else '%d' % (tbr if tbr else len(formats)))
986 'format_id': '-'.join(format_id),
987 'url': format_url(line.strip()),
990 'protocol': entry_protocol,
991 'preference': preference,
993 codecs = last_info.get('CODECS')
995 # TODO: looks like video codec is not always necessarily goes first
996 va_codecs = codecs.split(',')
998 f['vcodec'] = va_codecs[0].partition('.')[0]
999 if len(va_codecs) > 1 and va_codecs[1]:
1000 f['acodec'] = va_codecs[1].partition('.')[0]
1001 resolution = last_info.get('RESOLUTION')
1003 width_str, height_str = resolution.split('x')
1004 f['width'] = int(width_str)
1005 f['height'] = int(height_str)
1006 if last_media is not None:
1007 f['m3u8_media'] = last_media
1011 self._sort_formats(formats)
1015 def _xpath_ns(path, namespace=None):
1019 for c in path.split('/'):
1020 if not c or c == '.':
1023 out.append('{%s}%s' % (namespace, c))
1024 return '/'.join(out)
1026 def _extract_smil_formats(self, smil_url, video_id, fatal=True, f4m_params=None):
1027 smil = self._download_smil(smil_url, video_id, fatal=fatal)
1033 namespace = self._parse_smil_namespace(smil)
1035 return self._parse_smil_formats(
1036 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
1038 def _extract_smil_info(self, smil_url, video_id, fatal=True, f4m_params=None):
1039 smil = self._download_smil(smil_url, video_id, fatal=fatal)
1042 return self._parse_smil(smil, smil_url, video_id, f4m_params=f4m_params)
1044 def _download_smil(self, smil_url, video_id, fatal=True):
1045 return self._download_xml(
1046 smil_url, video_id, 'Downloading SMIL file',
1047 'Unable to download SMIL file', fatal=fatal)
1049 def _parse_smil(self, smil, smil_url, video_id, f4m_params=None):
1050 namespace = self._parse_smil_namespace(smil)
1052 formats = self._parse_smil_formats(
1053 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
1054 subtitles = self._parse_smil_subtitles(smil, namespace=namespace)
1056 video_id = os.path.splitext(url_basename(smil_url))[0]
1060 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
1061 name = meta.attrib.get('name')
1062 content = meta.attrib.get('content')
1063 if not name or not content:
1065 if not title and name == 'title':
1067 elif not description and name in ('description', 'abstract'):
1068 description = content
1069 elif not upload_date and name == 'date':
1070 upload_date = unified_strdate(content)
1073 'id': image.get('type'),
1074 'url': image.get('src'),
1075 'width': int_or_none(image.get('width')),
1076 'height': int_or_none(image.get('height')),
1077 } for image in smil.findall(self._xpath_ns('.//image', namespace)) if image.get('src')]
1081 'title': title or video_id,
1082 'description': description,
1083 'upload_date': upload_date,
1084 'thumbnails': thumbnails,
1086 'subtitles': subtitles,
1089 def _parse_smil_namespace(self, smil):
1090 return self._search_regex(
1091 r'(?i)^{([^}]+)?}smil$', smil.tag, 'namespace', default=None)
1093 def _parse_smil_formats(self, smil, smil_url, video_id, namespace=None, f4m_params=None, transform_rtmp_url=None):
1095 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
1096 b = meta.get('base') or meta.get('httpBase')
1105 videos = smil.findall(self._xpath_ns('.//video', namespace))
1106 for video in videos:
1107 src = video.get('src')
1111 bitrate = float_or_none(video.get('system-bitrate') or video.get('systemBitrate'), 1000)
1112 filesize = int_or_none(video.get('size') or video.get('fileSize'))
1113 width = int_or_none(video.get('width'))
1114 height = int_or_none(video.get('height'))
1115 proto = video.get('proto')
1116 ext = video.get('ext')
1117 src_ext = determine_ext(src)
1118 streamer = video.get('streamer') or base
1120 if proto == 'rtmp' or streamer.startswith('rtmp'):
1126 'format_id': 'rtmp-%d' % (rtmp_count if bitrate is None else bitrate),
1128 'filesize': filesize,
1132 if transform_rtmp_url:
1133 streamer, src = transform_rtmp_url(streamer, src)
1134 formats[-1].update({
1140 src_url = src if src.startswith('http') else compat_urlparse.urljoin(base, src)
1142 if proto == 'm3u8' or src_ext == 'm3u8':
1143 m3u8_formats = self._extract_m3u8_formats(
1144 src_url, video_id, ext or 'mp4', m3u8_id='hls', fatal=False)
1146 formats.extend(m3u8_formats)
1149 if src_ext == 'f4m':
1154 'plugin': 'flowplayer-3.2.0.1',
1156 f4m_url += '&' if '?' in f4m_url else '?'
1157 f4m_url += compat_urllib_parse.urlencode(f4m_params)
1158 f4m_formats = self._extract_f4m_formats(f4m_url, video_id, f4m_id='hds', fatal=False)
1160 formats.extend(f4m_formats)
1163 if src_url.startswith('http') and self._is_valid_url(src, video_id):
1167 'ext': ext or src_ext or 'flv',
1168 'format_id': 'http-%d' % (bitrate or http_count),
1170 'filesize': filesize,
1176 self._sort_formats(formats)
1180 def _parse_smil_subtitles(self, smil, namespace=None, subtitles_lang='en'):
1182 for num, textstream in enumerate(smil.findall(self._xpath_ns('.//textstream', namespace))):
1183 src = textstream.get('src')
1186 ext = textstream.get('ext') or determine_ext(src)
1188 type_ = textstream.get('type')
1192 'application/smptett+xml': 'tt',
1194 if type_ in SUBTITLES_TYPES:
1195 ext = SUBTITLES_TYPES[type_]
1196 lang = textstream.get('systemLanguage') or textstream.get('systemLanguageName') or textstream.get('lang') or subtitles_lang
1197 subtitles.setdefault(lang, []).append({
1203 def _extract_xspf_playlist(self, playlist_url, playlist_id, fatal=True):
1204 xspf = self._download_xml(
1205 playlist_url, playlist_id, 'Downloading xpsf playlist',
1206 'Unable to download xspf manifest', fatal=fatal)
1209 return self._parse_xspf(xspf, playlist_id)
1211 def _parse_xspf(self, playlist, playlist_id):
1213 'xspf': 'http://xspf.org/ns/0/',
1214 's1': 'http://static.streamone.nl/player/ns/0',
1218 for track in playlist.findall(xpath_with_ns('./xspf:trackList/xspf:track', NS_MAP)):
1220 track, xpath_with_ns('./xspf:title', NS_MAP), 'title', default=playlist_id)
1221 description = xpath_text(
1222 track, xpath_with_ns('./xspf:annotation', NS_MAP), 'description')
1223 thumbnail = xpath_text(
1224 track, xpath_with_ns('./xspf:image', NS_MAP), 'thumbnail')
1225 duration = float_or_none(
1226 xpath_text(track, xpath_with_ns('./xspf:duration', NS_MAP), 'duration'), 1000)
1229 'url': location.text,
1230 'format_id': location.get(xpath_with_ns('s1:label', NS_MAP)),
1231 'width': int_or_none(location.get(xpath_with_ns('s1:width', NS_MAP))),
1232 'height': int_or_none(location.get(xpath_with_ns('s1:height', NS_MAP))),
1233 } for location in track.findall(xpath_with_ns('./xspf:location', NS_MAP))]
1234 self._sort_formats(formats)
1239 'description': description,
1240 'thumbnail': thumbnail,
1241 'duration': duration,
1246 def _live_title(self, name):
1247 """ Generate the title for a live video """
1248 now = datetime.datetime.now()
1249 now_str = now.strftime("%Y-%m-%d %H:%M")
1250 return name + ' ' + now_str
1252 def _int(self, v, name, fatal=False, **kwargs):
1253 res = int_or_none(v, **kwargs)
1254 if 'get_attr' in kwargs:
1255 print(getattr(v, kwargs['get_attr']))
1257 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
1259 raise ExtractorError(msg)
1261 self._downloader.report_warning(msg)
1264 def _float(self, v, name, fatal=False, **kwargs):
1265 res = float_or_none(v, **kwargs)
1267 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
1269 raise ExtractorError(msg)
1271 self._downloader.report_warning(msg)
1274 def _set_cookie(self, domain, name, value, expire_time=None):
1275 cookie = compat_cookiejar.Cookie(
1276 0, name, value, None, None, domain, None,
1277 None, '/', True, False, expire_time, '', None, None, None)
1278 self._downloader.cookiejar.set_cookie(cookie)
1280 def _get_cookies(self, url):
1281 """ Return a compat_cookies.SimpleCookie with the cookies for the url """
1282 req = compat_urllib_request.Request(url)
1283 self._downloader.cookiejar.add_cookie_header(req)
1284 return compat_cookies.SimpleCookie(req.get_header('Cookie'))
1286 def get_testcases(self, include_onlymatching=False):
1287 t = getattr(self, '_TEST', None)
1289 assert not hasattr(self, '_TESTS'), \
1290 '%s has _TEST and _TESTS' % type(self).__name__
1293 tests = getattr(self, '_TESTS', [])
1295 if not include_onlymatching and t.get('only_matching', False):
1297 t['name'] = type(self).__name__[:-len('IE')]
1300 def is_suitable(self, age_limit):
1301 """ Test whether the extractor is generally suitable for the given
1302 age limit (i.e. pornographic sites are not, all others usually are) """
1304 any_restricted = False
1305 for tc in self.get_testcases(include_onlymatching=False):
1306 if 'playlist' in tc:
1307 tc = tc['playlist'][0]
1308 is_restricted = age_restricted(
1309 tc.get('info_dict', {}).get('age_limit'), age_limit)
1310 if not is_restricted:
1312 any_restricted = any_restricted or is_restricted
1313 return not any_restricted
1315 def extract_subtitles(self, *args, **kwargs):
1316 if (self._downloader.params.get('writesubtitles', False) or
1317 self._downloader.params.get('listsubtitles')):
1318 return self._get_subtitles(*args, **kwargs)
1321 def _get_subtitles(self, *args, **kwargs):
1322 raise NotImplementedError("This method must be implemented by subclasses")
1325 def _merge_subtitle_items(subtitle_list1, subtitle_list2):
1326 """ Merge subtitle items for one language. Items with duplicated URLs
1327 will be dropped. """
1328 list1_urls = set([item['url'] for item in subtitle_list1])
1329 ret = list(subtitle_list1)
1330 ret.extend([item for item in subtitle_list2 if item['url'] not in list1_urls])
1334 def _merge_subtitles(cls, subtitle_dict1, subtitle_dict2):
1335 """ Merge two subtitle dictionaries, language by language. """
1336 ret = dict(subtitle_dict1)
1337 for lang in subtitle_dict2:
1338 ret[lang] = cls._merge_subtitle_items(subtitle_dict1.get(lang, []), subtitle_dict2[lang])
1341 def extract_automatic_captions(self, *args, **kwargs):
1342 if (self._downloader.params.get('writeautomaticsub', False) or
1343 self._downloader.params.get('listsubtitles')):
1344 return self._get_automatic_captions(*args, **kwargs)
1347 def _get_automatic_captions(self, *args, **kwargs):
1348 raise NotImplementedError("This method must be implemented by subclasses")
1351 class SearchInfoExtractor(InfoExtractor):
1353 Base class for paged search queries extractors.
1354 They accept URLs in the format _SEARCH_KEY(|all|[0-9]):{query}
1355 Instances should define _SEARCH_KEY and _MAX_RESULTS.
1359 def _make_valid_url(cls):
1360 return r'%s(?P<prefix>|[1-9][0-9]*|all):(?P<query>[\s\S]+)' % cls._SEARCH_KEY
1363 def suitable(cls, url):
1364 return re.match(cls._make_valid_url(), url) is not None
1366 def _real_extract(self, query):
1367 mobj = re.match(self._make_valid_url(), query)
1369 raise ExtractorError('Invalid search query "%s"' % query)
1371 prefix = mobj.group('prefix')
1372 query = mobj.group('query')
1374 return self._get_n_results(query, 1)
1375 elif prefix == 'all':
1376 return self._get_n_results(query, self._MAX_RESULTS)
1380 raise ExtractorError('invalid download number %s for query "%s"' % (n, query))
1381 elif n > self._MAX_RESULTS:
1382 self._downloader.report_warning('%s returns max %i results (you requested %i)' % (self._SEARCH_KEY, self._MAX_RESULTS, n))
1383 n = self._MAX_RESULTS
1384 return self._get_n_results(query, n)
1386 def _get_n_results(self, query, n):
1387 """Get a specified number of results for a query"""
1388 raise NotImplementedError("This method must be implemented by subclasses")
1391 def SEARCH_KEY(self):
1392 return self._SEARCH_KEY