1 from __future__ import unicode_literals
13 import xml.etree.ElementTree
15 from ..compat import (
23 compat_urllib_parse_urlparse,
24 compat_urllib_request,
48 class InfoExtractor(object):
49 """Information Extractor class.
51 Information extractors are the classes that, given a URL, extract
52 information about the video (or videos) the URL refers to. This
53 information includes the real video URL, the video title, author and
54 others. The information is stored in a dictionary which is then
55 passed to the YoutubeDL. The YoutubeDL processes this
56 information possibly downloading the video to the file system, among
57 other possible outcomes.
59 The type field determines the type of the result.
60 By far the most common value (and the default if _type is missing) is
61 "video", which indicates a single video.
63 For a video, the dictionaries must include the following fields:
66 title: Video title, unescaped.
68 Additionally, it must contain either a formats entry or a url one:
70 formats: A list of dictionaries for each format available, ordered
71 from worst to best quality.
74 * url Mandatory. The URL of the video file
75 * ext Will be calculated from URL if missing
76 * format A human-readable description of the format
77 ("mp4 container with h264/opus").
78 Calculated from the format_id, width, height.
79 and format_note fields if missing.
80 * format_id A short description of the format
81 ("mp4_h264_opus" or "19").
82 Technically optional, but strongly recommended.
83 * format_note Additional info about the format
84 ("3D" or "DASH video")
85 * width Width of the video, if known
86 * height Height of the video, if known
87 * resolution Textual description of width and height
88 * tbr Average bitrate of audio and video in KBit/s
89 * abr Average audio bitrate in KBit/s
90 * acodec Name of the audio codec in use
91 * asr Audio sampling rate in Hertz
92 * vbr Average video bitrate in KBit/s
94 * vcodec Name of the video codec in use
95 * container Name of the container format
96 * filesize The number of bytes, if known in advance
97 * filesize_approx An estimate for the number of bytes
98 * player_url SWF Player URL (used for rtmpdump).
99 * protocol The protocol that will be used for the actual
100 download, lower-case.
101 "http", "https", "rtsp", "rtmp", "rtmpe",
102 "m3u8", or "m3u8_native".
103 * preference Order number of this format. If this field is
104 present and not None, the formats get sorted
105 by this field, regardless of all other values.
106 -1 for default (order by other properties),
107 -2 or smaller for less than default.
108 < -1000 to hide the format (if there is
109 another one which is strictly better)
110 * language_preference Is this in the correct requested
112 10 if it's what the URL is about,
113 -1 for default (don't know),
114 -10 otherwise, other values reserved for now.
115 * quality Order number of the video quality of this
116 format, irrespective of the file format.
117 -1 for default (order by other properties),
118 -2 or smaller for less than default.
119 * source_preference Order number for this video source
120 (quality takes higher priority)
121 -1 for default (order by other properties),
122 -2 or smaller for less than default.
123 * http_headers A dictionary of additional HTTP headers
124 to add to the request.
125 * stretched_ratio If given and not 1, indicates that the
126 video's pixels are not square.
127 width : height ratio as float.
128 * no_resume The server does not support resuming the
129 (HTTP or RTMP) download. Boolean.
131 url: Final video URL.
132 ext: Video filename extension.
133 format: The video format, defaults to ext (used for --get-format)
134 player_url: SWF Player URL (used for rtmpdump).
136 The following fields are optional:
138 alt_title: A secondary title of the video.
139 display_id An alternative identifier for the video, not necessarily
140 unique, but available before title. Typically, id is
141 something like "4234987", title "Dancing naked mole rats",
142 and display_id "dancing-naked-mole-rats"
143 thumbnails: A list of dictionaries, with the following entries:
144 * "id" (optional, string) - Thumbnail format ID
146 * "preference" (optional, int) - quality of the image
147 * "width" (optional, int)
148 * "height" (optional, int)
149 * "resolution" (optional, string "{width}x{height"},
151 thumbnail: Full URL to a video thumbnail image.
152 description: Full video description.
153 uploader: Full name of the video uploader.
154 creator: The main artist who created the video.
155 release_date: The date (YYYYMMDD) when the video was released.
156 timestamp: UNIX timestamp of the moment the video became available.
157 upload_date: Video upload date (YYYYMMDD).
158 If not explicitly set, calculated from timestamp.
159 uploader_id: Nickname or id of the video uploader.
160 location: Physical location where the video was filmed.
161 subtitles: The available subtitles as a dictionary in the format
162 {language: subformats}. "subformats" is a list sorted from
163 lower to higher preference, each element is a dictionary
164 with the "ext" entry and one of:
165 * "data": The subtitles file contents
166 * "url": A URL pointing to the subtitles file
167 automatic_captions: Like 'subtitles', used by the YoutubeIE for
168 automatically generated captions
169 duration: Length of the video in seconds, as an integer.
170 view_count: How many users have watched the video on the platform.
171 like_count: Number of positive ratings of the video
172 dislike_count: Number of negative ratings of the video
173 average_rating: Average rating give by users, the scale used depends on the webpage
174 comment_count: Number of comments on the video
175 comments: A list of comments, each with one or more of the following
176 properties (all but one of text or html optional):
177 * "author" - human-readable name of the comment author
178 * "author_id" - user ID of the comment author
180 * "html" - Comment as HTML
181 * "text" - Plain text of the comment
182 * "timestamp" - UNIX timestamp of comment
183 * "parent" - ID of the comment this one is replying to.
184 Set to "root" to indicate that this is a
185 comment to the original video.
186 age_limit: Age restriction for the video, as an integer (years)
187 webpage_url: The URL to the video webpage, if given to youtube-dl it
188 should allow to get the same result again. (It will be set
189 by YoutubeDL if it's missing)
190 categories: A list of categories that the video falls in, for example
192 tags: A list of tags assigned to the video, e.g. ["sweden", "pop music"]
193 is_live: True, False, or None (=unknown). Whether this video is a
194 live stream that goes on instead of a fixed-length video.
195 start_time: Time in seconds where the reproduction should start, as
196 specified in the URL.
197 end_time: Time in seconds where the reproduction should end, as
198 specified in the URL.
200 Unless mentioned otherwise, the fields should be Unicode strings.
202 Unless mentioned otherwise, None is equivalent to absence of information.
205 _type "playlist" indicates multiple videos.
206 There must be a key "entries", which is a list, an iterable, or a PagedList
207 object, each element of which is a valid dictionary by this specification.
209 Additionally, playlists can have "title", "description" and "id" attributes
210 with the same semantics as videos (see above).
213 _type "multi_video" indicates that there are multiple videos that
214 form a single show, for examples multiple acts of an opera or TV episode.
215 It must have an entries key like a playlist and contain all the keys
216 required for a video at the same time.
219 _type "url" indicates that the video must be extracted from another
220 location, possibly by a different extractor. Its only required key is:
221 "url" - the next URL to extract.
222 The key "ie_key" can be set to the class name (minus the trailing "IE",
223 e.g. "Youtube") if the extractor class is known in advance.
224 Additionally, the dictionary may have any properties of the resolved entity
225 known in advance, for example "title" if the title of the referred video is
229 _type "url_transparent" entities have the same specification as "url", but
230 indicate that the given additional information is more precise than the one
231 associated with the resolved URL.
232 This is useful when a site employs a video service that hosts the video and
233 its technical metadata, but that video service does not embed a useful
234 title, description etc.
237 Subclasses of this one should re-define the _real_initialize() and
238 _real_extract() methods and define a _VALID_URL regexp.
239 Probably, they should also be added to the list of extractors.
241 Finally, the _WORKING attribute should be set to False for broken IEs
242 in order to warn the users and skip the tests.
249 def __init__(self, downloader=None):
250 """Constructor. Receives an optional downloader."""
252 self.set_downloader(downloader)
255 def suitable(cls, url):
256 """Receives a URL and returns True if suitable for this IE."""
258 # This does not use has/getattr intentionally - we want to know whether
259 # we have cached the regexp for *this* class, whereas getattr would also
260 # match the superclass
261 if '_VALID_URL_RE' not in cls.__dict__:
262 cls._VALID_URL_RE = re.compile(cls._VALID_URL)
263 return cls._VALID_URL_RE.match(url) is not None
266 def _match_id(cls, url):
267 if '_VALID_URL_RE' not in cls.__dict__:
268 cls._VALID_URL_RE = re.compile(cls._VALID_URL)
269 m = cls._VALID_URL_RE.match(url)
275 """Getter method for _WORKING."""
278 def initialize(self):
279 """Initializes an instance (authentication, etc)."""
281 self._real_initialize()
284 def extract(self, url):
285 """Extracts URL information and returns it in list of dicts."""
288 return self._real_extract(url)
289 except ExtractorError:
291 except compat_http_client.IncompleteRead as e:
292 raise ExtractorError('A network error has occured.', cause=e, expected=True)
293 except (KeyError, StopIteration) as e:
294 raise ExtractorError('An extractor error has occured.', cause=e)
296 def set_downloader(self, downloader):
297 """Sets the downloader for this IE."""
298 self._downloader = downloader
300 def _real_initialize(self):
301 """Real initialization process. Redefine in subclasses."""
304 def _real_extract(self, url):
305 """Real extraction process. Redefine in subclasses."""
310 """A string for getting the InfoExtractor with get_info_extractor"""
311 return cls.__name__[:-2]
315 return type(self).__name__[:-2]
317 def _request_webpage(self, url_or_request, video_id, note=None, errnote=None, fatal=True):
318 """ Returns the response handle """
320 self.report_download_webpage(video_id)
321 elif note is not False:
323 self.to_screen('%s' % (note,))
325 self.to_screen('%s: %s' % (video_id, note))
327 return self._downloader.urlopen(url_or_request)
328 except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err:
332 errnote = 'Unable to download webpage'
333 errmsg = '%s: %s' % (errnote, compat_str(err))
335 raise ExtractorError(errmsg, sys.exc_info()[2], cause=err)
337 self._downloader.report_warning(errmsg)
340 def _download_webpage_handle(self, url_or_request, video_id, note=None, errnote=None, fatal=True, encoding=None):
341 """ Returns a tuple (page content as string, URL handle) """
342 # Strip hashes from the URL (#1038)
343 if isinstance(url_or_request, (compat_str, str)):
344 url_or_request = url_or_request.partition('#')[0]
346 urlh = self._request_webpage(url_or_request, video_id, note, errnote, fatal)
350 content = self._webpage_read_content(urlh, url_or_request, video_id, note, errnote, fatal, encoding=encoding)
351 return (content, urlh)
354 def _guess_encoding_from_content(content_type, webpage_bytes):
355 m = re.match(r'[a-zA-Z0-9_.-]+/[a-zA-Z0-9_.-]+\s*;\s*charset=(.+)', content_type)
357 encoding = m.group(1)
359 m = re.search(br'<meta[^>]+charset=[\'"]?([^\'")]+)[ /\'">]',
360 webpage_bytes[:1024])
362 encoding = m.group(1).decode('ascii')
363 elif webpage_bytes.startswith(b'\xff\xfe'):
370 def _webpage_read_content(self, urlh, url_or_request, video_id, note=None, errnote=None, fatal=True, prefix=None, encoding=None):
371 content_type = urlh.headers.get('Content-Type', '')
372 webpage_bytes = urlh.read()
373 if prefix is not None:
374 webpage_bytes = prefix + webpage_bytes
376 encoding = self._guess_encoding_from_content(content_type, webpage_bytes)
377 if self._downloader.params.get('dump_intermediate_pages', False):
379 url = url_or_request.get_full_url()
380 except AttributeError:
382 self.to_screen('Dumping request to ' + url)
383 dump = base64.b64encode(webpage_bytes).decode('ascii')
384 self._downloader.to_screen(dump)
385 if self._downloader.params.get('write_pages', False):
387 url = url_or_request.get_full_url()
388 except AttributeError:
390 basen = '%s_%s' % (video_id, url)
392 h = '___' + hashlib.md5(basen.encode('utf-8')).hexdigest()
393 basen = basen[:240 - len(h)] + h
394 raw_filename = basen + '.dump'
395 filename = sanitize_filename(raw_filename, restricted=True)
396 self.to_screen('Saving request to ' + filename)
397 # Working around MAX_PATH limitation on Windows (see
398 # http://msdn.microsoft.com/en-us/library/windows/desktop/aa365247(v=vs.85).aspx)
400 absfilepath = os.path.abspath(filename)
401 if len(absfilepath) > 259:
402 filename = '\\\\?\\' + absfilepath
403 with open(filename, 'wb') as outf:
404 outf.write(webpage_bytes)
407 content = webpage_bytes.decode(encoding, 'replace')
409 content = webpage_bytes.decode('utf-8', 'replace')
411 if ('<title>Access to this site is blocked</title>' in content and
412 'Websense' in content[:512]):
413 msg = 'Access to this webpage has been blocked by Websense filtering software in your network.'
414 blocked_iframe = self._html_search_regex(
415 r'<iframe src="([^"]+)"', content,
416 'Websense information URL', default=None)
418 msg += ' Visit %s for more details' % blocked_iframe
419 raise ExtractorError(msg, expected=True)
420 if '<title>The URL you requested has been blocked</title>' in content[:512]:
422 'Access to this webpage has been blocked by Indian censorship. '
423 'Use a VPN or proxy server (with --proxy) to route around it.')
424 block_msg = self._html_search_regex(
425 r'</h1><p>(.*?)</p>',
426 content, 'block message', default=None)
428 msg += ' (Message: "%s")' % block_msg.replace('\n', ' ')
429 raise ExtractorError(msg, expected=True)
433 def _download_webpage(self, url_or_request, video_id, note=None, errnote=None, fatal=True, tries=1, timeout=5, encoding=None):
434 """ Returns the data of the page as a string """
437 while success is False:
439 res = self._download_webpage_handle(url_or_request, video_id, note, errnote, fatal, encoding=encoding)
441 except compat_http_client.IncompleteRead as e:
443 if try_count >= tries:
445 self._sleep(timeout, video_id)
452 def _download_xml(self, url_or_request, video_id,
453 note='Downloading XML', errnote='Unable to download XML',
454 transform_source=None, fatal=True, encoding=None):
455 """Return the xml as an xml.etree.ElementTree.Element"""
456 xml_string = self._download_webpage(
457 url_or_request, video_id, note, errnote, fatal=fatal, encoding=encoding)
458 if xml_string is False:
461 xml_string = transform_source(xml_string)
462 return xml.etree.ElementTree.fromstring(xml_string.encode('utf-8'))
464 def _download_json(self, url_or_request, video_id,
465 note='Downloading JSON metadata',
466 errnote='Unable to download JSON metadata',
467 transform_source=None,
468 fatal=True, encoding=None):
469 json_string = self._download_webpage(
470 url_or_request, video_id, note, errnote, fatal=fatal,
472 if (not fatal) and json_string is False:
474 return self._parse_json(
475 json_string, video_id, transform_source=transform_source, fatal=fatal)
477 def _parse_json(self, json_string, video_id, transform_source=None, fatal=True):
479 json_string = transform_source(json_string)
481 return json.loads(json_string)
482 except ValueError as ve:
483 errmsg = '%s: Failed to parse JSON ' % video_id
485 raise ExtractorError(errmsg, cause=ve)
487 self.report_warning(errmsg + str(ve))
489 def report_warning(self, msg, video_id=None):
490 idstr = '' if video_id is None else '%s: ' % video_id
491 self._downloader.report_warning(
492 '[%s] %s%s' % (self.IE_NAME, idstr, msg))
494 def to_screen(self, msg):
495 """Print msg to screen, prefixing it with '[ie_name]'"""
496 self._downloader.to_screen('[%s] %s' % (self.IE_NAME, msg))
498 def report_extraction(self, id_or_name):
499 """Report information extraction."""
500 self.to_screen('%s: Extracting information' % id_or_name)
502 def report_download_webpage(self, video_id):
503 """Report webpage download."""
504 self.to_screen('%s: Downloading webpage' % video_id)
506 def report_age_confirmation(self):
507 """Report attempt to confirm age."""
508 self.to_screen('Confirming age')
510 def report_login(self):
511 """Report attempt to log in."""
512 self.to_screen('Logging in')
515 def raise_login_required(msg='This video is only available for registered users'):
516 raise ExtractorError(
517 '%s. Use --username and --password or --netrc to provide account credentials.' % msg,
521 def raise_geo_restricted(msg='This video is not available from your location due to geo restriction'):
522 raise ExtractorError(
523 '%s. You might want to use --proxy to workaround.' % msg,
526 # Methods for following #608
528 def url_result(url, ie=None, video_id=None, video_title=None):
529 """Returns a URL that points to a page that should be processed"""
530 # TODO: ie should be the class used for getting the info
531 video_info = {'_type': 'url',
534 if video_id is not None:
535 video_info['id'] = video_id
536 if video_title is not None:
537 video_info['title'] = video_title
541 def playlist_result(entries, playlist_id=None, playlist_title=None, playlist_description=None):
542 """Returns a playlist"""
543 video_info = {'_type': 'playlist',
546 video_info['id'] = playlist_id
548 video_info['title'] = playlist_title
549 if playlist_description:
550 video_info['description'] = playlist_description
553 def _search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
555 Perform a regex search on the given string, using a single or a list of
556 patterns returning the first matching group.
557 In case of failure return a default value or raise a WARNING or a
558 RegexNotFoundError, depending on fatal, specifying the field name.
560 if isinstance(pattern, (str, compat_str, compiled_regex_type)):
561 mobj = re.search(pattern, string, flags)
564 mobj = re.search(p, string, flags)
568 if not self._downloader.params.get('no_color') and os.name != 'nt' and sys.stderr.isatty():
569 _name = '\033[0;34m%s\033[0m' % name
575 # return the first matching group
576 return next(g for g in mobj.groups() if g is not None)
578 return mobj.group(group)
579 elif default is not NO_DEFAULT:
582 raise RegexNotFoundError('Unable to extract %s' % _name)
584 self._downloader.report_warning('unable to extract %s' % _name + bug_reports_message())
587 def _html_search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
589 Like _search_regex, but strips HTML tags and unescapes entities.
591 res = self._search_regex(pattern, string, name, default, fatal, flags, group)
593 return clean_html(res).strip()
597 def _get_login_info(self):
599 Get the login info as (username, password)
600 It will look in the netrc file using the _NETRC_MACHINE value
601 If there's no info available, return (None, None)
603 if self._downloader is None:
608 downloader_params = self._downloader.params
610 # Attempt to use provided username and password or .netrc data
611 if downloader_params.get('username', None) is not None:
612 username = downloader_params['username']
613 password = downloader_params['password']
614 elif downloader_params.get('usenetrc', False):
616 info = netrc.netrc().authenticators(self._NETRC_MACHINE)
621 raise netrc.NetrcParseError('No authenticators for %s' % self._NETRC_MACHINE)
622 except (IOError, netrc.NetrcParseError) as err:
623 self._downloader.report_warning('parsing .netrc: %s' % compat_str(err))
625 return (username, password)
627 def _get_tfa_info(self, note='two-factor verification code'):
629 Get the two-factor authentication info
630 TODO - asking the user will be required for sms/phone verify
631 currently just uses the command line option
632 If there's no info available, return None
634 if self._downloader is None:
636 downloader_params = self._downloader.params
638 if downloader_params.get('twofactor', None) is not None:
639 return downloader_params['twofactor']
641 return compat_getpass('Type %s and press [Return]: ' % note)
643 # Helper functions for extracting OpenGraph info
645 def _og_regexes(prop):
646 content_re = r'content=(?:"([^>]+?)"|\'([^>]+?)\')'
647 property_re = r'(?:name|property)=[\'"]og:%s[\'"]' % re.escape(prop)
648 template = r'<meta[^>]+?%s[^>]+?%s'
650 template % (property_re, content_re),
651 template % (content_re, property_re),
655 def _meta_regex(prop):
656 return r'''(?isx)<meta
657 (?=[^>]+(?:itemprop|name|property|id|http-equiv)=(["\']?)%s\1)
658 [^>]+?content=(["\'])(?P<content>.*?)\2''' % re.escape(prop)
660 def _og_search_property(self, prop, html, name=None, **kargs):
662 name = 'OpenGraph %s' % prop
663 escaped = self._search_regex(self._og_regexes(prop), html, name, flags=re.DOTALL, **kargs)
666 return unescapeHTML(escaped)
668 def _og_search_thumbnail(self, html, **kargs):
669 return self._og_search_property('image', html, 'thumbnail URL', fatal=False, **kargs)
671 def _og_search_description(self, html, **kargs):
672 return self._og_search_property('description', html, fatal=False, **kargs)
674 def _og_search_title(self, html, **kargs):
675 return self._og_search_property('title', html, **kargs)
677 def _og_search_video_url(self, html, name='video url', secure=True, **kargs):
678 regexes = self._og_regexes('video') + self._og_regexes('video:url')
680 regexes = self._og_regexes('video:secure_url') + regexes
681 return self._html_search_regex(regexes, html, name, **kargs)
683 def _og_search_url(self, html, **kargs):
684 return self._og_search_property('url', html, **kargs)
686 def _html_search_meta(self, name, html, display_name=None, fatal=False, **kwargs):
687 if display_name is None:
689 return self._html_search_regex(
690 self._meta_regex(name),
691 html, display_name, fatal=fatal, group='content', **kwargs)
693 def _dc_search_uploader(self, html):
694 return self._html_search_meta('dc.creator', html, 'uploader')
696 def _rta_search(self, html):
697 # See http://www.rtalabel.org/index.php?content=howtofaq#single
698 if re.search(r'(?ix)<meta\s+name="rating"\s+'
699 r' content="RTA-5042-1996-1400-1577-RTA"',
704 def _media_rating_search(self, html):
705 # See http://www.tjg-designs.com/WP/metadata-code-examples-adding-metadata-to-your-web-pages/
706 rating = self._html_search_meta('rating', html)
718 return RATING_TABLE.get(rating.lower(), None)
720 def _family_friendly_search(self, html):
721 # See http://schema.org/VideoObject
722 family_friendly = self._html_search_meta('isFamilyFriendly', html)
724 if not family_friendly:
733 return RATING_TABLE.get(family_friendly.lower(), None)
735 def _twitter_search_player(self, html):
736 return self._html_search_meta('twitter:player', html,
737 'twitter card player')
740 def _hidden_inputs(html):
741 html = re.sub(r'<!--(?:(?!<!--).)*-->', '', html)
743 for input in re.findall(r'(?i)<input([^>]+)>', html):
744 if not re.search(r'type=(["\'])(?:hidden|submit)\1', input):
746 name = re.search(r'name=(["\'])(?P<value>.+?)\1', input)
749 value = re.search(r'value=(["\'])(?P<value>.*?)\1', input)
752 hidden_inputs[name.group('value')] = value.group('value')
755 def _form_hidden_inputs(self, form_id, html):
756 form = self._search_regex(
757 r'(?is)<form[^>]+?id=(["\'])%s\1[^>]*>(?P<form>.+?)</form>' % form_id,
758 html, '%s form' % form_id, group='form')
759 return self._hidden_inputs(form)
761 def _sort_formats(self, formats, field_preference=None):
763 raise ExtractorError('No video formats found')
766 # TODO remove the following workaround
767 from ..utils import determine_ext
768 if not f.get('ext') and 'url' in f:
769 f['ext'] = determine_ext(f['url'])
771 if isinstance(field_preference, (list, tuple)):
772 return tuple(f.get(field) if f.get(field) is not None else -1 for field in field_preference)
774 preference = f.get('preference')
775 if preference is None:
776 proto = f.get('protocol')
778 proto = compat_urllib_parse_urlparse(f.get('url', '')).scheme
780 preference = 0 if proto in ['http', 'https'] else -0.1
781 if f.get('ext') in ['f4f', 'f4m']: # Not yet supported
784 if f.get('vcodec') == 'none': # audio only
785 if self._downloader.params.get('prefer_free_formats'):
786 ORDER = ['aac', 'mp3', 'm4a', 'webm', 'ogg', 'opus']
788 ORDER = ['webm', 'opus', 'ogg', 'mp3', 'aac', 'm4a']
791 audio_ext_preference = ORDER.index(f['ext'])
793 audio_ext_preference = -1
795 if self._downloader.params.get('prefer_free_formats'):
796 ORDER = ['flv', 'mp4', 'webm']
798 ORDER = ['webm', 'flv', 'mp4']
800 ext_preference = ORDER.index(f['ext'])
803 audio_ext_preference = 0
807 f.get('language_preference') if f.get('language_preference') is not None else -1,
808 f.get('quality') if f.get('quality') is not None else -1,
809 f.get('tbr') if f.get('tbr') is not None else -1,
810 f.get('filesize') if f.get('filesize') is not None else -1,
811 f.get('vbr') if f.get('vbr') is not None else -1,
812 f.get('height') if f.get('height') is not None else -1,
813 f.get('width') if f.get('width') is not None else -1,
815 f.get('abr') if f.get('abr') is not None else -1,
816 audio_ext_preference,
817 f.get('fps') if f.get('fps') is not None else -1,
818 f.get('filesize_approx') if f.get('filesize_approx') is not None else -1,
819 f.get('source_preference') if f.get('source_preference') is not None else -1,
820 f.get('format_id') if f.get('format_id') is not None else '',
822 formats.sort(key=_formats_key)
824 def _check_formats(self, formats, video_id):
827 lambda f: self._is_valid_url(
829 item='%s video format' % f.get('format_id') if f.get('format_id') else 'video'),
832 def _is_valid_url(self, url, video_id, item='video'):
833 url = self._proto_relative_url(url, scheme='http:')
834 # For now assume non HTTP(S) URLs always valid
835 if not (url.startswith('http://') or url.startswith('https://')):
838 self._request_webpage(url, video_id, 'Checking %s URL' % item)
840 except ExtractorError as e:
841 if isinstance(e.cause, compat_HTTPError):
843 '%s: %s URL is invalid, skipping' % (video_id, item))
847 def http_scheme(self):
848 """ Either "http:" or "https:", depending on the user's preferences """
851 if self._downloader.params.get('prefer_insecure', False)
854 def _proto_relative_url(self, url, scheme=None):
857 if url.startswith('//'):
859 scheme = self.http_scheme()
864 def _sleep(self, timeout, video_id, msg_template=None):
865 if msg_template is None:
866 msg_template = '%(video_id)s: Waiting for %(timeout)s seconds'
867 msg = msg_template % {'video_id': video_id, 'timeout': timeout}
871 def _extract_f4m_formats(self, manifest_url, video_id, preference=None, f4m_id=None,
872 transform_source=lambda s: fix_xml_ampersands(s).strip()):
873 manifest = self._download_xml(
874 manifest_url, video_id, 'Downloading f4m manifest',
875 'Unable to download f4m manifest',
876 # Some manifests may be malformed, e.g. prosiebensat1 generated manifests
877 # (see https://github.com/rg3/youtube-dl/issues/6215#issuecomment-121704244)
878 transform_source=transform_source)
881 manifest_version = '1.0'
882 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/1.0}media')
884 manifest_version = '2.0'
885 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/2.0}media')
886 for i, media_el in enumerate(media_nodes):
887 if manifest_version == '2.0':
888 media_url = media_el.attrib.get('href') or media_el.attrib.get('url')
892 media_url if media_url.startswith('http://') or media_url.startswith('https://')
893 else ('/'.join(manifest_url.split('/')[:-1]) + '/' + media_url))
894 # If media_url is itself a f4m manifest do the recursive extraction
895 # since bitrates in parent manifest (this one) and media_url manifest
896 # may differ leading to inability to resolve the format by requested
897 # bitrate in f4m downloader
898 if determine_ext(manifest_url) == 'f4m':
899 formats.extend(self._extract_f4m_formats(manifest_url, video_id, preference, f4m_id))
901 tbr = int_or_none(media_el.attrib.get('bitrate'))
903 'format_id': '-'.join(filter(None, [f4m_id, compat_str(i if tbr is None else tbr)])),
907 'width': int_or_none(media_el.attrib.get('width')),
908 'height': int_or_none(media_el.attrib.get('height')),
909 'preference': preference,
911 self._sort_formats(formats)
915 def _extract_m3u8_formats(self, m3u8_url, video_id, ext=None,
916 entry_protocol='m3u8', preference=None,
917 m3u8_id=None, note=None, errnote=None,
921 'format_id': '-'.join(filter(None, [m3u8_id, 'meta'])),
925 'preference': preference - 1 if preference else -1,
926 'resolution': 'multiple',
927 'format_note': 'Quality selection URL',
930 format_url = lambda u: (
932 if re.match(r'^https?://', u)
933 else compat_urlparse.urljoin(m3u8_url, u))
935 m3u8_doc = self._download_webpage(
937 note=note or 'Downloading m3u8 information',
938 errnote=errnote or 'Failed to download m3u8 information',
940 if m3u8_doc is False:
945 r'(?P<key>[a-zA-Z_-]+)=(?P<val>"[^"]+"|[^",]+)(?:,|$)')
946 for line in m3u8_doc.splitlines():
947 if line.startswith('#EXT-X-STREAM-INF:'):
949 for m in kv_rex.finditer(line):
951 if v.startswith('"'):
953 last_info[m.group('key')] = v
954 elif line.startswith('#EXT-X-MEDIA:'):
956 for m in kv_rex.finditer(line):
958 if v.startswith('"'):
960 last_media[m.group('key')] = v
961 elif line.startswith('#') or not line.strip():
964 if last_info is None:
965 formats.append({'url': format_url(line)})
967 tbr = int_or_none(last_info.get('BANDWIDTH'), scale=1000)
970 format_id.append(m3u8_id)
971 last_media_name = last_media.get('NAME') if last_media and last_media.get('TYPE') != 'SUBTITLES' else None
972 format_id.append(last_media_name if last_media_name else '%d' % (tbr if tbr else len(formats)))
974 'format_id': '-'.join(format_id),
975 'url': format_url(line.strip()),
978 'protocol': entry_protocol,
979 'preference': preference,
981 codecs = last_info.get('CODECS')
983 # TODO: looks like video codec is not always necessarily goes first
984 va_codecs = codecs.split(',')
986 f['vcodec'] = va_codecs[0].partition('.')[0]
987 if len(va_codecs) > 1 and va_codecs[1]:
988 f['acodec'] = va_codecs[1].partition('.')[0]
989 resolution = last_info.get('RESOLUTION')
991 width_str, height_str = resolution.split('x')
992 f['width'] = int(width_str)
993 f['height'] = int(height_str)
994 if last_media is not None:
995 f['m3u8_media'] = last_media
999 self._sort_formats(formats)
1003 def _xpath_ns(path, namespace=None):
1007 for c in path.split('/'):
1008 if not c or c == '.':
1011 out.append('{%s}%s' % (namespace, c))
1012 return '/'.join(out)
1014 def _extract_smil_formats(self, smil_url, video_id, fatal=True, f4m_params=None):
1015 smil = self._download_smil(smil_url, video_id, fatal=fatal)
1021 namespace = self._parse_smil_namespace(smil)
1023 return self._parse_smil_formats(
1024 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
1026 def _extract_smil_info(self, smil_url, video_id, fatal=True, f4m_params=None):
1027 smil = self._download_smil(smil_url, video_id, fatal=fatal)
1030 return self._parse_smil(smil, smil_url, video_id, f4m_params=f4m_params)
1032 def _download_smil(self, smil_url, video_id, fatal=True):
1033 return self._download_xml(
1034 smil_url, video_id, 'Downloading SMIL file',
1035 'Unable to download SMIL file', fatal=fatal)
1037 def _parse_smil(self, smil, smil_url, video_id, f4m_params=None):
1038 namespace = self._parse_smil_namespace(smil)
1040 formats = self._parse_smil_formats(
1041 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
1042 subtitles = self._parse_smil_subtitles(smil, namespace=namespace)
1044 video_id = os.path.splitext(url_basename(smil_url))[0]
1047 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
1048 name = meta.attrib.get('name')
1049 content = meta.attrib.get('content')
1050 if not name or not content:
1052 if not title and name == 'title':
1054 elif not description and name in ('description', 'abstract'):
1055 description = content
1058 'id': image.get('type'),
1059 'url': image.get('src'),
1060 'width': int_or_none(image.get('width')),
1061 'height': int_or_none(image.get('height')),
1062 } for image in smil.findall(self._xpath_ns('.//image', namespace)) if image.get('src')]
1066 'title': title or video_id,
1067 'description': description,
1068 'thumbnails': thumbnails,
1070 'subtitles': subtitles,
1073 def _parse_smil_namespace(self, smil):
1074 return self._search_regex(
1075 r'(?i)^{([^}]+)?}smil$', smil.tag, 'namespace', default=None)
1077 def _parse_smil_formats(self, smil, smil_url, video_id, namespace=None, f4m_params=None, transform_rtmp_url=None):
1079 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
1080 b = meta.get('base') or meta.get('httpBase')
1089 videos = smil.findall(self._xpath_ns('.//video', namespace))
1090 for video in videos:
1091 src = video.get('src')
1095 bitrate = float_or_none(video.get('system-bitrate') or video.get('systemBitrate'), 1000)
1096 filesize = int_or_none(video.get('size') or video.get('fileSize'))
1097 width = int_or_none(video.get('width'))
1098 height = int_or_none(video.get('height'))
1099 proto = video.get('proto')
1100 ext = video.get('ext')
1101 src_ext = determine_ext(src)
1102 streamer = video.get('streamer') or base
1104 if proto == 'rtmp' or streamer.startswith('rtmp'):
1110 'format_id': 'rtmp-%d' % (rtmp_count if bitrate is None else bitrate),
1112 'filesize': filesize,
1116 if transform_rtmp_url:
1117 streamer, src = transform_rtmp_url(streamer, src)
1118 formats[-1].update({
1124 src_url = src if src.startswith('http') else compat_urlparse.urljoin(base, src)
1126 if proto == 'm3u8' or src_ext == 'm3u8':
1127 formats.extend(self._extract_m3u8_formats(
1128 src_url, video_id, ext or 'mp4', m3u8_id='hls'))
1131 if src_ext == 'f4m':
1136 'plugin': 'flowplayer-3.2.0.1',
1138 f4m_url += '&' if '?' in f4m_url else '?'
1139 f4m_url += compat_urllib_parse.urlencode(f4m_params)
1140 formats.extend(self._extract_f4m_formats(f4m_url, video_id, f4m_id='hds'))
1143 if src_url.startswith('http'):
1147 'ext': ext or src_ext or 'flv',
1148 'format_id': 'http-%d' % (bitrate or http_count),
1150 'filesize': filesize,
1156 self._sort_formats(formats)
1160 def _parse_smil_subtitles(self, smil, namespace=None, subtitles_lang='en'):
1162 for num, textstream in enumerate(smil.findall(self._xpath_ns('.//textstream', namespace))):
1163 src = textstream.get('src')
1166 ext = textstream.get('ext') or determine_ext(src)
1168 type_ = textstream.get('type')
1172 'application/smptett+xml': 'tt',
1174 if type_ in SUBTITLES_TYPES:
1175 ext = SUBTITLES_TYPES[type_]
1176 lang = textstream.get('systemLanguage') or textstream.get('systemLanguageName') or textstream.get('lang') or subtitles_lang
1177 subtitles.setdefault(lang, []).append({
1183 def _extract_xspf_playlist(self, playlist_url, playlist_id, fatal=True):
1184 xspf = self._download_xml(
1185 playlist_url, playlist_id, 'Downloading xpsf playlist',
1186 'Unable to download xspf manifest', fatal=fatal)
1189 return self._parse_xspf(xspf, playlist_id)
1191 def _parse_xspf(self, playlist, playlist_id):
1193 'xspf': 'http://xspf.org/ns/0/',
1194 's1': 'http://static.streamone.nl/player/ns/0',
1198 for track in playlist.findall(xpath_with_ns('./xspf:trackList/xspf:track', NS_MAP)):
1200 track, xpath_with_ns('./xspf:title', NS_MAP), 'title', default=playlist_id)
1201 description = xpath_text(
1202 track, xpath_with_ns('./xspf:annotation', NS_MAP), 'description')
1203 thumbnail = xpath_text(
1204 track, xpath_with_ns('./xspf:image', NS_MAP), 'thumbnail')
1205 duration = float_or_none(
1206 xpath_text(track, xpath_with_ns('./xspf:duration', NS_MAP), 'duration'), 1000)
1209 'url': location.text,
1210 'format_id': location.get(xpath_with_ns('s1:label', NS_MAP)),
1211 'width': int_or_none(location.get(xpath_with_ns('s1:width', NS_MAP))),
1212 'height': int_or_none(location.get(xpath_with_ns('s1:height', NS_MAP))),
1213 } for location in track.findall(xpath_with_ns('./xspf:location', NS_MAP))]
1214 self._sort_formats(formats)
1219 'description': description,
1220 'thumbnail': thumbnail,
1221 'duration': duration,
1226 def _live_title(self, name):
1227 """ Generate the title for a live video """
1228 now = datetime.datetime.now()
1229 now_str = now.strftime("%Y-%m-%d %H:%M")
1230 return name + ' ' + now_str
1232 def _int(self, v, name, fatal=False, **kwargs):
1233 res = int_or_none(v, **kwargs)
1234 if 'get_attr' in kwargs:
1235 print(getattr(v, kwargs['get_attr']))
1237 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
1239 raise ExtractorError(msg)
1241 self._downloader.report_warning(msg)
1244 def _float(self, v, name, fatal=False, **kwargs):
1245 res = float_or_none(v, **kwargs)
1247 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
1249 raise ExtractorError(msg)
1251 self._downloader.report_warning(msg)
1254 def _set_cookie(self, domain, name, value, expire_time=None):
1255 cookie = compat_cookiejar.Cookie(
1256 0, name, value, None, None, domain, None,
1257 None, '/', True, False, expire_time, '', None, None, None)
1258 self._downloader.cookiejar.set_cookie(cookie)
1260 def _get_cookies(self, url):
1261 """ Return a compat_cookies.SimpleCookie with the cookies for the url """
1262 req = compat_urllib_request.Request(url)
1263 self._downloader.cookiejar.add_cookie_header(req)
1264 return compat_cookies.SimpleCookie(req.get_header('Cookie'))
1266 def get_testcases(self, include_onlymatching=False):
1267 t = getattr(self, '_TEST', None)
1269 assert not hasattr(self, '_TESTS'), \
1270 '%s has _TEST and _TESTS' % type(self).__name__
1273 tests = getattr(self, '_TESTS', [])
1275 if not include_onlymatching and t.get('only_matching', False):
1277 t['name'] = type(self).__name__[:-len('IE')]
1280 def is_suitable(self, age_limit):
1281 """ Test whether the extractor is generally suitable for the given
1282 age limit (i.e. pornographic sites are not, all others usually are) """
1284 any_restricted = False
1285 for tc in self.get_testcases(include_onlymatching=False):
1286 if 'playlist' in tc:
1287 tc = tc['playlist'][0]
1288 is_restricted = age_restricted(
1289 tc.get('info_dict', {}).get('age_limit'), age_limit)
1290 if not is_restricted:
1292 any_restricted = any_restricted or is_restricted
1293 return not any_restricted
1295 def extract_subtitles(self, *args, **kwargs):
1296 if (self._downloader.params.get('writesubtitles', False) or
1297 self._downloader.params.get('listsubtitles')):
1298 return self._get_subtitles(*args, **kwargs)
1301 def _get_subtitles(self, *args, **kwargs):
1302 raise NotImplementedError("This method must be implemented by subclasses")
1305 def _merge_subtitle_items(subtitle_list1, subtitle_list2):
1306 """ Merge subtitle items for one language. Items with duplicated URLs
1307 will be dropped. """
1308 list1_urls = set([item['url'] for item in subtitle_list1])
1309 ret = list(subtitle_list1)
1310 ret.extend([item for item in subtitle_list2 if item['url'] not in list1_urls])
1314 def _merge_subtitles(cls, subtitle_dict1, subtitle_dict2):
1315 """ Merge two subtitle dictionaries, language by language. """
1316 ret = dict(subtitle_dict1)
1317 for lang in subtitle_dict2:
1318 ret[lang] = cls._merge_subtitle_items(subtitle_dict1.get(lang, []), subtitle_dict2[lang])
1321 def extract_automatic_captions(self, *args, **kwargs):
1322 if (self._downloader.params.get('writeautomaticsub', False) or
1323 self._downloader.params.get('listsubtitles')):
1324 return self._get_automatic_captions(*args, **kwargs)
1327 def _get_automatic_captions(self, *args, **kwargs):
1328 raise NotImplementedError("This method must be implemented by subclasses")
1331 class SearchInfoExtractor(InfoExtractor):
1333 Base class for paged search queries extractors.
1334 They accept URLs in the format _SEARCH_KEY(|all|[0-9]):{query}
1335 Instances should define _SEARCH_KEY and _MAX_RESULTS.
1339 def _make_valid_url(cls):
1340 return r'%s(?P<prefix>|[1-9][0-9]*|all):(?P<query>[\s\S]+)' % cls._SEARCH_KEY
1343 def suitable(cls, url):
1344 return re.match(cls._make_valid_url(), url) is not None
1346 def _real_extract(self, query):
1347 mobj = re.match(self._make_valid_url(), query)
1349 raise ExtractorError('Invalid search query "%s"' % query)
1351 prefix = mobj.group('prefix')
1352 query = mobj.group('query')
1354 return self._get_n_results(query, 1)
1355 elif prefix == 'all':
1356 return self._get_n_results(query, self._MAX_RESULTS)
1360 raise ExtractorError('invalid download number %s for query "%s"' % (n, query))
1361 elif n > self._MAX_RESULTS:
1362 self._downloader.report_warning('%s returns max %i results (you requested %i)' % (self._SEARCH_KEY, self._MAX_RESULTS, n))
1363 n = self._MAX_RESULTS
1364 return self._get_n_results(query, n)
1366 def _get_n_results(self, query, n):
1367 """Get a specified number of results for a query"""
1368 raise NotImplementedError("This method must be implemented by subclasses")
1371 def SEARCH_KEY(self):
1372 return self._SEARCH_KEY