1 from __future__ import unicode_literals
13 import xml.etree.ElementTree
15 from ..compat import (
20 compat_urllib_parse_urlparse,
41 class InfoExtractor(object):
42 """Information Extractor class.
44 Information extractors are the classes that, given a URL, extract
45 information about the video (or videos) the URL refers to. This
46 information includes the real video URL, the video title, author and
47 others. The information is stored in a dictionary which is then
48 passed to the YoutubeDL. The YoutubeDL processes this
49 information possibly downloading the video to the file system, among
50 other possible outcomes.
52 The type field determines the type of the result.
53 By far the most common value (and the default if _type is missing) is
54 "video", which indicates a single video.
56 For a video, the dictionaries must include the following fields:
59 title: Video title, unescaped.
61 Additionally, it must contain either a formats entry or a url one:
63 formats: A list of dictionaries for each format available, ordered
64 from worst to best quality.
67 * url Mandatory. The URL of the video file
68 * ext Will be calculated from url if missing
69 * format A human-readable description of the format
70 ("mp4 container with h264/opus").
71 Calculated from the format_id, width, height.
72 and format_note fields if missing.
73 * format_id A short description of the format
74 ("mp4_h264_opus" or "19").
75 Technically optional, but strongly recommended.
76 * format_note Additional info about the format
77 ("3D" or "DASH video")
78 * width Width of the video, if known
79 * height Height of the video, if known
80 * resolution Textual description of width and height
81 * tbr Average bitrate of audio and video in KBit/s
82 * abr Average audio bitrate in KBit/s
83 * acodec Name of the audio codec in use
84 * asr Audio sampling rate in Hertz
85 * vbr Average video bitrate in KBit/s
87 * vcodec Name of the video codec in use
88 * container Name of the container format
89 * filesize The number of bytes, if known in advance
90 * filesize_approx An estimate for the number of bytes
91 * player_url SWF Player URL (used for rtmpdump).
92 * protocol The protocol that will be used for the actual
94 "http", "https", "rtsp", "rtmp", "rtmpe",
95 "m3u8", or "m3u8_native".
96 * preference Order number of this format. If this field is
97 present and not None, the formats get sorted
98 by this field, regardless of all other values.
99 -1 for default (order by other properties),
100 -2 or smaller for less than default.
101 < -1000 to hide the format (if there is
102 another one which is strictly better)
103 * language_preference Is this in the correct requested
105 10 if it's what the URL is about,
106 -1 for default (don't know),
107 -10 otherwise, other values reserved for now.
108 * quality Order number of the video quality of this
109 format, irrespective of the file format.
110 -1 for default (order by other properties),
111 -2 or smaller for less than default.
112 * source_preference Order number for this video source
113 (quality takes higher priority)
114 -1 for default (order by other properties),
115 -2 or smaller for less than default.
116 * http_headers A dictionary of additional HTTP headers
117 to add to the request.
118 * stretched_ratio If given and not 1, indicates that the
119 video's pixels are not square.
120 width : height ratio as float.
121 * no_resume The server does not support resuming the
122 (HTTP or RTMP) download. Boolean.
124 url: Final video URL.
125 ext: Video filename extension.
126 format: The video format, defaults to ext (used for --get-format)
127 player_url: SWF Player URL (used for rtmpdump).
129 The following fields are optional:
131 alt_title: A secondary title of the video.
132 display_id An alternative identifier for the video, not necessarily
133 unique, but available before title. Typically, id is
134 something like "4234987", title "Dancing naked mole rats",
135 and display_id "dancing-naked-mole-rats"
136 thumbnails: A list of dictionaries, with the following entries:
137 * "id" (optional, string) - Thumbnail format ID
139 * "preference" (optional, int) - quality of the image
140 * "width" (optional, int)
141 * "height" (optional, int)
142 * "resolution" (optional, string "{width}x{height"},
144 thumbnail: Full URL to a video thumbnail image.
145 description: Full video description.
146 uploader: Full name of the video uploader.
147 creator: The main artist who created the video.
148 timestamp: UNIX timestamp of the moment the video became available.
149 upload_date: Video upload date (YYYYMMDD).
150 If not explicitly set, calculated from timestamp.
151 uploader_id: Nickname or id of the video uploader.
152 location: Physical location where the video was filmed.
153 subtitles: The available subtitles as a dictionary in the format
154 {language: subformats}. "subformats" is a list sorted from
155 lower to higher preference, each element is a dictionary
156 with the "ext" entry and one of:
157 * "data": The subtitles file contents
158 * "url": A url pointing to the subtitles file
159 automatic_captions: Like 'subtitles', used by the YoutubeIE for
160 automatically generated captions
161 duration: Length of the video in seconds, as an integer.
162 view_count: How many users have watched the video on the platform.
163 like_count: Number of positive ratings of the video
164 dislike_count: Number of negative ratings of the video
165 average_rating: Average rating give by users, the scale used depends on the webpage
166 comment_count: Number of comments on the video
167 comments: A list of comments, each with one or more of the following
168 properties (all but one of text or html optional):
169 * "author" - human-readable name of the comment author
170 * "author_id" - user ID of the comment author
172 * "html" - Comment as HTML
173 * "text" - Plain text of the comment
174 * "timestamp" - UNIX timestamp of comment
175 * "parent" - ID of the comment this one is replying to.
176 Set to "root" to indicate that this is a
177 comment to the original video.
178 age_limit: Age restriction for the video, as an integer (years)
179 webpage_url: The url to the video webpage, if given to youtube-dl it
180 should allow to get the same result again. (It will be set
181 by YoutubeDL if it's missing)
182 categories: A list of categories that the video falls in, for example
184 is_live: True, False, or None (=unknown). Whether this video is a
185 live stream that goes on instead of a fixed-length video.
186 start_time: Time in seconds where the reproduction should start, as
187 specified in the url.
189 Unless mentioned otherwise, the fields should be Unicode strings.
191 Unless mentioned otherwise, None is equivalent to absence of information.
194 _type "playlist" indicates multiple videos.
195 There must be a key "entries", which is a list, an iterable, or a PagedList
196 object, each element of which is a valid dictionary by this specification.
198 Additionally, playlists can have "title" and "id" attributes with the same
199 semantics as videos (see above).
202 _type "multi_video" indicates that there are multiple videos that
203 form a single show, for examples multiple acts of an opera or TV episode.
204 It must have an entries key like a playlist and contain all the keys
205 required for a video at the same time.
208 _type "url" indicates that the video must be extracted from another
209 location, possibly by a different extractor. Its only required key is:
210 "url" - the next URL to extract.
211 The key "ie_key" can be set to the class name (minus the trailing "IE",
212 e.g. "Youtube") if the extractor class is known in advance.
213 Additionally, the dictionary may have any properties of the resolved entity
214 known in advance, for example "title" if the title of the referred video is
218 _type "url_transparent" entities have the same specification as "url", but
219 indicate that the given additional information is more precise than the one
220 associated with the resolved URL.
221 This is useful when a site employs a video service that hosts the video and
222 its technical metadata, but that video service does not embed a useful
223 title, description etc.
226 Subclasses of this one should re-define the _real_initialize() and
227 _real_extract() methods and define a _VALID_URL regexp.
228 Probably, they should also be added to the list of extractors.
230 Finally, the _WORKING attribute should be set to False for broken IEs
231 in order to warn the users and skip the tests.
238 def __init__(self, downloader=None):
239 """Constructor. Receives an optional downloader."""
241 self.set_downloader(downloader)
244 def suitable(cls, url):
245 """Receives a URL and returns True if suitable for this IE."""
247 # This does not use has/getattr intentionally - we want to know whether
248 # we have cached the regexp for *this* class, whereas getattr would also
249 # match the superclass
250 if '_VALID_URL_RE' not in cls.__dict__:
251 cls._VALID_URL_RE = re.compile(cls._VALID_URL)
252 return cls._VALID_URL_RE.match(url) is not None
255 def _match_id(cls, url):
256 if '_VALID_URL_RE' not in cls.__dict__:
257 cls._VALID_URL_RE = re.compile(cls._VALID_URL)
258 m = cls._VALID_URL_RE.match(url)
264 """Getter method for _WORKING."""
267 def initialize(self):
268 """Initializes an instance (authentication, etc)."""
270 self._real_initialize()
273 def extract(self, url):
274 """Extracts URL information and returns it in list of dicts."""
277 return self._real_extract(url)
278 except ExtractorError:
280 except compat_http_client.IncompleteRead as e:
281 raise ExtractorError('A network error has occured.', cause=e, expected=True)
282 except (KeyError, StopIteration) as e:
283 raise ExtractorError('An extractor error has occured.', cause=e)
285 def set_downloader(self, downloader):
286 """Sets the downloader for this IE."""
287 self._downloader = downloader
289 def _real_initialize(self):
290 """Real initialization process. Redefine in subclasses."""
293 def _real_extract(self, url):
294 """Real extraction process. Redefine in subclasses."""
299 """A string for getting the InfoExtractor with get_info_extractor"""
300 return cls.__name__[:-2]
304 return type(self).__name__[:-2]
306 def _request_webpage(self, url_or_request, video_id, note=None, errnote=None, fatal=True):
307 """ Returns the response handle """
309 self.report_download_webpage(video_id)
310 elif note is not False:
312 self.to_screen('%s' % (note,))
314 self.to_screen('%s: %s' % (video_id, note))
316 return self._downloader.urlopen(url_or_request)
317 except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err:
321 errnote = 'Unable to download webpage'
322 errmsg = '%s: %s' % (errnote, compat_str(err))
324 raise ExtractorError(errmsg, sys.exc_info()[2], cause=err)
326 self._downloader.report_warning(errmsg)
329 def _download_webpage_handle(self, url_or_request, video_id, note=None, errnote=None, fatal=True, encoding=None):
330 """ Returns a tuple (page content as string, URL handle) """
331 # Strip hashes from the URL (#1038)
332 if isinstance(url_or_request, (compat_str, str)):
333 url_or_request = url_or_request.partition('#')[0]
335 urlh = self._request_webpage(url_or_request, video_id, note, errnote, fatal)
339 content = self._webpage_read_content(urlh, url_or_request, video_id, note, errnote, fatal, encoding=encoding)
340 return (content, urlh)
343 def _guess_encoding_from_content(content_type, webpage_bytes):
344 m = re.match(r'[a-zA-Z0-9_.-]+/[a-zA-Z0-9_.-]+\s*;\s*charset=(.+)', content_type)
346 encoding = m.group(1)
348 m = re.search(br'<meta[^>]+charset=[\'"]?([^\'")]+)[ /\'">]',
349 webpage_bytes[:1024])
351 encoding = m.group(1).decode('ascii')
352 elif webpage_bytes.startswith(b'\xff\xfe'):
359 def _webpage_read_content(self, urlh, url_or_request, video_id, note=None, errnote=None, fatal=True, prefix=None, encoding=None):
360 content_type = urlh.headers.get('Content-Type', '')
361 webpage_bytes = urlh.read()
362 if prefix is not None:
363 webpage_bytes = prefix + webpage_bytes
365 encoding = self._guess_encoding_from_content(content_type, webpage_bytes)
366 if self._downloader.params.get('dump_intermediate_pages', False):
368 url = url_or_request.get_full_url()
369 except AttributeError:
371 self.to_screen('Dumping request to ' + url)
372 dump = base64.b64encode(webpage_bytes).decode('ascii')
373 self._downloader.to_screen(dump)
374 if self._downloader.params.get('write_pages', False):
376 url = url_or_request.get_full_url()
377 except AttributeError:
379 basen = '%s_%s' % (video_id, url)
381 h = '___' + hashlib.md5(basen.encode('utf-8')).hexdigest()
382 basen = basen[:240 - len(h)] + h
383 raw_filename = basen + '.dump'
384 filename = sanitize_filename(raw_filename, restricted=True)
385 self.to_screen('Saving request to ' + filename)
386 # Working around MAX_PATH limitation on Windows (see
387 # http://msdn.microsoft.com/en-us/library/windows/desktop/aa365247(v=vs.85).aspx)
389 absfilepath = os.path.abspath(filename)
390 if len(absfilepath) > 259:
391 filename = '\\\\?\\' + absfilepath
392 with open(filename, 'wb') as outf:
393 outf.write(webpage_bytes)
396 content = webpage_bytes.decode(encoding, 'replace')
398 content = webpage_bytes.decode('utf-8', 'replace')
400 if ('<title>Access to this site is blocked</title>' in content and
401 'Websense' in content[:512]):
402 msg = 'Access to this webpage has been blocked by Websense filtering software in your network.'
403 blocked_iframe = self._html_search_regex(
404 r'<iframe src="([^"]+)"', content,
405 'Websense information URL', default=None)
407 msg += ' Visit %s for more details' % blocked_iframe
408 raise ExtractorError(msg, expected=True)
409 if '<title>The URL you requested has been blocked</title>' in content[:512]:
411 'Access to this webpage has been blocked by Indian censorship. '
412 'Use a VPN or proxy server (with --proxy) to route around it.')
413 block_msg = self._html_search_regex(
414 r'</h1><p>(.*?)</p>',
415 content, 'block message', default=None)
417 msg += ' (Message: "%s")' % block_msg.replace('\n', ' ')
418 raise ExtractorError(msg, expected=True)
422 def _download_webpage(self, url_or_request, video_id, note=None, errnote=None, fatal=True, tries=1, timeout=5, encoding=None):
423 """ Returns the data of the page as a string """
426 while success is False:
428 res = self._download_webpage_handle(url_or_request, video_id, note, errnote, fatal, encoding=encoding)
430 except compat_http_client.IncompleteRead as e:
432 if try_count >= tries:
434 self._sleep(timeout, video_id)
441 def _download_xml(self, url_or_request, video_id,
442 note='Downloading XML', errnote='Unable to download XML',
443 transform_source=None, fatal=True, encoding=None):
444 """Return the xml as an xml.etree.ElementTree.Element"""
445 xml_string = self._download_webpage(
446 url_or_request, video_id, note, errnote, fatal=fatal, encoding=encoding)
447 if xml_string is False:
450 xml_string = transform_source(xml_string)
451 return xml.etree.ElementTree.fromstring(xml_string.encode('utf-8'))
453 def _download_json(self, url_or_request, video_id,
454 note='Downloading JSON metadata',
455 errnote='Unable to download JSON metadata',
456 transform_source=None,
457 fatal=True, encoding=None):
458 json_string = self._download_webpage(
459 url_or_request, video_id, note, errnote, fatal=fatal,
461 if (not fatal) and json_string is False:
463 return self._parse_json(
464 json_string, video_id, transform_source=transform_source, fatal=fatal)
466 def _parse_json(self, json_string, video_id, transform_source=None, fatal=True):
468 json_string = transform_source(json_string)
470 return json.loads(json_string)
471 except ValueError as ve:
472 errmsg = '%s: Failed to parse JSON ' % video_id
474 raise ExtractorError(errmsg, cause=ve)
476 self.report_warning(errmsg + str(ve))
478 def report_warning(self, msg, video_id=None):
479 idstr = '' if video_id is None else '%s: ' % video_id
480 self._downloader.report_warning(
481 '[%s] %s%s' % (self.IE_NAME, idstr, msg))
483 def to_screen(self, msg):
484 """Print msg to screen, prefixing it with '[ie_name]'"""
485 self._downloader.to_screen('[%s] %s' % (self.IE_NAME, msg))
487 def report_extraction(self, id_or_name):
488 """Report information extraction."""
489 self.to_screen('%s: Extracting information' % id_or_name)
491 def report_download_webpage(self, video_id):
492 """Report webpage download."""
493 self.to_screen('%s: Downloading webpage' % video_id)
495 def report_age_confirmation(self):
496 """Report attempt to confirm age."""
497 self.to_screen('Confirming age')
499 def report_login(self):
500 """Report attempt to log in."""
501 self.to_screen('Logging in')
503 # Methods for following #608
505 def url_result(url, ie=None, video_id=None, video_title=None):
506 """Returns a url that points to a page that should be processed"""
507 # TODO: ie should be the class used for getting the info
508 video_info = {'_type': 'url',
511 if video_id is not None:
512 video_info['id'] = video_id
513 if video_title is not None:
514 video_info['title'] = video_title
518 def playlist_result(entries, playlist_id=None, playlist_title=None, playlist_description=None):
519 """Returns a playlist"""
520 video_info = {'_type': 'playlist',
523 video_info['id'] = playlist_id
525 video_info['title'] = playlist_title
526 if playlist_description:
527 video_info['description'] = playlist_description
530 def _search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
532 Perform a regex search on the given string, using a single or a list of
533 patterns returning the first matching group.
534 In case of failure return a default value or raise a WARNING or a
535 RegexNotFoundError, depending on fatal, specifying the field name.
537 if isinstance(pattern, (str, compat_str, compiled_regex_type)):
538 mobj = re.search(pattern, string, flags)
541 mobj = re.search(p, string, flags)
545 if not self._downloader.params.get('no_color') and os.name != 'nt' and sys.stderr.isatty():
546 _name = '\033[0;34m%s\033[0m' % name
552 # return the first matching group
553 return next(g for g in mobj.groups() if g is not None)
555 return mobj.group(group)
556 elif default is not NO_DEFAULT:
559 raise RegexNotFoundError('Unable to extract %s' % _name)
561 self._downloader.report_warning('unable to extract %s' % _name + bug_reports_message())
564 def _html_search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
566 Like _search_regex, but strips HTML tags and unescapes entities.
568 res = self._search_regex(pattern, string, name, default, fatal, flags, group)
570 return clean_html(res).strip()
574 def _get_login_info(self):
576 Get the login info as (username, password)
577 It will look in the netrc file using the _NETRC_MACHINE value
578 If there's no info available, return (None, None)
580 if self._downloader is None:
585 downloader_params = self._downloader.params
587 # Attempt to use provided username and password or .netrc data
588 if downloader_params.get('username', None) is not None:
589 username = downloader_params['username']
590 password = downloader_params['password']
591 elif downloader_params.get('usenetrc', False):
593 info = netrc.netrc().authenticators(self._NETRC_MACHINE)
598 raise netrc.NetrcParseError('No authenticators for %s' % self._NETRC_MACHINE)
599 except (IOError, netrc.NetrcParseError) as err:
600 self._downloader.report_warning('parsing .netrc: %s' % compat_str(err))
602 return (username, password)
604 def _get_tfa_info(self):
606 Get the two-factor authentication info
607 TODO - asking the user will be required for sms/phone verify
608 currently just uses the command line option
609 If there's no info available, return None
611 if self._downloader is None:
613 downloader_params = self._downloader.params
615 if downloader_params.get('twofactor', None) is not None:
616 return downloader_params['twofactor']
620 # Helper functions for extracting OpenGraph info
622 def _og_regexes(prop):
623 content_re = r'content=(?:"([^>]+?)"|\'([^>]+?)\')'
624 property_re = r'(?:name|property)=[\'"]og:%s[\'"]' % re.escape(prop)
625 template = r'<meta[^>]+?%s[^>]+?%s'
627 template % (property_re, content_re),
628 template % (content_re, property_re),
631 def _og_search_property(self, prop, html, name=None, **kargs):
633 name = 'OpenGraph %s' % prop
634 escaped = self._search_regex(self._og_regexes(prop), html, name, flags=re.DOTALL, **kargs)
637 return unescapeHTML(escaped)
639 def _og_search_thumbnail(self, html, **kargs):
640 return self._og_search_property('image', html, 'thumbnail url', fatal=False, **kargs)
642 def _og_search_description(self, html, **kargs):
643 return self._og_search_property('description', html, fatal=False, **kargs)
645 def _og_search_title(self, html, **kargs):
646 return self._og_search_property('title', html, **kargs)
648 def _og_search_video_url(self, html, name='video url', secure=True, **kargs):
649 regexes = self._og_regexes('video') + self._og_regexes('video:url')
651 regexes = self._og_regexes('video:secure_url') + regexes
652 return self._html_search_regex(regexes, html, name, **kargs)
654 def _og_search_url(self, html, **kargs):
655 return self._og_search_property('url', html, **kargs)
657 def _html_search_meta(self, name, html, display_name=None, fatal=False, **kwargs):
658 if display_name is None:
660 return self._html_search_regex(
662 (?=[^>]+(?:itemprop|name|property)=(["\']?)%s\1)
663 [^>]+?content=(["\'])(?P<content>.*?)\2''' % re.escape(name),
664 html, display_name, fatal=fatal, group='content', **kwargs)
666 def _dc_search_uploader(self, html):
667 return self._html_search_meta('dc.creator', html, 'uploader')
669 def _rta_search(self, html):
670 # See http://www.rtalabel.org/index.php?content=howtofaq#single
671 if re.search(r'(?ix)<meta\s+name="rating"\s+'
672 r' content="RTA-5042-1996-1400-1577-RTA"',
677 def _media_rating_search(self, html):
678 # See http://www.tjg-designs.com/WP/metadata-code-examples-adding-metadata-to-your-web-pages/
679 rating = self._html_search_meta('rating', html)
691 return RATING_TABLE.get(rating.lower(), None)
693 def _family_friendly_search(self, html):
694 # See http://schema.org/VideoObject
695 family_friendly = self._html_search_meta('isFamilyFriendly', html)
697 if not family_friendly:
706 return RATING_TABLE.get(family_friendly.lower(), None)
708 def _twitter_search_player(self, html):
709 return self._html_search_meta('twitter:player', html,
710 'twitter card player')
713 def _hidden_inputs(html):
715 (input.group('name'), input.group('value')) for input in re.finditer(
718 type=(?P<q_hidden>["\'])hidden(?P=q_hidden)\s+
719 name=(?P<q_name>["\'])(?P<name>.+?)(?P=q_name)\s+
720 (?:id=(?P<q_id>["\']).+?(?P=q_id)\s+)?
721 value=(?P<q_value>["\'])(?P<value>.*?)(?P=q_value)
725 def _form_hidden_inputs(self, form_id, html):
726 form = self._search_regex(
727 r'(?s)<form[^>]+?id=(["\'])%s\1[^>]*>(?P<form>.+?)</form>' % form_id,
728 html, '%s form' % form_id, group='form')
729 return self._hidden_inputs(form)
731 def _sort_formats(self, formats, field_preference=None):
733 raise ExtractorError('No video formats found')
736 # TODO remove the following workaround
737 from ..utils import determine_ext
738 if not f.get('ext') and 'url' in f:
739 f['ext'] = determine_ext(f['url'])
741 if isinstance(field_preference, (list, tuple)):
742 return tuple(f.get(field) if f.get(field) is not None else -1 for field in field_preference)
744 preference = f.get('preference')
745 if preference is None:
746 proto = f.get('protocol')
748 proto = compat_urllib_parse_urlparse(f.get('url', '')).scheme
750 preference = 0 if proto in ['http', 'https'] else -0.1
751 if f.get('ext') in ['f4f', 'f4m']: # Not yet supported
754 if f.get('vcodec') == 'none': # audio only
755 if self._downloader.params.get('prefer_free_formats'):
756 ORDER = ['aac', 'mp3', 'm4a', 'webm', 'ogg', 'opus']
758 ORDER = ['webm', 'opus', 'ogg', 'mp3', 'aac', 'm4a']
761 audio_ext_preference = ORDER.index(f['ext'])
763 audio_ext_preference = -1
765 if self._downloader.params.get('prefer_free_formats'):
766 ORDER = ['flv', 'mp4', 'webm']
768 ORDER = ['webm', 'flv', 'mp4']
770 ext_preference = ORDER.index(f['ext'])
773 audio_ext_preference = 0
777 f.get('language_preference') if f.get('language_preference') is not None else -1,
778 f.get('quality') if f.get('quality') is not None else -1,
779 f.get('tbr') if f.get('tbr') is not None else -1,
780 f.get('filesize') if f.get('filesize') is not None else -1,
781 f.get('vbr') if f.get('vbr') is not None else -1,
782 f.get('height') if f.get('height') is not None else -1,
783 f.get('width') if f.get('width') is not None else -1,
785 f.get('abr') if f.get('abr') is not None else -1,
786 audio_ext_preference,
787 f.get('fps') if f.get('fps') is not None else -1,
788 f.get('filesize_approx') if f.get('filesize_approx') is not None else -1,
789 f.get('source_preference') if f.get('source_preference') is not None else -1,
790 f.get('format_id') if f.get('format_id') is not None else '',
792 formats.sort(key=_formats_key)
794 def _check_formats(self, formats, video_id):
797 lambda f: self._is_valid_url(
799 item='%s video format' % f.get('format_id') if f.get('format_id') else 'video'),
802 def _is_valid_url(self, url, video_id, item='video'):
803 url = self._proto_relative_url(url, scheme='http:')
804 # For now assume non HTTP(S) URLs always valid
805 if not (url.startswith('http://') or url.startswith('https://')):
808 self._request_webpage(url, video_id, 'Checking %s URL' % item)
810 except ExtractorError as e:
811 if isinstance(e.cause, compat_HTTPError):
813 '%s: %s URL is invalid, skipping' % (video_id, item))
817 def http_scheme(self):
818 """ Either "http:" or "https:", depending on the user's preferences """
821 if self._downloader.params.get('prefer_insecure', False)
824 def _proto_relative_url(self, url, scheme=None):
827 if url.startswith('//'):
829 scheme = self.http_scheme()
834 def _sleep(self, timeout, video_id, msg_template=None):
835 if msg_template is None:
836 msg_template = '%(video_id)s: Waiting for %(timeout)s seconds'
837 msg = msg_template % {'video_id': video_id, 'timeout': timeout}
841 def _extract_f4m_formats(self, manifest_url, video_id, preference=None, f4m_id=None,
842 transform_source=lambda s: fix_xml_ampersands(s).strip()):
843 manifest = self._download_xml(
844 manifest_url, video_id, 'Downloading f4m manifest',
845 'Unable to download f4m manifest',
846 # Some manifests may be malformed, e.g. prosiebensat1 generated manifests
847 # (see https://github.com/rg3/youtube-dl/issues/6215#issuecomment-121704244)
848 transform_source=transform_source)
851 manifest_version = '1.0'
852 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/1.0}media')
854 manifest_version = '2.0'
855 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/2.0}media')
856 for i, media_el in enumerate(media_nodes):
857 if manifest_version == '2.0':
858 media_url = media_el.attrib.get('href') or media_el.attrib.get('url')
862 media_url if media_url.startswith('http://') or media_url.startswith('https://')
863 else ('/'.join(manifest_url.split('/')[:-1]) + '/' + media_url))
864 # If media_url is itself a f4m manifest do the recursive extraction
865 # since bitrates in parent manifest (this one) and media_url manifest
866 # may differ leading to inability to resolve the format by requested
867 # bitrate in f4m downloader
868 if determine_ext(manifest_url) == 'f4m':
869 formats.extend(self._extract_f4m_formats(manifest_url, video_id, preference, f4m_id))
871 tbr = int_or_none(media_el.attrib.get('bitrate'))
873 'format_id': '-'.join(filter(None, [f4m_id, compat_str(i if tbr is None else tbr)])),
877 'width': int_or_none(media_el.attrib.get('width')),
878 'height': int_or_none(media_el.attrib.get('height')),
879 'preference': preference,
881 self._sort_formats(formats)
885 def _extract_m3u8_formats(self, m3u8_url, video_id, ext=None,
886 entry_protocol='m3u8', preference=None,
887 m3u8_id=None, note=None, errnote=None,
891 'format_id': '-'.join(filter(None, [m3u8_id, 'meta'])),
895 'preference': preference - 1 if preference else -1,
896 'resolution': 'multiple',
897 'format_note': 'Quality selection URL',
900 format_url = lambda u: (
902 if re.match(r'^https?://', u)
903 else compat_urlparse.urljoin(m3u8_url, u))
905 m3u8_doc = self._download_webpage(
907 note=note or 'Downloading m3u8 information',
908 errnote=errnote or 'Failed to download m3u8 information',
910 if m3u8_doc is False:
915 r'(?P<key>[a-zA-Z_-]+)=(?P<val>"[^"]+"|[^",]+)(?:,|$)')
916 for line in m3u8_doc.splitlines():
917 if line.startswith('#EXT-X-STREAM-INF:'):
919 for m in kv_rex.finditer(line):
921 if v.startswith('"'):
923 last_info[m.group('key')] = v
924 elif line.startswith('#EXT-X-MEDIA:'):
926 for m in kv_rex.finditer(line):
928 if v.startswith('"'):
930 last_media[m.group('key')] = v
931 elif line.startswith('#') or not line.strip():
934 if last_info is None:
935 formats.append({'url': format_url(line)})
937 tbr = int_or_none(last_info.get('BANDWIDTH'), scale=1000)
940 format_id.append(m3u8_id)
941 last_media_name = last_media.get('NAME') if last_media and last_media.get('TYPE') != 'SUBTITLES' else None
942 format_id.append(last_media_name if last_media_name else '%d' % (tbr if tbr else len(formats)))
944 'format_id': '-'.join(format_id),
945 'url': format_url(line.strip()),
948 'protocol': entry_protocol,
949 'preference': preference,
951 codecs = last_info.get('CODECS')
953 # TODO: looks like video codec is not always necessarily goes first
954 va_codecs = codecs.split(',')
956 f['vcodec'] = va_codecs[0].partition('.')[0]
957 if len(va_codecs) > 1 and va_codecs[1]:
958 f['acodec'] = va_codecs[1].partition('.')[0]
959 resolution = last_info.get('RESOLUTION')
961 width_str, height_str = resolution.split('x')
962 f['width'] = int(width_str)
963 f['height'] = int(height_str)
964 if last_media is not None:
965 f['m3u8_media'] = last_media
969 self._sort_formats(formats)
972 # TODO: improve extraction
973 def _extract_smil_formats(self, smil_url, video_id, fatal=True):
974 smil = self._download_xml(
975 smil_url, video_id, 'Downloading SMIL file',
976 'Unable to download SMIL file', fatal=fatal)
981 base = smil.find('./head/meta').get('base')
985 if smil.findall('./body/seq/video'):
986 video = smil.findall('./body/seq/video')[0]
987 fmts, rtmp_count = self._parse_smil_video(video, video_id, base, rtmp_count)
990 for video in smil.findall('./body/switch/video'):
991 fmts, rtmp_count = self._parse_smil_video(video, video_id, base, rtmp_count)
994 self._sort_formats(formats)
998 def _parse_smil_video(self, video, video_id, base, rtmp_count):
999 src = video.get('src')
1001 return [], rtmp_count
1002 bitrate = int_or_none(video.get('system-bitrate') or video.get('systemBitrate'), 1000)
1003 width = int_or_none(video.get('width'))
1004 height = int_or_none(video.get('height'))
1005 proto = video.get('proto')
1008 if base.startswith('rtmp'):
1010 elif base.startswith('http'):
1012 ext = video.get('ext')
1014 return self._extract_m3u8_formats(src, video_id, ext), rtmp_count
1015 elif proto == 'rtmp':
1017 streamer = video.get('streamer') or base
1022 'format_id': 'rtmp-%d' % (rtmp_count if bitrate is None else bitrate),
1027 elif proto.startswith('http'):
1030 'ext': ext or 'flv',
1036 def _live_title(self, name):
1037 """ Generate the title for a live video """
1038 now = datetime.datetime.now()
1039 now_str = now.strftime("%Y-%m-%d %H:%M")
1040 return name + ' ' + now_str
1042 def _int(self, v, name, fatal=False, **kwargs):
1043 res = int_or_none(v, **kwargs)
1044 if 'get_attr' in kwargs:
1045 print(getattr(v, kwargs['get_attr']))
1047 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
1049 raise ExtractorError(msg)
1051 self._downloader.report_warning(msg)
1054 def _float(self, v, name, fatal=False, **kwargs):
1055 res = float_or_none(v, **kwargs)
1057 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
1059 raise ExtractorError(msg)
1061 self._downloader.report_warning(msg)
1064 def _set_cookie(self, domain, name, value, expire_time=None):
1065 cookie = compat_cookiejar.Cookie(
1066 0, name, value, None, None, domain, None,
1067 None, '/', True, False, expire_time, '', None, None, None)
1068 self._downloader.cookiejar.set_cookie(cookie)
1070 def get_testcases(self, include_onlymatching=False):
1071 t = getattr(self, '_TEST', None)
1073 assert not hasattr(self, '_TESTS'), \
1074 '%s has _TEST and _TESTS' % type(self).__name__
1077 tests = getattr(self, '_TESTS', [])
1079 if not include_onlymatching and t.get('only_matching', False):
1081 t['name'] = type(self).__name__[:-len('IE')]
1084 def is_suitable(self, age_limit):
1085 """ Test whether the extractor is generally suitable for the given
1086 age limit (i.e. pornographic sites are not, all others usually are) """
1088 any_restricted = False
1089 for tc in self.get_testcases(include_onlymatching=False):
1090 if 'playlist' in tc:
1091 tc = tc['playlist'][0]
1092 is_restricted = age_restricted(
1093 tc.get('info_dict', {}).get('age_limit'), age_limit)
1094 if not is_restricted:
1096 any_restricted = any_restricted or is_restricted
1097 return not any_restricted
1099 def extract_subtitles(self, *args, **kwargs):
1100 if (self._downloader.params.get('writesubtitles', False) or
1101 self._downloader.params.get('listsubtitles')):
1102 return self._get_subtitles(*args, **kwargs)
1105 def _get_subtitles(self, *args, **kwargs):
1106 raise NotImplementedError("This method must be implemented by subclasses")
1108 def extract_automatic_captions(self, *args, **kwargs):
1109 if (self._downloader.params.get('writeautomaticsub', False) or
1110 self._downloader.params.get('listsubtitles')):
1111 return self._get_automatic_captions(*args, **kwargs)
1114 def _get_automatic_captions(self, *args, **kwargs):
1115 raise NotImplementedError("This method must be implemented by subclasses")
1118 class SearchInfoExtractor(InfoExtractor):
1120 Base class for paged search queries extractors.
1121 They accept urls in the format _SEARCH_KEY(|all|[0-9]):{query}
1122 Instances should define _SEARCH_KEY and _MAX_RESULTS.
1126 def _make_valid_url(cls):
1127 return r'%s(?P<prefix>|[1-9][0-9]*|all):(?P<query>[\s\S]+)' % cls._SEARCH_KEY
1130 def suitable(cls, url):
1131 return re.match(cls._make_valid_url(), url) is not None
1133 def _real_extract(self, query):
1134 mobj = re.match(self._make_valid_url(), query)
1136 raise ExtractorError('Invalid search query "%s"' % query)
1138 prefix = mobj.group('prefix')
1139 query = mobj.group('query')
1141 return self._get_n_results(query, 1)
1142 elif prefix == 'all':
1143 return self._get_n_results(query, self._MAX_RESULTS)
1147 raise ExtractorError('invalid download number %s for query "%s"' % (n, query))
1148 elif n > self._MAX_RESULTS:
1149 self._downloader.report_warning('%s returns max %i results (you requested %i)' % (self._SEARCH_KEY, self._MAX_RESULTS, n))
1150 n = self._MAX_RESULTS
1151 return self._get_n_results(query, n)
1153 def _get_n_results(self, query, n):
1154 """Get a specified number of results for a query"""
1155 raise NotImplementedError("This method must be implemented by subclasses")
1158 def SEARCH_KEY(self):
1159 return self._SEARCH_KEY