1 from __future__ import unicode_literals
13 import xml.etree.ElementTree
15 from ..compat import (
22 compat_urllib_parse_urlparse,
23 compat_urllib_request,
45 class InfoExtractor(object):
46 """Information Extractor class.
48 Information extractors are the classes that, given a URL, extract
49 information about the video (or videos) the URL refers to. This
50 information includes the real video URL, the video title, author and
51 others. The information is stored in a dictionary which is then
52 passed to the YoutubeDL. The YoutubeDL processes this
53 information possibly downloading the video to the file system, among
54 other possible outcomes.
56 The type field determines the type of the result.
57 By far the most common value (and the default if _type is missing) is
58 "video", which indicates a single video.
60 For a video, the dictionaries must include the following fields:
63 title: Video title, unescaped.
65 Additionally, it must contain either a formats entry or a url one:
67 formats: A list of dictionaries for each format available, ordered
68 from worst to best quality.
71 * url Mandatory. The URL of the video file
72 * ext Will be calculated from URL if missing
73 * format A human-readable description of the format
74 ("mp4 container with h264/opus").
75 Calculated from the format_id, width, height.
76 and format_note fields if missing.
77 * format_id A short description of the format
78 ("mp4_h264_opus" or "19").
79 Technically optional, but strongly recommended.
80 * format_note Additional info about the format
81 ("3D" or "DASH video")
82 * width Width of the video, if known
83 * height Height of the video, if known
84 * resolution Textual description of width and height
85 * tbr Average bitrate of audio and video in KBit/s
86 * abr Average audio bitrate in KBit/s
87 * acodec Name of the audio codec in use
88 * asr Audio sampling rate in Hertz
89 * vbr Average video bitrate in KBit/s
91 * vcodec Name of the video codec in use
92 * container Name of the container format
93 * filesize The number of bytes, if known in advance
94 * filesize_approx An estimate for the number of bytes
95 * player_url SWF Player URL (used for rtmpdump).
96 * protocol The protocol that will be used for the actual
98 "http", "https", "rtsp", "rtmp", "rtmpe",
99 "m3u8", or "m3u8_native".
100 * preference Order number of this format. If this field is
101 present and not None, the formats get sorted
102 by this field, regardless of all other values.
103 -1 for default (order by other properties),
104 -2 or smaller for less than default.
105 < -1000 to hide the format (if there is
106 another one which is strictly better)
107 * language_preference Is this in the correct requested
109 10 if it's what the URL is about,
110 -1 for default (don't know),
111 -10 otherwise, other values reserved for now.
112 * quality Order number of the video quality of this
113 format, irrespective of the file format.
114 -1 for default (order by other properties),
115 -2 or smaller for less than default.
116 * source_preference Order number for this video source
117 (quality takes higher priority)
118 -1 for default (order by other properties),
119 -2 or smaller for less than default.
120 * http_headers A dictionary of additional HTTP headers
121 to add to the request.
122 * stretched_ratio If given and not 1, indicates that the
123 video's pixels are not square.
124 width : height ratio as float.
125 * no_resume The server does not support resuming the
126 (HTTP or RTMP) download. Boolean.
128 url: Final video URL.
129 ext: Video filename extension.
130 format: The video format, defaults to ext (used for --get-format)
131 player_url: SWF Player URL (used for rtmpdump).
133 The following fields are optional:
135 alt_title: A secondary title of the video.
136 display_id An alternative identifier for the video, not necessarily
137 unique, but available before title. Typically, id is
138 something like "4234987", title "Dancing naked mole rats",
139 and display_id "dancing-naked-mole-rats"
140 thumbnails: A list of dictionaries, with the following entries:
141 * "id" (optional, string) - Thumbnail format ID
143 * "preference" (optional, int) - quality of the image
144 * "width" (optional, int)
145 * "height" (optional, int)
146 * "resolution" (optional, string "{width}x{height"},
148 thumbnail: Full URL to a video thumbnail image.
149 description: Full video description.
150 uploader: Full name of the video uploader.
151 creator: The main artist who created the video.
152 timestamp: UNIX timestamp of the moment the video became available.
153 upload_date: Video upload date (YYYYMMDD).
154 If not explicitly set, calculated from timestamp.
155 uploader_id: Nickname or id of the video uploader.
156 location: Physical location where the video was filmed.
157 subtitles: The available subtitles as a dictionary in the format
158 {language: subformats}. "subformats" is a list sorted from
159 lower to higher preference, each element is a dictionary
160 with the "ext" entry and one of:
161 * "data": The subtitles file contents
162 * "url": A URL pointing to the subtitles file
163 automatic_captions: Like 'subtitles', used by the YoutubeIE for
164 automatically generated captions
165 duration: Length of the video in seconds, as an integer.
166 view_count: How many users have watched the video on the platform.
167 like_count: Number of positive ratings of the video
168 dislike_count: Number of negative ratings of the video
169 average_rating: Average rating give by users, the scale used depends on the webpage
170 comment_count: Number of comments on the video
171 comments: A list of comments, each with one or more of the following
172 properties (all but one of text or html optional):
173 * "author" - human-readable name of the comment author
174 * "author_id" - user ID of the comment author
176 * "html" - Comment as HTML
177 * "text" - Plain text of the comment
178 * "timestamp" - UNIX timestamp of comment
179 * "parent" - ID of the comment this one is replying to.
180 Set to "root" to indicate that this is a
181 comment to the original video.
182 age_limit: Age restriction for the video, as an integer (years)
183 webpage_url: The URL to the video webpage, if given to youtube-dl it
184 should allow to get the same result again. (It will be set
185 by YoutubeDL if it's missing)
186 categories: A list of categories that the video falls in, for example
188 tags: A list of tags assigned to the video, e.g. ["sweden", "pop music"]
189 is_live: True, False, or None (=unknown). Whether this video is a
190 live stream that goes on instead of a fixed-length video.
191 start_time: Time in seconds where the reproduction should start, as
192 specified in the URL.
193 end_time: Time in seconds where the reproduction should end, as
194 specified in the URL.
196 Unless mentioned otherwise, the fields should be Unicode strings.
198 Unless mentioned otherwise, None is equivalent to absence of information.
201 _type "playlist" indicates multiple videos.
202 There must be a key "entries", which is a list, an iterable, or a PagedList
203 object, each element of which is a valid dictionary by this specification.
205 Additionally, playlists can have "title" and "id" attributes with the same
206 semantics as videos (see above).
209 _type "multi_video" indicates that there are multiple videos that
210 form a single show, for examples multiple acts of an opera or TV episode.
211 It must have an entries key like a playlist and contain all the keys
212 required for a video at the same time.
215 _type "url" indicates that the video must be extracted from another
216 location, possibly by a different extractor. Its only required key is:
217 "url" - the next URL to extract.
218 The key "ie_key" can be set to the class name (minus the trailing "IE",
219 e.g. "Youtube") if the extractor class is known in advance.
220 Additionally, the dictionary may have any properties of the resolved entity
221 known in advance, for example "title" if the title of the referred video is
225 _type "url_transparent" entities have the same specification as "url", but
226 indicate that the given additional information is more precise than the one
227 associated with the resolved URL.
228 This is useful when a site employs a video service that hosts the video and
229 its technical metadata, but that video service does not embed a useful
230 title, description etc.
233 Subclasses of this one should re-define the _real_initialize() and
234 _real_extract() methods and define a _VALID_URL regexp.
235 Probably, they should also be added to the list of extractors.
237 Finally, the _WORKING attribute should be set to False for broken IEs
238 in order to warn the users and skip the tests.
245 def __init__(self, downloader=None):
246 """Constructor. Receives an optional downloader."""
248 self.set_downloader(downloader)
251 def suitable(cls, url):
252 """Receives a URL and returns True if suitable for this IE."""
254 # This does not use has/getattr intentionally - we want to know whether
255 # we have cached the regexp for *this* class, whereas getattr would also
256 # match the superclass
257 if '_VALID_URL_RE' not in cls.__dict__:
258 cls._VALID_URL_RE = re.compile(cls._VALID_URL)
259 return cls._VALID_URL_RE.match(url) is not None
262 def _match_id(cls, url):
263 if '_VALID_URL_RE' not in cls.__dict__:
264 cls._VALID_URL_RE = re.compile(cls._VALID_URL)
265 m = cls._VALID_URL_RE.match(url)
271 """Getter method for _WORKING."""
274 def initialize(self):
275 """Initializes an instance (authentication, etc)."""
277 self._real_initialize()
280 def extract(self, url):
281 """Extracts URL information and returns it in list of dicts."""
284 return self._real_extract(url)
285 except ExtractorError:
287 except compat_http_client.IncompleteRead as e:
288 raise ExtractorError('A network error has occured.', cause=e, expected=True)
289 except (KeyError, StopIteration) as e:
290 raise ExtractorError('An extractor error has occured.', cause=e)
292 def set_downloader(self, downloader):
293 """Sets the downloader for this IE."""
294 self._downloader = downloader
296 def _real_initialize(self):
297 """Real initialization process. Redefine in subclasses."""
300 def _real_extract(self, url):
301 """Real extraction process. Redefine in subclasses."""
306 """A string for getting the InfoExtractor with get_info_extractor"""
307 return cls.__name__[:-2]
311 return type(self).__name__[:-2]
313 def _request_webpage(self, url_or_request, video_id, note=None, errnote=None, fatal=True):
314 """ Returns the response handle """
316 self.report_download_webpage(video_id)
317 elif note is not False:
319 self.to_screen('%s' % (note,))
321 self.to_screen('%s: %s' % (video_id, note))
323 return self._downloader.urlopen(url_or_request)
324 except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err:
328 errnote = 'Unable to download webpage'
329 errmsg = '%s: %s' % (errnote, compat_str(err))
331 raise ExtractorError(errmsg, sys.exc_info()[2], cause=err)
333 self._downloader.report_warning(errmsg)
336 def _download_webpage_handle(self, url_or_request, video_id, note=None, errnote=None, fatal=True, encoding=None):
337 """ Returns a tuple (page content as string, URL handle) """
338 # Strip hashes from the URL (#1038)
339 if isinstance(url_or_request, (compat_str, str)):
340 url_or_request = url_or_request.partition('#')[0]
342 urlh = self._request_webpage(url_or_request, video_id, note, errnote, fatal)
346 content = self._webpage_read_content(urlh, url_or_request, video_id, note, errnote, fatal, encoding=encoding)
347 return (content, urlh)
350 def _guess_encoding_from_content(content_type, webpage_bytes):
351 m = re.match(r'[a-zA-Z0-9_.-]+/[a-zA-Z0-9_.-]+\s*;\s*charset=(.+)', content_type)
353 encoding = m.group(1)
355 m = re.search(br'<meta[^>]+charset=[\'"]?([^\'")]+)[ /\'">]',
356 webpage_bytes[:1024])
358 encoding = m.group(1).decode('ascii')
359 elif webpage_bytes.startswith(b'\xff\xfe'):
366 def _webpage_read_content(self, urlh, url_or_request, video_id, note=None, errnote=None, fatal=True, prefix=None, encoding=None):
367 content_type = urlh.headers.get('Content-Type', '')
368 webpage_bytes = urlh.read()
369 if prefix is not None:
370 webpage_bytes = prefix + webpage_bytes
372 encoding = self._guess_encoding_from_content(content_type, webpage_bytes)
373 if self._downloader.params.get('dump_intermediate_pages', False):
375 url = url_or_request.get_full_url()
376 except AttributeError:
378 self.to_screen('Dumping request to ' + url)
379 dump = base64.b64encode(webpage_bytes).decode('ascii')
380 self._downloader.to_screen(dump)
381 if self._downloader.params.get('write_pages', False):
383 url = url_or_request.get_full_url()
384 except AttributeError:
386 basen = '%s_%s' % (video_id, url)
388 h = '___' + hashlib.md5(basen.encode('utf-8')).hexdigest()
389 basen = basen[:240 - len(h)] + h
390 raw_filename = basen + '.dump'
391 filename = sanitize_filename(raw_filename, restricted=True)
392 self.to_screen('Saving request to ' + filename)
393 # Working around MAX_PATH limitation on Windows (see
394 # http://msdn.microsoft.com/en-us/library/windows/desktop/aa365247(v=vs.85).aspx)
396 absfilepath = os.path.abspath(filename)
397 if len(absfilepath) > 259:
398 filename = '\\\\?\\' + absfilepath
399 with open(filename, 'wb') as outf:
400 outf.write(webpage_bytes)
403 content = webpage_bytes.decode(encoding, 'replace')
405 content = webpage_bytes.decode('utf-8', 'replace')
407 if ('<title>Access to this site is blocked</title>' in content and
408 'Websense' in content[:512]):
409 msg = 'Access to this webpage has been blocked by Websense filtering software in your network.'
410 blocked_iframe = self._html_search_regex(
411 r'<iframe src="([^"]+)"', content,
412 'Websense information URL', default=None)
414 msg += ' Visit %s for more details' % blocked_iframe
415 raise ExtractorError(msg, expected=True)
416 if '<title>The URL you requested has been blocked</title>' in content[:512]:
418 'Access to this webpage has been blocked by Indian censorship. '
419 'Use a VPN or proxy server (with --proxy) to route around it.')
420 block_msg = self._html_search_regex(
421 r'</h1><p>(.*?)</p>',
422 content, 'block message', default=None)
424 msg += ' (Message: "%s")' % block_msg.replace('\n', ' ')
425 raise ExtractorError(msg, expected=True)
429 def _download_webpage(self, url_or_request, video_id, note=None, errnote=None, fatal=True, tries=1, timeout=5, encoding=None):
430 """ Returns the data of the page as a string """
433 while success is False:
435 res = self._download_webpage_handle(url_or_request, video_id, note, errnote, fatal, encoding=encoding)
437 except compat_http_client.IncompleteRead as e:
439 if try_count >= tries:
441 self._sleep(timeout, video_id)
448 def _download_xml(self, url_or_request, video_id,
449 note='Downloading XML', errnote='Unable to download XML',
450 transform_source=None, fatal=True, encoding=None):
451 """Return the xml as an xml.etree.ElementTree.Element"""
452 xml_string = self._download_webpage(
453 url_or_request, video_id, note, errnote, fatal=fatal, encoding=encoding)
454 if xml_string is False:
457 xml_string = transform_source(xml_string)
458 return xml.etree.ElementTree.fromstring(xml_string.encode('utf-8'))
460 def _download_json(self, url_or_request, video_id,
461 note='Downloading JSON metadata',
462 errnote='Unable to download JSON metadata',
463 transform_source=None,
464 fatal=True, encoding=None):
465 json_string = self._download_webpage(
466 url_or_request, video_id, note, errnote, fatal=fatal,
468 if (not fatal) and json_string is False:
470 return self._parse_json(
471 json_string, video_id, transform_source=transform_source, fatal=fatal)
473 def _parse_json(self, json_string, video_id, transform_source=None, fatal=True):
475 json_string = transform_source(json_string)
477 return json.loads(json_string)
478 except ValueError as ve:
479 errmsg = '%s: Failed to parse JSON ' % video_id
481 raise ExtractorError(errmsg, cause=ve)
483 self.report_warning(errmsg + str(ve))
485 def report_warning(self, msg, video_id=None):
486 idstr = '' if video_id is None else '%s: ' % video_id
487 self._downloader.report_warning(
488 '[%s] %s%s' % (self.IE_NAME, idstr, msg))
490 def to_screen(self, msg):
491 """Print msg to screen, prefixing it with '[ie_name]'"""
492 self._downloader.to_screen('[%s] %s' % (self.IE_NAME, msg))
494 def report_extraction(self, id_or_name):
495 """Report information extraction."""
496 self.to_screen('%s: Extracting information' % id_or_name)
498 def report_download_webpage(self, video_id):
499 """Report webpage download."""
500 self.to_screen('%s: Downloading webpage' % video_id)
502 def report_age_confirmation(self):
503 """Report attempt to confirm age."""
504 self.to_screen('Confirming age')
506 def report_login(self):
507 """Report attempt to log in."""
508 self.to_screen('Logging in')
510 # Methods for following #608
512 def url_result(url, ie=None, video_id=None, video_title=None):
513 """Returns a URL that points to a page that should be processed"""
514 # TODO: ie should be the class used for getting the info
515 video_info = {'_type': 'url',
518 if video_id is not None:
519 video_info['id'] = video_id
520 if video_title is not None:
521 video_info['title'] = video_title
525 def playlist_result(entries, playlist_id=None, playlist_title=None, playlist_description=None):
526 """Returns a playlist"""
527 video_info = {'_type': 'playlist',
530 video_info['id'] = playlist_id
532 video_info['title'] = playlist_title
533 if playlist_description:
534 video_info['description'] = playlist_description
537 def _search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
539 Perform a regex search on the given string, using a single or a list of
540 patterns returning the first matching group.
541 In case of failure return a default value or raise a WARNING or a
542 RegexNotFoundError, depending on fatal, specifying the field name.
544 if isinstance(pattern, (str, compat_str, compiled_regex_type)):
545 mobj = re.search(pattern, string, flags)
548 mobj = re.search(p, string, flags)
552 if not self._downloader.params.get('no_color') and os.name != 'nt' and sys.stderr.isatty():
553 _name = '\033[0;34m%s\033[0m' % name
559 # return the first matching group
560 return next(g for g in mobj.groups() if g is not None)
562 return mobj.group(group)
563 elif default is not NO_DEFAULT:
566 raise RegexNotFoundError('Unable to extract %s' % _name)
568 self._downloader.report_warning('unable to extract %s' % _name + bug_reports_message())
571 def _html_search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
573 Like _search_regex, but strips HTML tags and unescapes entities.
575 res = self._search_regex(pattern, string, name, default, fatal, flags, group)
577 return clean_html(res).strip()
581 def _get_login_info(self):
583 Get the login info as (username, password)
584 It will look in the netrc file using the _NETRC_MACHINE value
585 If there's no info available, return (None, None)
587 if self._downloader is None:
592 downloader_params = self._downloader.params
594 # Attempt to use provided username and password or .netrc data
595 if downloader_params.get('username', None) is not None:
596 username = downloader_params['username']
597 password = downloader_params['password']
598 elif downloader_params.get('usenetrc', False):
600 info = netrc.netrc().authenticators(self._NETRC_MACHINE)
605 raise netrc.NetrcParseError('No authenticators for %s' % self._NETRC_MACHINE)
606 except (IOError, netrc.NetrcParseError) as err:
607 self._downloader.report_warning('parsing .netrc: %s' % compat_str(err))
609 return (username, password)
611 def _get_tfa_info(self):
613 Get the two-factor authentication info
614 TODO - asking the user will be required for sms/phone verify
615 currently just uses the command line option
616 If there's no info available, return None
618 if self._downloader is None:
620 downloader_params = self._downloader.params
622 if downloader_params.get('twofactor', None) is not None:
623 return downloader_params['twofactor']
627 # Helper functions for extracting OpenGraph info
629 def _og_regexes(prop):
630 content_re = r'content=(?:"([^>]+?)"|\'([^>]+?)\')'
631 property_re = r'(?:name|property)=[\'"]og:%s[\'"]' % re.escape(prop)
632 template = r'<meta[^>]+?%s[^>]+?%s'
634 template % (property_re, content_re),
635 template % (content_re, property_re),
639 def _meta_regex(prop):
640 return r'''(?isx)<meta
641 (?=[^>]+(?:itemprop|name|property)=(["\']?)%s\1)
642 [^>]+?content=(["\'])(?P<content>.*?)\2''' % re.escape(prop)
644 def _og_search_property(self, prop, html, name=None, **kargs):
646 name = 'OpenGraph %s' % prop
647 escaped = self._search_regex(self._og_regexes(prop), html, name, flags=re.DOTALL, **kargs)
650 return unescapeHTML(escaped)
652 def _og_search_thumbnail(self, html, **kargs):
653 return self._og_search_property('image', html, 'thumbnail URL', fatal=False, **kargs)
655 def _og_search_description(self, html, **kargs):
656 return self._og_search_property('description', html, fatal=False, **kargs)
658 def _og_search_title(self, html, **kargs):
659 return self._og_search_property('title', html, **kargs)
661 def _og_search_video_url(self, html, name='video url', secure=True, **kargs):
662 regexes = self._og_regexes('video') + self._og_regexes('video:url')
664 regexes = self._og_regexes('video:secure_url') + regexes
665 return self._html_search_regex(regexes, html, name, **kargs)
667 def _og_search_url(self, html, **kargs):
668 return self._og_search_property('url', html, **kargs)
670 def _html_search_meta(self, name, html, display_name=None, fatal=False, **kwargs):
671 if display_name is None:
673 return self._html_search_regex(
674 self._meta_regex(name),
675 html, display_name, fatal=fatal, group='content', **kwargs)
677 def _dc_search_uploader(self, html):
678 return self._html_search_meta('dc.creator', html, 'uploader')
680 def _rta_search(self, html):
681 # See http://www.rtalabel.org/index.php?content=howtofaq#single
682 if re.search(r'(?ix)<meta\s+name="rating"\s+'
683 r' content="RTA-5042-1996-1400-1577-RTA"',
688 def _media_rating_search(self, html):
689 # See http://www.tjg-designs.com/WP/metadata-code-examples-adding-metadata-to-your-web-pages/
690 rating = self._html_search_meta('rating', html)
702 return RATING_TABLE.get(rating.lower(), None)
704 def _family_friendly_search(self, html):
705 # See http://schema.org/VideoObject
706 family_friendly = self._html_search_meta('isFamilyFriendly', html)
708 if not family_friendly:
717 return RATING_TABLE.get(family_friendly.lower(), None)
719 def _twitter_search_player(self, html):
720 return self._html_search_meta('twitter:player', html,
721 'twitter card player')
724 def _hidden_inputs(html):
726 (input.group('name'), input.group('value')) for input in re.finditer(
729 type=(?P<q_hidden>["\'])hidden(?P=q_hidden)\s+
730 name=(?P<q_name>["\'])(?P<name>.+?)(?P=q_name)\s+
731 (?:id=(?P<q_id>["\']).+?(?P=q_id)\s+)?
732 value=(?P<q_value>["\'])(?P<value>.*?)(?P=q_value)
736 def _form_hidden_inputs(self, form_id, html):
737 form = self._search_regex(
738 r'(?s)<form[^>]+?id=(["\'])%s\1[^>]*>(?P<form>.+?)</form>' % form_id,
739 html, '%s form' % form_id, group='form')
740 return self._hidden_inputs(form)
742 def _sort_formats(self, formats, field_preference=None):
744 raise ExtractorError('No video formats found')
747 # TODO remove the following workaround
748 from ..utils import determine_ext
749 if not f.get('ext') and 'url' in f:
750 f['ext'] = determine_ext(f['url'])
752 if isinstance(field_preference, (list, tuple)):
753 return tuple(f.get(field) if f.get(field) is not None else -1 for field in field_preference)
755 preference = f.get('preference')
756 if preference is None:
757 proto = f.get('protocol')
759 proto = compat_urllib_parse_urlparse(f.get('url', '')).scheme
761 preference = 0 if proto in ['http', 'https'] else -0.1
762 if f.get('ext') in ['f4f', 'f4m']: # Not yet supported
765 if f.get('vcodec') == 'none': # audio only
766 if self._downloader.params.get('prefer_free_formats'):
767 ORDER = ['aac', 'mp3', 'm4a', 'webm', 'ogg', 'opus']
769 ORDER = ['webm', 'opus', 'ogg', 'mp3', 'aac', 'm4a']
772 audio_ext_preference = ORDER.index(f['ext'])
774 audio_ext_preference = -1
776 if self._downloader.params.get('prefer_free_formats'):
777 ORDER = ['flv', 'mp4', 'webm']
779 ORDER = ['webm', 'flv', 'mp4']
781 ext_preference = ORDER.index(f['ext'])
784 audio_ext_preference = 0
788 f.get('language_preference') if f.get('language_preference') is not None else -1,
789 f.get('quality') if f.get('quality') is not None else -1,
790 f.get('tbr') if f.get('tbr') is not None else -1,
791 f.get('filesize') if f.get('filesize') is not None else -1,
792 f.get('vbr') if f.get('vbr') is not None else -1,
793 f.get('height') if f.get('height') is not None else -1,
794 f.get('width') if f.get('width') is not None else -1,
796 f.get('abr') if f.get('abr') is not None else -1,
797 audio_ext_preference,
798 f.get('fps') if f.get('fps') is not None else -1,
799 f.get('filesize_approx') if f.get('filesize_approx') is not None else -1,
800 f.get('source_preference') if f.get('source_preference') is not None else -1,
801 f.get('format_id') if f.get('format_id') is not None else '',
803 formats.sort(key=_formats_key)
805 def _check_formats(self, formats, video_id):
808 lambda f: self._is_valid_url(
810 item='%s video format' % f.get('format_id') if f.get('format_id') else 'video'),
813 def _is_valid_url(self, url, video_id, item='video'):
814 url = self._proto_relative_url(url, scheme='http:')
815 # For now assume non HTTP(S) URLs always valid
816 if not (url.startswith('http://') or url.startswith('https://')):
819 self._request_webpage(url, video_id, 'Checking %s URL' % item)
821 except ExtractorError as e:
822 if isinstance(e.cause, compat_HTTPError):
824 '%s: %s URL is invalid, skipping' % (video_id, item))
828 def http_scheme(self):
829 """ Either "http:" or "https:", depending on the user's preferences """
832 if self._downloader.params.get('prefer_insecure', False)
835 def _proto_relative_url(self, url, scheme=None):
838 if url.startswith('//'):
840 scheme = self.http_scheme()
845 def _sleep(self, timeout, video_id, msg_template=None):
846 if msg_template is None:
847 msg_template = '%(video_id)s: Waiting for %(timeout)s seconds'
848 msg = msg_template % {'video_id': video_id, 'timeout': timeout}
852 def _extract_f4m_formats(self, manifest_url, video_id, preference=None, f4m_id=None,
853 transform_source=lambda s: fix_xml_ampersands(s).strip()):
854 manifest = self._download_xml(
855 manifest_url, video_id, 'Downloading f4m manifest',
856 'Unable to download f4m manifest',
857 # Some manifests may be malformed, e.g. prosiebensat1 generated manifests
858 # (see https://github.com/rg3/youtube-dl/issues/6215#issuecomment-121704244)
859 transform_source=transform_source)
862 manifest_version = '1.0'
863 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/1.0}media')
865 manifest_version = '2.0'
866 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/2.0}media')
867 for i, media_el in enumerate(media_nodes):
868 if manifest_version == '2.0':
869 media_url = media_el.attrib.get('href') or media_el.attrib.get('url')
873 media_url if media_url.startswith('http://') or media_url.startswith('https://')
874 else ('/'.join(manifest_url.split('/')[:-1]) + '/' + media_url))
875 # If media_url is itself a f4m manifest do the recursive extraction
876 # since bitrates in parent manifest (this one) and media_url manifest
877 # may differ leading to inability to resolve the format by requested
878 # bitrate in f4m downloader
879 if determine_ext(manifest_url) == 'f4m':
880 formats.extend(self._extract_f4m_formats(manifest_url, video_id, preference, f4m_id))
882 tbr = int_or_none(media_el.attrib.get('bitrate'))
884 'format_id': '-'.join(filter(None, [f4m_id, compat_str(i if tbr is None else tbr)])),
888 'width': int_or_none(media_el.attrib.get('width')),
889 'height': int_or_none(media_el.attrib.get('height')),
890 'preference': preference,
892 self._sort_formats(formats)
896 def _extract_m3u8_formats(self, m3u8_url, video_id, ext=None,
897 entry_protocol='m3u8', preference=None,
898 m3u8_id=None, note=None, errnote=None,
902 'format_id': '-'.join(filter(None, [m3u8_id, 'meta'])),
906 'preference': preference - 1 if preference else -1,
907 'resolution': 'multiple',
908 'format_note': 'Quality selection URL',
911 format_url = lambda u: (
913 if re.match(r'^https?://', u)
914 else compat_urlparse.urljoin(m3u8_url, u))
916 m3u8_doc = self._download_webpage(
918 note=note or 'Downloading m3u8 information',
919 errnote=errnote or 'Failed to download m3u8 information',
921 if m3u8_doc is False:
926 r'(?P<key>[a-zA-Z_-]+)=(?P<val>"[^"]+"|[^",]+)(?:,|$)')
927 for line in m3u8_doc.splitlines():
928 if line.startswith('#EXT-X-STREAM-INF:'):
930 for m in kv_rex.finditer(line):
932 if v.startswith('"'):
934 last_info[m.group('key')] = v
935 elif line.startswith('#EXT-X-MEDIA:'):
937 for m in kv_rex.finditer(line):
939 if v.startswith('"'):
941 last_media[m.group('key')] = v
942 elif line.startswith('#') or not line.strip():
945 if last_info is None:
946 formats.append({'url': format_url(line)})
948 tbr = int_or_none(last_info.get('BANDWIDTH'), scale=1000)
951 format_id.append(m3u8_id)
952 last_media_name = last_media.get('NAME') if last_media and last_media.get('TYPE') != 'SUBTITLES' else None
953 format_id.append(last_media_name if last_media_name else '%d' % (tbr if tbr else len(formats)))
955 'format_id': '-'.join(format_id),
956 'url': format_url(line.strip()),
959 'protocol': entry_protocol,
960 'preference': preference,
962 codecs = last_info.get('CODECS')
964 # TODO: looks like video codec is not always necessarily goes first
965 va_codecs = codecs.split(',')
967 f['vcodec'] = va_codecs[0].partition('.')[0]
968 if len(va_codecs) > 1 and va_codecs[1]:
969 f['acodec'] = va_codecs[1].partition('.')[0]
970 resolution = last_info.get('RESOLUTION')
972 width_str, height_str = resolution.split('x')
973 f['width'] = int(width_str)
974 f['height'] = int(height_str)
975 if last_media is not None:
976 f['m3u8_media'] = last_media
980 self._sort_formats(formats)
984 def _xpath_ns(path, namespace=None):
988 for c in path.split('/'):
989 if not c or c == '.':
992 out.append('{%s}%s' % (namespace, c))
995 def _extract_smil_formats(self, smil_url, video_id, fatal=True, f4m_params=None):
996 smil = self._download_smil(smil_url, video_id, fatal=fatal)
1002 namespace = self._search_regex(
1003 r'{([^}]+)?}smil', smil.tag, 'namespace', default=None)
1005 return self._parse_smil_formats(
1006 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
1008 def _extract_smil_info(self, smil_url, video_id, fatal=True, f4m_params=None):
1009 smil = self._download_smil(smil_url, video_id, fatal=fatal)
1012 return self._parse_smil(smil, smil_url, video_id, f4m_params=f4m_params)
1014 def _download_smil(self, smil_url, video_id, fatal=True):
1015 return self._download_xml(
1016 smil_url, video_id, 'Downloading SMIL file',
1017 'Unable to download SMIL file', fatal=fatal)
1019 def _parse_smil(self, smil, smil_url, video_id, f4m_params=None):
1020 namespace = self._search_regex(
1021 r'{([^}]+)?}smil', smil.tag, 'namespace', default=None)
1023 formats = self._parse_smil_formats(
1024 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
1025 subtitles = self._parse_smil_subtitles(smil, namespace=namespace)
1027 video_id = os.path.splitext(url_basename(smil_url))[0]
1030 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
1031 name = meta.attrib.get('name')
1032 content = meta.attrib.get('content')
1033 if not name or not content:
1035 if not title and name == 'title':
1037 elif not description and name in ('description', 'abstract'):
1038 description = content
1042 'title': title or video_id,
1043 'description': description,
1045 'subtitles': subtitles,
1048 def _parse_smil_formats(self, smil, smil_url, video_id, namespace=None, f4m_params=None):
1050 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
1051 b = meta.get('base') or meta.get('httpBase')
1060 videos = smil.findall(self._xpath_ns('.//video', namespace))
1061 for video in videos:
1062 src = video.get('src')
1066 bitrate = int_or_none(video.get('system-bitrate') or video.get('systemBitrate'), 1000)
1067 filesize = int_or_none(video.get('size') or video.get('fileSize'))
1068 width = int_or_none(video.get('width'))
1069 height = int_or_none(video.get('height'))
1070 proto = video.get('proto')
1071 ext = video.get('ext')
1072 src_ext = determine_ext(src)
1073 streamer = video.get('streamer') or base
1075 if proto == 'rtmp' or streamer.startswith('rtmp'):
1081 'format_id': 'rtmp-%d' % (rtmp_count if bitrate is None else bitrate),
1083 'filesize': filesize,
1089 src_url = src if src.startswith('http') else compat_urlparse.urljoin(base, src)
1091 if proto == 'm3u8' or src_ext == 'm3u8':
1092 formats.extend(self._extract_m3u8_formats(
1093 src_url, video_id, ext or 'mp4', m3u8_id='hls'))
1096 if src_ext == 'f4m':
1101 'plugin': 'flowplayer-3.2.0.1',
1103 f4m_url += '&' if '?' in f4m_url else '?'
1104 f4m_url += compat_urllib_parse.urlencode(f4m_params).encode('utf-8')
1105 formats.extend(self._extract_f4m_formats(f4m_url, video_id, f4m_id='hds'))
1108 if src_url.startswith('http'):
1112 'ext': ext or src_ext or 'flv',
1113 'format_id': 'http-%d' % (bitrate or http_count),
1115 'filesize': filesize,
1121 self._sort_formats(formats)
1125 def _parse_smil_subtitles(self, smil, namespace=None):
1127 for num, textstream in enumerate(smil.findall(self._xpath_ns('.//textstream', namespace))):
1128 src = textstream.get('src')
1131 ext = textstream.get('ext') or determine_ext(src)
1133 type_ = textstream.get('type')
1134 if type_ == 'text/srt':
1136 lang = textstream.get('systemLanguage') or textstream.get('systemLanguageName')
1137 subtitles.setdefault(lang, []).append({
1143 def _live_title(self, name):
1144 """ Generate the title for a live video """
1145 now = datetime.datetime.now()
1146 now_str = now.strftime("%Y-%m-%d %H:%M")
1147 return name + ' ' + now_str
1149 def _int(self, v, name, fatal=False, **kwargs):
1150 res = int_or_none(v, **kwargs)
1151 if 'get_attr' in kwargs:
1152 print(getattr(v, kwargs['get_attr']))
1154 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
1156 raise ExtractorError(msg)
1158 self._downloader.report_warning(msg)
1161 def _float(self, v, name, fatal=False, **kwargs):
1162 res = float_or_none(v, **kwargs)
1164 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
1166 raise ExtractorError(msg)
1168 self._downloader.report_warning(msg)
1171 def _set_cookie(self, domain, name, value, expire_time=None):
1172 cookie = compat_cookiejar.Cookie(
1173 0, name, value, None, None, domain, None,
1174 None, '/', True, False, expire_time, '', None, None, None)
1175 self._downloader.cookiejar.set_cookie(cookie)
1177 def _get_cookies(self, url):
1178 """ Return a compat_cookies.SimpleCookie with the cookies for the url """
1179 req = compat_urllib_request.Request(url)
1180 self._downloader.cookiejar.add_cookie_header(req)
1181 return compat_cookies.SimpleCookie(req.get_header('Cookie'))
1183 def get_testcases(self, include_onlymatching=False):
1184 t = getattr(self, '_TEST', None)
1186 assert not hasattr(self, '_TESTS'), \
1187 '%s has _TEST and _TESTS' % type(self).__name__
1190 tests = getattr(self, '_TESTS', [])
1192 if not include_onlymatching and t.get('only_matching', False):
1194 t['name'] = type(self).__name__[:-len('IE')]
1197 def is_suitable(self, age_limit):
1198 """ Test whether the extractor is generally suitable for the given
1199 age limit (i.e. pornographic sites are not, all others usually are) """
1201 any_restricted = False
1202 for tc in self.get_testcases(include_onlymatching=False):
1203 if 'playlist' in tc:
1204 tc = tc['playlist'][0]
1205 is_restricted = age_restricted(
1206 tc.get('info_dict', {}).get('age_limit'), age_limit)
1207 if not is_restricted:
1209 any_restricted = any_restricted or is_restricted
1210 return not any_restricted
1212 def extract_subtitles(self, *args, **kwargs):
1213 if (self._downloader.params.get('writesubtitles', False) or
1214 self._downloader.params.get('listsubtitles')):
1215 return self._get_subtitles(*args, **kwargs)
1218 def _get_subtitles(self, *args, **kwargs):
1219 raise NotImplementedError("This method must be implemented by subclasses")
1221 def extract_automatic_captions(self, *args, **kwargs):
1222 if (self._downloader.params.get('writeautomaticsub', False) or
1223 self._downloader.params.get('listsubtitles')):
1224 return self._get_automatic_captions(*args, **kwargs)
1227 def _get_automatic_captions(self, *args, **kwargs):
1228 raise NotImplementedError("This method must be implemented by subclasses")
1231 class SearchInfoExtractor(InfoExtractor):
1233 Base class for paged search queries extractors.
1234 They accept URLs in the format _SEARCH_KEY(|all|[0-9]):{query}
1235 Instances should define _SEARCH_KEY and _MAX_RESULTS.
1239 def _make_valid_url(cls):
1240 return r'%s(?P<prefix>|[1-9][0-9]*|all):(?P<query>[\s\S]+)' % cls._SEARCH_KEY
1243 def suitable(cls, url):
1244 return re.match(cls._make_valid_url(), url) is not None
1246 def _real_extract(self, query):
1247 mobj = re.match(self._make_valid_url(), query)
1249 raise ExtractorError('Invalid search query "%s"' % query)
1251 prefix = mobj.group('prefix')
1252 query = mobj.group('query')
1254 return self._get_n_results(query, 1)
1255 elif prefix == 'all':
1256 return self._get_n_results(query, self._MAX_RESULTS)
1260 raise ExtractorError('invalid download number %s for query "%s"' % (n, query))
1261 elif n > self._MAX_RESULTS:
1262 self._downloader.report_warning('%s returns max %i results (you requested %i)' % (self._SEARCH_KEY, self._MAX_RESULTS, n))
1263 n = self._MAX_RESULTS
1264 return self._get_n_results(query, n)
1266 def _get_n_results(self, query, n):
1267 """Get a specified number of results for a query"""
1268 raise NotImplementedError("This method must be implemented by subclasses")
1271 def SEARCH_KEY(self):
1272 return self._SEARCH_KEY