1 from __future__ import unicode_literals
13 import xml.etree.ElementTree
15 from ..compat import (
20 compat_urllib_parse_urlparse,
36 _NO_DEFAULT = object()
39 class InfoExtractor(object):
40 """Information Extractor class.
42 Information extractors are the classes that, given a URL, extract
43 information about the video (or videos) the URL refers to. This
44 information includes the real video URL, the video title, author and
45 others. The information is stored in a dictionary which is then
46 passed to the YoutubeDL. The YoutubeDL processes this
47 information possibly downloading the video to the file system, among
48 other possible outcomes.
50 The type field determines the the type of the result.
51 By far the most common value (and the default if _type is missing) is
52 "video", which indicates a single video.
54 For a video, the dictionaries must include the following fields:
57 title: Video title, unescaped.
59 Additionally, it must contain either a formats entry or a url one:
61 formats: A list of dictionaries for each format available, ordered
62 from worst to best quality.
65 * url Mandatory. The URL of the video file
66 * ext Will be calculated from url if missing
67 * format A human-readable description of the format
68 ("mp4 container with h264/opus").
69 Calculated from the format_id, width, height.
70 and format_note fields if missing.
71 * format_id A short description of the format
72 ("mp4_h264_opus" or "19").
73 Technically optional, but strongly recommended.
74 * format_note Additional info about the format
75 ("3D" or "DASH video")
76 * width Width of the video, if known
77 * height Height of the video, if known
78 * resolution Textual description of width and height
79 * tbr Average bitrate of audio and video in KBit/s
80 * abr Average audio bitrate in KBit/s
81 * acodec Name of the audio codec in use
82 * asr Audio sampling rate in Hertz
83 * vbr Average video bitrate in KBit/s
85 * vcodec Name of the video codec in use
86 * container Name of the container format
87 * filesize The number of bytes, if known in advance
88 * filesize_approx An estimate for the number of bytes
89 * player_url SWF Player URL (used for rtmpdump).
90 * protocol The protocol that will be used for the actual
92 "http", "https", "rtsp", "rtmp", "rtmpe",
93 "m3u8", or "m3u8_native".
94 * preference Order number of this format. If this field is
95 present and not None, the formats get sorted
96 by this field, regardless of all other values.
97 -1 for default (order by other properties),
98 -2 or smaller for less than default.
99 < -1000 to hide the format (if there is
100 another one which is strictly better)
101 * language_preference Is this in the correct requested
103 10 if it's what the URL is about,
104 -1 for default (don't know),
105 -10 otherwise, other values reserved for now.
106 * quality Order number of the video quality of this
107 format, irrespective of the file format.
108 -1 for default (order by other properties),
109 -2 or smaller for less than default.
110 * source_preference Order number for this video source
111 (quality takes higher priority)
112 -1 for default (order by other properties),
113 -2 or smaller for less than default.
114 * http_method HTTP method to use for the download.
115 * http_headers A dictionary of additional HTTP headers
116 to add to the request.
117 * http_post_data Additional data to send with a POST
119 * stretched_ratio If given and not 1, indicates that the
120 video's pixels are not square.
121 width : height ratio as float.
122 * no_resume The server does not support resuming the
123 (HTTP or RTMP) download. Boolean.
125 url: Final video URL.
126 ext: Video filename extension.
127 format: The video format, defaults to ext (used for --get-format)
128 player_url: SWF Player URL (used for rtmpdump).
130 The following fields are optional:
132 alt_title: A secondary title of the video.
133 display_id An alternative identifier for the video, not necessarily
134 unique, but available before title. Typically, id is
135 something like "4234987", title "Dancing naked mole rats",
136 and display_id "dancing-naked-mole-rats"
137 thumbnails: A list of dictionaries, with the following entries:
138 * "id" (optional, string) - Thumbnail format ID
140 * "preference" (optional, int) - quality of the image
141 * "width" (optional, int)
142 * "height" (optional, int)
143 * "resolution" (optional, string "{width}x{height"},
145 thumbnail: Full URL to a video thumbnail image.
146 description: Full video description.
147 uploader: Full name of the video uploader.
148 creator: The main artist who created the video.
149 timestamp: UNIX timestamp of the moment the video became available.
150 upload_date: Video upload date (YYYYMMDD).
151 If not explicitly set, calculated from timestamp.
152 uploader_id: Nickname or id of the video uploader.
153 location: Physical location where the video was filmed.
154 subtitles: The available subtitles as a dictionary in the format
155 {language: subformats}. "subformats" is a list sorted from
156 lower to higher preference, each element is a dictionary
157 with the "ext" entry and one of:
158 * "data": The subtitles file contents
159 * "url": A url pointing to the subtitles file
160 automatic_captions: Like 'subtitles', used by the YoutubeIE for
161 automatically generated captions
162 duration: Length of the video in seconds, as an integer.
163 view_count: How many users have watched the video on the platform.
164 like_count: Number of positive ratings of the video
165 dislike_count: Number of negative ratings of the video
166 average_rating: Average rating give by users, the scale used depends on the webpage
167 comment_count: Number of comments on the video
168 comments: A list of comments, each with one or more of the following
169 properties (all but one of text or html optional):
170 * "author" - human-readable name of the comment author
171 * "author_id" - user ID of the comment author
173 * "html" - Comment as HTML
174 * "text" - Plain text of the comment
175 * "timestamp" - UNIX timestamp of comment
176 * "parent" - ID of the comment this one is replying to.
177 Set to "root" to indicate that this is a
178 comment to the original video.
179 age_limit: Age restriction for the video, as an integer (years)
180 webpage_url: The url to the video webpage, if given to youtube-dl it
181 should allow to get the same result again. (It will be set
182 by YoutubeDL if it's missing)
183 categories: A list of categories that the video falls in, for example
185 is_live: True, False, or None (=unknown). Whether this video is a
186 live stream that goes on instead of a fixed-length video.
188 Unless mentioned otherwise, the fields should be Unicode strings.
190 Unless mentioned otherwise, None is equivalent to absence of information.
193 _type "playlist" indicates multiple videos.
194 There must be a key "entries", which is a list, an iterable, or a PagedList
195 object, each element of which is a valid dictionary by this specification.
197 Additionally, playlists can have "title" and "id" attributes with the same
198 semantics as videos (see above).
201 _type "multi_video" indicates that there are multiple videos that
202 form a single show, for examples multiple acts of an opera or TV episode.
203 It must have an entries key like a playlist and contain all the keys
204 required for a video at the same time.
207 _type "url" indicates that the video must be extracted from another
208 location, possibly by a different extractor. Its only required key is:
209 "url" - the next URL to extract.
210 The key "ie_key" can be set to the class name (minus the trailing "IE",
211 e.g. "Youtube") if the extractor class is known in advance.
212 Additionally, the dictionary may have any properties of the resolved entity
213 known in advance, for example "title" if the title of the referred video is
217 _type "url_transparent" entities have the same specification as "url", but
218 indicate that the given additional information is more precise than the one
219 associated with the resolved URL.
220 This is useful when a site employs a video service that hosts the video and
221 its technical metadata, but that video service does not embed a useful
222 title, description etc.
225 Subclasses of this one should re-define the _real_initialize() and
226 _real_extract() methods and define a _VALID_URL regexp.
227 Probably, they should also be added to the list of extractors.
229 Finally, the _WORKING attribute should be set to False for broken IEs
230 in order to warn the users and skip the tests.
237 def __init__(self, downloader=None):
238 """Constructor. Receives an optional downloader."""
240 self.set_downloader(downloader)
243 def suitable(cls, url):
244 """Receives a URL and returns True if suitable for this IE."""
246 # This does not use has/getattr intentionally - we want to know whether
247 # we have cached the regexp for *this* class, whereas getattr would also
248 # match the superclass
249 if '_VALID_URL_RE' not in cls.__dict__:
250 cls._VALID_URL_RE = re.compile(cls._VALID_URL)
251 return cls._VALID_URL_RE.match(url) is not None
254 def _match_id(cls, url):
255 if '_VALID_URL_RE' not in cls.__dict__:
256 cls._VALID_URL_RE = re.compile(cls._VALID_URL)
257 m = cls._VALID_URL_RE.match(url)
263 """Getter method for _WORKING."""
266 def initialize(self):
267 """Initializes an instance (authentication, etc)."""
269 self._real_initialize()
272 def extract(self, url):
273 """Extracts URL information and returns it in list of dicts."""
276 return self._real_extract(url)
277 except ExtractorError:
279 except compat_http_client.IncompleteRead as e:
280 raise ExtractorError('A network error has occured.', cause=e, expected=True)
281 except (KeyError, StopIteration) as e:
282 raise ExtractorError('An extractor error has occured.', cause=e)
284 def set_downloader(self, downloader):
285 """Sets the downloader for this IE."""
286 self._downloader = downloader
288 def _real_initialize(self):
289 """Real initialization process. Redefine in subclasses."""
292 def _real_extract(self, url):
293 """Real extraction process. Redefine in subclasses."""
298 """A string for getting the InfoExtractor with get_info_extractor"""
299 return cls.__name__[:-2]
303 return type(self).__name__[:-2]
305 def _request_webpage(self, url_or_request, video_id, note=None, errnote=None, fatal=True):
306 """ Returns the response handle """
308 self.report_download_webpage(video_id)
309 elif note is not False:
311 self.to_screen('%s' % (note,))
313 self.to_screen('%s: %s' % (video_id, note))
315 return self._downloader.urlopen(url_or_request)
316 except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err:
320 errnote = 'Unable to download webpage'
321 errmsg = '%s: %s' % (errnote, compat_str(err))
323 raise ExtractorError(errmsg, sys.exc_info()[2], cause=err)
325 self._downloader.report_warning(errmsg)
328 def _download_webpage_handle(self, url_or_request, video_id, note=None, errnote=None, fatal=True):
329 """ Returns a tuple (page content as string, URL handle) """
330 # Strip hashes from the URL (#1038)
331 if isinstance(url_or_request, (compat_str, str)):
332 url_or_request = url_or_request.partition('#')[0]
334 urlh = self._request_webpage(url_or_request, video_id, note, errnote, fatal)
338 content = self._webpage_read_content(urlh, url_or_request, video_id, note, errnote, fatal)
339 return (content, urlh)
341 def _webpage_read_content(self, urlh, url_or_request, video_id, note=None, errnote=None, fatal=True, prefix=None):
342 content_type = urlh.headers.get('Content-Type', '')
343 webpage_bytes = urlh.read()
344 if prefix is not None:
345 webpage_bytes = prefix + webpage_bytes
346 m = re.match(r'[a-zA-Z0-9_.-]+/[a-zA-Z0-9_.-]+\s*;\s*charset=(.+)', content_type)
348 encoding = m.group(1)
350 m = re.search(br'<meta[^>]+charset=[\'"]?([^\'")]+)[ /\'">]',
351 webpage_bytes[:1024])
353 encoding = m.group(1).decode('ascii')
354 elif webpage_bytes.startswith(b'\xff\xfe'):
358 if self._downloader.params.get('dump_intermediate_pages', False):
360 url = url_or_request.get_full_url()
361 except AttributeError:
363 self.to_screen('Dumping request to ' + url)
364 dump = base64.b64encode(webpage_bytes).decode('ascii')
365 self._downloader.to_screen(dump)
366 if self._downloader.params.get('write_pages', False):
368 url = url_or_request.get_full_url()
369 except AttributeError:
371 basen = '%s_%s' % (video_id, url)
373 h = '___' + hashlib.md5(basen.encode('utf-8')).hexdigest()
374 basen = basen[:240 - len(h)] + h
375 raw_filename = basen + '.dump'
376 filename = sanitize_filename(raw_filename, restricted=True)
377 self.to_screen('Saving request to ' + filename)
378 # Working around MAX_PATH limitation on Windows (see
379 # http://msdn.microsoft.com/en-us/library/windows/desktop/aa365247(v=vs.85).aspx)
381 absfilepath = os.path.abspath(filename)
382 if len(absfilepath) > 259:
383 filename = '\\\\?\\' + absfilepath
384 with open(filename, 'wb') as outf:
385 outf.write(webpage_bytes)
388 content = webpage_bytes.decode(encoding, 'replace')
390 content = webpage_bytes.decode('utf-8', 'replace')
392 if ('<title>Access to this site is blocked</title>' in content and
393 'Websense' in content[:512]):
394 msg = 'Access to this webpage has been blocked by Websense filtering software in your network.'
395 blocked_iframe = self._html_search_regex(
396 r'<iframe src="([^"]+)"', content,
397 'Websense information URL', default=None)
399 msg += ' Visit %s for more details' % blocked_iframe
400 raise ExtractorError(msg, expected=True)
404 def _download_webpage(self, url_or_request, video_id, note=None, errnote=None, fatal=True, tries=1, timeout=5):
405 """ Returns the data of the page as a string """
408 while success is False:
410 res = self._download_webpage_handle(url_or_request, video_id, note, errnote, fatal)
412 except compat_http_client.IncompleteRead as e:
414 if try_count >= tries:
416 self._sleep(timeout, video_id)
423 def _download_xml(self, url_or_request, video_id,
424 note='Downloading XML', errnote='Unable to download XML',
425 transform_source=None, fatal=True):
426 """Return the xml as an xml.etree.ElementTree.Element"""
427 xml_string = self._download_webpage(
428 url_or_request, video_id, note, errnote, fatal=fatal)
429 if xml_string is False:
432 xml_string = transform_source(xml_string)
433 return xml.etree.ElementTree.fromstring(xml_string.encode('utf-8'))
435 def _download_json(self, url_or_request, video_id,
436 note='Downloading JSON metadata',
437 errnote='Unable to download JSON metadata',
438 transform_source=None,
440 json_string = self._download_webpage(
441 url_or_request, video_id, note, errnote, fatal=fatal)
442 if (not fatal) and json_string is False:
444 return self._parse_json(
445 json_string, video_id, transform_source=transform_source, fatal=fatal)
447 def _parse_json(self, json_string, video_id, transform_source=None, fatal=True):
449 json_string = transform_source(json_string)
451 return json.loads(json_string)
452 except ValueError as ve:
453 errmsg = '%s: Failed to parse JSON ' % video_id
455 raise ExtractorError(errmsg, cause=ve)
457 self.report_warning(errmsg + str(ve))
459 def report_warning(self, msg, video_id=None):
460 idstr = '' if video_id is None else '%s: ' % video_id
461 self._downloader.report_warning(
462 '[%s] %s%s' % (self.IE_NAME, idstr, msg))
464 def to_screen(self, msg):
465 """Print msg to screen, prefixing it with '[ie_name]'"""
466 self._downloader.to_screen('[%s] %s' % (self.IE_NAME, msg))
468 def report_extraction(self, id_or_name):
469 """Report information extraction."""
470 self.to_screen('%s: Extracting information' % id_or_name)
472 def report_download_webpage(self, video_id):
473 """Report webpage download."""
474 self.to_screen('%s: Downloading webpage' % video_id)
476 def report_age_confirmation(self):
477 """Report attempt to confirm age."""
478 self.to_screen('Confirming age')
480 def report_login(self):
481 """Report attempt to log in."""
482 self.to_screen('Logging in')
484 # Methods for following #608
486 def url_result(url, ie=None, video_id=None):
487 """Returns a url that points to a page that should be processed"""
488 # TODO: ie should be the class used for getting the info
489 video_info = {'_type': 'url',
492 if video_id is not None:
493 video_info['id'] = video_id
497 def playlist_result(entries, playlist_id=None, playlist_title=None, playlist_description=None):
498 """Returns a playlist"""
499 video_info = {'_type': 'playlist',
502 video_info['id'] = playlist_id
504 video_info['title'] = playlist_title
505 if playlist_description:
506 video_info['description'] = playlist_description
509 def _search_regex(self, pattern, string, name, default=_NO_DEFAULT, fatal=True, flags=0, group=None):
511 Perform a regex search on the given string, using a single or a list of
512 patterns returning the first matching group.
513 In case of failure return a default value or raise a WARNING or a
514 RegexNotFoundError, depending on fatal, specifying the field name.
516 if isinstance(pattern, (str, compat_str, compiled_regex_type)):
517 mobj = re.search(pattern, string, flags)
520 mobj = re.search(p, string, flags)
524 if not self._downloader.params.get('no_color') and os.name != 'nt' and sys.stderr.isatty():
525 _name = '\033[0;34m%s\033[0m' % name
531 # return the first matching group
532 return next(g for g in mobj.groups() if g is not None)
534 return mobj.group(group)
535 elif default is not _NO_DEFAULT:
538 raise RegexNotFoundError('Unable to extract %s' % _name)
540 self._downloader.report_warning('unable to extract %s; '
541 'please report this issue on http://yt-dl.org/bug' % _name)
544 def _html_search_regex(self, pattern, string, name, default=_NO_DEFAULT, fatal=True, flags=0, group=None):
546 Like _search_regex, but strips HTML tags and unescapes entities.
548 res = self._search_regex(pattern, string, name, default, fatal, flags, group)
550 return clean_html(res).strip()
554 def _get_login_info(self):
556 Get the the login info as (username, password)
557 It will look in the netrc file using the _NETRC_MACHINE value
558 If there's no info available, return (None, None)
560 if self._downloader is None:
565 downloader_params = self._downloader.params
567 # Attempt to use provided username and password or .netrc data
568 if downloader_params.get('username', None) is not None:
569 username = downloader_params['username']
570 password = downloader_params['password']
571 elif downloader_params.get('usenetrc', False):
573 info = netrc.netrc().authenticators(self._NETRC_MACHINE)
578 raise netrc.NetrcParseError('No authenticators for %s' % self._NETRC_MACHINE)
579 except (IOError, netrc.NetrcParseError) as err:
580 self._downloader.report_warning('parsing .netrc: %s' % compat_str(err))
582 return (username, password)
584 def _get_tfa_info(self):
586 Get the two-factor authentication info
587 TODO - asking the user will be required for sms/phone verify
588 currently just uses the command line option
589 If there's no info available, return None
591 if self._downloader is None:
593 downloader_params = self._downloader.params
595 if downloader_params.get('twofactor', None) is not None:
596 return downloader_params['twofactor']
600 # Helper functions for extracting OpenGraph info
602 def _og_regexes(prop):
603 content_re = r'content=(?:"([^>]+?)"|\'([^>]+?)\')'
604 property_re = r'(?:name|property)=[\'"]og:%s[\'"]' % re.escape(prop)
605 template = r'<meta[^>]+?%s[^>]+?%s'
607 template % (property_re, content_re),
608 template % (content_re, property_re),
611 def _og_search_property(self, prop, html, name=None, **kargs):
613 name = 'OpenGraph %s' % prop
614 escaped = self._search_regex(self._og_regexes(prop), html, name, flags=re.DOTALL, **kargs)
617 return unescapeHTML(escaped)
619 def _og_search_thumbnail(self, html, **kargs):
620 return self._og_search_property('image', html, 'thumbnail url', fatal=False, **kargs)
622 def _og_search_description(self, html, **kargs):
623 return self._og_search_property('description', html, fatal=False, **kargs)
625 def _og_search_title(self, html, **kargs):
626 return self._og_search_property('title', html, **kargs)
628 def _og_search_video_url(self, html, name='video url', secure=True, **kargs):
629 regexes = self._og_regexes('video') + self._og_regexes('video:url')
631 regexes = self._og_regexes('video:secure_url') + regexes
632 return self._html_search_regex(regexes, html, name, **kargs)
634 def _og_search_url(self, html, **kargs):
635 return self._og_search_property('url', html, **kargs)
637 def _html_search_meta(self, name, html, display_name=None, fatal=False, **kwargs):
638 if display_name is None:
640 return self._html_search_regex(
642 (?=[^>]+(?:itemprop|name|property)=(["\']?)%s\1)
643 [^>]+?content=(["\'])(?P<content>.*?)\2''' % re.escape(name),
644 html, display_name, fatal=fatal, group='content', **kwargs)
646 def _dc_search_uploader(self, html):
647 return self._html_search_meta('dc.creator', html, 'uploader')
649 def _rta_search(self, html):
650 # See http://www.rtalabel.org/index.php?content=howtofaq#single
651 if re.search(r'(?ix)<meta\s+name="rating"\s+'
652 r' content="RTA-5042-1996-1400-1577-RTA"',
657 def _media_rating_search(self, html):
658 # See http://www.tjg-designs.com/WP/metadata-code-examples-adding-metadata-to-your-web-pages/
659 rating = self._html_search_meta('rating', html)
671 return RATING_TABLE.get(rating.lower(), None)
673 def _family_friendly_search(self, html):
674 # See http://schema.org/VideoObject
675 family_friendly = self._html_search_meta('isFamilyFriendly', html)
677 if not family_friendly:
686 return RATING_TABLE.get(family_friendly.lower(), None)
688 def _twitter_search_player(self, html):
689 return self._html_search_meta('twitter:player', html,
690 'twitter card player')
692 def _sort_formats(self, formats):
694 raise ExtractorError('No video formats found')
697 # TODO remove the following workaround
698 from ..utils import determine_ext
699 if not f.get('ext') and 'url' in f:
700 f['ext'] = determine_ext(f['url'])
702 preference = f.get('preference')
703 if preference is None:
704 proto = f.get('protocol')
706 proto = compat_urllib_parse_urlparse(f.get('url', '')).scheme
708 preference = 0 if proto in ['http', 'https'] else -0.1
709 if f.get('ext') in ['f4f', 'f4m']: # Not yet supported
712 if f.get('vcodec') == 'none': # audio only
713 if self._downloader.params.get('prefer_free_formats'):
714 ORDER = ['aac', 'mp3', 'm4a', 'webm', 'ogg', 'opus']
716 ORDER = ['webm', 'opus', 'ogg', 'mp3', 'aac', 'm4a']
719 audio_ext_preference = ORDER.index(f['ext'])
721 audio_ext_preference = -1
723 if self._downloader.params.get('prefer_free_formats'):
724 ORDER = ['flv', 'mp4', 'webm']
726 ORDER = ['webm', 'flv', 'mp4']
728 ext_preference = ORDER.index(f['ext'])
731 audio_ext_preference = 0
735 f.get('language_preference') if f.get('language_preference') is not None else -1,
736 f.get('quality') if f.get('quality') is not None else -1,
737 f.get('tbr') if f.get('tbr') is not None else -1,
738 f.get('vbr') if f.get('vbr') is not None else -1,
739 f.get('height') if f.get('height') is not None else -1,
740 f.get('width') if f.get('width') is not None else -1,
742 f.get('abr') if f.get('abr') is not None else -1,
743 audio_ext_preference,
744 f.get('fps') if f.get('fps') is not None else -1,
745 f.get('filesize') if f.get('filesize') is not None else -1,
746 f.get('filesize_approx') if f.get('filesize_approx') is not None else -1,
747 f.get('source_preference') if f.get('source_preference') is not None else -1,
750 formats.sort(key=_formats_key)
752 def _check_formats(self, formats, video_id):
755 lambda f: self._is_valid_url(
757 item='%s video format' % f.get('format_id') if f.get('format_id') else 'video'),
760 def _is_valid_url(self, url, video_id, item='video'):
762 self._request_webpage(
763 HEADRequest(url), video_id,
764 'Checking %s URL' % item)
766 except ExtractorError as e:
767 if isinstance(e.cause, compat_HTTPError):
769 '%s URL is invalid, skipping' % item, video_id)
773 def http_scheme(self):
774 """ Either "http:" or "https:", depending on the user's preferences """
777 if self._downloader.params.get('prefer_insecure', False)
780 def _proto_relative_url(self, url, scheme=None):
783 if url.startswith('//'):
785 scheme = self.http_scheme()
790 def _sleep(self, timeout, video_id, msg_template=None):
791 if msg_template is None:
792 msg_template = '%(video_id)s: Waiting for %(timeout)s seconds'
793 msg = msg_template % {'video_id': video_id, 'timeout': timeout}
797 def _extract_f4m_formats(self, manifest_url, video_id, preference=None, f4m_id=None):
798 manifest = self._download_xml(
799 manifest_url, video_id, 'Downloading f4m manifest',
800 'Unable to download f4m manifest')
803 manifest_version = '1.0'
804 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/1.0}media')
806 manifest_version = '2.0'
807 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/2.0}media')
808 for i, media_el in enumerate(media_nodes):
809 if manifest_version == '2.0':
810 manifest_url = ('/'.join(manifest_url.split('/')[:-1]) + '/'
811 + (media_el.attrib.get('href') or media_el.attrib.get('url')))
812 tbr = int_or_none(media_el.attrib.get('bitrate'))
814 'format_id': '-'.join(filter(None, [f4m_id, 'f4m-%d' % (i if tbr is None else tbr)])),
818 'width': int_or_none(media_el.attrib.get('width')),
819 'height': int_or_none(media_el.attrib.get('height')),
820 'preference': preference,
822 self._sort_formats(formats)
826 def _extract_m3u8_formats(self, m3u8_url, video_id, ext=None,
827 entry_protocol='m3u8', preference=None,
831 'format_id': '-'.join(filter(None, [m3u8_id, 'm3u8-meta'])),
836 'resolution': 'multiple',
837 'format_note': 'Quality selection URL',
840 format_url = lambda u: (
842 if re.match(r'^https?://', u)
843 else compat_urlparse.urljoin(m3u8_url, u))
845 m3u8_doc = self._download_webpage(
847 note='Downloading m3u8 information',
848 errnote='Failed to download m3u8 information')
851 r'(?P<key>[a-zA-Z_-]+)=(?P<val>"[^"]+"|[^",]+)(?:,|$)')
852 for line in m3u8_doc.splitlines():
853 if line.startswith('#EXT-X-STREAM-INF:'):
855 for m in kv_rex.finditer(line):
857 if v.startswith('"'):
859 last_info[m.group('key')] = v
860 elif line.startswith('#') or not line.strip():
863 if last_info is None:
864 formats.append({'url': format_url(line)})
866 tbr = int_or_none(last_info.get('BANDWIDTH'), scale=1000)
868 'format_id': '-'.join(filter(None, [m3u8_id, 'm3u8-%d' % (tbr if tbr else len(formats))])),
869 'url': format_url(line.strip()),
872 'protocol': entry_protocol,
873 'preference': preference,
875 codecs = last_info.get('CODECS')
877 # TODO: looks like video codec is not always necessarily goes first
878 va_codecs = codecs.split(',')
880 f['vcodec'] = va_codecs[0].partition('.')[0]
881 if len(va_codecs) > 1 and va_codecs[1]:
882 f['acodec'] = va_codecs[1].partition('.')[0]
883 resolution = last_info.get('RESOLUTION')
885 width_str, height_str = resolution.split('x')
886 f['width'] = int(width_str)
887 f['height'] = int(height_str)
890 self._sort_formats(formats)
893 # TODO: improve extraction
894 def _extract_smil_formats(self, smil_url, video_id, fatal=True):
895 smil = self._download_xml(
896 smil_url, video_id, 'Downloading SMIL file',
897 'Unable to download SMIL file', fatal=fatal)
902 base = smil.find('./head/meta').get('base')
906 for video in smil.findall('./body/switch/video'):
907 src = video.get('src')
910 bitrate = int_or_none(video.get('system-bitrate') or video.get('systemBitrate'), 1000)
911 width = int_or_none(video.get('width'))
912 height = int_or_none(video.get('height'))
913 proto = video.get('proto')
916 if base.startswith('rtmp'):
918 elif base.startswith('http'):
920 ext = video.get('ext')
922 formats.extend(self._extract_m3u8_formats(src, video_id, ext))
923 elif proto == 'rtmp':
925 streamer = video.get('streamer') or base
930 'format_id': 'rtmp-%d' % (rtmp_count if bitrate is None else bitrate),
935 self._sort_formats(formats)
939 def _live_title(self, name):
940 """ Generate the title for a live video """
941 now = datetime.datetime.now()
942 now_str = now.strftime("%Y-%m-%d %H:%M")
943 return name + ' ' + now_str
945 def _int(self, v, name, fatal=False, **kwargs):
946 res = int_or_none(v, **kwargs)
947 if 'get_attr' in kwargs:
948 print(getattr(v, kwargs['get_attr']))
950 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
952 raise ExtractorError(msg)
954 self._downloader.report_warning(msg)
957 def _float(self, v, name, fatal=False, **kwargs):
958 res = float_or_none(v, **kwargs)
960 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
962 raise ExtractorError(msg)
964 self._downloader.report_warning(msg)
967 def _set_cookie(self, domain, name, value, expire_time=None):
968 cookie = compat_cookiejar.Cookie(
969 0, name, value, None, None, domain, None,
970 None, '/', True, False, expire_time, '', None, None, None)
971 self._downloader.cookiejar.set_cookie(cookie)
973 def get_testcases(self, include_onlymatching=False):
974 t = getattr(self, '_TEST', None)
976 assert not hasattr(self, '_TESTS'), \
977 '%s has _TEST and _TESTS' % type(self).__name__
980 tests = getattr(self, '_TESTS', [])
982 if not include_onlymatching and t.get('only_matching', False):
984 t['name'] = type(self).__name__[:-len('IE')]
987 def is_suitable(self, age_limit):
988 """ Test whether the extractor is generally suitable for the given
989 age limit (i.e. pornographic sites are not, all others usually are) """
991 any_restricted = False
992 for tc in self.get_testcases(include_onlymatching=False):
994 tc = tc['playlist'][0]
995 is_restricted = age_restricted(
996 tc.get('info_dict', {}).get('age_limit'), age_limit)
997 if not is_restricted:
999 any_restricted = any_restricted or is_restricted
1000 return not any_restricted
1002 def extract_subtitles(self, *args, **kwargs):
1003 if (self._downloader.params.get('writesubtitles', False) or
1004 self._downloader.params.get('listsubtitles')):
1005 return self._get_subtitles(*args, **kwargs)
1008 def _get_subtitles(self, *args, **kwargs):
1009 raise NotImplementedError("This method must be implemented by subclasses")
1011 def extract_automatic_captions(self, *args, **kwargs):
1012 if (self._downloader.params.get('writeautomaticsub', False) or
1013 self._downloader.params.get('listsubtitles')):
1014 return self._get_automatic_captions(*args, **kwargs)
1017 def _get_automatic_captions(self, *args, **kwargs):
1018 raise NotImplementedError("This method must be implemented by subclasses")
1021 class SearchInfoExtractor(InfoExtractor):
1023 Base class for paged search queries extractors.
1024 They accept urls in the format _SEARCH_KEY(|all|[0-9]):{query}
1025 Instances should define _SEARCH_KEY and _MAX_RESULTS.
1029 def _make_valid_url(cls):
1030 return r'%s(?P<prefix>|[1-9][0-9]*|all):(?P<query>[\s\S]+)' % cls._SEARCH_KEY
1033 def suitable(cls, url):
1034 return re.match(cls._make_valid_url(), url) is not None
1036 def _real_extract(self, query):
1037 mobj = re.match(self._make_valid_url(), query)
1039 raise ExtractorError('Invalid search query "%s"' % query)
1041 prefix = mobj.group('prefix')
1042 query = mobj.group('query')
1044 return self._get_n_results(query, 1)
1045 elif prefix == 'all':
1046 return self._get_n_results(query, self._MAX_RESULTS)
1050 raise ExtractorError('invalid download number %s for query "%s"' % (n, query))
1051 elif n > self._MAX_RESULTS:
1052 self._downloader.report_warning('%s returns max %i results (you requested %i)' % (self._SEARCH_KEY, self._MAX_RESULTS, n))
1053 n = self._MAX_RESULTS
1054 return self._get_n_results(query, n)
1056 def _get_n_results(self, query, n):
1057 """Get a specified number of results for a query"""
1058 raise NotImplementedError("This method must be implemented by subclasses")
1061 def SEARCH_KEY(self):
1062 return self._SEARCH_KEY