1 from __future__ import unicode_literals
13 import xml.etree.ElementTree
15 from ..compat import (
20 compat_urllib_parse_urlparse,
35 _NO_DEFAULT = object()
38 class InfoExtractor(object):
39 """Information Extractor class.
41 Information extractors are the classes that, given a URL, extract
42 information about the video (or videos) the URL refers to. This
43 information includes the real video URL, the video title, author and
44 others. The information is stored in a dictionary which is then
45 passed to the YoutubeDL. The YoutubeDL processes this
46 information possibly downloading the video to the file system, among
47 other possible outcomes.
49 The type field determines the the type of the result.
50 By far the most common value (and the default if _type is missing) is
51 "video", which indicates a single video.
53 For a video, the dictionaries must include the following fields:
56 title: Video title, unescaped.
58 Additionally, it must contain either a formats entry or a url one:
60 formats: A list of dictionaries for each format available, ordered
61 from worst to best quality.
64 * url Mandatory. The URL of the video file
65 * ext Will be calculated from url if missing
66 * format A human-readable description of the format
67 ("mp4 container with h264/opus").
68 Calculated from the format_id, width, height.
69 and format_note fields if missing.
70 * format_id A short description of the format
71 ("mp4_h264_opus" or "19").
72 Technically optional, but strongly recommended.
73 * format_note Additional info about the format
74 ("3D" or "DASH video")
75 * width Width of the video, if known
76 * height Height of the video, if known
77 * resolution Textual description of width and height
78 * tbr Average bitrate of audio and video in KBit/s
79 * abr Average audio bitrate in KBit/s
80 * acodec Name of the audio codec in use
81 * asr Audio sampling rate in Hertz
82 * vbr Average video bitrate in KBit/s
84 * vcodec Name of the video codec in use
85 * container Name of the container format
86 * filesize The number of bytes, if known in advance
87 * filesize_approx An estimate for the number of bytes
88 * player_url SWF Player URL (used for rtmpdump).
89 * protocol The protocol that will be used for the actual
91 "http", "https", "rtsp", "rtmp", "rtmpe",
92 "m3u8", or "m3u8_native".
93 * preference Order number of this format. If this field is
94 present and not None, the formats get sorted
95 by this field, regardless of all other values.
96 -1 for default (order by other properties),
97 -2 or smaller for less than default.
98 < -1000 to hide the format (if there is
99 another one which is strictly better)
100 * language_preference Is this in the correct requested
102 10 if it's what the URL is about,
103 -1 for default (don't know),
104 -10 otherwise, other values reserved for now.
105 * quality Order number of the video quality of this
106 format, irrespective of the file format.
107 -1 for default (order by other properties),
108 -2 or smaller for less than default.
109 * source_preference Order number for this video source
110 (quality takes higher priority)
111 -1 for default (order by other properties),
112 -2 or smaller for less than default.
113 * http_method HTTP method to use for the download.
114 * http_headers A dictionary of additional HTTP headers
115 to add to the request.
116 * http_post_data Additional data to send with a POST
118 * stretched_ratio If given and not 1, indicates that the
119 video's pixels are not square.
120 width : height ratio as float.
121 * no_resume The server does not support resuming the
122 (HTTP or RTMP) download. Boolean.
124 url: Final video URL.
125 ext: Video filename extension.
126 format: The video format, defaults to ext (used for --get-format)
127 player_url: SWF Player URL (used for rtmpdump).
129 The following fields are optional:
131 alt_title: A secondary title of the video.
132 display_id An alternative identifier for the video, not necessarily
133 unique, but available before title. Typically, id is
134 something like "4234987", title "Dancing naked mole rats",
135 and display_id "dancing-naked-mole-rats"
136 thumbnails: A list of dictionaries, with the following entries:
137 * "id" (optional, string) - Thumbnail format ID
139 * "preference" (optional, int) - quality of the image
140 * "width" (optional, int)
141 * "height" (optional, int)
142 * "resolution" (optional, string "{width}x{height"},
144 thumbnail: Full URL to a video thumbnail image.
145 description: Full video description.
146 uploader: Full name of the video uploader.
147 creator: The main artist who created the video.
148 timestamp: UNIX timestamp of the moment the video became available.
149 upload_date: Video upload date (YYYYMMDD).
150 If not explicitly set, calculated from timestamp.
151 uploader_id: Nickname or id of the video uploader.
152 location: Physical location where the video was filmed.
153 subtitles: The subtitle file contents as a dictionary in the format
154 {language: subtitles}.
155 duration: Length of the video in seconds, as an integer.
156 view_count: How many users have watched the video on the platform.
157 like_count: Number of positive ratings of the video
158 dislike_count: Number of negative ratings of the video
159 average_rating: Average rating give by users, the scale used depends on the webpage
160 comment_count: Number of comments on the video
161 comments: A list of comments, each with one or more of the following
162 properties (all but one of text or html optional):
163 * "author" - human-readable name of the comment author
164 * "author_id" - user ID of the comment author
166 * "html" - Comment as HTML
167 * "text" - Plain text of the comment
168 * "timestamp" - UNIX timestamp of comment
169 * "parent" - ID of the comment this one is replying to.
170 Set to "root" to indicate that this is a
171 comment to the original video.
172 age_limit: Age restriction for the video, as an integer (years)
173 webpage_url: The url to the video webpage, if given to youtube-dl it
174 should allow to get the same result again. (It will be set
175 by YoutubeDL if it's missing)
176 categories: A list of categories that the video falls in, for example
178 is_live: True, False, or None (=unknown). Whether this video is a
179 live stream that goes on instead of a fixed-length video.
181 Unless mentioned otherwise, the fields should be Unicode strings.
183 Unless mentioned otherwise, None is equivalent to absence of information.
186 _type "playlist" indicates multiple videos.
187 There must be a key "entries", which is a list, an iterable, or a PagedList
188 object, each element of which is a valid dictionary by this specification.
190 Additionally, playlists can have "title" and "id" attributes with the same
191 semantics as videos (see above).
194 _type "multi_video" indicates that there are multiple videos that
195 form a single show, for examples multiple acts of an opera or TV episode.
196 It must have an entries key like a playlist and contain all the keys
197 required for a video at the same time.
200 _type "url" indicates that the video must be extracted from another
201 location, possibly by a different extractor. Its only required key is:
202 "url" - the next URL to extract.
203 The key "ie_key" can be set to the class name (minus the trailing "IE",
204 e.g. "Youtube") if the extractor class is known in advance.
205 Additionally, the dictionary may have any properties of the resolved entity
206 known in advance, for example "title" if the title of the referred video is
210 _type "url_transparent" entities have the same specification as "url", but
211 indicate that the given additional information is more precise than the one
212 associated with the resolved URL.
213 This is useful when a site employs a video service that hosts the video and
214 its technical metadata, but that video service does not embed a useful
215 title, description etc.
218 Subclasses of this one should re-define the _real_initialize() and
219 _real_extract() methods and define a _VALID_URL regexp.
220 Probably, they should also be added to the list of extractors.
222 Finally, the _WORKING attribute should be set to False for broken IEs
223 in order to warn the users and skip the tests.
230 def __init__(self, downloader=None):
231 """Constructor. Receives an optional downloader."""
233 self.set_downloader(downloader)
236 def suitable(cls, url):
237 """Receives a URL and returns True if suitable for this IE."""
239 # This does not use has/getattr intentionally - we want to know whether
240 # we have cached the regexp for *this* class, whereas getattr would also
241 # match the superclass
242 if '_VALID_URL_RE' not in cls.__dict__:
243 cls._VALID_URL_RE = re.compile(cls._VALID_URL)
244 return cls._VALID_URL_RE.match(url) is not None
247 def _match_id(cls, url):
248 if '_VALID_URL_RE' not in cls.__dict__:
249 cls._VALID_URL_RE = re.compile(cls._VALID_URL)
250 m = cls._VALID_URL_RE.match(url)
256 """Getter method for _WORKING."""
259 def initialize(self):
260 """Initializes an instance (authentication, etc)."""
262 self._real_initialize()
265 def extract(self, url):
266 """Extracts URL information and returns it in list of dicts."""
269 return self._real_extract(url)
270 except ExtractorError:
272 except compat_http_client.IncompleteRead as e:
273 raise ExtractorError('A network error has occured.', cause=e, expected=True)
274 except (KeyError, StopIteration) as e:
275 raise ExtractorError('An extractor error has occured.', cause=e)
277 def set_downloader(self, downloader):
278 """Sets the downloader for this IE."""
279 self._downloader = downloader
281 def _real_initialize(self):
282 """Real initialization process. Redefine in subclasses."""
285 def _real_extract(self, url):
286 """Real extraction process. Redefine in subclasses."""
291 """A string for getting the InfoExtractor with get_info_extractor"""
292 return cls.__name__[:-2]
296 return type(self).__name__[:-2]
298 def _request_webpage(self, url_or_request, video_id, note=None, errnote=None, fatal=True):
299 """ Returns the response handle """
301 self.report_download_webpage(video_id)
302 elif note is not False:
304 self.to_screen('%s' % (note,))
306 self.to_screen('%s: %s' % (video_id, note))
308 return self._downloader.urlopen(url_or_request)
309 except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err:
313 errnote = 'Unable to download webpage'
314 errmsg = '%s: %s' % (errnote, compat_str(err))
316 raise ExtractorError(errmsg, sys.exc_info()[2], cause=err)
318 self._downloader.report_warning(errmsg)
321 def _download_webpage_handle(self, url_or_request, video_id, note=None, errnote=None, fatal=True):
322 """ Returns a tuple (page content as string, URL handle) """
323 # Strip hashes from the URL (#1038)
324 if isinstance(url_or_request, (compat_str, str)):
325 url_or_request = url_or_request.partition('#')[0]
327 urlh = self._request_webpage(url_or_request, video_id, note, errnote, fatal)
331 content = self._webpage_read_content(urlh, url_or_request, video_id, note, errnote, fatal)
332 return (content, urlh)
334 def _webpage_read_content(self, urlh, url_or_request, video_id, note=None, errnote=None, fatal=True, prefix=None):
335 content_type = urlh.headers.get('Content-Type', '')
336 webpage_bytes = urlh.read()
337 if prefix is not None:
338 webpage_bytes = prefix + webpage_bytes
339 m = re.match(r'[a-zA-Z0-9_.-]+/[a-zA-Z0-9_.-]+\s*;\s*charset=(.+)', content_type)
341 encoding = m.group(1)
343 m = re.search(br'<meta[^>]+charset=[\'"]?([^\'")]+)[ /\'">]',
344 webpage_bytes[:1024])
346 encoding = m.group(1).decode('ascii')
347 elif webpage_bytes.startswith(b'\xff\xfe'):
351 if self._downloader.params.get('dump_intermediate_pages', False):
353 url = url_or_request.get_full_url()
354 except AttributeError:
356 self.to_screen('Dumping request to ' + url)
357 dump = base64.b64encode(webpage_bytes).decode('ascii')
358 self._downloader.to_screen(dump)
359 if self._downloader.params.get('write_pages', False):
361 url = url_or_request.get_full_url()
362 except AttributeError:
364 basen = '%s_%s' % (video_id, url)
366 h = '___' + hashlib.md5(basen.encode('utf-8')).hexdigest()
367 basen = basen[:240 - len(h)] + h
368 raw_filename = basen + '.dump'
369 filename = sanitize_filename(raw_filename, restricted=True)
370 self.to_screen('Saving request to ' + filename)
371 # Working around MAX_PATH limitation on Windows (see
372 # http://msdn.microsoft.com/en-us/library/windows/desktop/aa365247(v=vs.85).aspx)
374 absfilepath = os.path.abspath(filename)
375 if len(absfilepath) > 259:
376 filename = '\\\\?\\' + absfilepath
377 with open(filename, 'wb') as outf:
378 outf.write(webpage_bytes)
381 content = webpage_bytes.decode(encoding, 'replace')
383 content = webpage_bytes.decode('utf-8', 'replace')
385 if ('<title>Access to this site is blocked</title>' in content and
386 'Websense' in content[:512]):
387 msg = 'Access to this webpage has been blocked by Websense filtering software in your network.'
388 blocked_iframe = self._html_search_regex(
389 r'<iframe src="([^"]+)"', content,
390 'Websense information URL', default=None)
392 msg += ' Visit %s for more details' % blocked_iframe
393 raise ExtractorError(msg, expected=True)
397 def _download_webpage(self, url_or_request, video_id, note=None, errnote=None, fatal=True, tries=1, timeout=5):
398 """ Returns the data of the page as a string """
401 while success is False:
403 res = self._download_webpage_handle(url_or_request, video_id, note, errnote, fatal)
405 except compat_http_client.IncompleteRead as e:
407 if try_count >= tries:
409 self._sleep(timeout, video_id)
416 def _download_xml(self, url_or_request, video_id,
417 note='Downloading XML', errnote='Unable to download XML',
418 transform_source=None, fatal=True):
419 """Return the xml as an xml.etree.ElementTree.Element"""
420 xml_string = self._download_webpage(
421 url_or_request, video_id, note, errnote, fatal=fatal)
422 if xml_string is False:
425 xml_string = transform_source(xml_string)
426 return xml.etree.ElementTree.fromstring(xml_string.encode('utf-8'))
428 def _download_json(self, url_or_request, video_id,
429 note='Downloading JSON metadata',
430 errnote='Unable to download JSON metadata',
431 transform_source=None,
433 json_string = self._download_webpage(
434 url_or_request, video_id, note, errnote, fatal=fatal)
435 if (not fatal) and json_string is False:
437 return self._parse_json(
438 json_string, video_id, transform_source=transform_source, fatal=fatal)
440 def _parse_json(self, json_string, video_id, transform_source=None, fatal=True):
442 json_string = transform_source(json_string)
444 return json.loads(json_string)
445 except ValueError as ve:
446 errmsg = '%s: Failed to parse JSON ' % video_id
448 raise ExtractorError(errmsg, cause=ve)
450 self.report_warning(errmsg + str(ve))
452 def report_warning(self, msg, video_id=None):
453 idstr = '' if video_id is None else '%s: ' % video_id
454 self._downloader.report_warning(
455 '[%s] %s%s' % (self.IE_NAME, idstr, msg))
457 def to_screen(self, msg):
458 """Print msg to screen, prefixing it with '[ie_name]'"""
459 self._downloader.to_screen('[%s] %s' % (self.IE_NAME, msg))
461 def report_extraction(self, id_or_name):
462 """Report information extraction."""
463 self.to_screen('%s: Extracting information' % id_or_name)
465 def report_download_webpage(self, video_id):
466 """Report webpage download."""
467 self.to_screen('%s: Downloading webpage' % video_id)
469 def report_age_confirmation(self):
470 """Report attempt to confirm age."""
471 self.to_screen('Confirming age')
473 def report_login(self):
474 """Report attempt to log in."""
475 self.to_screen('Logging in')
477 # Methods for following #608
479 def url_result(url, ie=None, video_id=None):
480 """Returns a url that points to a page that should be processed"""
481 # TODO: ie should be the class used for getting the info
482 video_info = {'_type': 'url',
485 if video_id is not None:
486 video_info['id'] = video_id
490 def playlist_result(entries, playlist_id=None, playlist_title=None, playlist_description=None):
491 """Returns a playlist"""
492 video_info = {'_type': 'playlist',
495 video_info['id'] = playlist_id
497 video_info['title'] = playlist_title
498 if playlist_description:
499 video_info['description'] = playlist_description
502 def _search_regex(self, pattern, string, name, default=_NO_DEFAULT, fatal=True, flags=0, group=None):
504 Perform a regex search on the given string, using a single or a list of
505 patterns returning the first matching group.
506 In case of failure return a default value or raise a WARNING or a
507 RegexNotFoundError, depending on fatal, specifying the field name.
509 if isinstance(pattern, (str, compat_str, compiled_regex_type)):
510 mobj = re.search(pattern, string, flags)
513 mobj = re.search(p, string, flags)
517 if not self._downloader.params.get('no_color') and os.name != 'nt' and sys.stderr.isatty():
518 _name = '\033[0;34m%s\033[0m' % name
524 # return the first matching group
525 return next(g for g in mobj.groups() if g is not None)
527 return mobj.group(group)
528 elif default is not _NO_DEFAULT:
531 raise RegexNotFoundError('Unable to extract %s' % _name)
533 self._downloader.report_warning('unable to extract %s; '
534 'please report this issue on http://yt-dl.org/bug' % _name)
537 def _html_search_regex(self, pattern, string, name, default=_NO_DEFAULT, fatal=True, flags=0, group=None):
539 Like _search_regex, but strips HTML tags and unescapes entities.
541 res = self._search_regex(pattern, string, name, default, fatal, flags, group)
543 return clean_html(res).strip()
547 def _get_login_info(self):
549 Get the the login info as (username, password)
550 It will look in the netrc file using the _NETRC_MACHINE value
551 If there's no info available, return (None, None)
553 if self._downloader is None:
558 downloader_params = self._downloader.params
560 # Attempt to use provided username and password or .netrc data
561 if downloader_params.get('username', None) is not None:
562 username = downloader_params['username']
563 password = downloader_params['password']
564 elif downloader_params.get('usenetrc', False):
566 info = netrc.netrc().authenticators(self._NETRC_MACHINE)
571 raise netrc.NetrcParseError('No authenticators for %s' % self._NETRC_MACHINE)
572 except (IOError, netrc.NetrcParseError) as err:
573 self._downloader.report_warning('parsing .netrc: %s' % compat_str(err))
575 return (username, password)
577 def _get_tfa_info(self):
579 Get the two-factor authentication info
580 TODO - asking the user will be required for sms/phone verify
581 currently just uses the command line option
582 If there's no info available, return None
584 if self._downloader is None:
586 downloader_params = self._downloader.params
588 if downloader_params.get('twofactor', None) is not None:
589 return downloader_params['twofactor']
593 # Helper functions for extracting OpenGraph info
595 def _og_regexes(prop):
596 content_re = r'content=(?:"([^>]+?)"|\'([^>]+?)\')'
597 property_re = r'(?:name|property)=[\'"]og:%s[\'"]' % re.escape(prop)
598 template = r'<meta[^>]+?%s[^>]+?%s'
600 template % (property_re, content_re),
601 template % (content_re, property_re),
604 def _og_search_property(self, prop, html, name=None, **kargs):
606 name = 'OpenGraph %s' % prop
607 escaped = self._search_regex(self._og_regexes(prop), html, name, flags=re.DOTALL, **kargs)
610 return unescapeHTML(escaped)
612 def _og_search_thumbnail(self, html, **kargs):
613 return self._og_search_property('image', html, 'thumbnail url', fatal=False, **kargs)
615 def _og_search_description(self, html, **kargs):
616 return self._og_search_property('description', html, fatal=False, **kargs)
618 def _og_search_title(self, html, **kargs):
619 return self._og_search_property('title', html, **kargs)
621 def _og_search_video_url(self, html, name='video url', secure=True, **kargs):
622 regexes = self._og_regexes('video') + self._og_regexes('video:url')
624 regexes = self._og_regexes('video:secure_url') + regexes
625 return self._html_search_regex(regexes, html, name, **kargs)
627 def _og_search_url(self, html, **kargs):
628 return self._og_search_property('url', html, **kargs)
630 def _html_search_meta(self, name, html, display_name=None, fatal=False, **kwargs):
631 if display_name is None:
633 return self._html_search_regex(
635 (?=[^>]+(?:itemprop|name|property)=(["\']?)%s\1)
636 [^>]+?content=(["\'])(?P<content>.*?)\2''' % re.escape(name),
637 html, display_name, fatal=fatal, group='content', **kwargs)
639 def _dc_search_uploader(self, html):
640 return self._html_search_meta('dc.creator', html, 'uploader')
642 def _rta_search(self, html):
643 # See http://www.rtalabel.org/index.php?content=howtofaq#single
644 if re.search(r'(?ix)<meta\s+name="rating"\s+'
645 r' content="RTA-5042-1996-1400-1577-RTA"',
650 def _media_rating_search(self, html):
651 # See http://www.tjg-designs.com/WP/metadata-code-examples-adding-metadata-to-your-web-pages/
652 rating = self._html_search_meta('rating', html)
664 return RATING_TABLE.get(rating.lower(), None)
666 def _family_friendly_search(self, html):
667 # See http://schema.org/VideoObject
668 family_friendly = self._html_search_meta('isFamilyFriendly', html)
670 if not family_friendly:
679 return RATING_TABLE.get(family_friendly.lower(), None)
681 def _twitter_search_player(self, html):
682 return self._html_search_meta('twitter:player', html,
683 'twitter card player')
685 def _sort_formats(self, formats):
687 raise ExtractorError('No video formats found')
690 # TODO remove the following workaround
691 from ..utils import determine_ext
692 if not f.get('ext') and 'url' in f:
693 f['ext'] = determine_ext(f['url'])
695 preference = f.get('preference')
696 if preference is None:
697 proto = f.get('protocol')
699 proto = compat_urllib_parse_urlparse(f.get('url', '')).scheme
701 preference = 0 if proto in ['http', 'https'] else -0.1
702 if f.get('ext') in ['f4f', 'f4m']: # Not yet supported
705 if f.get('vcodec') == 'none': # audio only
706 if self._downloader.params.get('prefer_free_formats'):
707 ORDER = ['aac', 'mp3', 'm4a', 'webm', 'ogg', 'opus']
709 ORDER = ['webm', 'opus', 'ogg', 'mp3', 'aac', 'm4a']
712 audio_ext_preference = ORDER.index(f['ext'])
714 audio_ext_preference = -1
716 if self._downloader.params.get('prefer_free_formats'):
717 ORDER = ['flv', 'mp4', 'webm']
719 ORDER = ['webm', 'flv', 'mp4']
721 ext_preference = ORDER.index(f['ext'])
724 audio_ext_preference = 0
728 f.get('language_preference') if f.get('language_preference') is not None else -1,
729 f.get('quality') if f.get('quality') is not None else -1,
730 f.get('tbr') if f.get('tbr') is not None else -1,
731 f.get('filesize') if f.get('filesize') is not None else -1,
732 f.get('vbr') if f.get('vbr') is not None else -1,
733 f.get('height') if f.get('height') is not None else -1,
734 f.get('width') if f.get('width') is not None else -1,
736 f.get('abr') if f.get('abr') is not None else -1,
737 audio_ext_preference,
738 f.get('fps') if f.get('fps') is not None else -1,
739 f.get('filesize_approx') if f.get('filesize_approx') is not None else -1,
740 f.get('source_preference') if f.get('source_preference') is not None else -1,
743 formats.sort(key=_formats_key)
745 def _check_formats(self, formats, video_id):
748 lambda f: self._is_valid_url(
750 item='%s video format' % f.get('format_id') if f.get('format_id') else 'video'),
753 def _is_valid_url(self, url, video_id, item='video'):
755 self._request_webpage(url, video_id, 'Checking %s URL' % item)
757 except ExtractorError as e:
758 if isinstance(e.cause, compat_HTTPError):
760 '%s URL is invalid, skipping' % item, video_id)
764 def http_scheme(self):
765 """ Either "http:" or "https:", depending on the user's preferences """
768 if self._downloader.params.get('prefer_insecure', False)
771 def _proto_relative_url(self, url, scheme=None):
774 if url.startswith('//'):
776 scheme = self.http_scheme()
781 def _sleep(self, timeout, video_id, msg_template=None):
782 if msg_template is None:
783 msg_template = '%(video_id)s: Waiting for %(timeout)s seconds'
784 msg = msg_template % {'video_id': video_id, 'timeout': timeout}
788 def _extract_f4m_formats(self, manifest_url, video_id, preference=None, f4m_id=None):
789 manifest = self._download_xml(
790 manifest_url, video_id, 'Downloading f4m manifest',
791 'Unable to download f4m manifest')
794 manifest_version = '1.0'
795 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/1.0}media')
797 manifest_version = '2.0'
798 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/2.0}media')
799 for i, media_el in enumerate(media_nodes):
800 if manifest_version == '2.0':
801 manifest_url = ('/'.join(manifest_url.split('/')[:-1]) + '/'
802 + (media_el.attrib.get('href') or media_el.attrib.get('url')))
803 tbr = int_or_none(media_el.attrib.get('bitrate'))
805 'format_id': '-'.join(filter(None, [f4m_id, 'f4m-%d' % (i if tbr is None else tbr)])),
809 'width': int_or_none(media_el.attrib.get('width')),
810 'height': int_or_none(media_el.attrib.get('height')),
811 'preference': preference,
813 self._sort_formats(formats)
817 def _extract_m3u8_formats(self, m3u8_url, video_id, ext=None,
818 entry_protocol='m3u8', preference=None,
822 'format_id': '-'.join(filter(None, [m3u8_id, 'm3u8-meta'])),
827 'resolution': 'multiple',
828 'format_note': 'Quality selection URL',
831 format_url = lambda u: (
833 if re.match(r'^https?://', u)
834 else compat_urlparse.urljoin(m3u8_url, u))
836 m3u8_doc = self._download_webpage(
838 note='Downloading m3u8 information',
839 errnote='Failed to download m3u8 information')
843 r'(?P<key>[a-zA-Z_-]+)=(?P<val>"[^"]+"|[^",]+)(?:,|$)')
844 for line in m3u8_doc.splitlines():
845 if line.startswith('#EXT-X-STREAM-INF:'):
847 for m in kv_rex.finditer(line):
849 if v.startswith('"'):
851 last_info[m.group('key')] = v
852 elif line.startswith('#EXT-X-MEDIA:'):
854 for m in kv_rex.finditer(line):
856 if v.startswith('"'):
858 last_media[m.group('key')] = v
859 elif line.startswith('#') or not line.strip():
862 if last_info is None:
863 formats.append({'url': format_url(line)})
865 tbr = int_or_none(last_info.get('BANDWIDTH'), scale=1000)
867 'format_id': '-'.join(filter(None, [m3u8_id, 'm3u8-%d' % (tbr if tbr else len(formats))])),
868 'url': format_url(line.strip()),
871 'protocol': entry_protocol,
872 'preference': preference,
874 codecs = last_info.get('CODECS')
876 # TODO: looks like video codec is not always necessarily goes first
877 va_codecs = codecs.split(',')
879 f['vcodec'] = va_codecs[0].partition('.')[0]
880 if len(va_codecs) > 1 and va_codecs[1]:
881 f['acodec'] = va_codecs[1].partition('.')[0]
882 resolution = last_info.get('RESOLUTION')
884 width_str, height_str = resolution.split('x')
885 f['width'] = int(width_str)
886 f['height'] = int(height_str)
887 if last_media is not None:
888 f['m3u8_media'] = last_media
892 self._sort_formats(formats)
895 # TODO: improve extraction
896 def _extract_smil_formats(self, smil_url, video_id, fatal=True):
897 smil = self._download_xml(
898 smil_url, video_id, 'Downloading SMIL file',
899 'Unable to download SMIL file', fatal=fatal)
904 base = smil.find('./head/meta').get('base')
908 for video in smil.findall('./body/switch/video'):
909 src = video.get('src')
912 bitrate = int_or_none(video.get('system-bitrate') or video.get('systemBitrate'), 1000)
913 width = int_or_none(video.get('width'))
914 height = int_or_none(video.get('height'))
915 proto = video.get('proto')
918 if base.startswith('rtmp'):
920 elif base.startswith('http'):
922 ext = video.get('ext')
924 formats.extend(self._extract_m3u8_formats(src, video_id, ext))
925 elif proto == 'rtmp':
927 streamer = video.get('streamer') or base
932 'format_id': 'rtmp-%d' % (rtmp_count if bitrate is None else bitrate),
937 self._sort_formats(formats)
941 def _live_title(self, name):
942 """ Generate the title for a live video """
943 now = datetime.datetime.now()
944 now_str = now.strftime("%Y-%m-%d %H:%M")
945 return name + ' ' + now_str
947 def _int(self, v, name, fatal=False, **kwargs):
948 res = int_or_none(v, **kwargs)
949 if 'get_attr' in kwargs:
950 print(getattr(v, kwargs['get_attr']))
952 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
954 raise ExtractorError(msg)
956 self._downloader.report_warning(msg)
959 def _float(self, v, name, fatal=False, **kwargs):
960 res = float_or_none(v, **kwargs)
962 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
964 raise ExtractorError(msg)
966 self._downloader.report_warning(msg)
969 def _set_cookie(self, domain, name, value, expire_time=None):
970 cookie = compat_cookiejar.Cookie(
971 0, name, value, None, None, domain, None,
972 None, '/', True, False, expire_time, '', None, None, None)
973 self._downloader.cookiejar.set_cookie(cookie)
975 def get_testcases(self, include_onlymatching=False):
976 t = getattr(self, '_TEST', None)
978 assert not hasattr(self, '_TESTS'), \
979 '%s has _TEST and _TESTS' % type(self).__name__
982 tests = getattr(self, '_TESTS', [])
984 if not include_onlymatching and t.get('only_matching', False):
986 t['name'] = type(self).__name__[:-len('IE')]
989 def is_suitable(self, age_limit):
990 """ Test whether the extractor is generally suitable for the given
991 age limit (i.e. pornographic sites are not, all others usually are) """
993 any_restricted = False
994 for tc in self.get_testcases(include_onlymatching=False):
996 tc = tc['playlist'][0]
997 is_restricted = age_restricted(
998 tc.get('info_dict', {}).get('age_limit'), age_limit)
999 if not is_restricted:
1001 any_restricted = any_restricted or is_restricted
1002 return not any_restricted
1005 class SearchInfoExtractor(InfoExtractor):
1007 Base class for paged search queries extractors.
1008 They accept urls in the format _SEARCH_KEY(|all|[0-9]):{query}
1009 Instances should define _SEARCH_KEY and _MAX_RESULTS.
1013 def _make_valid_url(cls):
1014 return r'%s(?P<prefix>|[1-9][0-9]*|all):(?P<query>[\s\S]+)' % cls._SEARCH_KEY
1017 def suitable(cls, url):
1018 return re.match(cls._make_valid_url(), url) is not None
1020 def _real_extract(self, query):
1021 mobj = re.match(self._make_valid_url(), query)
1023 raise ExtractorError('Invalid search query "%s"' % query)
1025 prefix = mobj.group('prefix')
1026 query = mobj.group('query')
1028 return self._get_n_results(query, 1)
1029 elif prefix == 'all':
1030 return self._get_n_results(query, self._MAX_RESULTS)
1034 raise ExtractorError('invalid download number %s for query "%s"' % (n, query))
1035 elif n > self._MAX_RESULTS:
1036 self._downloader.report_warning('%s returns max %i results (you requested %i)' % (self._SEARCH_KEY, self._MAX_RESULTS, n))
1037 n = self._MAX_RESULTS
1038 return self._get_n_results(query, n)
1040 def _get_n_results(self, query, n):
1041 """Get a specified number of results for a query"""
1042 raise NotImplementedError("This method must be implemented by subclasses")
1045 def SEARCH_KEY(self):
1046 return self._SEARCH_KEY