1 from __future__ import unicode_literals
13 import xml.etree.ElementTree
18 compat_urllib_parse_urlparse,
31 _NO_DEFAULT = object()
34 class InfoExtractor(object):
35 """Information Extractor class.
37 Information extractors are the classes that, given a URL, extract
38 information about the video (or videos) the URL refers to. This
39 information includes the real video URL, the video title, author and
40 others. The information is stored in a dictionary which is then
41 passed to the FileDownloader. The FileDownloader processes this
42 information possibly downloading the video to the file system, among
43 other possible outcomes.
45 The dictionaries must include the following fields:
48 title: Video title, unescaped.
50 Additionally, it must contain either a formats entry or a url one:
52 formats: A list of dictionaries for each format available, ordered
53 from worst to best quality.
56 * url Mandatory. The URL of the video file
57 * ext Will be calculated from url if missing
58 * format A human-readable description of the format
59 ("mp4 container with h264/opus").
60 Calculated from the format_id, width, height.
61 and format_note fields if missing.
62 * format_id A short description of the format
63 ("mp4_h264_opus" or "19").
64 Technically optional, but strongly recommended.
65 * format_note Additional info about the format
66 ("3D" or "DASH video")
67 * width Width of the video, if known
68 * height Height of the video, if known
69 * resolution Textual description of width and height
70 * tbr Average bitrate of audio and video in KBit/s
71 * abr Average audio bitrate in KBit/s
72 * acodec Name of the audio codec in use
73 * asr Audio sampling rate in Hertz
74 * vbr Average video bitrate in KBit/s
75 * vcodec Name of the video codec in use
76 * container Name of the container format
77 * filesize The number of bytes, if known in advance
78 * filesize_approx An estimate for the number of bytes
79 * player_url SWF Player URL (used for rtmpdump).
80 * protocol The protocol that will be used for the actual
82 "http", "https", "rtsp", "rtmp", "m3u8" or so.
83 * preference Order number of this format. If this field is
84 present and not None, the formats get sorted
85 by this field, regardless of all other values.
86 -1 for default (order by other properties),
87 -2 or smaller for less than default.
88 * quality Order number of the video quality of this
89 format, irrespective of the file format.
90 -1 for default (order by other properties),
91 -2 or smaller for less than default.
92 * http_referer HTTP Referer header value to set.
93 * http_method HTTP method to use for the download.
94 * http_headers A dictionary of additional HTTP headers
95 to add to the request.
96 * http_post_data Additional data to send with a POST
99 ext: Video filename extension.
100 format: The video format, defaults to ext (used for --get-format)
101 player_url: SWF Player URL (used for rtmpdump).
103 The following fields are optional:
105 display_id An alternative identifier for the video, not necessarily
106 unique, but available before title. Typically, id is
107 something like "4234987", title "Dancing naked mole rats",
108 and display_id "dancing-naked-mole-rats"
109 thumbnails: A list of dictionaries, with the following entries:
111 * "width" (optional, int)
112 * "height" (optional, int)
113 * "resolution" (optional, string "{width}x{height"},
115 thumbnail: Full URL to a video thumbnail image.
116 description: One-line video description.
117 uploader: Full name of the video uploader.
118 timestamp: UNIX timestamp of the moment the video became available.
119 upload_date: Video upload date (YYYYMMDD).
120 If not explicitly set, calculated from timestamp.
121 uploader_id: Nickname or id of the video uploader.
122 location: Physical location where the video was filmed.
123 subtitles: The subtitle file contents as a dictionary in the format
124 {language: subtitles}.
125 duration: Length of the video in seconds, as an integer.
126 view_count: How many users have watched the video on the platform.
127 like_count: Number of positive ratings of the video
128 dislike_count: Number of negative ratings of the video
129 comment_count: Number of comments on the video
130 age_limit: Age restriction for the video, as an integer (years)
131 webpage_url: The url to the video webpage, if given to youtube-dl it
132 should allow to get the same result again. (It will be set
133 by YoutubeDL if it's missing)
134 categories: A list of categories that the video falls in, for example
136 is_live: True, False, or None (=unknown). Whether this video is a
137 live stream that goes on instead of a fixed-length video.
139 Unless mentioned otherwise, the fields should be Unicode strings.
141 Subclasses of this one should re-define the _real_initialize() and
142 _real_extract() methods and define a _VALID_URL regexp.
143 Probably, they should also be added to the list of extractors.
145 Finally, the _WORKING attribute should be set to False for broken IEs
146 in order to warn the users and skip the tests.
153 def __init__(self, downloader=None):
154 """Constructor. Receives an optional downloader."""
156 self.set_downloader(downloader)
159 def suitable(cls, url):
160 """Receives a URL and returns True if suitable for this IE."""
162 # This does not use has/getattr intentionally - we want to know whether
163 # we have cached the regexp for *this* class, whereas getattr would also
164 # match the superclass
165 if '_VALID_URL_RE' not in cls.__dict__:
166 cls._VALID_URL_RE = re.compile(cls._VALID_URL)
167 return cls._VALID_URL_RE.match(url) is not None
170 def _match_id(cls, url):
171 if '_VALID_URL_RE' not in cls.__dict__:
172 cls._VALID_URL_RE = re.compile(cls._VALID_URL)
173 m = cls._VALID_URL_RE.match(url)
179 """Getter method for _WORKING."""
182 def initialize(self):
183 """Initializes an instance (authentication, etc)."""
185 self._real_initialize()
188 def extract(self, url):
189 """Extracts URL information and returns it in list of dicts."""
191 return self._real_extract(url)
193 def set_downloader(self, downloader):
194 """Sets the downloader for this IE."""
195 self._downloader = downloader
197 def _real_initialize(self):
198 """Real initialization process. Redefine in subclasses."""
201 def _real_extract(self, url):
202 """Real extraction process. Redefine in subclasses."""
207 """A string for getting the InfoExtractor with get_info_extractor"""
208 return cls.__name__[:-2]
212 return type(self).__name__[:-2]
214 def _request_webpage(self, url_or_request, video_id, note=None, errnote=None, fatal=True):
215 """ Returns the response handle """
217 self.report_download_webpage(video_id)
218 elif note is not False:
220 self.to_screen('%s' % (note,))
222 self.to_screen('%s: %s' % (video_id, note))
224 return self._downloader.urlopen(url_or_request)
225 except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err:
229 errnote = 'Unable to download webpage'
230 errmsg = '%s: %s' % (errnote, compat_str(err))
232 raise ExtractorError(errmsg, sys.exc_info()[2], cause=err)
234 self._downloader.report_warning(errmsg)
237 def _download_webpage_handle(self, url_or_request, video_id, note=None, errnote=None, fatal=True):
238 """ Returns a tuple (page content as string, URL handle) """
240 # Strip hashes from the URL (#1038)
241 if isinstance(url_or_request, (compat_str, str)):
242 url_or_request = url_or_request.partition('#')[0]
244 urlh = self._request_webpage(url_or_request, video_id, note, errnote, fatal)
248 content_type = urlh.headers.get('Content-Type', '')
249 webpage_bytes = urlh.read()
250 m = re.match(r'[a-zA-Z0-9_.-]+/[a-zA-Z0-9_.-]+\s*;\s*charset=(.+)', content_type)
252 encoding = m.group(1)
254 m = re.search(br'<meta[^>]+charset=[\'"]?([^\'")]+)[ /\'">]',
255 webpage_bytes[:1024])
257 encoding = m.group(1).decode('ascii')
258 elif webpage_bytes.startswith(b'\xff\xfe'):
262 if self._downloader.params.get('dump_intermediate_pages', False):
264 url = url_or_request.get_full_url()
265 except AttributeError:
267 self.to_screen('Dumping request to ' + url)
268 dump = base64.b64encode(webpage_bytes).decode('ascii')
269 self._downloader.to_screen(dump)
270 if self._downloader.params.get('write_pages', False):
272 url = url_or_request.get_full_url()
273 except AttributeError:
275 basen = '%s_%s' % (video_id, url)
277 h = '___' + hashlib.md5(basen.encode('utf-8')).hexdigest()
278 basen = basen[:240 - len(h)] + h
279 raw_filename = basen + '.dump'
280 filename = sanitize_filename(raw_filename, restricted=True)
281 self.to_screen('Saving request to ' + filename)
282 with open(filename, 'wb') as outf:
283 outf.write(webpage_bytes)
286 content = webpage_bytes.decode(encoding, 'replace')
288 content = webpage_bytes.decode('utf-8', 'replace')
290 if ('<title>Access to this site is blocked</title>' in content and
291 'Websense' in content[:512]):
292 msg = 'Access to this webpage has been blocked by Websense filtering software in your network.'
293 blocked_iframe = self._html_search_regex(
294 r'<iframe src="([^"]+)"', content,
295 'Websense information URL', default=None)
297 msg += ' Visit %s for more details' % blocked_iframe
298 raise ExtractorError(msg, expected=True)
300 return (content, urlh)
302 def _download_webpage(self, url_or_request, video_id, note=None, errnote=None, fatal=True):
303 """ Returns the data of the page as a string """
304 res = self._download_webpage_handle(url_or_request, video_id, note, errnote, fatal)
311 def _download_xml(self, url_or_request, video_id,
312 note='Downloading XML', errnote='Unable to download XML',
313 transform_source=None, fatal=True):
314 """Return the xml as an xml.etree.ElementTree.Element"""
315 xml_string = self._download_webpage(
316 url_or_request, video_id, note, errnote, fatal=fatal)
317 if xml_string is False:
320 xml_string = transform_source(xml_string)
321 return xml.etree.ElementTree.fromstring(xml_string.encode('utf-8'))
323 def _download_json(self, url_or_request, video_id,
324 note='Downloading JSON metadata',
325 errnote='Unable to download JSON metadata',
326 transform_source=None,
328 json_string = self._download_webpage(
329 url_or_request, video_id, note, errnote, fatal=fatal)
330 if (not fatal) and json_string is False:
333 json_string = transform_source(json_string)
335 return json.loads(json_string)
336 except ValueError as ve:
337 raise ExtractorError('Failed to download JSON', cause=ve)
339 def report_warning(self, msg, video_id=None):
340 idstr = '' if video_id is None else '%s: ' % video_id
341 self._downloader.report_warning(
342 '[%s] %s%s' % (self.IE_NAME, idstr, msg))
344 def to_screen(self, msg):
345 """Print msg to screen, prefixing it with '[ie_name]'"""
346 self._downloader.to_screen('[%s] %s' % (self.IE_NAME, msg))
348 def report_extraction(self, id_or_name):
349 """Report information extraction."""
350 self.to_screen('%s: Extracting information' % id_or_name)
352 def report_download_webpage(self, video_id):
353 """Report webpage download."""
354 self.to_screen('%s: Downloading webpage' % video_id)
356 def report_age_confirmation(self):
357 """Report attempt to confirm age."""
358 self.to_screen('Confirming age')
360 def report_login(self):
361 """Report attempt to log in."""
362 self.to_screen('Logging in')
364 #Methods for following #608
366 def url_result(url, ie=None, video_id=None):
367 """Returns a url that points to a page that should be processed"""
368 #TODO: ie should be the class used for getting the info
369 video_info = {'_type': 'url',
372 if video_id is not None:
373 video_info['id'] = video_id
376 def playlist_result(entries, playlist_id=None, playlist_title=None):
377 """Returns a playlist"""
378 video_info = {'_type': 'playlist',
381 video_info['id'] = playlist_id
383 video_info['title'] = playlist_title
386 def _search_regex(self, pattern, string, name, default=_NO_DEFAULT, fatal=True, flags=0):
388 Perform a regex search on the given string, using a single or a list of
389 patterns returning the first matching group.
390 In case of failure return a default value or raise a WARNING or a
391 RegexNotFoundError, depending on fatal, specifying the field name.
393 if isinstance(pattern, (str, compat_str, compiled_regex_type)):
394 mobj = re.search(pattern, string, flags)
397 mobj = re.search(p, string, flags)
401 if os.name != 'nt' and sys.stderr.isatty():
402 _name = '\033[0;34m%s\033[0m' % name
407 # return the first matching group
408 return next(g for g in mobj.groups() if g is not None)
409 elif default is not _NO_DEFAULT:
412 raise RegexNotFoundError('Unable to extract %s' % _name)
414 self._downloader.report_warning('unable to extract %s; '
415 'please report this issue on http://yt-dl.org/bug' % _name)
418 def _html_search_regex(self, pattern, string, name, default=_NO_DEFAULT, fatal=True, flags=0):
420 Like _search_regex, but strips HTML tags and unescapes entities.
422 res = self._search_regex(pattern, string, name, default, fatal, flags)
424 return clean_html(res).strip()
428 def _get_login_info(self):
430 Get the the login info as (username, password)
431 It will look in the netrc file using the _NETRC_MACHINE value
432 If there's no info available, return (None, None)
434 if self._downloader is None:
439 downloader_params = self._downloader.params
441 # Attempt to use provided username and password or .netrc data
442 if downloader_params.get('username', None) is not None:
443 username = downloader_params['username']
444 password = downloader_params['password']
445 elif downloader_params.get('usenetrc', False):
447 info = netrc.netrc().authenticators(self._NETRC_MACHINE)
452 raise netrc.NetrcParseError('No authenticators for %s' % self._NETRC_MACHINE)
453 except (IOError, netrc.NetrcParseError) as err:
454 self._downloader.report_warning('parsing .netrc: %s' % compat_str(err))
456 return (username, password)
458 def _get_tfa_info(self):
460 Get the two-factor authentication info
461 TODO - asking the user will be required for sms/phone verify
462 currently just uses the command line option
463 If there's no info available, return None
465 if self._downloader is None:
467 downloader_params = self._downloader.params
469 if downloader_params.get('twofactor', None) is not None:
470 return downloader_params['twofactor']
474 # Helper functions for extracting OpenGraph info
476 def _og_regexes(prop):
477 content_re = r'content=(?:"([^>]+?)"|\'([^>]+?)\')'
478 property_re = r'(?:name|property)=[\'"]og:%s[\'"]' % re.escape(prop)
479 template = r'<meta[^>]+?%s[^>]+?%s'
481 template % (property_re, content_re),
482 template % (content_re, property_re),
485 def _og_search_property(self, prop, html, name=None, **kargs):
487 name = 'OpenGraph %s' % prop
488 escaped = self._search_regex(self._og_regexes(prop), html, name, flags=re.DOTALL, **kargs)
491 return unescapeHTML(escaped)
493 def _og_search_thumbnail(self, html, **kargs):
494 return self._og_search_property('image', html, 'thumbnail url', fatal=False, **kargs)
496 def _og_search_description(self, html, **kargs):
497 return self._og_search_property('description', html, fatal=False, **kargs)
499 def _og_search_title(self, html, **kargs):
500 return self._og_search_property('title', html, **kargs)
502 def _og_search_video_url(self, html, name='video url', secure=True, **kargs):
503 regexes = self._og_regexes('video') + self._og_regexes('video:url')
505 regexes = self._og_regexes('video:secure_url') + regexes
506 return self._html_search_regex(regexes, html, name, **kargs)
508 def _og_search_url(self, html, **kargs):
509 return self._og_search_property('url', html, **kargs)
511 def _html_search_meta(self, name, html, display_name=None, fatal=False, **kwargs):
512 if display_name is None:
514 return self._html_search_regex(
516 (?=[^>]+(?:itemprop|name|property)=["\']?%s["\']?)
517 [^>]+content=["\']([^"\']+)["\']''' % re.escape(name),
518 html, display_name, fatal=fatal, **kwargs)
520 def _dc_search_uploader(self, html):
521 return self._html_search_meta('dc.creator', html, 'uploader')
523 def _rta_search(self, html):
524 # See http://www.rtalabel.org/index.php?content=howtofaq#single
525 if re.search(r'(?ix)<meta\s+name="rating"\s+'
526 r' content="RTA-5042-1996-1400-1577-RTA"',
531 def _media_rating_search(self, html):
532 # See http://www.tjg-designs.com/WP/metadata-code-examples-adding-metadata-to-your-web-pages/
533 rating = self._html_search_meta('rating', html)
545 return RATING_TABLE.get(rating.lower(), None)
547 def _twitter_search_player(self, html):
548 return self._html_search_meta('twitter:player', html,
549 'twitter card player')
551 def _sort_formats(self, formats):
553 raise ExtractorError('No video formats found')
556 # TODO remove the following workaround
557 from ..utils import determine_ext
558 if not f.get('ext') and 'url' in f:
559 f['ext'] = determine_ext(f['url'])
561 preference = f.get('preference')
562 if preference is None:
563 proto = f.get('protocol')
565 proto = compat_urllib_parse_urlparse(f.get('url', '')).scheme
567 preference = 0 if proto in ['http', 'https'] else -0.1
568 if f.get('ext') in ['f4f', 'f4m']: # Not yet supported
571 if f.get('vcodec') == 'none': # audio only
572 if self._downloader.params.get('prefer_free_formats'):
573 ORDER = ['aac', 'mp3', 'm4a', 'webm', 'ogg', 'opus']
575 ORDER = ['webm', 'opus', 'ogg', 'mp3', 'aac', 'm4a']
578 audio_ext_preference = ORDER.index(f['ext'])
580 audio_ext_preference = -1
582 if self._downloader.params.get('prefer_free_formats'):
583 ORDER = ['flv', 'mp4', 'webm']
585 ORDER = ['webm', 'flv', 'mp4']
587 ext_preference = ORDER.index(f['ext'])
590 audio_ext_preference = 0
594 f.get('quality') if f.get('quality') is not None else -1,
595 f.get('height') if f.get('height') is not None else -1,
596 f.get('width') if f.get('width') is not None else -1,
598 f.get('tbr') if f.get('tbr') is not None else -1,
599 f.get('vbr') if f.get('vbr') is not None else -1,
600 f.get('abr') if f.get('abr') is not None else -1,
601 audio_ext_preference,
602 f.get('filesize') if f.get('filesize') is not None else -1,
603 f.get('filesize_approx') if f.get('filesize_approx') is not None else -1,
606 formats.sort(key=_formats_key)
608 def http_scheme(self):
609 """ Either "https:" or "https:", depending on the user's preferences """
612 if self._downloader.params.get('prefer_insecure', False)
615 def _proto_relative_url(self, url, scheme=None):
618 if url.startswith('//'):
620 scheme = self.http_scheme()
625 def _sleep(self, timeout, video_id, msg_template=None):
626 if msg_template is None:
627 msg_template = '%(video_id)s: Waiting for %(timeout)s seconds'
628 msg = msg_template % {'video_id': video_id, 'timeout': timeout}
632 def _extract_f4m_formats(self, manifest_url, video_id):
633 manifest = self._download_xml(
634 manifest_url, video_id, 'Downloading f4m manifest',
635 'Unable to download f4m manifest')
638 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/1.0}media')
639 for i, media_el in enumerate(media_nodes):
640 tbr = int_or_none(media_el.attrib.get('bitrate'))
641 format_id = 'f4m-%d' % (i if tbr is None else tbr)
643 'format_id': format_id,
647 'width': int_or_none(media_el.attrib.get('width')),
648 'height': int_or_none(media_el.attrib.get('height')),
650 self._sort_formats(formats)
654 def _extract_m3u8_formats(self, m3u8_url, video_id, ext=None,
655 entry_protocol='m3u8', preference=None):
658 'format_id': 'm3u8-meta',
663 'resolution': 'multiple',
664 'format_note': 'Quality selection URL',
667 format_url = lambda u: (
669 if re.match(r'^https?://', u)
670 else compat_urlparse.urljoin(m3u8_url, u))
672 m3u8_doc = self._download_webpage(m3u8_url, video_id)
675 r'(?P<key>[a-zA-Z_-]+)=(?P<val>"[^"]+"|[^",]+)(?:,|$)')
676 for line in m3u8_doc.splitlines():
677 if line.startswith('#EXT-X-STREAM-INF:'):
679 for m in kv_rex.finditer(line):
681 if v.startswith('"'):
683 last_info[m.group('key')] = v
684 elif line.startswith('#') or not line.strip():
687 if last_info is None:
688 formats.append({'url': format_url(line)})
690 tbr = int_or_none(last_info.get('BANDWIDTH'), scale=1000)
693 'format_id': 'm3u8-%d' % (tbr if tbr else len(formats)),
694 'url': format_url(line.strip()),
697 'protocol': entry_protocol,
698 'preference': preference,
700 codecs = last_info.get('CODECS')
702 # TODO: looks like video codec is not always necessarily goes first
703 va_codecs = codecs.split(',')
705 f['vcodec'] = va_codecs[0].partition('.')[0]
706 if len(va_codecs) > 1 and va_codecs[1]:
707 f['acodec'] = va_codecs[1].partition('.')[0]
708 resolution = last_info.get('RESOLUTION')
710 width_str, height_str = resolution.split('x')
711 f['width'] = int(width_str)
712 f['height'] = int(height_str)
715 self._sort_formats(formats)
718 def _live_title(self, name):
719 """ Generate the title for a live video """
720 now = datetime.datetime.now()
721 now_str = now.strftime("%Y-%m-%d %H:%M")
722 return name + ' ' + now_str
724 def _int(self, v, name, fatal=False, **kwargs):
725 res = int_or_none(v, **kwargs)
726 if 'get_attr' in kwargs:
727 print(getattr(v, kwargs['get_attr']))
729 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
731 raise ExtractorError(msg)
733 self._downloader.report_warning(msg)
736 def _float(self, v, name, fatal=False, **kwargs):
737 res = float_or_none(v, **kwargs)
739 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
741 raise ExtractorError(msg)
743 self._downloader.report_warning(msg)
747 class SearchInfoExtractor(InfoExtractor):
749 Base class for paged search queries extractors.
750 They accept urls in the format _SEARCH_KEY(|all|[0-9]):{query}
751 Instances should define _SEARCH_KEY and _MAX_RESULTS.
755 def _make_valid_url(cls):
756 return r'%s(?P<prefix>|[1-9][0-9]*|all):(?P<query>[\s\S]+)' % cls._SEARCH_KEY
759 def suitable(cls, url):
760 return re.match(cls._make_valid_url(), url) is not None
762 def _real_extract(self, query):
763 mobj = re.match(self._make_valid_url(), query)
765 raise ExtractorError('Invalid search query "%s"' % query)
767 prefix = mobj.group('prefix')
768 query = mobj.group('query')
770 return self._get_n_results(query, 1)
771 elif prefix == 'all':
772 return self._get_n_results(query, self._MAX_RESULTS)
776 raise ExtractorError('invalid download number %s for query "%s"' % (n, query))
777 elif n > self._MAX_RESULTS:
778 self._downloader.report_warning('%s returns max %i results (you requested %i)' % (self._SEARCH_KEY, self._MAX_RESULTS, n))
779 n = self._MAX_RESULTS
780 return self._get_n_results(query, n)
782 def _get_n_results(self, query, n):
783 """Get a specified number of results for a query"""
784 raise NotImplementedError("This method must be implemented by subclasses")
787 def SEARCH_KEY(self):
788 return self._SEARCH_KEY