1 from __future__ import unicode_literals
13 import xml.etree.ElementTree
15 from ..compat import (
20 compat_urllib_parse_urlparse,
36 _NO_DEFAULT = object()
39 class InfoExtractor(object):
40 """Information Extractor class.
42 Information extractors are the classes that, given a URL, extract
43 information about the video (or videos) the URL refers to. This
44 information includes the real video URL, the video title, author and
45 others. The information is stored in a dictionary which is then
46 passed to the YoutubeDL. The YoutubeDL processes this
47 information possibly downloading the video to the file system, among
48 other possible outcomes.
50 The type field determines the the type of the result.
51 By far the most common value (and the default if _type is missing) is
52 "video", which indicates a single video.
54 For a video, the dictionaries must include the following fields:
57 title: Video title, unescaped.
59 Additionally, it must contain either a formats entry or a url one:
61 formats: A list of dictionaries for each format available, ordered
62 from worst to best quality.
65 * url Mandatory. The URL of the video file
66 * ext Will be calculated from url if missing
67 * format A human-readable description of the format
68 ("mp4 container with h264/opus").
69 Calculated from the format_id, width, height.
70 and format_note fields if missing.
71 * format_id A short description of the format
72 ("mp4_h264_opus" or "19").
73 Technically optional, but strongly recommended.
74 * format_note Additional info about the format
75 ("3D" or "DASH video")
76 * width Width of the video, if known
77 * height Height of the video, if known
78 * resolution Textual description of width and height
79 * tbr Average bitrate of audio and video in KBit/s
80 * abr Average audio bitrate in KBit/s
81 * acodec Name of the audio codec in use
82 * asr Audio sampling rate in Hertz
83 * vbr Average video bitrate in KBit/s
85 * vcodec Name of the video codec in use
86 * container Name of the container format
87 * filesize The number of bytes, if known in advance
88 * filesize_approx An estimate for the number of bytes
89 * player_url SWF Player URL (used for rtmpdump).
90 * protocol The protocol that will be used for the actual
92 "http", "https", "rtsp", "rtmp", "rtmpe",
93 "m3u8", or "m3u8_native".
94 * preference Order number of this format. If this field is
95 present and not None, the formats get sorted
96 by this field, regardless of all other values.
97 -1 for default (order by other properties),
98 -2 or smaller for less than default.
99 < -1000 to hide the format (if there is
100 another one which is strictly better)
101 * language_preference Is this in the correct requested
103 10 if it's what the URL is about,
104 -1 for default (don't know),
105 -10 otherwise, other values reserved for now.
106 * quality Order number of the video quality of this
107 format, irrespective of the file format.
108 -1 for default (order by other properties),
109 -2 or smaller for less than default.
110 * source_preference Order number for this video source
111 (quality takes higher priority)
112 -1 for default (order by other properties),
113 -2 or smaller for less than default.
114 * http_method HTTP method to use for the download.
115 * http_headers A dictionary of additional HTTP headers
116 to add to the request.
117 * http_post_data Additional data to send with a POST
119 * stretched_ratio If given and not 1, indicates that the
120 video's pixels are not square.
121 width : height ratio as float.
122 * no_resume The server does not support resuming the
123 (HTTP or RTMP) download. Boolean.
125 url: Final video URL.
126 ext: Video filename extension.
127 format: The video format, defaults to ext (used for --get-format)
128 player_url: SWF Player URL (used for rtmpdump).
130 The following fields are optional:
132 alt_title: A secondary title of the video.
133 display_id An alternative identifier for the video, not necessarily
134 unique, but available before title. Typically, id is
135 something like "4234987", title "Dancing naked mole rats",
136 and display_id "dancing-naked-mole-rats"
137 thumbnails: A list of dictionaries, with the following entries:
138 * "id" (optional, string) - Thumbnail format ID
140 * "preference" (optional, int) - quality of the image
141 * "width" (optional, int)
142 * "height" (optional, int)
143 * "resolution" (optional, string "{width}x{height"},
145 thumbnail: Full URL to a video thumbnail image.
146 description: Full video description.
147 uploader: Full name of the video uploader.
148 creator: The main artist who created the video.
149 timestamp: UNIX timestamp of the moment the video became available.
150 upload_date: Video upload date (YYYYMMDD).
151 If not explicitly set, calculated from timestamp.
152 uploader_id: Nickname or id of the video uploader.
153 location: Physical location where the video was filmed.
154 subtitles: The available subtitles as a dictionary in the format
155 {language: subformats}. "subformats" is a list sorted from
156 lower to higher preference, each element is a dictionary
157 with the "ext" entry and one of:
158 * "data": The subtitles file contents
159 * "url": A url pointing to the subtitles file
160 automatic_captions: Like 'subtitles', used by the YoutubeIE for
161 automatically generated captions
162 duration: Length of the video in seconds, as an integer.
163 view_count: How many users have watched the video on the platform.
164 like_count: Number of positive ratings of the video
165 dislike_count: Number of negative ratings of the video
166 average_rating: Average rating give by users, the scale used depends on the webpage
167 comment_count: Number of comments on the video
168 comments: A list of comments, each with one or more of the following
169 properties (all but one of text or html optional):
170 * "author" - human-readable name of the comment author
171 * "author_id" - user ID of the comment author
173 * "html" - Comment as HTML
174 * "text" - Plain text of the comment
175 * "timestamp" - UNIX timestamp of comment
176 * "parent" - ID of the comment this one is replying to.
177 Set to "root" to indicate that this is a
178 comment to the original video.
179 age_limit: Age restriction for the video, as an integer (years)
180 webpage_url: The url to the video webpage, if given to youtube-dl it
181 should allow to get the same result again. (It will be set
182 by YoutubeDL if it's missing)
183 categories: A list of categories that the video falls in, for example
185 is_live: True, False, or None (=unknown). Whether this video is a
186 live stream that goes on instead of a fixed-length video.
188 Unless mentioned otherwise, the fields should be Unicode strings.
190 Unless mentioned otherwise, None is equivalent to absence of information.
193 _type "playlist" indicates multiple videos.
194 There must be a key "entries", which is a list, an iterable, or a PagedList
195 object, each element of which is a valid dictionary by this specification.
197 Additionally, playlists can have "title" and "id" attributes with the same
198 semantics as videos (see above).
201 _type "multi_video" indicates that there are multiple videos that
202 form a single show, for examples multiple acts of an opera or TV episode.
203 It must have an entries key like a playlist and contain all the keys
204 required for a video at the same time.
207 _type "url" indicates that the video must be extracted from another
208 location, possibly by a different extractor. Its only required key is:
209 "url" - the next URL to extract.
210 The key "ie_key" can be set to the class name (minus the trailing "IE",
211 e.g. "Youtube") if the extractor class is known in advance.
212 Additionally, the dictionary may have any properties of the resolved entity
213 known in advance, for example "title" if the title of the referred video is
217 _type "url_transparent" entities have the same specification as "url", but
218 indicate that the given additional information is more precise than the one
219 associated with the resolved URL.
220 This is useful when a site employs a video service that hosts the video and
221 its technical metadata, but that video service does not embed a useful
222 title, description etc.
225 Subclasses of this one should re-define the _real_initialize() and
226 _real_extract() methods and define a _VALID_URL regexp.
227 Probably, they should also be added to the list of extractors.
229 Finally, the _WORKING attribute should be set to False for broken IEs
230 in order to warn the users and skip the tests.
237 def __init__(self, downloader=None):
238 """Constructor. Receives an optional downloader."""
240 self.set_downloader(downloader)
243 def suitable(cls, url):
244 """Receives a URL and returns True if suitable for this IE."""
246 # This does not use has/getattr intentionally - we want to know whether
247 # we have cached the regexp for *this* class, whereas getattr would also
248 # match the superclass
249 if '_VALID_URL_RE' not in cls.__dict__:
250 cls._VALID_URL_RE = re.compile(cls._VALID_URL)
251 return cls._VALID_URL_RE.match(url) is not None
254 def _match_id(cls, url):
255 if '_VALID_URL_RE' not in cls.__dict__:
256 cls._VALID_URL_RE = re.compile(cls._VALID_URL)
257 m = cls._VALID_URL_RE.match(url)
263 """Getter method for _WORKING."""
266 def initialize(self):
267 """Initializes an instance (authentication, etc)."""
269 self._real_initialize()
272 def extract(self, url):
273 """Extracts URL information and returns it in list of dicts."""
276 return self._real_extract(url)
277 except ExtractorError:
279 except compat_http_client.IncompleteRead as e:
280 raise ExtractorError('A network error has occured.', cause=e, expected=True)
281 except (KeyError, StopIteration) as e:
282 raise ExtractorError('An extractor error has occured.', cause=e)
284 def set_downloader(self, downloader):
285 """Sets the downloader for this IE."""
286 self._downloader = downloader
288 def _real_initialize(self):
289 """Real initialization process. Redefine in subclasses."""
292 def _real_extract(self, url):
293 """Real extraction process. Redefine in subclasses."""
298 """A string for getting the InfoExtractor with get_info_extractor"""
299 return cls.__name__[:-2]
303 return type(self).__name__[:-2]
305 def _request_webpage(self, url_or_request, video_id, note=None, errnote=None, fatal=True):
306 """ Returns the response handle """
308 self.report_download_webpage(video_id)
309 elif note is not False:
311 self.to_screen('%s' % (note,))
313 self.to_screen('%s: %s' % (video_id, note))
315 return self._downloader.urlopen(url_or_request)
316 except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err:
320 errnote = 'Unable to download webpage'
321 errmsg = '%s: %s' % (errnote, compat_str(err))
323 raise ExtractorError(errmsg, sys.exc_info()[2], cause=err)
325 self._downloader.report_warning(errmsg)
328 def _download_webpage_handle(self, url_or_request, video_id, note=None, errnote=None, fatal=True, encoding=None):
329 """ Returns a tuple (page content as string, URL handle) """
330 # Strip hashes from the URL (#1038)
331 if isinstance(url_or_request, (compat_str, str)):
332 url_or_request = url_or_request.partition('#')[0]
334 urlh = self._request_webpage(url_or_request, video_id, note, errnote, fatal)
338 content = self._webpage_read_content(urlh, url_or_request, video_id, note, errnote, fatal, encoding=encoding)
339 return (content, urlh)
342 def _guess_encoding_from_content(content_type, webpage_bytes):
343 m = re.match(r'[a-zA-Z0-9_.-]+/[a-zA-Z0-9_.-]+\s*;\s*charset=(.+)', content_type)
345 encoding = m.group(1)
347 m = re.search(br'<meta[^>]+charset=[\'"]?([^\'")]+)[ /\'">]',
348 webpage_bytes[:1024])
350 encoding = m.group(1).decode('ascii')
351 elif webpage_bytes.startswith(b'\xff\xfe'):
358 def _webpage_read_content(self, urlh, url_or_request, video_id, note=None, errnote=None, fatal=True, prefix=None, encoding=None):
359 content_type = urlh.headers.get('Content-Type', '')
360 webpage_bytes = urlh.read()
361 if prefix is not None:
362 webpage_bytes = prefix + webpage_bytes
364 encoding = self._guess_encoding_from_content(content_type, webpage_bytes)
365 if self._downloader.params.get('dump_intermediate_pages', False):
367 url = url_or_request.get_full_url()
368 except AttributeError:
370 self.to_screen('Dumping request to ' + url)
371 dump = base64.b64encode(webpage_bytes).decode('ascii')
372 self._downloader.to_screen(dump)
373 if self._downloader.params.get('write_pages', False):
375 url = url_or_request.get_full_url()
376 except AttributeError:
378 basen = '%s_%s' % (video_id, url)
380 h = '___' + hashlib.md5(basen.encode('utf-8')).hexdigest()
381 basen = basen[:240 - len(h)] + h
382 raw_filename = basen + '.dump'
383 filename = sanitize_filename(raw_filename, restricted=True)
384 self.to_screen('Saving request to ' + filename)
385 # Working around MAX_PATH limitation on Windows (see
386 # http://msdn.microsoft.com/en-us/library/windows/desktop/aa365247(v=vs.85).aspx)
388 absfilepath = os.path.abspath(filename)
389 if len(absfilepath) > 259:
390 filename = '\\\\?\\' + absfilepath
391 with open(filename, 'wb') as outf:
392 outf.write(webpage_bytes)
395 content = webpage_bytes.decode(encoding, 'replace')
397 content = webpage_bytes.decode('utf-8', 'replace')
399 if ('<title>Access to this site is blocked</title>' in content and
400 'Websense' in content[:512]):
401 msg = 'Access to this webpage has been blocked by Websense filtering software in your network.'
402 blocked_iframe = self._html_search_regex(
403 r'<iframe src="([^"]+)"', content,
404 'Websense information URL', default=None)
406 msg += ' Visit %s for more details' % blocked_iframe
407 raise ExtractorError(msg, expected=True)
408 if '<title>The URL you requested has been blocked</title>' in content[:512]:
410 'Access to this webpage has been blocked by Indian censorship. '
411 'Use a VPN or proxy server (with --proxy) to route around it.')
412 block_msg = self._html_search_regex(
413 r'</h1><p>(.*?)</p>',
414 content, 'block message', default=None)
416 msg += ' (Message: "%s")' % block_msg.replace('\n', ' ')
417 raise ExtractorError(msg, expected=True)
421 def _download_webpage(self, url_or_request, video_id, note=None, errnote=None, fatal=True, tries=1, timeout=5, encoding=None):
422 """ Returns the data of the page as a string """
425 while success is False:
427 res = self._download_webpage_handle(url_or_request, video_id, note, errnote, fatal, encoding=encoding)
429 except compat_http_client.IncompleteRead as e:
431 if try_count >= tries:
433 self._sleep(timeout, video_id)
440 def _download_xml(self, url_or_request, video_id,
441 note='Downloading XML', errnote='Unable to download XML',
442 transform_source=None, fatal=True, encoding=None):
443 """Return the xml as an xml.etree.ElementTree.Element"""
444 xml_string = self._download_webpage(
445 url_or_request, video_id, note, errnote, fatal=fatal, encoding=encoding)
446 if xml_string is False:
449 xml_string = transform_source(xml_string)
450 return xml.etree.ElementTree.fromstring(xml_string.encode('utf-8'))
452 def _download_json(self, url_or_request, video_id,
453 note='Downloading JSON metadata',
454 errnote='Unable to download JSON metadata',
455 transform_source=None,
456 fatal=True, encoding=None):
457 json_string = self._download_webpage(
458 url_or_request, video_id, note, errnote, fatal=fatal,
460 if (not fatal) and json_string is False:
462 return self._parse_json(
463 json_string, video_id, transform_source=transform_source, fatal=fatal)
465 def _parse_json(self, json_string, video_id, transform_source=None, fatal=True):
467 json_string = transform_source(json_string)
469 return json.loads(json_string)
470 except ValueError as ve:
471 errmsg = '%s: Failed to parse JSON ' % video_id
473 raise ExtractorError(errmsg, cause=ve)
475 self.report_warning(errmsg + str(ve))
477 def report_warning(self, msg, video_id=None):
478 idstr = '' if video_id is None else '%s: ' % video_id
479 self._downloader.report_warning(
480 '[%s] %s%s' % (self.IE_NAME, idstr, msg))
482 def to_screen(self, msg):
483 """Print msg to screen, prefixing it with '[ie_name]'"""
484 self._downloader.to_screen('[%s] %s' % (self.IE_NAME, msg))
486 def report_extraction(self, id_or_name):
487 """Report information extraction."""
488 self.to_screen('%s: Extracting information' % id_or_name)
490 def report_download_webpage(self, video_id):
491 """Report webpage download."""
492 self.to_screen('%s: Downloading webpage' % video_id)
494 def report_age_confirmation(self):
495 """Report attempt to confirm age."""
496 self.to_screen('Confirming age')
498 def report_login(self):
499 """Report attempt to log in."""
500 self.to_screen('Logging in')
502 # Methods for following #608
504 def url_result(url, ie=None, video_id=None, video_title=None):
505 """Returns a url that points to a page that should be processed"""
506 # TODO: ie should be the class used for getting the info
507 video_info = {'_type': 'url',
510 if video_id is not None:
511 video_info['id'] = video_id
512 if video_title is not None:
513 video_info['title'] = video_title
517 def playlist_result(entries, playlist_id=None, playlist_title=None, playlist_description=None):
518 """Returns a playlist"""
519 video_info = {'_type': 'playlist',
522 video_info['id'] = playlist_id
524 video_info['title'] = playlist_title
525 if playlist_description:
526 video_info['description'] = playlist_description
529 def _search_regex(self, pattern, string, name, default=_NO_DEFAULT, fatal=True, flags=0, group=None):
531 Perform a regex search on the given string, using a single or a list of
532 patterns returning the first matching group.
533 In case of failure return a default value or raise a WARNING or a
534 RegexNotFoundError, depending on fatal, specifying the field name.
536 if isinstance(pattern, (str, compat_str, compiled_regex_type)):
537 mobj = re.search(pattern, string, flags)
540 mobj = re.search(p, string, flags)
544 if not self._downloader.params.get('no_color') and os.name != 'nt' and sys.stderr.isatty():
545 _name = '\033[0;34m%s\033[0m' % name
551 # return the first matching group
552 return next(g for g in mobj.groups() if g is not None)
554 return mobj.group(group)
555 elif default is not _NO_DEFAULT:
558 raise RegexNotFoundError('Unable to extract %s' % _name)
560 self._downloader.report_warning('unable to extract %s' % _name + bug_reports_message())
563 def _html_search_regex(self, pattern, string, name, default=_NO_DEFAULT, fatal=True, flags=0, group=None):
565 Like _search_regex, but strips HTML tags and unescapes entities.
567 res = self._search_regex(pattern, string, name, default, fatal, flags, group)
569 return clean_html(res).strip()
573 def _get_login_info(self):
575 Get the the login info as (username, password)
576 It will look in the netrc file using the _NETRC_MACHINE value
577 If there's no info available, return (None, None)
579 if self._downloader is None:
584 downloader_params = self._downloader.params
586 # Attempt to use provided username and password or .netrc data
587 if downloader_params.get('username', None) is not None:
588 username = downloader_params['username']
589 password = downloader_params['password']
590 elif downloader_params.get('usenetrc', False):
592 info = netrc.netrc().authenticators(self._NETRC_MACHINE)
597 raise netrc.NetrcParseError('No authenticators for %s' % self._NETRC_MACHINE)
598 except (IOError, netrc.NetrcParseError) as err:
599 self._downloader.report_warning('parsing .netrc: %s' % compat_str(err))
601 return (username, password)
603 def _get_tfa_info(self):
605 Get the two-factor authentication info
606 TODO - asking the user will be required for sms/phone verify
607 currently just uses the command line option
608 If there's no info available, return None
610 if self._downloader is None:
612 downloader_params = self._downloader.params
614 if downloader_params.get('twofactor', None) is not None:
615 return downloader_params['twofactor']
619 # Helper functions for extracting OpenGraph info
621 def _og_regexes(prop):
622 content_re = r'content=(?:"([^>]+?)"|\'([^>]+?)\')'
623 property_re = r'(?:name|property)=[\'"]og:%s[\'"]' % re.escape(prop)
624 template = r'<meta[^>]+?%s[^>]+?%s'
626 template % (property_re, content_re),
627 template % (content_re, property_re),
630 def _og_search_property(self, prop, html, name=None, **kargs):
632 name = 'OpenGraph %s' % prop
633 escaped = self._search_regex(self._og_regexes(prop), html, name, flags=re.DOTALL, **kargs)
636 return unescapeHTML(escaped)
638 def _og_search_thumbnail(self, html, **kargs):
639 return self._og_search_property('image', html, 'thumbnail url', fatal=False, **kargs)
641 def _og_search_description(self, html, **kargs):
642 return self._og_search_property('description', html, fatal=False, **kargs)
644 def _og_search_title(self, html, **kargs):
645 return self._og_search_property('title', html, **kargs)
647 def _og_search_video_url(self, html, name='video url', secure=True, **kargs):
648 regexes = self._og_regexes('video') + self._og_regexes('video:url')
650 regexes = self._og_regexes('video:secure_url') + regexes
651 return self._html_search_regex(regexes, html, name, **kargs)
653 def _og_search_url(self, html, **kargs):
654 return self._og_search_property('url', html, **kargs)
656 def _html_search_meta(self, name, html, display_name=None, fatal=False, **kwargs):
657 if display_name is None:
659 return self._html_search_regex(
661 (?=[^>]+(?:itemprop|name|property)=(["\']?)%s\1)
662 [^>]+?content=(["\'])(?P<content>.*?)\2''' % re.escape(name),
663 html, display_name, fatal=fatal, group='content', **kwargs)
665 def _dc_search_uploader(self, html):
666 return self._html_search_meta('dc.creator', html, 'uploader')
668 def _rta_search(self, html):
669 # See http://www.rtalabel.org/index.php?content=howtofaq#single
670 if re.search(r'(?ix)<meta\s+name="rating"\s+'
671 r' content="RTA-5042-1996-1400-1577-RTA"',
676 def _media_rating_search(self, html):
677 # See http://www.tjg-designs.com/WP/metadata-code-examples-adding-metadata-to-your-web-pages/
678 rating = self._html_search_meta('rating', html)
690 return RATING_TABLE.get(rating.lower(), None)
692 def _family_friendly_search(self, html):
693 # See http://schema.org/VideoObject
694 family_friendly = self._html_search_meta('isFamilyFriendly', html)
696 if not family_friendly:
705 return RATING_TABLE.get(family_friendly.lower(), None)
707 def _twitter_search_player(self, html):
708 return self._html_search_meta('twitter:player', html,
709 'twitter card player')
711 def _sort_formats(self, formats):
713 raise ExtractorError('No video formats found')
716 # TODO remove the following workaround
717 from ..utils import determine_ext
718 if not f.get('ext') and 'url' in f:
719 f['ext'] = determine_ext(f['url'])
721 preference = f.get('preference')
722 if preference is None:
723 proto = f.get('protocol')
725 proto = compat_urllib_parse_urlparse(f.get('url', '')).scheme
727 preference = 0 if proto in ['http', 'https'] else -0.1
728 if f.get('ext') in ['f4f', 'f4m']: # Not yet supported
731 if f.get('vcodec') == 'none': # audio only
732 if self._downloader.params.get('prefer_free_formats'):
733 ORDER = ['aac', 'mp3', 'm4a', 'webm', 'ogg', 'opus']
735 ORDER = ['webm', 'opus', 'ogg', 'mp3', 'aac', 'm4a']
738 audio_ext_preference = ORDER.index(f['ext'])
740 audio_ext_preference = -1
742 if self._downloader.params.get('prefer_free_formats'):
743 ORDER = ['flv', 'mp4', 'webm']
745 ORDER = ['webm', 'flv', 'mp4']
747 ext_preference = ORDER.index(f['ext'])
750 audio_ext_preference = 0
754 f.get('language_preference') if f.get('language_preference') is not None else -1,
755 f.get('quality') if f.get('quality') is not None else -1,
756 f.get('tbr') if f.get('tbr') is not None else -1,
757 f.get('filesize') if f.get('filesize') is not None else -1,
758 f.get('vbr') if f.get('vbr') is not None else -1,
759 f.get('height') if f.get('height') is not None else -1,
760 f.get('width') if f.get('width') is not None else -1,
762 f.get('abr') if f.get('abr') is not None else -1,
763 audio_ext_preference,
764 f.get('fps') if f.get('fps') is not None else -1,
765 f.get('filesize_approx') if f.get('filesize_approx') is not None else -1,
766 f.get('source_preference') if f.get('source_preference') is not None else -1,
769 formats.sort(key=_formats_key)
771 def _check_formats(self, formats, video_id):
774 lambda f: self._is_valid_url(
776 item='%s video format' % f.get('format_id') if f.get('format_id') else 'video'),
779 def _is_valid_url(self, url, video_id, item='video'):
780 url = self._proto_relative_url(url, scheme='http:')
781 # For now assume non HTTP(S) URLs always valid
782 if not (url.startswith('http://') or url.startswith('https://')):
785 self._request_webpage(url, video_id, 'Checking %s URL' % item)
787 except ExtractorError as e:
788 if isinstance(e.cause, compat_HTTPError):
790 '%s URL is invalid, skipping' % item, video_id)
794 def http_scheme(self):
795 """ Either "http:" or "https:", depending on the user's preferences """
798 if self._downloader.params.get('prefer_insecure', False)
801 def _proto_relative_url(self, url, scheme=None):
804 if url.startswith('//'):
806 scheme = self.http_scheme()
811 def _sleep(self, timeout, video_id, msg_template=None):
812 if msg_template is None:
813 msg_template = '%(video_id)s: Waiting for %(timeout)s seconds'
814 msg = msg_template % {'video_id': video_id, 'timeout': timeout}
818 def _extract_f4m_formats(self, manifest_url, video_id, preference=None, f4m_id=None):
819 manifest = self._download_xml(
820 manifest_url, video_id, 'Downloading f4m manifest',
821 'Unable to download f4m manifest')
824 manifest_version = '1.0'
825 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/1.0}media')
827 manifest_version = '2.0'
828 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/2.0}media')
829 for i, media_el in enumerate(media_nodes):
830 if manifest_version == '2.0':
831 manifest_url = ('/'.join(manifest_url.split('/')[:-1]) + '/' +
832 (media_el.attrib.get('href') or media_el.attrib.get('url')))
833 tbr = int_or_none(media_el.attrib.get('bitrate'))
835 'format_id': '-'.join(filter(None, [f4m_id, compat_str(i if tbr is None else tbr)])),
839 'width': int_or_none(media_el.attrib.get('width')),
840 'height': int_or_none(media_el.attrib.get('height')),
841 'preference': preference,
843 self._sort_formats(formats)
847 def _extract_m3u8_formats(self, m3u8_url, video_id, ext=None,
848 entry_protocol='m3u8', preference=None,
852 'format_id': '-'.join(filter(None, [m3u8_id, 'meta'])),
856 'preference': preference - 1 if preference else -1,
857 'resolution': 'multiple',
858 'format_note': 'Quality selection URL',
861 format_url = lambda u: (
863 if re.match(r'^https?://', u)
864 else compat_urlparse.urljoin(m3u8_url, u))
866 m3u8_doc = self._download_webpage(
868 note='Downloading m3u8 information',
869 errnote='Failed to download m3u8 information')
873 r'(?P<key>[a-zA-Z_-]+)=(?P<val>"[^"]+"|[^",]+)(?:,|$)')
874 for line in m3u8_doc.splitlines():
875 if line.startswith('#EXT-X-STREAM-INF:'):
877 for m in kv_rex.finditer(line):
879 if v.startswith('"'):
881 last_info[m.group('key')] = v
882 elif line.startswith('#EXT-X-MEDIA:'):
884 for m in kv_rex.finditer(line):
886 if v.startswith('"'):
888 last_media[m.group('key')] = v
889 elif line.startswith('#') or not line.strip():
892 if last_info is None:
893 formats.append({'url': format_url(line)})
895 tbr = int_or_none(last_info.get('BANDWIDTH'), scale=1000)
898 format_id.append(m3u8_id)
899 last_media_name = last_media.get('NAME') if last_media else None
900 format_id.append(last_media_name if last_media_name else '%d' % (tbr if tbr else len(formats)))
902 'format_id': '-'.join(format_id),
903 'url': format_url(line.strip()),
906 'protocol': entry_protocol,
907 'preference': preference,
909 codecs = last_info.get('CODECS')
911 # TODO: looks like video codec is not always necessarily goes first
912 va_codecs = codecs.split(',')
914 f['vcodec'] = va_codecs[0].partition('.')[0]
915 if len(va_codecs) > 1 and va_codecs[1]:
916 f['acodec'] = va_codecs[1].partition('.')[0]
917 resolution = last_info.get('RESOLUTION')
919 width_str, height_str = resolution.split('x')
920 f['width'] = int(width_str)
921 f['height'] = int(height_str)
922 if last_media is not None:
923 f['m3u8_media'] = last_media
927 self._sort_formats(formats)
930 # TODO: improve extraction
931 def _extract_smil_formats(self, smil_url, video_id, fatal=True):
932 smil = self._download_xml(
933 smil_url, video_id, 'Downloading SMIL file',
934 'Unable to download SMIL file', fatal=fatal)
939 base = smil.find('./head/meta').get('base')
943 if smil.findall('./body/seq/video'):
944 video = smil.findall('./body/seq/video')[0]
945 fmts, rtmp_count = self._parse_smil_video(video, video_id, base, rtmp_count)
948 for video in smil.findall('./body/switch/video'):
949 fmts, rtmp_count = self._parse_smil_video(video, video_id, base, rtmp_count)
952 self._sort_formats(formats)
956 def _parse_smil_video(self, video, video_id, base, rtmp_count):
957 src = video.get('src')
959 return ([], rtmp_count)
960 bitrate = int_or_none(video.get('system-bitrate') or video.get('systemBitrate'), 1000)
961 width = int_or_none(video.get('width'))
962 height = int_or_none(video.get('height'))
963 proto = video.get('proto')
966 if base.startswith('rtmp'):
968 elif base.startswith('http'):
970 ext = video.get('ext')
972 return (self._extract_m3u8_formats(src, video_id, ext), rtmp_count)
973 elif proto == 'rtmp':
975 streamer = video.get('streamer') or base
980 'format_id': 'rtmp-%d' % (rtmp_count if bitrate is None else bitrate),
985 elif proto.startswith('http'):
994 def _live_title(self, name):
995 """ Generate the title for a live video """
996 now = datetime.datetime.now()
997 now_str = now.strftime("%Y-%m-%d %H:%M")
998 return name + ' ' + now_str
1000 def _int(self, v, name, fatal=False, **kwargs):
1001 res = int_or_none(v, **kwargs)
1002 if 'get_attr' in kwargs:
1003 print(getattr(v, kwargs['get_attr']))
1005 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
1007 raise ExtractorError(msg)
1009 self._downloader.report_warning(msg)
1012 def _float(self, v, name, fatal=False, **kwargs):
1013 res = float_or_none(v, **kwargs)
1015 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
1017 raise ExtractorError(msg)
1019 self._downloader.report_warning(msg)
1022 def _set_cookie(self, domain, name, value, expire_time=None):
1023 cookie = compat_cookiejar.Cookie(
1024 0, name, value, None, None, domain, None,
1025 None, '/', True, False, expire_time, '', None, None, None)
1026 self._downloader.cookiejar.set_cookie(cookie)
1028 def get_testcases(self, include_onlymatching=False):
1029 t = getattr(self, '_TEST', None)
1031 assert not hasattr(self, '_TESTS'), \
1032 '%s has _TEST and _TESTS' % type(self).__name__
1035 tests = getattr(self, '_TESTS', [])
1037 if not include_onlymatching and t.get('only_matching', False):
1039 t['name'] = type(self).__name__[:-len('IE')]
1042 def is_suitable(self, age_limit):
1043 """ Test whether the extractor is generally suitable for the given
1044 age limit (i.e. pornographic sites are not, all others usually are) """
1046 any_restricted = False
1047 for tc in self.get_testcases(include_onlymatching=False):
1048 if 'playlist' in tc:
1049 tc = tc['playlist'][0]
1050 is_restricted = age_restricted(
1051 tc.get('info_dict', {}).get('age_limit'), age_limit)
1052 if not is_restricted:
1054 any_restricted = any_restricted or is_restricted
1055 return not any_restricted
1057 def extract_subtitles(self, *args, **kwargs):
1058 if (self._downloader.params.get('writesubtitles', False) or
1059 self._downloader.params.get('listsubtitles')):
1060 return self._get_subtitles(*args, **kwargs)
1063 def _get_subtitles(self, *args, **kwargs):
1064 raise NotImplementedError("This method must be implemented by subclasses")
1066 def extract_automatic_captions(self, *args, **kwargs):
1067 if (self._downloader.params.get('writeautomaticsub', False) or
1068 self._downloader.params.get('listsubtitles')):
1069 return self._get_automatic_captions(*args, **kwargs)
1072 def _get_automatic_captions(self, *args, **kwargs):
1073 raise NotImplementedError("This method must be implemented by subclasses")
1075 def _subtitles_timecode(self, seconds):
1076 return '%02d:%02d:%02d.%03d' % (seconds / 3600, (seconds % 3600) / 60, seconds % 60, (seconds % 1) * 1000)
1079 class SearchInfoExtractor(InfoExtractor):
1081 Base class for paged search queries extractors.
1082 They accept urls in the format _SEARCH_KEY(|all|[0-9]):{query}
1083 Instances should define _SEARCH_KEY and _MAX_RESULTS.
1087 def _make_valid_url(cls):
1088 return r'%s(?P<prefix>|[1-9][0-9]*|all):(?P<query>[\s\S]+)' % cls._SEARCH_KEY
1091 def suitable(cls, url):
1092 return re.match(cls._make_valid_url(), url) is not None
1094 def _real_extract(self, query):
1095 mobj = re.match(self._make_valid_url(), query)
1097 raise ExtractorError('Invalid search query "%s"' % query)
1099 prefix = mobj.group('prefix')
1100 query = mobj.group('query')
1102 return self._get_n_results(query, 1)
1103 elif prefix == 'all':
1104 return self._get_n_results(query, self._MAX_RESULTS)
1108 raise ExtractorError('invalid download number %s for query "%s"' % (n, query))
1109 elif n > self._MAX_RESULTS:
1110 self._downloader.report_warning('%s returns max %i results (you requested %i)' % (self._SEARCH_KEY, self._MAX_RESULTS, n))
1111 n = self._MAX_RESULTS
1112 return self._get_n_results(query, n)
1114 def _get_n_results(self, query, n):
1115 """Get a specified number of results for a query"""
1116 raise NotImplementedError("This method must be implemented by subclasses")
1119 def SEARCH_KEY(self):
1120 return self._SEARCH_KEY