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 available subtitles as a dictionary in the format
154 {language: subformats}. "subformats" is a list sorted from
155 lower to higher preference, each element is a dictionary
156 with the "ext" entry and one of:
157 * "data": The subtitles file contents
158 * "url": A url pointing to the subtitles file
159 automatic_captions: Like 'subtitles', used by the YoutubeIE for
160 automatically generated captions
161 duration: Length of the video in seconds, as an integer.
162 view_count: How many users have watched the video on the platform.
163 like_count: Number of positive ratings of the video
164 dislike_count: Number of negative ratings of the video
165 average_rating: Average rating give by users, the scale used depends on the webpage
166 comment_count: Number of comments on the video
167 comments: A list of comments, each with one or more of the following
168 properties (all but one of text or html optional):
169 * "author" - human-readable name of the comment author
170 * "author_id" - user ID of the comment author
172 * "html" - Comment as HTML
173 * "text" - Plain text of the comment
174 * "timestamp" - UNIX timestamp of comment
175 * "parent" - ID of the comment this one is replying to.
176 Set to "root" to indicate that this is a
177 comment to the original video.
178 age_limit: Age restriction for the video, as an integer (years)
179 webpage_url: The url to the video webpage, if given to youtube-dl it
180 should allow to get the same result again. (It will be set
181 by YoutubeDL if it's missing)
182 categories: A list of categories that the video falls in, for example
184 is_live: True, False, or None (=unknown). Whether this video is a
185 live stream that goes on instead of a fixed-length video.
187 Unless mentioned otherwise, the fields should be Unicode strings.
189 Unless mentioned otherwise, None is equivalent to absence of information.
192 _type "playlist" indicates multiple videos.
193 There must be a key "entries", which is a list, an iterable, or a PagedList
194 object, each element of which is a valid dictionary by this specification.
196 Additionally, playlists can have "title" and "id" attributes with the same
197 semantics as videos (see above).
200 _type "multi_video" indicates that there are multiple videos that
201 form a single show, for examples multiple acts of an opera or TV episode.
202 It must have an entries key like a playlist and contain all the keys
203 required for a video at the same time.
206 _type "url" indicates that the video must be extracted from another
207 location, possibly by a different extractor. Its only required key is:
208 "url" - the next URL to extract.
209 The key "ie_key" can be set to the class name (minus the trailing "IE",
210 e.g. "Youtube") if the extractor class is known in advance.
211 Additionally, the dictionary may have any properties of the resolved entity
212 known in advance, for example "title" if the title of the referred video is
216 _type "url_transparent" entities have the same specification as "url", but
217 indicate that the given additional information is more precise than the one
218 associated with the resolved URL.
219 This is useful when a site employs a video service that hosts the video and
220 its technical metadata, but that video service does not embed a useful
221 title, description etc.
224 Subclasses of this one should re-define the _real_initialize() and
225 _real_extract() methods and define a _VALID_URL regexp.
226 Probably, they should also be added to the list of extractors.
228 Finally, the _WORKING attribute should be set to False for broken IEs
229 in order to warn the users and skip the tests.
236 def __init__(self, downloader=None):
237 """Constructor. Receives an optional downloader."""
239 self.set_downloader(downloader)
242 def suitable(cls, url):
243 """Receives a URL and returns True if suitable for this IE."""
245 # This does not use has/getattr intentionally - we want to know whether
246 # we have cached the regexp for *this* class, whereas getattr would also
247 # match the superclass
248 if '_VALID_URL_RE' not in cls.__dict__:
249 cls._VALID_URL_RE = re.compile(cls._VALID_URL)
250 return cls._VALID_URL_RE.match(url) is not None
253 def _match_id(cls, url):
254 if '_VALID_URL_RE' not in cls.__dict__:
255 cls._VALID_URL_RE = re.compile(cls._VALID_URL)
256 m = cls._VALID_URL_RE.match(url)
262 """Getter method for _WORKING."""
265 def initialize(self):
266 """Initializes an instance (authentication, etc)."""
268 self._real_initialize()
271 def extract(self, url):
272 """Extracts URL information and returns it in list of dicts."""
275 return self._real_extract(url)
276 except ExtractorError:
278 except compat_http_client.IncompleteRead as e:
279 raise ExtractorError('A network error has occured.', cause=e, expected=True)
280 except (KeyError, StopIteration) as e:
281 raise ExtractorError('An extractor error has occured.', cause=e)
283 def set_downloader(self, downloader):
284 """Sets the downloader for this IE."""
285 self._downloader = downloader
287 def _real_initialize(self):
288 """Real initialization process. Redefine in subclasses."""
291 def _real_extract(self, url):
292 """Real extraction process. Redefine in subclasses."""
297 """A string for getting the InfoExtractor with get_info_extractor"""
298 return cls.__name__[:-2]
302 return type(self).__name__[:-2]
304 def _request_webpage(self, url_or_request, video_id, note=None, errnote=None, fatal=True):
305 """ Returns the response handle """
307 self.report_download_webpage(video_id)
308 elif note is not False:
310 self.to_screen('%s' % (note,))
312 self.to_screen('%s: %s' % (video_id, note))
314 return self._downloader.urlopen(url_or_request)
315 except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err:
319 errnote = 'Unable to download webpage'
320 errmsg = '%s: %s' % (errnote, compat_str(err))
322 raise ExtractorError(errmsg, sys.exc_info()[2], cause=err)
324 self._downloader.report_warning(errmsg)
327 def _download_webpage_handle(self, url_or_request, video_id, note=None, errnote=None, fatal=True):
328 """ Returns a tuple (page content as string, URL handle) """
329 # Strip hashes from the URL (#1038)
330 if isinstance(url_or_request, (compat_str, str)):
331 url_or_request = url_or_request.partition('#')[0]
333 urlh = self._request_webpage(url_or_request, video_id, note, errnote, fatal)
337 content = self._webpage_read_content(urlh, url_or_request, video_id, note, errnote, fatal)
338 return (content, urlh)
340 def _webpage_read_content(self, urlh, url_or_request, video_id, note=None, errnote=None, fatal=True, prefix=None):
341 content_type = urlh.headers.get('Content-Type', '')
342 webpage_bytes = urlh.read()
343 if prefix is not None:
344 webpage_bytes = prefix + webpage_bytes
345 m = re.match(r'[a-zA-Z0-9_.-]+/[a-zA-Z0-9_.-]+\s*;\s*charset=(.+)', content_type)
347 encoding = m.group(1)
349 m = re.search(br'<meta[^>]+charset=[\'"]?([^\'")]+)[ /\'">]',
350 webpage_bytes[:1024])
352 encoding = m.group(1).decode('ascii')
353 elif webpage_bytes.startswith(b'\xff\xfe'):
357 if self._downloader.params.get('dump_intermediate_pages', False):
359 url = url_or_request.get_full_url()
360 except AttributeError:
362 self.to_screen('Dumping request to ' + url)
363 dump = base64.b64encode(webpage_bytes).decode('ascii')
364 self._downloader.to_screen(dump)
365 if self._downloader.params.get('write_pages', False):
367 url = url_or_request.get_full_url()
368 except AttributeError:
370 basen = '%s_%s' % (video_id, url)
372 h = '___' + hashlib.md5(basen.encode('utf-8')).hexdigest()
373 basen = basen[:240 - len(h)] + h
374 raw_filename = basen + '.dump'
375 filename = sanitize_filename(raw_filename, restricted=True)
376 self.to_screen('Saving request to ' + filename)
377 # Working around MAX_PATH limitation on Windows (see
378 # http://msdn.microsoft.com/en-us/library/windows/desktop/aa365247(v=vs.85).aspx)
380 absfilepath = os.path.abspath(filename)
381 if len(absfilepath) > 259:
382 filename = '\\\\?\\' + absfilepath
383 with open(filename, 'wb') as outf:
384 outf.write(webpage_bytes)
387 content = webpage_bytes.decode(encoding, 'replace')
389 content = webpage_bytes.decode('utf-8', 'replace')
391 if ('<title>Access to this site is blocked</title>' in content and
392 'Websense' in content[:512]):
393 msg = 'Access to this webpage has been blocked by Websense filtering software in your network.'
394 blocked_iframe = self._html_search_regex(
395 r'<iframe src="([^"]+)"', content,
396 'Websense information URL', default=None)
398 msg += ' Visit %s for more details' % blocked_iframe
399 raise ExtractorError(msg, expected=True)
400 if '<title>The URL you requested has been blocked</title>' in content[:512]:
402 'Access to this webpage has been blocked by Indian censorship. '
403 'Use a VPN or proxy server (with --proxy) to route around it.')
404 block_msg = self._html_search_regex(
405 r'</h1><p>(.*?)</p>',
406 content, 'block message', default=None)
408 msg += ' (Message: "%s")' % block_msg.replace('\n', ' ')
409 raise ExtractorError(msg, expected=True)
413 def _download_webpage(self, url_or_request, video_id, note=None, errnote=None, fatal=True, tries=1, timeout=5):
414 """ Returns the data of the page as a string """
417 while success is False:
419 res = self._download_webpage_handle(url_or_request, video_id, note, errnote, fatal)
421 except compat_http_client.IncompleteRead as e:
423 if try_count >= tries:
425 self._sleep(timeout, video_id)
432 def _download_xml(self, url_or_request, video_id,
433 note='Downloading XML', errnote='Unable to download XML',
434 transform_source=None, fatal=True):
435 """Return the xml as an xml.etree.ElementTree.Element"""
436 xml_string = self._download_webpage(
437 url_or_request, video_id, note, errnote, fatal=fatal)
438 if xml_string is False:
441 xml_string = transform_source(xml_string)
442 return xml.etree.ElementTree.fromstring(xml_string.encode('utf-8'))
444 def _download_json(self, url_or_request, video_id,
445 note='Downloading JSON metadata',
446 errnote='Unable to download JSON metadata',
447 transform_source=None,
449 json_string = self._download_webpage(
450 url_or_request, video_id, note, errnote, fatal=fatal)
451 if (not fatal) and json_string is False:
453 return self._parse_json(
454 json_string, video_id, transform_source=transform_source, fatal=fatal)
456 def _parse_json(self, json_string, video_id, transform_source=None, fatal=True):
458 json_string = transform_source(json_string)
460 return json.loads(json_string)
461 except ValueError as ve:
462 errmsg = '%s: Failed to parse JSON ' % video_id
464 raise ExtractorError(errmsg, cause=ve)
466 self.report_warning(errmsg + str(ve))
468 def report_warning(self, msg, video_id=None):
469 idstr = '' if video_id is None else '%s: ' % video_id
470 self._downloader.report_warning(
471 '[%s] %s%s' % (self.IE_NAME, idstr, msg))
473 def to_screen(self, msg):
474 """Print msg to screen, prefixing it with '[ie_name]'"""
475 self._downloader.to_screen('[%s] %s' % (self.IE_NAME, msg))
477 def report_extraction(self, id_or_name):
478 """Report information extraction."""
479 self.to_screen('%s: Extracting information' % id_or_name)
481 def report_download_webpage(self, video_id):
482 """Report webpage download."""
483 self.to_screen('%s: Downloading webpage' % video_id)
485 def report_age_confirmation(self):
486 """Report attempt to confirm age."""
487 self.to_screen('Confirming age')
489 def report_login(self):
490 """Report attempt to log in."""
491 self.to_screen('Logging in')
493 # Methods for following #608
495 def url_result(url, ie=None, video_id=None, video_title=None):
496 """Returns a url that points to a page that should be processed"""
497 # TODO: ie should be the class used for getting the info
498 video_info = {'_type': 'url',
501 if video_id is not None:
502 video_info['id'] = video_id
503 if video_title is not None:
504 video_info['title'] = video_title
508 def playlist_result(entries, playlist_id=None, playlist_title=None, playlist_description=None):
509 """Returns a playlist"""
510 video_info = {'_type': 'playlist',
513 video_info['id'] = playlist_id
515 video_info['title'] = playlist_title
516 if playlist_description:
517 video_info['description'] = playlist_description
520 def _search_regex(self, pattern, string, name, default=_NO_DEFAULT, fatal=True, flags=0, group=None):
522 Perform a regex search on the given string, using a single or a list of
523 patterns returning the first matching group.
524 In case of failure return a default value or raise a WARNING or a
525 RegexNotFoundError, depending on fatal, specifying the field name.
527 if isinstance(pattern, (str, compat_str, compiled_regex_type)):
528 mobj = re.search(pattern, string, flags)
531 mobj = re.search(p, string, flags)
535 if not self._downloader.params.get('no_color') and os.name != 'nt' and sys.stderr.isatty():
536 _name = '\033[0;34m%s\033[0m' % name
542 # return the first matching group
543 return next(g for g in mobj.groups() if g is not None)
545 return mobj.group(group)
546 elif default is not _NO_DEFAULT:
549 raise RegexNotFoundError('Unable to extract %s' % _name)
551 self._downloader.report_warning('unable to extract %s; '
552 'please report this issue on http://yt-dl.org/bug' % _name)
555 def _html_search_regex(self, pattern, string, name, default=_NO_DEFAULT, fatal=True, flags=0, group=None):
557 Like _search_regex, but strips HTML tags and unescapes entities.
559 res = self._search_regex(pattern, string, name, default, fatal, flags, group)
561 return clean_html(res).strip()
565 def _get_login_info(self):
567 Get the the login info as (username, password)
568 It will look in the netrc file using the _NETRC_MACHINE value
569 If there's no info available, return (None, None)
571 if self._downloader is None:
576 downloader_params = self._downloader.params
578 # Attempt to use provided username and password or .netrc data
579 if downloader_params.get('username', None) is not None:
580 username = downloader_params['username']
581 password = downloader_params['password']
582 elif downloader_params.get('usenetrc', False):
584 info = netrc.netrc().authenticators(self._NETRC_MACHINE)
589 raise netrc.NetrcParseError('No authenticators for %s' % self._NETRC_MACHINE)
590 except (IOError, netrc.NetrcParseError) as err:
591 self._downloader.report_warning('parsing .netrc: %s' % compat_str(err))
593 return (username, password)
595 def _get_tfa_info(self):
597 Get the two-factor authentication info
598 TODO - asking the user will be required for sms/phone verify
599 currently just uses the command line option
600 If there's no info available, return None
602 if self._downloader is None:
604 downloader_params = self._downloader.params
606 if downloader_params.get('twofactor', None) is not None:
607 return downloader_params['twofactor']
611 # Helper functions for extracting OpenGraph info
613 def _og_regexes(prop):
614 content_re = r'content=(?:"([^>]+?)"|\'([^>]+?)\')'
615 property_re = r'(?:name|property)=[\'"]og:%s[\'"]' % re.escape(prop)
616 template = r'<meta[^>]+?%s[^>]+?%s'
618 template % (property_re, content_re),
619 template % (content_re, property_re),
622 def _og_search_property(self, prop, html, name=None, **kargs):
624 name = 'OpenGraph %s' % prop
625 escaped = self._search_regex(self._og_regexes(prop), html, name, flags=re.DOTALL, **kargs)
628 return unescapeHTML(escaped)
630 def _og_search_thumbnail(self, html, **kargs):
631 return self._og_search_property('image', html, 'thumbnail url', fatal=False, **kargs)
633 def _og_search_description(self, html, **kargs):
634 return self._og_search_property('description', html, fatal=False, **kargs)
636 def _og_search_title(self, html, **kargs):
637 return self._og_search_property('title', html, **kargs)
639 def _og_search_video_url(self, html, name='video url', secure=True, **kargs):
640 regexes = self._og_regexes('video') + self._og_regexes('video:url')
642 regexes = self._og_regexes('video:secure_url') + regexes
643 return self._html_search_regex(regexes, html, name, **kargs)
645 def _og_search_url(self, html, **kargs):
646 return self._og_search_property('url', html, **kargs)
648 def _html_search_meta(self, name, html, display_name=None, fatal=False, **kwargs):
649 if display_name is None:
651 return self._html_search_regex(
653 (?=[^>]+(?:itemprop|name|property)=(["\']?)%s\1)
654 [^>]+?content=(["\'])(?P<content>.*?)\2''' % re.escape(name),
655 html, display_name, fatal=fatal, group='content', **kwargs)
657 def _dc_search_uploader(self, html):
658 return self._html_search_meta('dc.creator', html, 'uploader')
660 def _rta_search(self, html):
661 # See http://www.rtalabel.org/index.php?content=howtofaq#single
662 if re.search(r'(?ix)<meta\s+name="rating"\s+'
663 r' content="RTA-5042-1996-1400-1577-RTA"',
668 def _media_rating_search(self, html):
669 # See http://www.tjg-designs.com/WP/metadata-code-examples-adding-metadata-to-your-web-pages/
670 rating = self._html_search_meta('rating', html)
682 return RATING_TABLE.get(rating.lower(), None)
684 def _family_friendly_search(self, html):
685 # See http://schema.org/VideoObject
686 family_friendly = self._html_search_meta('isFamilyFriendly', html)
688 if not family_friendly:
697 return RATING_TABLE.get(family_friendly.lower(), None)
699 def _twitter_search_player(self, html):
700 return self._html_search_meta('twitter:player', html,
701 'twitter card player')
703 def _sort_formats(self, formats):
705 raise ExtractorError('No video formats found')
708 # TODO remove the following workaround
709 from ..utils import determine_ext
710 if not f.get('ext') and 'url' in f:
711 f['ext'] = determine_ext(f['url'])
713 preference = f.get('preference')
714 if preference is None:
715 proto = f.get('protocol')
717 proto = compat_urllib_parse_urlparse(f.get('url', '')).scheme
719 preference = 0 if proto in ['http', 'https'] else -0.1
720 if f.get('ext') in ['f4f', 'f4m']: # Not yet supported
723 if f.get('vcodec') == 'none': # audio only
724 if self._downloader.params.get('prefer_free_formats'):
725 ORDER = ['aac', 'mp3', 'm4a', 'webm', 'ogg', 'opus']
727 ORDER = ['webm', 'opus', 'ogg', 'mp3', 'aac', 'm4a']
730 audio_ext_preference = ORDER.index(f['ext'])
732 audio_ext_preference = -1
734 if self._downloader.params.get('prefer_free_formats'):
735 ORDER = ['flv', 'mp4', 'webm']
737 ORDER = ['webm', 'flv', 'mp4']
739 ext_preference = ORDER.index(f['ext'])
742 audio_ext_preference = 0
746 f.get('language_preference') if f.get('language_preference') is not None else -1,
747 f.get('quality') if f.get('quality') is not None else -1,
748 f.get('tbr') if f.get('tbr') is not None else -1,
749 f.get('filesize') if f.get('filesize') is not None else -1,
750 f.get('vbr') if f.get('vbr') is not None else -1,
751 f.get('height') if f.get('height') is not None else -1,
752 f.get('width') if f.get('width') is not None else -1,
754 f.get('abr') if f.get('abr') is not None else -1,
755 audio_ext_preference,
756 f.get('fps') if f.get('fps') is not None else -1,
757 f.get('filesize_approx') if f.get('filesize_approx') is not None else -1,
758 f.get('source_preference') if f.get('source_preference') is not None else -1,
761 formats.sort(key=_formats_key)
763 def _check_formats(self, formats, video_id):
766 lambda f: self._is_valid_url(
768 item='%s video format' % f.get('format_id') if f.get('format_id') else 'video'),
771 def _is_valid_url(self, url, video_id, item='video'):
772 url = self._proto_relative_url(url, scheme='http:')
773 # For now assume non HTTP(S) URLs always valid
774 if not (url.startswith('http://') or url.startswith('https://')):
777 self._request_webpage(url, video_id, 'Checking %s URL' % item)
779 except ExtractorError as e:
780 if isinstance(e.cause, compat_HTTPError):
782 '%s URL is invalid, skipping' % item, video_id)
786 def http_scheme(self):
787 """ Either "http:" or "https:", depending on the user's preferences """
790 if self._downloader.params.get('prefer_insecure', False)
793 def _proto_relative_url(self, url, scheme=None):
796 if url.startswith('//'):
798 scheme = self.http_scheme()
803 def _sleep(self, timeout, video_id, msg_template=None):
804 if msg_template is None:
805 msg_template = '%(video_id)s: Waiting for %(timeout)s seconds'
806 msg = msg_template % {'video_id': video_id, 'timeout': timeout}
810 def _extract_f4m_formats(self, manifest_url, video_id, preference=None, f4m_id=None):
811 manifest = self._download_xml(
812 manifest_url, video_id, 'Downloading f4m manifest',
813 'Unable to download f4m manifest')
816 manifest_version = '1.0'
817 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/1.0}media')
819 manifest_version = '2.0'
820 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/2.0}media')
821 for i, media_el in enumerate(media_nodes):
822 if manifest_version == '2.0':
823 manifest_url = ('/'.join(manifest_url.split('/')[:-1]) + '/' +
824 (media_el.attrib.get('href') or media_el.attrib.get('url')))
825 tbr = int_or_none(media_el.attrib.get('bitrate'))
827 'format_id': '-'.join(filter(None, [f4m_id, compat_str(i if tbr is None else tbr)])),
831 'width': int_or_none(media_el.attrib.get('width')),
832 'height': int_or_none(media_el.attrib.get('height')),
833 'preference': preference,
835 self._sort_formats(formats)
839 def _extract_m3u8_formats(self, m3u8_url, video_id, ext=None,
840 entry_protocol='m3u8', preference=None,
844 'format_id': '-'.join(filter(None, [m3u8_id, 'meta'])),
848 'preference': preference - 1 if preference else -1,
849 'resolution': 'multiple',
850 'format_note': 'Quality selection URL',
853 format_url = lambda u: (
855 if re.match(r'^https?://', u)
856 else compat_urlparse.urljoin(m3u8_url, u))
858 m3u8_doc = self._download_webpage(
860 note='Downloading m3u8 information',
861 errnote='Failed to download m3u8 information')
865 r'(?P<key>[a-zA-Z_-]+)=(?P<val>"[^"]+"|[^",]+)(?:,|$)')
866 for line in m3u8_doc.splitlines():
867 if line.startswith('#EXT-X-STREAM-INF:'):
869 for m in kv_rex.finditer(line):
871 if v.startswith('"'):
873 last_info[m.group('key')] = v
874 elif line.startswith('#EXT-X-MEDIA:'):
876 for m in kv_rex.finditer(line):
878 if v.startswith('"'):
880 last_media[m.group('key')] = v
881 elif line.startswith('#') or not line.strip():
884 if last_info is None:
885 formats.append({'url': format_url(line)})
887 tbr = int_or_none(last_info.get('BANDWIDTH'), scale=1000)
890 format_id.append(m3u8_id)
891 last_media_name = last_media.get('NAME') if last_media else None
892 format_id.append(last_media_name if last_media_name else '%d' % (tbr if tbr else len(formats)))
894 'format_id': '-'.join(format_id),
895 'url': format_url(line.strip()),
898 'protocol': entry_protocol,
899 'preference': preference,
901 codecs = last_info.get('CODECS')
903 # TODO: looks like video codec is not always necessarily goes first
904 va_codecs = codecs.split(',')
906 f['vcodec'] = va_codecs[0].partition('.')[0]
907 if len(va_codecs) > 1 and va_codecs[1]:
908 f['acodec'] = va_codecs[1].partition('.')[0]
909 resolution = last_info.get('RESOLUTION')
911 width_str, height_str = resolution.split('x')
912 f['width'] = int(width_str)
913 f['height'] = int(height_str)
914 if last_media is not None:
915 f['m3u8_media'] = last_media
919 self._sort_formats(formats)
922 # TODO: improve extraction
923 def _extract_smil_formats(self, smil_url, video_id, fatal=True):
924 smil = self._download_xml(
925 smil_url, video_id, 'Downloading SMIL file',
926 'Unable to download SMIL file', fatal=fatal)
931 base = smil.find('./head/meta').get('base')
935 if smil.findall('./body/seq/video'):
936 video = smil.findall('./body/seq/video')[0]
937 fmts, rtmp_count = self._parse_smil_video(video, video_id, base, rtmp_count)
940 for video in smil.findall('./body/switch/video'):
941 fmts, rtmp_count = self._parse_smil_video(video, video_id, base, rtmp_count)
944 self._sort_formats(formats)
948 def _parse_smil_video(self, video, video_id, base, rtmp_count):
949 src = video.get('src')
951 return ([], rtmp_count)
952 bitrate = int_or_none(video.get('system-bitrate') or video.get('systemBitrate'), 1000)
953 width = int_or_none(video.get('width'))
954 height = int_or_none(video.get('height'))
955 proto = video.get('proto')
958 if base.startswith('rtmp'):
960 elif base.startswith('http'):
962 ext = video.get('ext')
964 return (self._extract_m3u8_formats(src, video_id, ext), rtmp_count)
965 elif proto == 'rtmp':
967 streamer = video.get('streamer') or base
972 'format_id': 'rtmp-%d' % (rtmp_count if bitrate is None else bitrate),
977 elif proto.startswith('http'):
986 def _live_title(self, name):
987 """ Generate the title for a live video """
988 now = datetime.datetime.now()
989 now_str = now.strftime("%Y-%m-%d %H:%M")
990 return name + ' ' + now_str
992 def _int(self, v, name, fatal=False, **kwargs):
993 res = int_or_none(v, **kwargs)
994 if 'get_attr' in kwargs:
995 print(getattr(v, kwargs['get_attr']))
997 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
999 raise ExtractorError(msg)
1001 self._downloader.report_warning(msg)
1004 def _float(self, v, name, fatal=False, **kwargs):
1005 res = float_or_none(v, **kwargs)
1007 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
1009 raise ExtractorError(msg)
1011 self._downloader.report_warning(msg)
1014 def _set_cookie(self, domain, name, value, expire_time=None):
1015 cookie = compat_cookiejar.Cookie(
1016 0, name, value, None, None, domain, None,
1017 None, '/', True, False, expire_time, '', None, None, None)
1018 self._downloader.cookiejar.set_cookie(cookie)
1020 def get_testcases(self, include_onlymatching=False):
1021 t = getattr(self, '_TEST', None)
1023 assert not hasattr(self, '_TESTS'), \
1024 '%s has _TEST and _TESTS' % type(self).__name__
1027 tests = getattr(self, '_TESTS', [])
1029 if not include_onlymatching and t.get('only_matching', False):
1031 t['name'] = type(self).__name__[:-len('IE')]
1034 def is_suitable(self, age_limit):
1035 """ Test whether the extractor is generally suitable for the given
1036 age limit (i.e. pornographic sites are not, all others usually are) """
1038 any_restricted = False
1039 for tc in self.get_testcases(include_onlymatching=False):
1040 if 'playlist' in tc:
1041 tc = tc['playlist'][0]
1042 is_restricted = age_restricted(
1043 tc.get('info_dict', {}).get('age_limit'), age_limit)
1044 if not is_restricted:
1046 any_restricted = any_restricted or is_restricted
1047 return not any_restricted
1049 def extract_subtitles(self, *args, **kwargs):
1050 if (self._downloader.params.get('writesubtitles', False) or
1051 self._downloader.params.get('listsubtitles')):
1052 return self._get_subtitles(*args, **kwargs)
1055 def _get_subtitles(self, *args, **kwargs):
1056 raise NotImplementedError("This method must be implemented by subclasses")
1058 def extract_automatic_captions(self, *args, **kwargs):
1059 if (self._downloader.params.get('writeautomaticsub', False) or
1060 self._downloader.params.get('listsubtitles')):
1061 return self._get_automatic_captions(*args, **kwargs)
1064 def _get_automatic_captions(self, *args, **kwargs):
1065 raise NotImplementedError("This method must be implemented by subclasses")
1067 def _subtitles_timecode(self, seconds):
1068 return '%02d:%02d:%02d.%03d' % (seconds / 3600, (seconds % 3600) / 60, seconds % 60, (seconds % 1) * 1000)
1071 class SearchInfoExtractor(InfoExtractor):
1073 Base class for paged search queries extractors.
1074 They accept urls in the format _SEARCH_KEY(|all|[0-9]):{query}
1075 Instances should define _SEARCH_KEY and _MAX_RESULTS.
1079 def _make_valid_url(cls):
1080 return r'%s(?P<prefix>|[1-9][0-9]*|all):(?P<query>[\s\S]+)' % cls._SEARCH_KEY
1083 def suitable(cls, url):
1084 return re.match(cls._make_valid_url(), url) is not None
1086 def _real_extract(self, query):
1087 mobj = re.match(self._make_valid_url(), query)
1089 raise ExtractorError('Invalid search query "%s"' % query)
1091 prefix = mobj.group('prefix')
1092 query = mobj.group('query')
1094 return self._get_n_results(query, 1)
1095 elif prefix == 'all':
1096 return self._get_n_results(query, self._MAX_RESULTS)
1100 raise ExtractorError('invalid download number %s for query "%s"' % (n, query))
1101 elif n > self._MAX_RESULTS:
1102 self._downloader.report_warning('%s returns max %i results (you requested %i)' % (self._SEARCH_KEY, self._MAX_RESULTS, n))
1103 n = self._MAX_RESULTS
1104 return self._get_n_results(query, n)
1106 def _get_n_results(self, query, n):
1107 """Get a specified number of results for a query"""
1108 raise NotImplementedError("This method must be implemented by subclasses")
1111 def SEARCH_KEY(self):
1112 return self._SEARCH_KEY