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.
188 end_time: Time in seconds where the reproduction should end, as
189 specified in the URL.
191 Unless mentioned otherwise, the fields should be Unicode strings.
193 Unless mentioned otherwise, None is equivalent to absence of information.
196 _type "playlist" indicates multiple videos.
197 There must be a key "entries", which is a list, an iterable, or a PagedList
198 object, each element of which is a valid dictionary by this specification.
200 Additionally, playlists can have "title" and "id" attributes with the same
201 semantics as videos (see above).
204 _type "multi_video" indicates that there are multiple videos that
205 form a single show, for examples multiple acts of an opera or TV episode.
206 It must have an entries key like a playlist and contain all the keys
207 required for a video at the same time.
210 _type "url" indicates that the video must be extracted from another
211 location, possibly by a different extractor. Its only required key is:
212 "url" - the next URL to extract.
213 The key "ie_key" can be set to the class name (minus the trailing "IE",
214 e.g. "Youtube") if the extractor class is known in advance.
215 Additionally, the dictionary may have any properties of the resolved entity
216 known in advance, for example "title" if the title of the referred video is
220 _type "url_transparent" entities have the same specification as "url", but
221 indicate that the given additional information is more precise than the one
222 associated with the resolved URL.
223 This is useful when a site employs a video service that hosts the video and
224 its technical metadata, but that video service does not embed a useful
225 title, description etc.
228 Subclasses of this one should re-define the _real_initialize() and
229 _real_extract() methods and define a _VALID_URL regexp.
230 Probably, they should also be added to the list of extractors.
232 Finally, the _WORKING attribute should be set to False for broken IEs
233 in order to warn the users and skip the tests.
240 def __init__(self, downloader=None):
241 """Constructor. Receives an optional downloader."""
243 self.set_downloader(downloader)
246 def suitable(cls, url):
247 """Receives a URL and returns True if suitable for this IE."""
249 # This does not use has/getattr intentionally - we want to know whether
250 # we have cached the regexp for *this* class, whereas getattr would also
251 # match the superclass
252 if '_VALID_URL_RE' not in cls.__dict__:
253 cls._VALID_URL_RE = re.compile(cls._VALID_URL)
254 return cls._VALID_URL_RE.match(url) is not None
257 def _match_id(cls, url):
258 if '_VALID_URL_RE' not in cls.__dict__:
259 cls._VALID_URL_RE = re.compile(cls._VALID_URL)
260 m = cls._VALID_URL_RE.match(url)
266 """Getter method for _WORKING."""
269 def initialize(self):
270 """Initializes an instance (authentication, etc)."""
272 self._real_initialize()
275 def extract(self, url):
276 """Extracts URL information and returns it in list of dicts."""
279 return self._real_extract(url)
280 except ExtractorError:
282 except compat_http_client.IncompleteRead as e:
283 raise ExtractorError('A network error has occured.', cause=e, expected=True)
284 except (KeyError, StopIteration) as e:
285 raise ExtractorError('An extractor error has occured.', cause=e)
287 def set_downloader(self, downloader):
288 """Sets the downloader for this IE."""
289 self._downloader = downloader
291 def _real_initialize(self):
292 """Real initialization process. Redefine in subclasses."""
295 def _real_extract(self, url):
296 """Real extraction process. Redefine in subclasses."""
301 """A string for getting the InfoExtractor with get_info_extractor"""
302 return cls.__name__[:-2]
306 return type(self).__name__[:-2]
308 def _request_webpage(self, url_or_request, video_id, note=None, errnote=None, fatal=True):
309 """ Returns the response handle """
311 self.report_download_webpage(video_id)
312 elif note is not False:
314 self.to_screen('%s' % (note,))
316 self.to_screen('%s: %s' % (video_id, note))
318 return self._downloader.urlopen(url_or_request)
319 except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err:
323 errnote = 'Unable to download webpage'
324 errmsg = '%s: %s' % (errnote, compat_str(err))
326 raise ExtractorError(errmsg, sys.exc_info()[2], cause=err)
328 self._downloader.report_warning(errmsg)
331 def _download_webpage_handle(self, url_or_request, video_id, note=None, errnote=None, fatal=True, encoding=None):
332 """ Returns a tuple (page content as string, URL handle) """
333 # Strip hashes from the URL (#1038)
334 if isinstance(url_or_request, (compat_str, str)):
335 url_or_request = url_or_request.partition('#')[0]
337 urlh = self._request_webpage(url_or_request, video_id, note, errnote, fatal)
341 content = self._webpage_read_content(urlh, url_or_request, video_id, note, errnote, fatal, encoding=encoding)
342 return (content, urlh)
345 def _guess_encoding_from_content(content_type, webpage_bytes):
346 m = re.match(r'[a-zA-Z0-9_.-]+/[a-zA-Z0-9_.-]+\s*;\s*charset=(.+)', content_type)
348 encoding = m.group(1)
350 m = re.search(br'<meta[^>]+charset=[\'"]?([^\'")]+)[ /\'">]',
351 webpage_bytes[:1024])
353 encoding = m.group(1).decode('ascii')
354 elif webpage_bytes.startswith(b'\xff\xfe'):
361 def _webpage_read_content(self, urlh, url_or_request, video_id, note=None, errnote=None, fatal=True, prefix=None, encoding=None):
362 content_type = urlh.headers.get('Content-Type', '')
363 webpage_bytes = urlh.read()
364 if prefix is not None:
365 webpage_bytes = prefix + webpage_bytes
367 encoding = self._guess_encoding_from_content(content_type, webpage_bytes)
368 if self._downloader.params.get('dump_intermediate_pages', False):
370 url = url_or_request.get_full_url()
371 except AttributeError:
373 self.to_screen('Dumping request to ' + url)
374 dump = base64.b64encode(webpage_bytes).decode('ascii')
375 self._downloader.to_screen(dump)
376 if self._downloader.params.get('write_pages', False):
378 url = url_or_request.get_full_url()
379 except AttributeError:
381 basen = '%s_%s' % (video_id, url)
383 h = '___' + hashlib.md5(basen.encode('utf-8')).hexdigest()
384 basen = basen[:240 - len(h)] + h
385 raw_filename = basen + '.dump'
386 filename = sanitize_filename(raw_filename, restricted=True)
387 self.to_screen('Saving request to ' + filename)
388 # Working around MAX_PATH limitation on Windows (see
389 # http://msdn.microsoft.com/en-us/library/windows/desktop/aa365247(v=vs.85).aspx)
391 absfilepath = os.path.abspath(filename)
392 if len(absfilepath) > 259:
393 filename = '\\\\?\\' + absfilepath
394 with open(filename, 'wb') as outf:
395 outf.write(webpage_bytes)
398 content = webpage_bytes.decode(encoding, 'replace')
400 content = webpage_bytes.decode('utf-8', 'replace')
402 if ('<title>Access to this site is blocked</title>' in content and
403 'Websense' in content[:512]):
404 msg = 'Access to this webpage has been blocked by Websense filtering software in your network.'
405 blocked_iframe = self._html_search_regex(
406 r'<iframe src="([^"]+)"', content,
407 'Websense information URL', default=None)
409 msg += ' Visit %s for more details' % blocked_iframe
410 raise ExtractorError(msg, expected=True)
411 if '<title>The URL you requested has been blocked</title>' in content[:512]:
413 'Access to this webpage has been blocked by Indian censorship. '
414 'Use a VPN or proxy server (with --proxy) to route around it.')
415 block_msg = self._html_search_regex(
416 r'</h1><p>(.*?)</p>',
417 content, 'block message', default=None)
419 msg += ' (Message: "%s")' % block_msg.replace('\n', ' ')
420 raise ExtractorError(msg, expected=True)
424 def _download_webpage(self, url_or_request, video_id, note=None, errnote=None, fatal=True, tries=1, timeout=5, encoding=None):
425 """ Returns the data of the page as a string """
428 while success is False:
430 res = self._download_webpage_handle(url_or_request, video_id, note, errnote, fatal, encoding=encoding)
432 except compat_http_client.IncompleteRead as e:
434 if try_count >= tries:
436 self._sleep(timeout, video_id)
443 def _download_xml(self, url_or_request, video_id,
444 note='Downloading XML', errnote='Unable to download XML',
445 transform_source=None, fatal=True, encoding=None):
446 """Return the xml as an xml.etree.ElementTree.Element"""
447 xml_string = self._download_webpage(
448 url_or_request, video_id, note, errnote, fatal=fatal, encoding=encoding)
449 if xml_string is False:
452 xml_string = transform_source(xml_string)
453 return xml.etree.ElementTree.fromstring(xml_string.encode('utf-8'))
455 def _download_json(self, url_or_request, video_id,
456 note='Downloading JSON metadata',
457 errnote='Unable to download JSON metadata',
458 transform_source=None,
459 fatal=True, encoding=None):
460 json_string = self._download_webpage(
461 url_or_request, video_id, note, errnote, fatal=fatal,
463 if (not fatal) and json_string is False:
465 return self._parse_json(
466 json_string, video_id, transform_source=transform_source, fatal=fatal)
468 def _parse_json(self, json_string, video_id, transform_source=None, fatal=True):
470 json_string = transform_source(json_string)
472 return json.loads(json_string)
473 except ValueError as ve:
474 errmsg = '%s: Failed to parse JSON ' % video_id
476 raise ExtractorError(errmsg, cause=ve)
478 self.report_warning(errmsg + str(ve))
480 def report_warning(self, msg, video_id=None):
481 idstr = '' if video_id is None else '%s: ' % video_id
482 self._downloader.report_warning(
483 '[%s] %s%s' % (self.IE_NAME, idstr, msg))
485 def to_screen(self, msg):
486 """Print msg to screen, prefixing it with '[ie_name]'"""
487 self._downloader.to_screen('[%s] %s' % (self.IE_NAME, msg))
489 def report_extraction(self, id_or_name):
490 """Report information extraction."""
491 self.to_screen('%s: Extracting information' % id_or_name)
493 def report_download_webpage(self, video_id):
494 """Report webpage download."""
495 self.to_screen('%s: Downloading webpage' % video_id)
497 def report_age_confirmation(self):
498 """Report attempt to confirm age."""
499 self.to_screen('Confirming age')
501 def report_login(self):
502 """Report attempt to log in."""
503 self.to_screen('Logging in')
505 # Methods for following #608
507 def url_result(url, ie=None, video_id=None, video_title=None):
508 """Returns a URL that points to a page that should be processed"""
509 # TODO: ie should be the class used for getting the info
510 video_info = {'_type': 'url',
513 if video_id is not None:
514 video_info['id'] = video_id
515 if video_title is not None:
516 video_info['title'] = video_title
520 def playlist_result(entries, playlist_id=None, playlist_title=None, playlist_description=None):
521 """Returns a playlist"""
522 video_info = {'_type': 'playlist',
525 video_info['id'] = playlist_id
527 video_info['title'] = playlist_title
528 if playlist_description:
529 video_info['description'] = playlist_description
532 def _search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
534 Perform a regex search on the given string, using a single or a list of
535 patterns returning the first matching group.
536 In case of failure return a default value or raise a WARNING or a
537 RegexNotFoundError, depending on fatal, specifying the field name.
539 if isinstance(pattern, (str, compat_str, compiled_regex_type)):
540 mobj = re.search(pattern, string, flags)
543 mobj = re.search(p, string, flags)
547 if not self._downloader.params.get('no_color') and os.name != 'nt' and sys.stderr.isatty():
548 _name = '\033[0;34m%s\033[0m' % name
554 # return the first matching group
555 return next(g for g in mobj.groups() if g is not None)
557 return mobj.group(group)
558 elif default is not NO_DEFAULT:
561 raise RegexNotFoundError('Unable to extract %s' % _name)
563 self._downloader.report_warning('unable to extract %s' % _name + bug_reports_message())
566 def _html_search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
568 Like _search_regex, but strips HTML tags and unescapes entities.
570 res = self._search_regex(pattern, string, name, default, fatal, flags, group)
572 return clean_html(res).strip()
576 def _get_login_info(self):
578 Get the login info as (username, password)
579 It will look in the netrc file using the _NETRC_MACHINE value
580 If there's no info available, return (None, None)
582 if self._downloader is None:
587 downloader_params = self._downloader.params
589 # Attempt to use provided username and password or .netrc data
590 if downloader_params.get('username', None) is not None:
591 username = downloader_params['username']
592 password = downloader_params['password']
593 elif downloader_params.get('usenetrc', False):
595 info = netrc.netrc().authenticators(self._NETRC_MACHINE)
600 raise netrc.NetrcParseError('No authenticators for %s' % self._NETRC_MACHINE)
601 except (IOError, netrc.NetrcParseError) as err:
602 self._downloader.report_warning('parsing .netrc: %s' % compat_str(err))
604 return (username, password)
606 def _get_tfa_info(self):
608 Get the two-factor authentication info
609 TODO - asking the user will be required for sms/phone verify
610 currently just uses the command line option
611 If there's no info available, return None
613 if self._downloader is None:
615 downloader_params = self._downloader.params
617 if downloader_params.get('twofactor', None) is not None:
618 return downloader_params['twofactor']
622 # Helper functions for extracting OpenGraph info
624 def _og_regexes(prop):
625 content_re = r'content=(?:"([^>]+?)"|\'([^>]+?)\')'
626 property_re = r'(?:name|property)=[\'"]og:%s[\'"]' % re.escape(prop)
627 template = r'<meta[^>]+?%s[^>]+?%s'
629 template % (property_re, content_re),
630 template % (content_re, property_re),
633 def _og_search_property(self, prop, html, name=None, **kargs):
635 name = 'OpenGraph %s' % prop
636 escaped = self._search_regex(self._og_regexes(prop), html, name, flags=re.DOTALL, **kargs)
639 return unescapeHTML(escaped)
641 def _og_search_thumbnail(self, html, **kargs):
642 return self._og_search_property('image', html, 'thumbnail URL', fatal=False, **kargs)
644 def _og_search_description(self, html, **kargs):
645 return self._og_search_property('description', html, fatal=False, **kargs)
647 def _og_search_title(self, html, **kargs):
648 return self._og_search_property('title', html, **kargs)
650 def _og_search_video_url(self, html, name='video url', secure=True, **kargs):
651 regexes = self._og_regexes('video') + self._og_regexes('video:url')
653 regexes = self._og_regexes('video:secure_url') + regexes
654 return self._html_search_regex(regexes, html, name, **kargs)
656 def _og_search_url(self, html, **kargs):
657 return self._og_search_property('url', html, **kargs)
659 def _html_search_meta(self, name, html, display_name=None, fatal=False, **kwargs):
660 if display_name is None:
662 return self._html_search_regex(
664 (?=[^>]+(?:itemprop|name|property)=(["\']?)%s\1)
665 [^>]+?content=(["\'])(?P<content>.*?)\2''' % re.escape(name),
666 html, display_name, fatal=fatal, group='content', **kwargs)
668 def _dc_search_uploader(self, html):
669 return self._html_search_meta('dc.creator', html, 'uploader')
671 def _rta_search(self, html):
672 # See http://www.rtalabel.org/index.php?content=howtofaq#single
673 if re.search(r'(?ix)<meta\s+name="rating"\s+'
674 r' content="RTA-5042-1996-1400-1577-RTA"',
679 def _media_rating_search(self, html):
680 # See http://www.tjg-designs.com/WP/metadata-code-examples-adding-metadata-to-your-web-pages/
681 rating = self._html_search_meta('rating', html)
693 return RATING_TABLE.get(rating.lower(), None)
695 def _family_friendly_search(self, html):
696 # See http://schema.org/VideoObject
697 family_friendly = self._html_search_meta('isFamilyFriendly', html)
699 if not family_friendly:
708 return RATING_TABLE.get(family_friendly.lower(), None)
710 def _twitter_search_player(self, html):
711 return self._html_search_meta('twitter:player', html,
712 'twitter card player')
715 def _hidden_inputs(html):
717 (input.group('name'), input.group('value')) for input in re.finditer(
720 type=(?P<q_hidden>["\'])hidden(?P=q_hidden)\s+
721 name=(?P<q_name>["\'])(?P<name>.+?)(?P=q_name)\s+
722 (?:id=(?P<q_id>["\']).+?(?P=q_id)\s+)?
723 value=(?P<q_value>["\'])(?P<value>.*?)(?P=q_value)
727 def _form_hidden_inputs(self, form_id, html):
728 form = self._search_regex(
729 r'(?s)<form[^>]+?id=(["\'])%s\1[^>]*>(?P<form>.+?)</form>' % form_id,
730 html, '%s form' % form_id, group='form')
731 return self._hidden_inputs(form)
733 def _sort_formats(self, formats, field_preference=None):
735 raise ExtractorError('No video formats found')
738 # TODO remove the following workaround
739 from ..utils import determine_ext
740 if not f.get('ext') and 'url' in f:
741 f['ext'] = determine_ext(f['url'])
743 if isinstance(field_preference, (list, tuple)):
744 return tuple(f.get(field) if f.get(field) is not None else -1 for field in field_preference)
746 preference = f.get('preference')
747 if preference is None:
748 proto = f.get('protocol')
750 proto = compat_urllib_parse_urlparse(f.get('url', '')).scheme
752 preference = 0 if proto in ['http', 'https'] else -0.1
753 if f.get('ext') in ['f4f', 'f4m']: # Not yet supported
756 if f.get('vcodec') == 'none': # audio only
757 if self._downloader.params.get('prefer_free_formats'):
758 ORDER = ['aac', 'mp3', 'm4a', 'webm', 'ogg', 'opus']
760 ORDER = ['webm', 'opus', 'ogg', 'mp3', 'aac', 'm4a']
763 audio_ext_preference = ORDER.index(f['ext'])
765 audio_ext_preference = -1
767 if self._downloader.params.get('prefer_free_formats'):
768 ORDER = ['flv', 'mp4', 'webm']
770 ORDER = ['webm', 'flv', 'mp4']
772 ext_preference = ORDER.index(f['ext'])
775 audio_ext_preference = 0
779 f.get('language_preference') if f.get('language_preference') is not None else -1,
780 f.get('quality') if f.get('quality') is not None else -1,
781 f.get('tbr') if f.get('tbr') is not None else -1,
782 f.get('filesize') if f.get('filesize') is not None else -1,
783 f.get('vbr') if f.get('vbr') is not None else -1,
784 f.get('height') if f.get('height') is not None else -1,
785 f.get('width') if f.get('width') is not None else -1,
787 f.get('abr') if f.get('abr') is not None else -1,
788 audio_ext_preference,
789 f.get('fps') if f.get('fps') is not None else -1,
790 f.get('filesize_approx') if f.get('filesize_approx') is not None else -1,
791 f.get('source_preference') if f.get('source_preference') is not None else -1,
792 f.get('format_id') if f.get('format_id') is not None else '',
794 formats.sort(key=_formats_key)
796 def _check_formats(self, formats, video_id):
799 lambda f: self._is_valid_url(
801 item='%s video format' % f.get('format_id') if f.get('format_id') else 'video'),
804 def _is_valid_url(self, url, video_id, item='video'):
805 url = self._proto_relative_url(url, scheme='http:')
806 # For now assume non HTTP(S) URLs always valid
807 if not (url.startswith('http://') or url.startswith('https://')):
810 self._request_webpage(url, video_id, 'Checking %s URL' % item)
812 except ExtractorError as e:
813 if isinstance(e.cause, compat_HTTPError):
815 '%s: %s URL is invalid, skipping' % (video_id, item))
819 def http_scheme(self):
820 """ Either "http:" or "https:", depending on the user's preferences """
823 if self._downloader.params.get('prefer_insecure', False)
826 def _proto_relative_url(self, url, scheme=None):
829 if url.startswith('//'):
831 scheme = self.http_scheme()
836 def _sleep(self, timeout, video_id, msg_template=None):
837 if msg_template is None:
838 msg_template = '%(video_id)s: Waiting for %(timeout)s seconds'
839 msg = msg_template % {'video_id': video_id, 'timeout': timeout}
843 def _extract_f4m_formats(self, manifest_url, video_id, preference=None, f4m_id=None,
844 transform_source=lambda s: fix_xml_ampersands(s).strip()):
845 manifest = self._download_xml(
846 manifest_url, video_id, 'Downloading f4m manifest',
847 'Unable to download f4m manifest',
848 # Some manifests may be malformed, e.g. prosiebensat1 generated manifests
849 # (see https://github.com/rg3/youtube-dl/issues/6215#issuecomment-121704244)
850 transform_source=transform_source)
853 manifest_version = '1.0'
854 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/1.0}media')
856 manifest_version = '2.0'
857 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/2.0}media')
858 for i, media_el in enumerate(media_nodes):
859 if manifest_version == '2.0':
860 media_url = media_el.attrib.get('href') or media_el.attrib.get('url')
864 media_url if media_url.startswith('http://') or media_url.startswith('https://')
865 else ('/'.join(manifest_url.split('/')[:-1]) + '/' + media_url))
866 # If media_url is itself a f4m manifest do the recursive extraction
867 # since bitrates in parent manifest (this one) and media_url manifest
868 # may differ leading to inability to resolve the format by requested
869 # bitrate in f4m downloader
870 if determine_ext(manifest_url) == 'f4m':
871 formats.extend(self._extract_f4m_formats(manifest_url, video_id, preference, f4m_id))
873 tbr = int_or_none(media_el.attrib.get('bitrate'))
875 'format_id': '-'.join(filter(None, [f4m_id, compat_str(i if tbr is None else tbr)])),
879 'width': int_or_none(media_el.attrib.get('width')),
880 'height': int_or_none(media_el.attrib.get('height')),
881 'preference': preference,
883 self._sort_formats(formats)
887 def _extract_m3u8_formats(self, m3u8_url, video_id, ext=None,
888 entry_protocol='m3u8', preference=None,
889 m3u8_id=None, note=None, errnote=None,
893 'format_id': '-'.join(filter(None, [m3u8_id, 'meta'])),
897 'preference': preference - 1 if preference else -1,
898 'resolution': 'multiple',
899 'format_note': 'Quality selection URL',
902 format_url = lambda u: (
904 if re.match(r'^https?://', u)
905 else compat_urlparse.urljoin(m3u8_url, u))
907 m3u8_doc = self._download_webpage(
909 note=note or 'Downloading m3u8 information',
910 errnote=errnote or 'Failed to download m3u8 information',
912 if m3u8_doc is False:
917 r'(?P<key>[a-zA-Z_-]+)=(?P<val>"[^"]+"|[^",]+)(?:,|$)')
918 for line in m3u8_doc.splitlines():
919 if line.startswith('#EXT-X-STREAM-INF:'):
921 for m in kv_rex.finditer(line):
923 if v.startswith('"'):
925 last_info[m.group('key')] = v
926 elif line.startswith('#EXT-X-MEDIA:'):
928 for m in kv_rex.finditer(line):
930 if v.startswith('"'):
932 last_media[m.group('key')] = v
933 elif line.startswith('#') or not line.strip():
936 if last_info is None:
937 formats.append({'url': format_url(line)})
939 tbr = int_or_none(last_info.get('BANDWIDTH'), scale=1000)
942 format_id.append(m3u8_id)
943 last_media_name = last_media.get('NAME') if last_media and last_media.get('TYPE') != 'SUBTITLES' else None
944 format_id.append(last_media_name if last_media_name else '%d' % (tbr if tbr else len(formats)))
946 'format_id': '-'.join(format_id),
947 'url': format_url(line.strip()),
950 'protocol': entry_protocol,
951 'preference': preference,
953 codecs = last_info.get('CODECS')
955 # TODO: looks like video codec is not always necessarily goes first
956 va_codecs = codecs.split(',')
958 f['vcodec'] = va_codecs[0].partition('.')[0]
959 if len(va_codecs) > 1 and va_codecs[1]:
960 f['acodec'] = va_codecs[1].partition('.')[0]
961 resolution = last_info.get('RESOLUTION')
963 width_str, height_str = resolution.split('x')
964 f['width'] = int(width_str)
965 f['height'] = int(height_str)
966 if last_media is not None:
967 f['m3u8_media'] = last_media
971 self._sort_formats(formats)
974 # TODO: improve extraction
975 def _extract_smil_formats(self, smil_url, video_id, fatal=True):
976 smil = self._download_xml(
977 smil_url, video_id, 'Downloading SMIL file',
978 'Unable to download SMIL file', fatal=fatal)
983 base = smil.find('./head/meta').get('base')
987 if smil.findall('./body/seq/video'):
988 video = smil.findall('./body/seq/video')[0]
989 fmts, rtmp_count = self._parse_smil_video(video, video_id, base, rtmp_count)
992 for video in smil.findall('./body/switch/video'):
993 fmts, rtmp_count = self._parse_smil_video(video, video_id, base, rtmp_count)
996 self._sort_formats(formats)
1000 def _parse_smil_video(self, video, video_id, base, rtmp_count):
1001 src = video.get('src')
1003 return [], rtmp_count
1004 bitrate = int_or_none(video.get('system-bitrate') or video.get('systemBitrate'), 1000)
1005 width = int_or_none(video.get('width'))
1006 height = int_or_none(video.get('height'))
1007 proto = video.get('proto')
1010 if base.startswith('rtmp'):
1012 elif base.startswith('http'):
1014 ext = video.get('ext')
1016 return self._extract_m3u8_formats(src, video_id, ext), rtmp_count
1017 elif proto == 'rtmp':
1019 streamer = video.get('streamer') or base
1024 'format_id': 'rtmp-%d' % (rtmp_count if bitrate is None else bitrate),
1029 elif proto.startswith('http'):
1032 'ext': ext or 'flv',
1038 def _live_title(self, name):
1039 """ Generate the title for a live video """
1040 now = datetime.datetime.now()
1041 now_str = now.strftime("%Y-%m-%d %H:%M")
1042 return name + ' ' + now_str
1044 def _int(self, v, name, fatal=False, **kwargs):
1045 res = int_or_none(v, **kwargs)
1046 if 'get_attr' in kwargs:
1047 print(getattr(v, kwargs['get_attr']))
1049 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
1051 raise ExtractorError(msg)
1053 self._downloader.report_warning(msg)
1056 def _float(self, v, name, fatal=False, **kwargs):
1057 res = float_or_none(v, **kwargs)
1059 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
1061 raise ExtractorError(msg)
1063 self._downloader.report_warning(msg)
1066 def _set_cookie(self, domain, name, value, expire_time=None):
1067 cookie = compat_cookiejar.Cookie(
1068 0, name, value, None, None, domain, None,
1069 None, '/', True, False, expire_time, '', None, None, None)
1070 self._downloader.cookiejar.set_cookie(cookie)
1072 def get_testcases(self, include_onlymatching=False):
1073 t = getattr(self, '_TEST', None)
1075 assert not hasattr(self, '_TESTS'), \
1076 '%s has _TEST and _TESTS' % type(self).__name__
1079 tests = getattr(self, '_TESTS', [])
1081 if not include_onlymatching and t.get('only_matching', False):
1083 t['name'] = type(self).__name__[:-len('IE')]
1086 def is_suitable(self, age_limit):
1087 """ Test whether the extractor is generally suitable for the given
1088 age limit (i.e. pornographic sites are not, all others usually are) """
1090 any_restricted = False
1091 for tc in self.get_testcases(include_onlymatching=False):
1092 if 'playlist' in tc:
1093 tc = tc['playlist'][0]
1094 is_restricted = age_restricted(
1095 tc.get('info_dict', {}).get('age_limit'), age_limit)
1096 if not is_restricted:
1098 any_restricted = any_restricted or is_restricted
1099 return not any_restricted
1101 def extract_subtitles(self, *args, **kwargs):
1102 if (self._downloader.params.get('writesubtitles', False) or
1103 self._downloader.params.get('listsubtitles')):
1104 return self._get_subtitles(*args, **kwargs)
1107 def _get_subtitles(self, *args, **kwargs):
1108 raise NotImplementedError("This method must be implemented by subclasses")
1110 def extract_automatic_captions(self, *args, **kwargs):
1111 if (self._downloader.params.get('writeautomaticsub', False) or
1112 self._downloader.params.get('listsubtitles')):
1113 return self._get_automatic_captions(*args, **kwargs)
1116 def _get_automatic_captions(self, *args, **kwargs):
1117 raise NotImplementedError("This method must be implemented by subclasses")
1120 class SearchInfoExtractor(InfoExtractor):
1122 Base class for paged search queries extractors.
1123 They accept URLs in the format _SEARCH_KEY(|all|[0-9]):{query}
1124 Instances should define _SEARCH_KEY and _MAX_RESULTS.
1128 def _make_valid_url(cls):
1129 return r'%s(?P<prefix>|[1-9][0-9]*|all):(?P<query>[\s\S]+)' % cls._SEARCH_KEY
1132 def suitable(cls, url):
1133 return re.match(cls._make_valid_url(), url) is not None
1135 def _real_extract(self, query):
1136 mobj = re.match(self._make_valid_url(), query)
1138 raise ExtractorError('Invalid search query "%s"' % query)
1140 prefix = mobj.group('prefix')
1141 query = mobj.group('query')
1143 return self._get_n_results(query, 1)
1144 elif prefix == 'all':
1145 return self._get_n_results(query, self._MAX_RESULTS)
1149 raise ExtractorError('invalid download number %s for query "%s"' % (n, query))
1150 elif n > self._MAX_RESULTS:
1151 self._downloader.report_warning('%s returns max %i results (you requested %i)' % (self._SEARCH_KEY, self._MAX_RESULTS, n))
1152 n = self._MAX_RESULTS
1153 return self._get_n_results(query, n)
1155 def _get_n_results(self, query, n):
1156 """Get a specified number of results for a query"""
1157 raise NotImplementedError("This method must be implemented by subclasses")
1160 def SEARCH_KEY(self):
1161 return self._SEARCH_KEY