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 average_rating: Average rating give by users, the scale used depends on the webpage
176 comment_count: Number of comments on the video
177 comments: A list of comments, each with one or more of the following
178 properties (all but one of text or html optional):
179 * "author" - human-readable name of the comment author
180 * "author_id" - user ID of the comment author
182 * "html" - Comment as HTML
183 * "text" - Plain text of the comment
184 * "timestamp" - UNIX timestamp of comment
185 * "parent" - ID of the comment this one is replying to.
186 Set to "root" to indicate that this is a
187 comment to the original video.
188 age_limit: Age restriction for the video, as an integer (years)
189 webpage_url: The URL to the video webpage, if given to youtube-dl it
190 should allow to get the same result again. (It will be set
191 by YoutubeDL if it's missing)
192 categories: A list of categories that the video falls in, for example
194 tags: A list of tags assigned to the video, e.g. ["sweden", "pop music"]
195 is_live: True, False, or None (=unknown). Whether this video is a
196 live stream that goes on instead of a fixed-length video.
197 start_time: Time in seconds where the reproduction should start, as
198 specified in the URL.
199 end_time: Time in seconds where the reproduction should end, as
200 specified in the URL.
202 Unless mentioned otherwise, the fields should be Unicode strings.
204 Unless mentioned otherwise, None is equivalent to absence of information.
207 _type "playlist" indicates multiple videos.
208 There must be a key "entries", which is a list, an iterable, or a PagedList
209 object, each element of which is a valid dictionary by this specification.
211 Additionally, playlists can have "title", "description" and "id" attributes
212 with the same semantics as videos (see above).
215 _type "multi_video" indicates that there are multiple videos that
216 form a single show, for examples multiple acts of an opera or TV episode.
217 It must have an entries key like a playlist and contain all the keys
218 required for a video at the same time.
221 _type "url" indicates that the video must be extracted from another
222 location, possibly by a different extractor. Its only required key is:
223 "url" - the next URL to extract.
224 The key "ie_key" can be set to the class name (minus the trailing "IE",
225 e.g. "Youtube") if the extractor class is known in advance.
226 Additionally, the dictionary may have any properties of the resolved entity
227 known in advance, for example "title" if the title of the referred video is
231 _type "url_transparent" entities have the same specification as "url", but
232 indicate that the given additional information is more precise than the one
233 associated with the resolved URL.
234 This is useful when a site employs a video service that hosts the video and
235 its technical metadata, but that video service does not embed a useful
236 title, description etc.
239 Subclasses of this one should re-define the _real_initialize() and
240 _real_extract() methods and define a _VALID_URL regexp.
241 Probably, they should also be added to the list of extractors.
243 Finally, the _WORKING attribute should be set to False for broken IEs
244 in order to warn the users and skip the tests.
251 def __init__(self, downloader=None):
252 """Constructor. Receives an optional downloader."""
254 self.set_downloader(downloader)
257 def suitable(cls, url):
258 """Receives a URL and returns True if suitable for this IE."""
260 # This does not use has/getattr intentionally - we want to know whether
261 # we have cached the regexp for *this* class, whereas getattr would also
262 # match the superclass
263 if '_VALID_URL_RE' not in cls.__dict__:
264 cls._VALID_URL_RE = re.compile(cls._VALID_URL)
265 return cls._VALID_URL_RE.match(url) is not None
268 def _match_id(cls, url):
269 if '_VALID_URL_RE' not in cls.__dict__:
270 cls._VALID_URL_RE = re.compile(cls._VALID_URL)
271 m = cls._VALID_URL_RE.match(url)
277 """Getter method for _WORKING."""
280 def initialize(self):
281 """Initializes an instance (authentication, etc)."""
283 self._real_initialize()
286 def extract(self, url):
287 """Extracts URL information and returns it in list of dicts."""
290 return self._real_extract(url)
291 except ExtractorError:
293 except compat_http_client.IncompleteRead as e:
294 raise ExtractorError('A network error has occured.', cause=e, expected=True)
295 except (KeyError, StopIteration) as e:
296 raise ExtractorError('An extractor error has occured.', cause=e)
298 def set_downloader(self, downloader):
299 """Sets the downloader for this IE."""
300 self._downloader = downloader
302 def _real_initialize(self):
303 """Real initialization process. Redefine in subclasses."""
306 def _real_extract(self, url):
307 """Real extraction process. Redefine in subclasses."""
312 """A string for getting the InfoExtractor with get_info_extractor"""
313 return cls.__name__[:-2]
317 return type(self).__name__[:-2]
319 def _request_webpage(self, url_or_request, video_id, note=None, errnote=None, fatal=True):
320 """ Returns the response handle """
322 self.report_download_webpage(video_id)
323 elif note is not False:
325 self.to_screen('%s' % (note,))
327 self.to_screen('%s: %s' % (video_id, note))
329 return self._downloader.urlopen(url_or_request)
330 except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err:
334 errnote = 'Unable to download webpage'
335 errmsg = '%s: %s' % (errnote, compat_str(err))
337 raise ExtractorError(errmsg, sys.exc_info()[2], cause=err)
339 self._downloader.report_warning(errmsg)
342 def _download_webpage_handle(self, url_or_request, video_id, note=None, errnote=None, fatal=True, encoding=None):
343 """ Returns a tuple (page content as string, URL handle) """
344 # Strip hashes from the URL (#1038)
345 if isinstance(url_or_request, (compat_str, str)):
346 url_or_request = url_or_request.partition('#')[0]
348 urlh = self._request_webpage(url_or_request, video_id, note, errnote, fatal)
352 content = self._webpage_read_content(urlh, url_or_request, video_id, note, errnote, fatal, encoding=encoding)
353 return (content, urlh)
356 def _guess_encoding_from_content(content_type, webpage_bytes):
357 m = re.match(r'[a-zA-Z0-9_.-]+/[a-zA-Z0-9_.-]+\s*;\s*charset=(.+)', content_type)
359 encoding = m.group(1)
361 m = re.search(br'<meta[^>]+charset=[\'"]?([^\'")]+)[ /\'">]',
362 webpage_bytes[:1024])
364 encoding = m.group(1).decode('ascii')
365 elif webpage_bytes.startswith(b'\xff\xfe'):
372 def _webpage_read_content(self, urlh, url_or_request, video_id, note=None, errnote=None, fatal=True, prefix=None, encoding=None):
373 content_type = urlh.headers.get('Content-Type', '')
374 webpage_bytes = urlh.read()
375 if prefix is not None:
376 webpage_bytes = prefix + webpage_bytes
378 encoding = self._guess_encoding_from_content(content_type, webpage_bytes)
379 if self._downloader.params.get('dump_intermediate_pages', False):
381 url = url_or_request.get_full_url()
382 except AttributeError:
384 self.to_screen('Dumping request to ' + url)
385 dump = base64.b64encode(webpage_bytes).decode('ascii')
386 self._downloader.to_screen(dump)
387 if self._downloader.params.get('write_pages', False):
389 url = url_or_request.get_full_url()
390 except AttributeError:
392 basen = '%s_%s' % (video_id, url)
394 h = '___' + hashlib.md5(basen.encode('utf-8')).hexdigest()
395 basen = basen[:240 - len(h)] + h
396 raw_filename = basen + '.dump'
397 filename = sanitize_filename(raw_filename, restricted=True)
398 self.to_screen('Saving request to ' + filename)
399 # Working around MAX_PATH limitation on Windows (see
400 # http://msdn.microsoft.com/en-us/library/windows/desktop/aa365247(v=vs.85).aspx)
402 absfilepath = os.path.abspath(filename)
403 if len(absfilepath) > 259:
404 filename = '\\\\?\\' + absfilepath
405 with open(filename, 'wb') as outf:
406 outf.write(webpage_bytes)
409 content = webpage_bytes.decode(encoding, 'replace')
411 content = webpage_bytes.decode('utf-8', 'replace')
413 if ('<title>Access to this site is blocked</title>' in content and
414 'Websense' in content[:512]):
415 msg = 'Access to this webpage has been blocked by Websense filtering software in your network.'
416 blocked_iframe = self._html_search_regex(
417 r'<iframe src="([^"]+)"', content,
418 'Websense information URL', default=None)
420 msg += ' Visit %s for more details' % blocked_iframe
421 raise ExtractorError(msg, expected=True)
422 if '<title>The URL you requested has been blocked</title>' in content[:512]:
424 'Access to this webpage has been blocked by Indian censorship. '
425 'Use a VPN or proxy server (with --proxy) to route around it.')
426 block_msg = self._html_search_regex(
427 r'</h1><p>(.*?)</p>',
428 content, 'block message', default=None)
430 msg += ' (Message: "%s")' % block_msg.replace('\n', ' ')
431 raise ExtractorError(msg, expected=True)
435 def _download_webpage(self, url_or_request, video_id, note=None, errnote=None, fatal=True, tries=1, timeout=5, encoding=None):
436 """ Returns the data of the page as a string """
439 while success is False:
441 res = self._download_webpage_handle(url_or_request, video_id, note, errnote, fatal, encoding=encoding)
443 except compat_http_client.IncompleteRead as e:
445 if try_count >= tries:
447 self._sleep(timeout, video_id)
454 def _download_xml(self, url_or_request, video_id,
455 note='Downloading XML', errnote='Unable to download XML',
456 transform_source=None, fatal=True, encoding=None):
457 """Return the xml as an xml.etree.ElementTree.Element"""
458 xml_string = self._download_webpage(
459 url_or_request, video_id, note, errnote, fatal=fatal, encoding=encoding)
460 if xml_string is False:
463 xml_string = transform_source(xml_string)
464 return xml.etree.ElementTree.fromstring(xml_string.encode('utf-8'))
466 def _download_json(self, url_or_request, video_id,
467 note='Downloading JSON metadata',
468 errnote='Unable to download JSON metadata',
469 transform_source=None,
470 fatal=True, encoding=None):
471 json_string = self._download_webpage(
472 url_or_request, video_id, note, errnote, fatal=fatal,
474 if (not fatal) and json_string is False:
476 return self._parse_json(
477 json_string, video_id, transform_source=transform_source, fatal=fatal)
479 def _parse_json(self, json_string, video_id, transform_source=None, fatal=True):
481 json_string = transform_source(json_string)
483 return json.loads(json_string)
484 except ValueError as ve:
485 errmsg = '%s: Failed to parse JSON ' % video_id
487 raise ExtractorError(errmsg, cause=ve)
489 self.report_warning(errmsg + str(ve))
491 def report_warning(self, msg, video_id=None):
492 idstr = '' if video_id is None else '%s: ' % video_id
493 self._downloader.report_warning(
494 '[%s] %s%s' % (self.IE_NAME, idstr, msg))
496 def to_screen(self, msg):
497 """Print msg to screen, prefixing it with '[ie_name]'"""
498 self._downloader.to_screen('[%s] %s' % (self.IE_NAME, msg))
500 def report_extraction(self, id_or_name):
501 """Report information extraction."""
502 self.to_screen('%s: Extracting information' % id_or_name)
504 def report_download_webpage(self, video_id):
505 """Report webpage download."""
506 self.to_screen('%s: Downloading webpage' % video_id)
508 def report_age_confirmation(self):
509 """Report attempt to confirm age."""
510 self.to_screen('Confirming age')
512 def report_login(self):
513 """Report attempt to log in."""
514 self.to_screen('Logging in')
517 def raise_login_required(msg='This video is only available for registered users'):
518 raise ExtractorError(
519 '%s. Use --username and --password or --netrc to provide account credentials.' % msg,
523 def raise_geo_restricted(msg='This video is not available from your location due to geo restriction'):
524 raise ExtractorError(
525 '%s. You might want to use --proxy to workaround.' % msg,
528 # Methods for following #608
530 def url_result(url, ie=None, video_id=None, video_title=None):
531 """Returns a URL that points to a page that should be processed"""
532 # TODO: ie should be the class used for getting the info
533 video_info = {'_type': 'url',
536 if video_id is not None:
537 video_info['id'] = video_id
538 if video_title is not None:
539 video_info['title'] = video_title
543 def playlist_result(entries, playlist_id=None, playlist_title=None, playlist_description=None):
544 """Returns a playlist"""
545 video_info = {'_type': 'playlist',
548 video_info['id'] = playlist_id
550 video_info['title'] = playlist_title
551 if playlist_description:
552 video_info['description'] = playlist_description
555 def _search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
557 Perform a regex search on the given string, using a single or a list of
558 patterns returning the first matching group.
559 In case of failure return a default value or raise a WARNING or a
560 RegexNotFoundError, depending on fatal, specifying the field name.
562 if isinstance(pattern, (str, compat_str, compiled_regex_type)):
563 mobj = re.search(pattern, string, flags)
566 mobj = re.search(p, string, flags)
570 if not self._downloader.params.get('no_color') and os.name != 'nt' and sys.stderr.isatty():
571 _name = '\033[0;34m%s\033[0m' % name
577 # return the first matching group
578 return next(g for g in mobj.groups() if g is not None)
580 return mobj.group(group)
581 elif default is not NO_DEFAULT:
584 raise RegexNotFoundError('Unable to extract %s' % _name)
586 self._downloader.report_warning('unable to extract %s' % _name + bug_reports_message())
589 def _html_search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
591 Like _search_regex, but strips HTML tags and unescapes entities.
593 res = self._search_regex(pattern, string, name, default, fatal, flags, group)
595 return clean_html(res).strip()
599 def _get_login_info(self):
601 Get the login info as (username, password)
602 It will look in the netrc file using the _NETRC_MACHINE value
603 If there's no info available, return (None, None)
605 if self._downloader is None:
610 downloader_params = self._downloader.params
612 # Attempt to use provided username and password or .netrc data
613 if downloader_params.get('username', None) is not None:
614 username = downloader_params['username']
615 password = downloader_params['password']
616 elif downloader_params.get('usenetrc', False):
618 info = netrc.netrc().authenticators(self._NETRC_MACHINE)
623 raise netrc.NetrcParseError('No authenticators for %s' % self._NETRC_MACHINE)
624 except (IOError, netrc.NetrcParseError) as err:
625 self._downloader.report_warning('parsing .netrc: %s' % compat_str(err))
627 return (username, password)
629 def _get_tfa_info(self, note='two-factor verification code'):
631 Get the two-factor authentication info
632 TODO - asking the user will be required for sms/phone verify
633 currently just uses the command line option
634 If there's no info available, return None
636 if self._downloader is None:
638 downloader_params = self._downloader.params
640 if downloader_params.get('twofactor', None) is not None:
641 return downloader_params['twofactor']
643 return compat_getpass('Type %s and press [Return]: ' % note)
645 # Helper functions for extracting OpenGraph info
647 def _og_regexes(prop):
648 content_re = r'content=(?:"([^>]+?)"|\'([^>]+?)\')'
649 property_re = r'(?:name|property)=[\'"]og:%s[\'"]' % re.escape(prop)
650 template = r'<meta[^>]+?%s[^>]+?%s'
652 template % (property_re, content_re),
653 template % (content_re, property_re),
657 def _meta_regex(prop):
658 return r'''(?isx)<meta
659 (?=[^>]+(?:itemprop|name|property|id|http-equiv)=(["\']?)%s\1)
660 [^>]+?content=(["\'])(?P<content>.*?)\2''' % re.escape(prop)
662 def _og_search_property(self, prop, html, name=None, **kargs):
664 name = 'OpenGraph %s' % prop
665 escaped = self._search_regex(self._og_regexes(prop), html, name, flags=re.DOTALL, **kargs)
668 return unescapeHTML(escaped)
670 def _og_search_thumbnail(self, html, **kargs):
671 return self._og_search_property('image', html, 'thumbnail URL', fatal=False, **kargs)
673 def _og_search_description(self, html, **kargs):
674 return self._og_search_property('description', html, fatal=False, **kargs)
676 def _og_search_title(self, html, **kargs):
677 return self._og_search_property('title', html, **kargs)
679 def _og_search_video_url(self, html, name='video url', secure=True, **kargs):
680 regexes = self._og_regexes('video') + self._og_regexes('video:url')
682 regexes = self._og_regexes('video:secure_url') + regexes
683 return self._html_search_regex(regexes, html, name, **kargs)
685 def _og_search_url(self, html, **kargs):
686 return self._og_search_property('url', html, **kargs)
688 def _html_search_meta(self, name, html, display_name=None, fatal=False, **kwargs):
689 if display_name is None:
691 return self._html_search_regex(
692 self._meta_regex(name),
693 html, display_name, fatal=fatal, group='content', **kwargs)
695 def _dc_search_uploader(self, html):
696 return self._html_search_meta('dc.creator', html, 'uploader')
698 def _rta_search(self, html):
699 # See http://www.rtalabel.org/index.php?content=howtofaq#single
700 if re.search(r'(?ix)<meta\s+name="rating"\s+'
701 r' content="RTA-5042-1996-1400-1577-RTA"',
706 def _media_rating_search(self, html):
707 # See http://www.tjg-designs.com/WP/metadata-code-examples-adding-metadata-to-your-web-pages/
708 rating = self._html_search_meta('rating', html)
720 return RATING_TABLE.get(rating.lower(), None)
722 def _family_friendly_search(self, html):
723 # See http://schema.org/VideoObject
724 family_friendly = self._html_search_meta('isFamilyFriendly', html)
726 if not family_friendly:
735 return RATING_TABLE.get(family_friendly.lower(), None)
737 def _twitter_search_player(self, html):
738 return self._html_search_meta('twitter:player', html,
739 'twitter card player')
742 def _hidden_inputs(html):
743 html = re.sub(r'<!--(?:(?!<!--).)*-->', '', html)
745 for input in re.findall(r'(?i)<input([^>]+)>', html):
746 if not re.search(r'type=(["\'])(?:hidden|submit)\1', input):
748 name = re.search(r'name=(["\'])(?P<value>.+?)\1', input)
751 value = re.search(r'value=(["\'])(?P<value>.*?)\1', input)
754 hidden_inputs[name.group('value')] = value.group('value')
757 def _form_hidden_inputs(self, form_id, html):
758 form = self._search_regex(
759 r'(?is)<form[^>]+?id=(["\'])%s\1[^>]*>(?P<form>.+?)</form>' % form_id,
760 html, '%s form' % form_id, group='form')
761 return self._hidden_inputs(form)
763 def _sort_formats(self, formats, field_preference=None):
765 raise ExtractorError('No video formats found')
768 # TODO remove the following workaround
769 from ..utils import determine_ext
770 if not f.get('ext') and 'url' in f:
771 f['ext'] = determine_ext(f['url'])
773 if isinstance(field_preference, (list, tuple)):
774 return tuple(f.get(field) if f.get(field) is not None else -1 for field in field_preference)
776 preference = f.get('preference')
777 if preference is None:
778 proto = f.get('protocol')
780 proto = compat_urllib_parse_urlparse(f.get('url', '')).scheme
782 preference = 0 if proto in ['http', 'https'] else -0.1
783 if f.get('ext') in ['f4f', 'f4m']: # Not yet supported
786 if f.get('vcodec') == 'none': # audio only
787 if self._downloader.params.get('prefer_free_formats'):
788 ORDER = ['aac', 'mp3', 'm4a', 'webm', 'ogg', 'opus']
790 ORDER = ['webm', 'opus', 'ogg', 'mp3', 'aac', 'm4a']
793 audio_ext_preference = ORDER.index(f['ext'])
795 audio_ext_preference = -1
797 if self._downloader.params.get('prefer_free_formats'):
798 ORDER = ['flv', 'mp4', 'webm']
800 ORDER = ['webm', 'flv', 'mp4']
802 ext_preference = ORDER.index(f['ext'])
805 audio_ext_preference = 0
809 f.get('language_preference') if f.get('language_preference') is not None else -1,
810 f.get('quality') if f.get('quality') is not None else -1,
811 f.get('tbr') if f.get('tbr') is not None else -1,
812 f.get('filesize') if f.get('filesize') is not None else -1,
813 f.get('vbr') if f.get('vbr') is not None else -1,
814 f.get('height') if f.get('height') is not None else -1,
815 f.get('width') if f.get('width') is not None else -1,
817 f.get('abr') if f.get('abr') is not None else -1,
818 audio_ext_preference,
819 f.get('fps') if f.get('fps') is not None else -1,
820 f.get('filesize_approx') if f.get('filesize_approx') is not None else -1,
821 f.get('source_preference') if f.get('source_preference') is not None else -1,
822 f.get('format_id') if f.get('format_id') is not None else '',
824 formats.sort(key=_formats_key)
826 def _check_formats(self, formats, video_id):
829 lambda f: self._is_valid_url(
831 item='%s video format' % f.get('format_id') if f.get('format_id') else 'video'),
834 def _is_valid_url(self, url, video_id, item='video'):
835 url = self._proto_relative_url(url, scheme='http:')
836 # For now assume non HTTP(S) URLs always valid
837 if not (url.startswith('http://') or url.startswith('https://')):
840 self._request_webpage(url, video_id, 'Checking %s URL' % item)
842 except ExtractorError as e:
843 if isinstance(e.cause, compat_HTTPError):
845 '%s: %s URL is invalid, skipping' % (video_id, item))
849 def http_scheme(self):
850 """ Either "http:" or "https:", depending on the user's preferences """
853 if self._downloader.params.get('prefer_insecure', False)
856 def _proto_relative_url(self, url, scheme=None):
859 if url.startswith('//'):
861 scheme = self.http_scheme()
866 def _sleep(self, timeout, video_id, msg_template=None):
867 if msg_template is None:
868 msg_template = '%(video_id)s: Waiting for %(timeout)s seconds'
869 msg = msg_template % {'video_id': video_id, 'timeout': timeout}
873 def _extract_f4m_formats(self, manifest_url, video_id, preference=None, f4m_id=None,
874 transform_source=lambda s: fix_xml_ampersands(s).strip(),
876 manifest = self._download_xml(
877 manifest_url, video_id, 'Downloading f4m manifest',
878 'Unable to download f4m manifest',
879 # Some manifests may be malformed, e.g. prosiebensat1 generated manifests
880 # (see https://github.com/rg3/youtube-dl/issues/6215#issuecomment-121704244)
881 transform_source=transform_source,
884 if manifest is False:
888 manifest_version = '1.0'
889 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/1.0}media')
891 manifest_version = '2.0'
892 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/2.0}media')
893 for i, media_el in enumerate(media_nodes):
894 if manifest_version == '2.0':
895 media_url = media_el.attrib.get('href') or media_el.attrib.get('url')
899 media_url if media_url.startswith('http://') or media_url.startswith('https://')
900 else ('/'.join(manifest_url.split('/')[:-1]) + '/' + media_url))
901 # If media_url is itself a f4m manifest do the recursive extraction
902 # since bitrates in parent manifest (this one) and media_url manifest
903 # may differ leading to inability to resolve the format by requested
904 # bitrate in f4m downloader
905 if determine_ext(manifest_url) == 'f4m':
906 f4m_formats = self._extract_f4m_formats(
907 manifest_url, video_id, preference, f4m_id, fatal=fatal)
909 formats.extend(f4m_formats)
911 tbr = int_or_none(media_el.attrib.get('bitrate'))
913 'format_id': '-'.join(filter(None, [f4m_id, compat_str(i if tbr is None else tbr)])),
917 'width': int_or_none(media_el.attrib.get('width')),
918 'height': int_or_none(media_el.attrib.get('height')),
919 'preference': preference,
921 self._sort_formats(formats)
925 def _extract_m3u8_formats(self, m3u8_url, video_id, ext=None,
926 entry_protocol='m3u8', preference=None,
927 m3u8_id=None, note=None, errnote=None,
931 'format_id': '-'.join(filter(None, [m3u8_id, 'meta'])),
935 'preference': preference - 1 if preference else -1,
936 'resolution': 'multiple',
937 'format_note': 'Quality selection URL',
940 format_url = lambda u: (
942 if re.match(r'^https?://', u)
943 else compat_urlparse.urljoin(m3u8_url, u))
945 m3u8_doc = self._download_webpage(
947 note=note or 'Downloading m3u8 information',
948 errnote=errnote or 'Failed to download m3u8 information',
950 if m3u8_doc is False:
955 r'(?P<key>[a-zA-Z_-]+)=(?P<val>"[^"]+"|[^",]+)(?:,|$)')
956 for line in m3u8_doc.splitlines():
957 if line.startswith('#EXT-X-STREAM-INF:'):
959 for m in kv_rex.finditer(line):
961 if v.startswith('"'):
963 last_info[m.group('key')] = v
964 elif line.startswith('#EXT-X-MEDIA:'):
966 for m in kv_rex.finditer(line):
968 if v.startswith('"'):
970 last_media[m.group('key')] = v
971 elif line.startswith('#') or not line.strip():
974 if last_info is None:
975 formats.append({'url': format_url(line)})
977 tbr = int_or_none(last_info.get('BANDWIDTH'), scale=1000)
980 format_id.append(m3u8_id)
981 last_media_name = last_media.get('NAME') if last_media and last_media.get('TYPE') != 'SUBTITLES' else None
982 format_id.append(last_media_name if last_media_name else '%d' % (tbr if tbr else len(formats)))
984 'format_id': '-'.join(format_id),
985 'url': format_url(line.strip()),
988 'protocol': entry_protocol,
989 'preference': preference,
991 codecs = last_info.get('CODECS')
993 # TODO: looks like video codec is not always necessarily goes first
994 va_codecs = codecs.split(',')
996 f['vcodec'] = va_codecs[0].partition('.')[0]
997 if len(va_codecs) > 1 and va_codecs[1]:
998 f['acodec'] = va_codecs[1].partition('.')[0]
999 resolution = last_info.get('RESOLUTION')
1001 width_str, height_str = resolution.split('x')
1002 f['width'] = int(width_str)
1003 f['height'] = int(height_str)
1004 if last_media is not None:
1005 f['m3u8_media'] = last_media
1009 self._sort_formats(formats)
1013 def _xpath_ns(path, namespace=None):
1017 for c in path.split('/'):
1018 if not c or c == '.':
1021 out.append('{%s}%s' % (namespace, c))
1022 return '/'.join(out)
1024 def _extract_smil_formats(self, smil_url, video_id, fatal=True, f4m_params=None):
1025 smil = self._download_smil(smil_url, video_id, fatal=fatal)
1031 namespace = self._parse_smil_namespace(smil)
1033 return self._parse_smil_formats(
1034 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
1036 def _extract_smil_info(self, smil_url, video_id, fatal=True, f4m_params=None):
1037 smil = self._download_smil(smil_url, video_id, fatal=fatal)
1040 return self._parse_smil(smil, smil_url, video_id, f4m_params=f4m_params)
1042 def _download_smil(self, smil_url, video_id, fatal=True):
1043 return self._download_xml(
1044 smil_url, video_id, 'Downloading SMIL file',
1045 'Unable to download SMIL file', fatal=fatal)
1047 def _parse_smil(self, smil, smil_url, video_id, f4m_params=None):
1048 namespace = self._parse_smil_namespace(smil)
1050 formats = self._parse_smil_formats(
1051 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
1052 subtitles = self._parse_smil_subtitles(smil, namespace=namespace)
1054 video_id = os.path.splitext(url_basename(smil_url))[0]
1058 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
1059 name = meta.attrib.get('name')
1060 content = meta.attrib.get('content')
1061 if not name or not content:
1063 if not title and name == 'title':
1065 elif not description and name in ('description', 'abstract'):
1066 description = content
1067 elif not upload_date and name == 'date':
1068 upload_date = unified_strdate(content)
1071 'id': image.get('type'),
1072 'url': image.get('src'),
1073 'width': int_or_none(image.get('width')),
1074 'height': int_or_none(image.get('height')),
1075 } for image in smil.findall(self._xpath_ns('.//image', namespace)) if image.get('src')]
1079 'title': title or video_id,
1080 'description': description,
1081 'upload_date': upload_date,
1082 'thumbnails': thumbnails,
1084 'subtitles': subtitles,
1087 def _parse_smil_namespace(self, smil):
1088 return self._search_regex(
1089 r'(?i)^{([^}]+)?}smil$', smil.tag, 'namespace', default=None)
1091 def _parse_smil_formats(self, smil, smil_url, video_id, namespace=None, f4m_params=None, transform_rtmp_url=None):
1093 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
1094 b = meta.get('base') or meta.get('httpBase')
1103 videos = smil.findall(self._xpath_ns('.//video', namespace))
1104 for video in videos:
1105 src = video.get('src')
1109 bitrate = float_or_none(video.get('system-bitrate') or video.get('systemBitrate'), 1000)
1110 filesize = int_or_none(video.get('size') or video.get('fileSize'))
1111 width = int_or_none(video.get('width'))
1112 height = int_or_none(video.get('height'))
1113 proto = video.get('proto')
1114 ext = video.get('ext')
1115 src_ext = determine_ext(src)
1116 streamer = video.get('streamer') or base
1118 if proto == 'rtmp' or streamer.startswith('rtmp'):
1124 'format_id': 'rtmp-%d' % (rtmp_count if bitrate is None else bitrate),
1126 'filesize': filesize,
1130 if transform_rtmp_url:
1131 streamer, src = transform_rtmp_url(streamer, src)
1132 formats[-1].update({
1138 src_url = src if src.startswith('http') else compat_urlparse.urljoin(base, src)
1140 if proto == 'm3u8' or src_ext == 'm3u8':
1141 m3u8_formats = self._extract_m3u8_formats(
1142 src_url, video_id, ext or 'mp4', m3u8_id='hls', fatal=False)
1144 formats.extend(m3u8_formats)
1147 if src_ext == 'f4m':
1152 'plugin': 'flowplayer-3.2.0.1',
1154 f4m_url += '&' if '?' in f4m_url else '?'
1155 f4m_url += compat_urllib_parse.urlencode(f4m_params)
1156 f4m_formats = self._extract_f4m_formats(f4m_url, video_id, f4m_id='hds', fatal=False)
1158 formats.extend(f4m_formats)
1161 if src_url.startswith('http') and self._is_valid_url(src, video_id):
1165 'ext': ext or src_ext or 'flv',
1166 'format_id': 'http-%d' % (bitrate or http_count),
1168 'filesize': filesize,
1174 self._sort_formats(formats)
1178 def _parse_smil_subtitles(self, smil, namespace=None, subtitles_lang='en'):
1180 for num, textstream in enumerate(smil.findall(self._xpath_ns('.//textstream', namespace))):
1181 src = textstream.get('src')
1184 ext = textstream.get('ext') or determine_ext(src)
1186 type_ = textstream.get('type')
1190 'application/smptett+xml': 'tt',
1192 if type_ in SUBTITLES_TYPES:
1193 ext = SUBTITLES_TYPES[type_]
1194 lang = textstream.get('systemLanguage') or textstream.get('systemLanguageName') or textstream.get('lang') or subtitles_lang
1195 subtitles.setdefault(lang, []).append({
1201 def _extract_xspf_playlist(self, playlist_url, playlist_id, fatal=True):
1202 xspf = self._download_xml(
1203 playlist_url, playlist_id, 'Downloading xpsf playlist',
1204 'Unable to download xspf manifest', fatal=fatal)
1207 return self._parse_xspf(xspf, playlist_id)
1209 def _parse_xspf(self, playlist, playlist_id):
1211 'xspf': 'http://xspf.org/ns/0/',
1212 's1': 'http://static.streamone.nl/player/ns/0',
1216 for track in playlist.findall(xpath_with_ns('./xspf:trackList/xspf:track', NS_MAP)):
1218 track, xpath_with_ns('./xspf:title', NS_MAP), 'title', default=playlist_id)
1219 description = xpath_text(
1220 track, xpath_with_ns('./xspf:annotation', NS_MAP), 'description')
1221 thumbnail = xpath_text(
1222 track, xpath_with_ns('./xspf:image', NS_MAP), 'thumbnail')
1223 duration = float_or_none(
1224 xpath_text(track, xpath_with_ns('./xspf:duration', NS_MAP), 'duration'), 1000)
1227 'url': location.text,
1228 'format_id': location.get(xpath_with_ns('s1:label', NS_MAP)),
1229 'width': int_or_none(location.get(xpath_with_ns('s1:width', NS_MAP))),
1230 'height': int_or_none(location.get(xpath_with_ns('s1:height', NS_MAP))),
1231 } for location in track.findall(xpath_with_ns('./xspf:location', NS_MAP))]
1232 self._sort_formats(formats)
1237 'description': description,
1238 'thumbnail': thumbnail,
1239 'duration': duration,
1244 def _live_title(self, name):
1245 """ Generate the title for a live video """
1246 now = datetime.datetime.now()
1247 now_str = now.strftime("%Y-%m-%d %H:%M")
1248 return name + ' ' + now_str
1250 def _int(self, v, name, fatal=False, **kwargs):
1251 res = int_or_none(v, **kwargs)
1252 if 'get_attr' in kwargs:
1253 print(getattr(v, kwargs['get_attr']))
1255 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
1257 raise ExtractorError(msg)
1259 self._downloader.report_warning(msg)
1262 def _float(self, v, name, fatal=False, **kwargs):
1263 res = float_or_none(v, **kwargs)
1265 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
1267 raise ExtractorError(msg)
1269 self._downloader.report_warning(msg)
1272 def _set_cookie(self, domain, name, value, expire_time=None):
1273 cookie = compat_cookiejar.Cookie(
1274 0, name, value, None, None, domain, None,
1275 None, '/', True, False, expire_time, '', None, None, None)
1276 self._downloader.cookiejar.set_cookie(cookie)
1278 def _get_cookies(self, url):
1279 """ Return a compat_cookies.SimpleCookie with the cookies for the url """
1280 req = compat_urllib_request.Request(url)
1281 self._downloader.cookiejar.add_cookie_header(req)
1282 return compat_cookies.SimpleCookie(req.get_header('Cookie'))
1284 def get_testcases(self, include_onlymatching=False):
1285 t = getattr(self, '_TEST', None)
1287 assert not hasattr(self, '_TESTS'), \
1288 '%s has _TEST and _TESTS' % type(self).__name__
1291 tests = getattr(self, '_TESTS', [])
1293 if not include_onlymatching and t.get('only_matching', False):
1295 t['name'] = type(self).__name__[:-len('IE')]
1298 def is_suitable(self, age_limit):
1299 """ Test whether the extractor is generally suitable for the given
1300 age limit (i.e. pornographic sites are not, all others usually are) """
1302 any_restricted = False
1303 for tc in self.get_testcases(include_onlymatching=False):
1304 if 'playlist' in tc:
1305 tc = tc['playlist'][0]
1306 is_restricted = age_restricted(
1307 tc.get('info_dict', {}).get('age_limit'), age_limit)
1308 if not is_restricted:
1310 any_restricted = any_restricted or is_restricted
1311 return not any_restricted
1313 def extract_subtitles(self, *args, **kwargs):
1314 if (self._downloader.params.get('writesubtitles', False) or
1315 self._downloader.params.get('listsubtitles')):
1316 return self._get_subtitles(*args, **kwargs)
1319 def _get_subtitles(self, *args, **kwargs):
1320 raise NotImplementedError("This method must be implemented by subclasses")
1323 def _merge_subtitle_items(subtitle_list1, subtitle_list2):
1324 """ Merge subtitle items for one language. Items with duplicated URLs
1325 will be dropped. """
1326 list1_urls = set([item['url'] for item in subtitle_list1])
1327 ret = list(subtitle_list1)
1328 ret.extend([item for item in subtitle_list2 if item['url'] not in list1_urls])
1332 def _merge_subtitles(cls, subtitle_dict1, subtitle_dict2):
1333 """ Merge two subtitle dictionaries, language by language. """
1334 ret = dict(subtitle_dict1)
1335 for lang in subtitle_dict2:
1336 ret[lang] = cls._merge_subtitle_items(subtitle_dict1.get(lang, []), subtitle_dict2[lang])
1339 def extract_automatic_captions(self, *args, **kwargs):
1340 if (self._downloader.params.get('writeautomaticsub', False) or
1341 self._downloader.params.get('listsubtitles')):
1342 return self._get_automatic_captions(*args, **kwargs)
1345 def _get_automatic_captions(self, *args, **kwargs):
1346 raise NotImplementedError("This method must be implemented by subclasses")
1349 class SearchInfoExtractor(InfoExtractor):
1351 Base class for paged search queries extractors.
1352 They accept URLs in the format _SEARCH_KEY(|all|[0-9]):{query}
1353 Instances should define _SEARCH_KEY and _MAX_RESULTS.
1357 def _make_valid_url(cls):
1358 return r'%s(?P<prefix>|[1-9][0-9]*|all):(?P<query>[\s\S]+)' % cls._SEARCH_KEY
1361 def suitable(cls, url):
1362 return re.match(cls._make_valid_url(), url) is not None
1364 def _real_extract(self, query):
1365 mobj = re.match(self._make_valid_url(), query)
1367 raise ExtractorError('Invalid search query "%s"' % query)
1369 prefix = mobj.group('prefix')
1370 query = mobj.group('query')
1372 return self._get_n_results(query, 1)
1373 elif prefix == 'all':
1374 return self._get_n_results(query, self._MAX_RESULTS)
1378 raise ExtractorError('invalid download number %s for query "%s"' % (n, query))
1379 elif n > self._MAX_RESULTS:
1380 self._downloader.report_warning('%s returns max %i results (you requested %i)' % (self._SEARCH_KEY, self._MAX_RESULTS, n))
1381 n = self._MAX_RESULTS
1382 return self._get_n_results(query, n)
1384 def _get_n_results(self, query, n):
1385 """Get a specified number of results for a query"""
1386 raise NotImplementedError("This method must be implemented by subclasses")
1389 def SEARCH_KEY(self):
1390 return self._SEARCH_KEY