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 duration: Length of the video in seconds, as an integer.
161 view_count: How many users have watched the video on the platform.
162 like_count: Number of positive ratings of the video
163 dislike_count: Number of negative ratings of the video
164 average_rating: Average rating give by users, the scale used depends on the webpage
165 comment_count: Number of comments on the video
166 comments: A list of comments, each with one or more of the following
167 properties (all but one of text or html optional):
168 * "author" - human-readable name of the comment author
169 * "author_id" - user ID of the comment author
171 * "html" - Comment as HTML
172 * "text" - Plain text of the comment
173 * "timestamp" - UNIX timestamp of comment
174 * "parent" - ID of the comment this one is replying to.
175 Set to "root" to indicate that this is a
176 comment to the original video.
177 age_limit: Age restriction for the video, as an integer (years)
178 webpage_url: The url to the video webpage, if given to youtube-dl it
179 should allow to get the same result again. (It will be set
180 by YoutubeDL if it's missing)
181 categories: A list of categories that the video falls in, for example
183 is_live: True, False, or None (=unknown). Whether this video is a
184 live stream that goes on instead of a fixed-length video.
186 Unless mentioned otherwise, the fields should be Unicode strings.
188 Unless mentioned otherwise, None is equivalent to absence of information.
191 _type "playlist" indicates multiple videos.
192 There must be a key "entries", which is a list, an iterable, or a PagedList
193 object, each element of which is a valid dictionary by this specification.
195 Additionally, playlists can have "title" and "id" attributes with the same
196 semantics as videos (see above).
199 _type "multi_video" indicates that there are multiple videos that
200 form a single show, for examples multiple acts of an opera or TV episode.
201 It must have an entries key like a playlist and contain all the keys
202 required for a video at the same time.
205 _type "url" indicates that the video must be extracted from another
206 location, possibly by a different extractor. Its only required key is:
207 "url" - the next URL to extract.
208 The key "ie_key" can be set to the class name (minus the trailing "IE",
209 e.g. "Youtube") if the extractor class is known in advance.
210 Additionally, the dictionary may have any properties of the resolved entity
211 known in advance, for example "title" if the title of the referred video is
215 _type "url_transparent" entities have the same specification as "url", but
216 indicate that the given additional information is more precise than the one
217 associated with the resolved URL.
218 This is useful when a site employs a video service that hosts the video and
219 its technical metadata, but that video service does not embed a useful
220 title, description etc.
223 Subclasses of this one should re-define the _real_initialize() and
224 _real_extract() methods and define a _VALID_URL regexp.
225 Probably, they should also be added to the list of extractors.
227 Finally, the _WORKING attribute should be set to False for broken IEs
228 in order to warn the users and skip the tests.
235 def __init__(self, downloader=None):
236 """Constructor. Receives an optional downloader."""
238 self.set_downloader(downloader)
241 def suitable(cls, url):
242 """Receives a URL and returns True if suitable for this IE."""
244 # This does not use has/getattr intentionally - we want to know whether
245 # we have cached the regexp for *this* class, whereas getattr would also
246 # match the superclass
247 if '_VALID_URL_RE' not in cls.__dict__:
248 cls._VALID_URL_RE = re.compile(cls._VALID_URL)
249 return cls._VALID_URL_RE.match(url) is not None
252 def _match_id(cls, url):
253 if '_VALID_URL_RE' not in cls.__dict__:
254 cls._VALID_URL_RE = re.compile(cls._VALID_URL)
255 m = cls._VALID_URL_RE.match(url)
261 """Getter method for _WORKING."""
264 def initialize(self):
265 """Initializes an instance (authentication, etc)."""
267 self._real_initialize()
270 def extract(self, url):
271 """Extracts URL information and returns it in list of dicts."""
274 return self._real_extract(url)
275 except ExtractorError:
277 except compat_http_client.IncompleteRead as e:
278 raise ExtractorError('A network error has occured.', cause=e, expected=True)
279 except (KeyError, StopIteration) as e:
280 raise ExtractorError('An extractor error has occured.', cause=e)
282 def set_downloader(self, downloader):
283 """Sets the downloader for this IE."""
284 self._downloader = downloader
286 def _real_initialize(self):
287 """Real initialization process. Redefine in subclasses."""
290 def _real_extract(self, url):
291 """Real extraction process. Redefine in subclasses."""
296 """A string for getting the InfoExtractor with get_info_extractor"""
297 return cls.__name__[:-2]
301 return type(self).__name__[:-2]
303 def _request_webpage(self, url_or_request, video_id, note=None, errnote=None, fatal=True):
304 """ Returns the response handle """
306 self.report_download_webpage(video_id)
307 elif note is not False:
309 self.to_screen('%s' % (note,))
311 self.to_screen('%s: %s' % (video_id, note))
313 return self._downloader.urlopen(url_or_request)
314 except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err:
318 errnote = 'Unable to download webpage'
319 errmsg = '%s: %s' % (errnote, compat_str(err))
321 raise ExtractorError(errmsg, sys.exc_info()[2], cause=err)
323 self._downloader.report_warning(errmsg)
326 def _download_webpage_handle(self, url_or_request, video_id, note=None, errnote=None, fatal=True):
327 """ Returns a tuple (page content as string, URL handle) """
328 # Strip hashes from the URL (#1038)
329 if isinstance(url_or_request, (compat_str, str)):
330 url_or_request = url_or_request.partition('#')[0]
332 urlh = self._request_webpage(url_or_request, video_id, note, errnote, fatal)
336 content = self._webpage_read_content(urlh, url_or_request, video_id, note, errnote, fatal)
337 return (content, urlh)
339 def _webpage_read_content(self, urlh, url_or_request, video_id, note=None, errnote=None, fatal=True, prefix=None):
340 content_type = urlh.headers.get('Content-Type', '')
341 webpage_bytes = urlh.read()
342 if prefix is not None:
343 webpage_bytes = prefix + webpage_bytes
344 m = re.match(r'[a-zA-Z0-9_.-]+/[a-zA-Z0-9_.-]+\s*;\s*charset=(.+)', content_type)
346 encoding = m.group(1)
348 m = re.search(br'<meta[^>]+charset=[\'"]?([^\'")]+)[ /\'">]',
349 webpage_bytes[:1024])
351 encoding = m.group(1).decode('ascii')
352 elif webpage_bytes.startswith(b'\xff\xfe'):
356 if self._downloader.params.get('dump_intermediate_pages', False):
358 url = url_or_request.get_full_url()
359 except AttributeError:
361 self.to_screen('Dumping request to ' + url)
362 dump = base64.b64encode(webpage_bytes).decode('ascii')
363 self._downloader.to_screen(dump)
364 if self._downloader.params.get('write_pages', False):
366 url = url_or_request.get_full_url()
367 except AttributeError:
369 basen = '%s_%s' % (video_id, url)
371 h = '___' + hashlib.md5(basen.encode('utf-8')).hexdigest()
372 basen = basen[:240 - len(h)] + h
373 raw_filename = basen + '.dump'
374 filename = sanitize_filename(raw_filename, restricted=True)
375 self.to_screen('Saving request to ' + filename)
376 # Working around MAX_PATH limitation on Windows (see
377 # http://msdn.microsoft.com/en-us/library/windows/desktop/aa365247(v=vs.85).aspx)
379 absfilepath = os.path.abspath(filename)
380 if len(absfilepath) > 259:
381 filename = '\\\\?\\' + absfilepath
382 with open(filename, 'wb') as outf:
383 outf.write(webpage_bytes)
386 content = webpage_bytes.decode(encoding, 'replace')
388 content = webpage_bytes.decode('utf-8', 'replace')
390 if ('<title>Access to this site is blocked</title>' in content and
391 'Websense' in content[:512]):
392 msg = 'Access to this webpage has been blocked by Websense filtering software in your network.'
393 blocked_iframe = self._html_search_regex(
394 r'<iframe src="([^"]+)"', content,
395 'Websense information URL', default=None)
397 msg += ' Visit %s for more details' % blocked_iframe
398 raise ExtractorError(msg, expected=True)
402 def _download_webpage(self, url_or_request, video_id, note=None, errnote=None, fatal=True, tries=1, timeout=5):
403 """ Returns the data of the page as a string """
406 while success is False:
408 res = self._download_webpage_handle(url_or_request, video_id, note, errnote, fatal)
410 except compat_http_client.IncompleteRead as e:
412 if try_count >= tries:
414 self._sleep(timeout, video_id)
421 def _download_xml(self, url_or_request, video_id,
422 note='Downloading XML', errnote='Unable to download XML',
423 transform_source=None, fatal=True):
424 """Return the xml as an xml.etree.ElementTree.Element"""
425 xml_string = self._download_webpage(
426 url_or_request, video_id, note, errnote, fatal=fatal)
427 if xml_string is False:
430 xml_string = transform_source(xml_string)
431 return xml.etree.ElementTree.fromstring(xml_string.encode('utf-8'))
433 def _download_json(self, url_or_request, video_id,
434 note='Downloading JSON metadata',
435 errnote='Unable to download JSON metadata',
436 transform_source=None,
438 json_string = self._download_webpage(
439 url_or_request, video_id, note, errnote, fatal=fatal)
440 if (not fatal) and json_string is False:
442 return self._parse_json(
443 json_string, video_id, transform_source=transform_source, fatal=fatal)
445 def _parse_json(self, json_string, video_id, transform_source=None, fatal=True):
447 json_string = transform_source(json_string)
449 return json.loads(json_string)
450 except ValueError as ve:
451 errmsg = '%s: Failed to parse JSON ' % video_id
453 raise ExtractorError(errmsg, cause=ve)
455 self.report_warning(errmsg + str(ve))
457 def report_warning(self, msg, video_id=None):
458 idstr = '' if video_id is None else '%s: ' % video_id
459 self._downloader.report_warning(
460 '[%s] %s%s' % (self.IE_NAME, idstr, msg))
462 def to_screen(self, msg):
463 """Print msg to screen, prefixing it with '[ie_name]'"""
464 self._downloader.to_screen('[%s] %s' % (self.IE_NAME, msg))
466 def report_extraction(self, id_or_name):
467 """Report information extraction."""
468 self.to_screen('%s: Extracting information' % id_or_name)
470 def report_download_webpage(self, video_id):
471 """Report webpage download."""
472 self.to_screen('%s: Downloading webpage' % video_id)
474 def report_age_confirmation(self):
475 """Report attempt to confirm age."""
476 self.to_screen('Confirming age')
478 def report_login(self):
479 """Report attempt to log in."""
480 self.to_screen('Logging in')
482 # Methods for following #608
484 def url_result(url, ie=None, video_id=None):
485 """Returns a url that points to a page that should be processed"""
486 # TODO: ie should be the class used for getting the info
487 video_info = {'_type': 'url',
490 if video_id is not None:
491 video_info['id'] = video_id
495 def playlist_result(entries, playlist_id=None, playlist_title=None, playlist_description=None):
496 """Returns a playlist"""
497 video_info = {'_type': 'playlist',
500 video_info['id'] = playlist_id
502 video_info['title'] = playlist_title
503 if playlist_description:
504 video_info['description'] = playlist_description
507 def _search_regex(self, pattern, string, name, default=_NO_DEFAULT, fatal=True, flags=0, group=None):
509 Perform a regex search on the given string, using a single or a list of
510 patterns returning the first matching group.
511 In case of failure return a default value or raise a WARNING or a
512 RegexNotFoundError, depending on fatal, specifying the field name.
514 if isinstance(pattern, (str, compat_str, compiled_regex_type)):
515 mobj = re.search(pattern, string, flags)
518 mobj = re.search(p, string, flags)
522 if not self._downloader.params.get('no_color') and os.name != 'nt' and sys.stderr.isatty():
523 _name = '\033[0;34m%s\033[0m' % name
529 # return the first matching group
530 return next(g for g in mobj.groups() if g is not None)
532 return mobj.group(group)
533 elif default is not _NO_DEFAULT:
536 raise RegexNotFoundError('Unable to extract %s' % _name)
538 self._downloader.report_warning('unable to extract %s; '
539 'please report this issue on http://yt-dl.org/bug' % _name)
542 def _html_search_regex(self, pattern, string, name, default=_NO_DEFAULT, fatal=True, flags=0, group=None):
544 Like _search_regex, but strips HTML tags and unescapes entities.
546 res = self._search_regex(pattern, string, name, default, fatal, flags, group)
548 return clean_html(res).strip()
552 def _get_login_info(self):
554 Get the the login info as (username, password)
555 It will look in the netrc file using the _NETRC_MACHINE value
556 If there's no info available, return (None, None)
558 if self._downloader is None:
563 downloader_params = self._downloader.params
565 # Attempt to use provided username and password or .netrc data
566 if downloader_params.get('username', None) is not None:
567 username = downloader_params['username']
568 password = downloader_params['password']
569 elif downloader_params.get('usenetrc', False):
571 info = netrc.netrc().authenticators(self._NETRC_MACHINE)
576 raise netrc.NetrcParseError('No authenticators for %s' % self._NETRC_MACHINE)
577 except (IOError, netrc.NetrcParseError) as err:
578 self._downloader.report_warning('parsing .netrc: %s' % compat_str(err))
580 return (username, password)
582 def _get_tfa_info(self):
584 Get the two-factor authentication info
585 TODO - asking the user will be required for sms/phone verify
586 currently just uses the command line option
587 If there's no info available, return None
589 if self._downloader is None:
591 downloader_params = self._downloader.params
593 if downloader_params.get('twofactor', None) is not None:
594 return downloader_params['twofactor']
598 # Helper functions for extracting OpenGraph info
600 def _og_regexes(prop):
601 content_re = r'content=(?:"([^>]+?)"|\'([^>]+?)\')'
602 property_re = r'(?:name|property)=[\'"]og:%s[\'"]' % re.escape(prop)
603 template = r'<meta[^>]+?%s[^>]+?%s'
605 template % (property_re, content_re),
606 template % (content_re, property_re),
609 def _og_search_property(self, prop, html, name=None, **kargs):
611 name = 'OpenGraph %s' % prop
612 escaped = self._search_regex(self._og_regexes(prop), html, name, flags=re.DOTALL, **kargs)
615 return unescapeHTML(escaped)
617 def _og_search_thumbnail(self, html, **kargs):
618 return self._og_search_property('image', html, 'thumbnail url', fatal=False, **kargs)
620 def _og_search_description(self, html, **kargs):
621 return self._og_search_property('description', html, fatal=False, **kargs)
623 def _og_search_title(self, html, **kargs):
624 return self._og_search_property('title', html, **kargs)
626 def _og_search_video_url(self, html, name='video url', secure=True, **kargs):
627 regexes = self._og_regexes('video') + self._og_regexes('video:url')
629 regexes = self._og_regexes('video:secure_url') + regexes
630 return self._html_search_regex(regexes, html, name, **kargs)
632 def _og_search_url(self, html, **kargs):
633 return self._og_search_property('url', html, **kargs)
635 def _html_search_meta(self, name, html, display_name=None, fatal=False, **kwargs):
636 if display_name is None:
638 return self._html_search_regex(
640 (?=[^>]+(?:itemprop|name|property)=(["\']?)%s\1)
641 [^>]+?content=(["\'])(?P<content>.*?)\2''' % re.escape(name),
642 html, display_name, fatal=fatal, group='content', **kwargs)
644 def _dc_search_uploader(self, html):
645 return self._html_search_meta('dc.creator', html, 'uploader')
647 def _rta_search(self, html):
648 # See http://www.rtalabel.org/index.php?content=howtofaq#single
649 if re.search(r'(?ix)<meta\s+name="rating"\s+'
650 r' content="RTA-5042-1996-1400-1577-RTA"',
655 def _media_rating_search(self, html):
656 # See http://www.tjg-designs.com/WP/metadata-code-examples-adding-metadata-to-your-web-pages/
657 rating = self._html_search_meta('rating', html)
669 return RATING_TABLE.get(rating.lower(), None)
671 def _family_friendly_search(self, html):
672 # See http://schema.org/VideoObject
673 family_friendly = self._html_search_meta('isFamilyFriendly', html)
675 if not family_friendly:
684 return RATING_TABLE.get(family_friendly.lower(), None)
686 def _twitter_search_player(self, html):
687 return self._html_search_meta('twitter:player', html,
688 'twitter card player')
690 def _sort_formats(self, formats):
692 raise ExtractorError('No video formats found')
695 # TODO remove the following workaround
696 from ..utils import determine_ext
697 if not f.get('ext') and 'url' in f:
698 f['ext'] = determine_ext(f['url'])
700 preference = f.get('preference')
701 if preference is None:
702 proto = f.get('protocol')
704 proto = compat_urllib_parse_urlparse(f.get('url', '')).scheme
706 preference = 0 if proto in ['http', 'https'] else -0.1
707 if f.get('ext') in ['f4f', 'f4m']: # Not yet supported
710 if f.get('vcodec') == 'none': # audio only
711 if self._downloader.params.get('prefer_free_formats'):
712 ORDER = ['aac', 'mp3', 'm4a', 'webm', 'ogg', 'opus']
714 ORDER = ['webm', 'opus', 'ogg', 'mp3', 'aac', 'm4a']
717 audio_ext_preference = ORDER.index(f['ext'])
719 audio_ext_preference = -1
721 if self._downloader.params.get('prefer_free_formats'):
722 ORDER = ['flv', 'mp4', 'webm']
724 ORDER = ['webm', 'flv', 'mp4']
726 ext_preference = ORDER.index(f['ext'])
729 audio_ext_preference = 0
733 f.get('language_preference') if f.get('language_preference') is not None else -1,
734 f.get('quality') if f.get('quality') is not None else -1,
735 f.get('tbr') if f.get('tbr') is not None else -1,
736 f.get('vbr') if f.get('vbr') is not None else -1,
737 f.get('height') if f.get('height') is not None else -1,
738 f.get('width') if f.get('width') is not None else -1,
740 f.get('abr') if f.get('abr') is not None else -1,
741 audio_ext_preference,
742 f.get('fps') if f.get('fps') is not None else -1,
743 f.get('filesize') if f.get('filesize') is not None else -1,
744 f.get('filesize_approx') if f.get('filesize_approx') is not None else -1,
745 f.get('source_preference') if f.get('source_preference') is not None else -1,
748 formats.sort(key=_formats_key)
750 def _check_formats(self, formats, video_id):
753 lambda f: self._is_valid_url(
755 item='%s video format' % f.get('format_id') if f.get('format_id') else 'video'),
758 def _is_valid_url(self, url, video_id, item='video'):
760 self._request_webpage(
761 HEADRequest(url), video_id,
762 'Checking %s URL' % item)
764 except ExtractorError as e:
765 if isinstance(e.cause, compat_HTTPError):
767 '%s URL is invalid, skipping' % item, video_id)
771 def http_scheme(self):
772 """ Either "http:" or "https:", depending on the user's preferences """
775 if self._downloader.params.get('prefer_insecure', False)
778 def _proto_relative_url(self, url, scheme=None):
781 if url.startswith('//'):
783 scheme = self.http_scheme()
788 def _sleep(self, timeout, video_id, msg_template=None):
789 if msg_template is None:
790 msg_template = '%(video_id)s: Waiting for %(timeout)s seconds'
791 msg = msg_template % {'video_id': video_id, 'timeout': timeout}
795 def _extract_f4m_formats(self, manifest_url, video_id, preference=None, f4m_id=None):
796 manifest = self._download_xml(
797 manifest_url, video_id, 'Downloading f4m manifest',
798 'Unable to download f4m manifest')
801 manifest_version = '1.0'
802 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/1.0}media')
804 manifest_version = '2.0'
805 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/2.0}media')
806 for i, media_el in enumerate(media_nodes):
807 if manifest_version == '2.0':
808 manifest_url = ('/'.join(manifest_url.split('/')[:-1]) + '/'
809 + (media_el.attrib.get('href') or media_el.attrib.get('url')))
810 tbr = int_or_none(media_el.attrib.get('bitrate'))
812 'format_id': '-'.join(filter(None, [f4m_id, 'f4m-%d' % (i if tbr is None else tbr)])),
816 'width': int_or_none(media_el.attrib.get('width')),
817 'height': int_or_none(media_el.attrib.get('height')),
818 'preference': preference,
820 self._sort_formats(formats)
824 def _extract_m3u8_formats(self, m3u8_url, video_id, ext=None,
825 entry_protocol='m3u8', preference=None,
829 'format_id': '-'.join(filter(None, [m3u8_id, 'm3u8-meta'])),
834 'resolution': 'multiple',
835 'format_note': 'Quality selection URL',
838 format_url = lambda u: (
840 if re.match(r'^https?://', u)
841 else compat_urlparse.urljoin(m3u8_url, u))
843 m3u8_doc = self._download_webpage(
845 note='Downloading m3u8 information',
846 errnote='Failed to download m3u8 information')
849 r'(?P<key>[a-zA-Z_-]+)=(?P<val>"[^"]+"|[^",]+)(?:,|$)')
850 for line in m3u8_doc.splitlines():
851 if line.startswith('#EXT-X-STREAM-INF:'):
853 for m in kv_rex.finditer(line):
855 if v.startswith('"'):
857 last_info[m.group('key')] = v
858 elif line.startswith('#') or not line.strip():
861 if last_info is None:
862 formats.append({'url': format_url(line)})
864 tbr = int_or_none(last_info.get('BANDWIDTH'), scale=1000)
866 'format_id': '-'.join(filter(None, [m3u8_id, 'm3u8-%d' % (tbr if tbr else len(formats))])),
867 'url': format_url(line.strip()),
870 'protocol': entry_protocol,
871 'preference': preference,
873 codecs = last_info.get('CODECS')
875 # TODO: looks like video codec is not always necessarily goes first
876 va_codecs = codecs.split(',')
878 f['vcodec'] = va_codecs[0].partition('.')[0]
879 if len(va_codecs) > 1 and va_codecs[1]:
880 f['acodec'] = va_codecs[1].partition('.')[0]
881 resolution = last_info.get('RESOLUTION')
883 width_str, height_str = resolution.split('x')
884 f['width'] = int(width_str)
885 f['height'] = int(height_str)
888 self._sort_formats(formats)
891 # TODO: improve extraction
892 def _extract_smil_formats(self, smil_url, video_id, fatal=True):
893 smil = self._download_xml(
894 smil_url, video_id, 'Downloading SMIL file',
895 'Unable to download SMIL file', fatal=fatal)
900 base = smil.find('./head/meta').get('base')
904 for video in smil.findall('./body/switch/video'):
905 src = video.get('src')
908 bitrate = int_or_none(video.get('system-bitrate') or video.get('systemBitrate'), 1000)
909 width = int_or_none(video.get('width'))
910 height = int_or_none(video.get('height'))
911 proto = video.get('proto')
914 if base.startswith('rtmp'):
916 elif base.startswith('http'):
918 ext = video.get('ext')
920 formats.extend(self._extract_m3u8_formats(src, video_id, ext))
921 elif proto == 'rtmp':
923 streamer = video.get('streamer') or base
928 'format_id': 'rtmp-%d' % (rtmp_count if bitrate is None else bitrate),
933 self._sort_formats(formats)
937 def _live_title(self, name):
938 """ Generate the title for a live video """
939 now = datetime.datetime.now()
940 now_str = now.strftime("%Y-%m-%d %H:%M")
941 return name + ' ' + now_str
943 def _int(self, v, name, fatal=False, **kwargs):
944 res = int_or_none(v, **kwargs)
945 if 'get_attr' in kwargs:
946 print(getattr(v, kwargs['get_attr']))
948 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
950 raise ExtractorError(msg)
952 self._downloader.report_warning(msg)
955 def _float(self, v, name, fatal=False, **kwargs):
956 res = float_or_none(v, **kwargs)
958 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
960 raise ExtractorError(msg)
962 self._downloader.report_warning(msg)
965 def _set_cookie(self, domain, name, value, expire_time=None):
966 cookie = compat_cookiejar.Cookie(
967 0, name, value, None, None, domain, None,
968 None, '/', True, False, expire_time, '', None, None, None)
969 self._downloader.cookiejar.set_cookie(cookie)
971 def get_testcases(self, include_onlymatching=False):
972 t = getattr(self, '_TEST', None)
974 assert not hasattr(self, '_TESTS'), \
975 '%s has _TEST and _TESTS' % type(self).__name__
978 tests = getattr(self, '_TESTS', [])
980 if not include_onlymatching and t.get('only_matching', False):
982 t['name'] = type(self).__name__[:-len('IE')]
985 def is_suitable(self, age_limit):
986 """ Test whether the extractor is generally suitable for the given
987 age limit (i.e. pornographic sites are not, all others usually are) """
989 any_restricted = False
990 for tc in self.get_testcases(include_onlymatching=False):
992 tc = tc['playlist'][0]
993 is_restricted = age_restricted(
994 tc.get('info_dict', {}).get('age_limit'), age_limit)
995 if not is_restricted:
997 any_restricted = any_restricted or is_restricted
998 return not any_restricted
1000 def extract_subtitles(self, *args, **kwargs):
1002 list_subtitles = self._downloader.params.get('listsubtitles')
1003 if self._downloader.params.get('writesubtitles', False) or list_subtitles:
1004 subtitles.update(self._get_subtitles(*args, **kwargs))
1007 def _get_subtitles(self, *args, **kwargs):
1008 raise NotImplementedError("This method must be implemented by subclasses")
1011 class SearchInfoExtractor(InfoExtractor):
1013 Base class for paged search queries extractors.
1014 They accept urls in the format _SEARCH_KEY(|all|[0-9]):{query}
1015 Instances should define _SEARCH_KEY and _MAX_RESULTS.
1019 def _make_valid_url(cls):
1020 return r'%s(?P<prefix>|[1-9][0-9]*|all):(?P<query>[\s\S]+)' % cls._SEARCH_KEY
1023 def suitable(cls, url):
1024 return re.match(cls._make_valid_url(), url) is not None
1026 def _real_extract(self, query):
1027 mobj = re.match(self._make_valid_url(), query)
1029 raise ExtractorError('Invalid search query "%s"' % query)
1031 prefix = mobj.group('prefix')
1032 query = mobj.group('query')
1034 return self._get_n_results(query, 1)
1035 elif prefix == 'all':
1036 return self._get_n_results(query, self._MAX_RESULTS)
1040 raise ExtractorError('invalid download number %s for query "%s"' % (n, query))
1041 elif n > self._MAX_RESULTS:
1042 self._downloader.report_warning('%s returns max %i results (you requested %i)' % (self._SEARCH_KEY, self._MAX_RESULTS, n))
1043 n = self._MAX_RESULTS
1044 return self._get_n_results(query, n)
1046 def _get_n_results(self, query, n):
1047 """Get a specified number of results for a query"""
1048 raise NotImplementedError("This method must be implemented by subclasses")
1051 def SEARCH_KEY(self):
1052 return self._SEARCH_KEY