X-Git-Url: https://git.rapsys.eu/youtubedl/blobdiff_plain/9ed7fe4fe4c445eb7d9f3197bb300d0db8f1807a..50ac49cc58db6cb928094be89415983aca6bbd34:/youtube_dl/extractor/common.py?ds=sidebyside diff --git a/youtube_dl/extractor/common.py b/youtube_dl/extractor/common.py index eb9bfa3..9427ff4 100644 --- a/youtube_dl/extractor/common.py +++ b/youtube_dl/extractor/common.py @@ -10,19 +10,22 @@ import re import socket import sys import time +import math from ..compat import ( compat_cookiejar, compat_cookies, + compat_etree_fromstring, compat_getpass, compat_http_client, + compat_os_name, + compat_str, compat_urllib_error, - compat_urllib_parse, - compat_urllib_parse_urlparse, + compat_urllib_parse_urlencode, + compat_urllib_request, compat_urlparse, - compat_str, - compat_etree_fromstring, ) +from ..downloader.f4m import remove_encrypted_media from ..utils import ( NO_DEFAULT, age_restricted, @@ -30,18 +33,30 @@ from ..utils import ( clean_html, compiled_regex_type, determine_ext, + error_to_compat_str, ExtractorError, fix_xml_ampersands, float_or_none, int_or_none, + parse_iso8601, RegexNotFoundError, sanitize_filename, sanitized_Request, unescapeHTML, unified_strdate, + unified_timestamp, url_basename, + xpath_element, xpath_text, xpath_with_ns, + determine_protocol, + parse_duration, + mimetype2ext, + update_Request, + update_url_query, + parse_m3u8_attributes, + extract_attributes, + parse_codecs, ) @@ -99,7 +114,7 @@ class InfoExtractor(object): * protocol The protocol that will be used for the actual download, lower-case. "http", "https", "rtsp", "rtmp", "rtmpe", - "m3u8", or "m3u8_native". + "m3u8", "m3u8_native" or "http_dash_segments". * preference Order number of this format. If this field is present and not None, the formats get sorted by this field, regardless of all other values. @@ -107,8 +122,9 @@ class InfoExtractor(object): -2 or smaller for less than default. < -1000 to hide the format (if there is another one which is strictly better) - * language_preference Is this in the correct requested - language? + * language Language code, e.g. "de" or "en-US". + * language_preference Is this in the language mentioned in + the URL? 10 if it's what the URL is about, -1 for default (don't know), -10 otherwise, other values reserved for now. @@ -148,15 +164,18 @@ class InfoExtractor(object): * "height" (optional, int) * "resolution" (optional, string "{width}x{height"}, deprecated) + * "filesize" (optional, int) thumbnail: Full URL to a video thumbnail image. description: Full video description. uploader: Full name of the video uploader. - creator: The main artist who created the video. + license: License name the video is licensed under. + creator: The creator of the video. release_date: The date (YYYYMMDD) when the video was released. timestamp: UNIX timestamp of the moment the video became available. upload_date: Video upload date (YYYYMMDD). If not explicitly set, calculated from timestamp. uploader_id: Nickname or id of the video uploader. + uploader_url: Full URL to a personal webpage of the video uploader. location: Physical location where the video was filmed. subtitles: The available subtitles as a dictionary in the format {language: subformats}. "subformats" is a list sorted from @@ -167,7 +186,7 @@ class InfoExtractor(object): "ext" will be calculated from URL if missing automatic_captions: Like 'subtitles', used by the YoutubeIE for automatically generated captions - duration: Length of the video in seconds, as an integer. + duration: Length of the video in seconds, as an integer or float. view_count: How many users have watched the video on the platform. like_count: Number of positive ratings of the video dislike_count: Number of negative ratings of the video @@ -199,6 +218,44 @@ class InfoExtractor(object): end_time: Time in seconds where the reproduction should end, as specified in the URL. + The following fields should only be used when the video belongs to some logical + chapter or section: + + chapter: Name or title of the chapter the video belongs to. + chapter_number: Number of the chapter the video belongs to, as an integer. + chapter_id: Id of the chapter the video belongs to, as a unicode string. + + The following fields should only be used when the video is an episode of some + series or programme: + + series: Title of the series or programme the video episode belongs to. + season: Title of the season the video episode belongs to. + season_number: Number of the season the video episode belongs to, as an integer. + season_id: Id of the season the video episode belongs to, as a unicode string. + episode: Title of the video episode. Unlike mandatory video title field, + this field should denote the exact title of the video episode + without any kind of decoration. + episode_number: Number of the video episode within a season, as an integer. + episode_id: Id of the video episode, as a unicode string. + + The following fields should only be used when the media is a track or a part of + a music album: + + track: Title of the track. + track_number: Number of the track within an album or a disc, as an integer. + track_id: Id of the track (useful in case of custom indexing, e.g. 6.iii), + as a unicode string. + artist: Artist(s) of the track. + genre: Genre(s) of the track. + album: Title of the album the track belongs to. + album_type: Type of the album (e.g. "Demo", "Full-length", "Split", "Compilation", etc). + album_artist: List of all artists appeared on the album (e.g. + "Ash Borer / Fell Voices" or "Various Artists", useful for splits + and compilations). + disc_number: Number of the disc or other physical medium the track belongs to, + as an integer. + release_year: Year (YYYY) when the album was released. + Unless mentioned otherwise, the fields should be Unicode strings. Unless mentioned otherwise, None is equivalent to absence of information. @@ -291,9 +348,9 @@ class InfoExtractor(object): except ExtractorError: raise except compat_http_client.IncompleteRead as e: - raise ExtractorError('A network error has occured.', cause=e, expected=True) + raise ExtractorError('A network error has occurred.', cause=e, expected=True) except (KeyError, StopIteration) as e: - raise ExtractorError('An extractor error has occured.', cause=e) + raise ExtractorError('An extractor error has occurred.', cause=e) def set_downloader(self, downloader): """Sets the downloader for this IE.""" @@ -316,7 +373,7 @@ class InfoExtractor(object): def IE_NAME(self): return compat_str(type(self).__name__[:-2]) - def _request_webpage(self, url_or_request, video_id, note=None, errnote=None, fatal=True): + def _request_webpage(self, url_or_request, video_id, note=None, errnote=None, fatal=True, data=None, headers={}, query={}): """ Returns the response handle """ if note is None: self.report_download_webpage(video_id) @@ -325,6 +382,14 @@ class InfoExtractor(object): self.to_screen('%s' % (note,)) else: self.to_screen('%s: %s' % (video_id, note)) + if isinstance(url_or_request, compat_urllib_request.Request): + url_or_request = update_Request( + url_or_request, data=data, headers=headers, query=query) + else: + if query: + url_or_request = update_url_query(url_or_request, query) + if data is not None or headers: + url_or_request = sanitized_Request(url_or_request, data, headers) try: return self._downloader.urlopen(url_or_request) except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err: @@ -332,20 +397,21 @@ class InfoExtractor(object): return False if errnote is None: errnote = 'Unable to download webpage' - errmsg = '%s: %s' % (errnote, compat_str(err)) + + errmsg = '%s: %s' % (errnote, error_to_compat_str(err)) if fatal: raise ExtractorError(errmsg, sys.exc_info()[2], cause=err) else: self._downloader.report_warning(errmsg) return False - def _download_webpage_handle(self, url_or_request, video_id, note=None, errnote=None, fatal=True, encoding=None): + def _download_webpage_handle(self, url_or_request, video_id, note=None, errnote=None, fatal=True, encoding=None, data=None, headers={}, query={}): """ Returns a tuple (page content as string, URL handle) """ # Strip hashes from the URL (#1038) if isinstance(url_or_request, (compat_str, str)): url_or_request = url_or_request.partition('#')[0] - urlh = self._request_webpage(url_or_request, video_id, note, errnote, fatal) + urlh = self._request_webpage(url_or_request, video_id, note, errnote, fatal, data=data, headers=headers, query=query) if urlh is False: assert not fatal return False @@ -398,7 +464,7 @@ class InfoExtractor(object): self.to_screen('Saving request to ' + filename) # Working around MAX_PATH limitation on Windows (see # http://msdn.microsoft.com/en-us/library/windows/desktop/aa365247(v=vs.85).aspx) - if os.name == 'nt': + if compat_os_name == 'nt': absfilepath = os.path.abspath(filename) if len(absfilepath) > 259: filename = '\\\\?\\' + absfilepath @@ -432,13 +498,13 @@ class InfoExtractor(object): return content - def _download_webpage(self, url_or_request, video_id, note=None, errnote=None, fatal=True, tries=1, timeout=5, encoding=None): + def _download_webpage(self, url_or_request, video_id, note=None, errnote=None, fatal=True, tries=1, timeout=5, encoding=None, data=None, headers={}, query={}): """ Returns the data of the page as a string """ success = False try_count = 0 while success is False: try: - res = self._download_webpage_handle(url_or_request, video_id, note, errnote, fatal, encoding=encoding) + res = self._download_webpage_handle(url_or_request, video_id, note, errnote, fatal, encoding=encoding, data=data, headers=headers, query=query) success = True except compat_http_client.IncompleteRead as e: try_count += 1 @@ -453,10 +519,10 @@ class InfoExtractor(object): def _download_xml(self, url_or_request, video_id, note='Downloading XML', errnote='Unable to download XML', - transform_source=None, fatal=True, encoding=None): + transform_source=None, fatal=True, encoding=None, data=None, headers={}, query={}): """Return the xml as an xml.etree.ElementTree.Element""" xml_string = self._download_webpage( - url_or_request, video_id, note, errnote, fatal=fatal, encoding=encoding) + url_or_request, video_id, note, errnote, fatal=fatal, encoding=encoding, data=data, headers=headers, query=query) if xml_string is False: return xml_string if transform_source: @@ -467,10 +533,10 @@ class InfoExtractor(object): note='Downloading JSON metadata', errnote='Unable to download JSON metadata', transform_source=None, - fatal=True, encoding=None): + fatal=True, encoding=None, data=None, headers={}, query={}): json_string = self._download_webpage( url_or_request, video_id, note, errnote, fatal=fatal, - encoding=encoding) + encoding=encoding, data=data, headers=headers, query=query) if (not fatal) and json_string is False: return None return self._parse_json( @@ -567,7 +633,7 @@ class InfoExtractor(object): if mobj: break - if not self._downloader.params.get('no_color') and os.name != 'nt' and sys.stderr.isatty(): + if not self._downloader.params.get('no_color') and compat_os_name != 'nt' and sys.stderr.isatty(): _name = '\033[0;34m%s\033[0m' % name else: _name = name @@ -596,6 +662,24 @@ class InfoExtractor(object): else: return res + def _get_netrc_login_info(self, netrc_machine=None): + username = None + password = None + netrc_machine = netrc_machine or self._NETRC_MACHINE + + if self._downloader.params.get('usenetrc', False): + try: + info = netrc.netrc().authenticators(netrc_machine) + if info is not None: + username = info[0] + password = info[2] + else: + raise netrc.NetrcParseError('No authenticators for %s' % netrc_machine) + except (IOError, netrc.NetrcParseError) as err: + self._downloader.report_warning('parsing .netrc: %s' % error_to_compat_str(err)) + + return (username, password) + def _get_login_info(self): """ Get the login info as (username, password) @@ -610,19 +694,11 @@ class InfoExtractor(object): downloader_params = self._downloader.params # Attempt to use provided username and password or .netrc data - if downloader_params.get('username', None) is not None: + if downloader_params.get('username') is not None: username = downloader_params['username'] password = downloader_params['password'] - elif downloader_params.get('usenetrc', False): - try: - info = netrc.netrc().authenticators(self._NETRC_MACHINE) - if info is not None: - username = info[0] - password = info[2] - else: - raise netrc.NetrcParseError('No authenticators for %s' % self._NETRC_MACHINE) - except (IOError, netrc.NetrcParseError) as err: - self._downloader.report_warning('parsing .netrc: %s' % compat_str(err)) + else: + username, password = self._get_netrc_login_info() return (username, password) @@ -637,7 +713,7 @@ class InfoExtractor(object): return None downloader_params = self._downloader.params - if downloader_params.get('twofactor', None) is not None: + if downloader_params.get('twofactor') is not None: return downloader_params['twofactor'] return compat_getpass('Type %s and press [Return]: ' % note) @@ -661,9 +737,14 @@ class InfoExtractor(object): [^>]+?content=(["\'])(?P.*?)\2''' % re.escape(prop) def _og_search_property(self, prop, html, name=None, **kargs): + if not isinstance(prop, (list, tuple)): + prop = [prop] if name is None: - name = 'OpenGraph %s' % prop - escaped = self._search_regex(self._og_regexes(prop), html, name, flags=re.DOTALL, **kargs) + name = 'OpenGraph %s' % prop[0] + og_regexes = [] + for p in prop: + og_regexes.extend(self._og_regexes(p)) + escaped = self._search_regex(og_regexes, html, name, flags=re.DOTALL, **kargs) if escaped is None: return None return unescapeHTML(escaped) @@ -687,10 +768,12 @@ class InfoExtractor(object): return self._og_search_property('url', html, **kargs) def _html_search_meta(self, name, html, display_name=None, fatal=False, **kwargs): + if not isinstance(name, (list, tuple)): + name = [name] if display_name is None: - display_name = name + display_name = name[0] return self._html_search_regex( - self._meta_regex(name), + [self._meta_regex(n) for n in name], html, display_name, fatal=fatal, group='content', **kwargs) def _dc_search_uploader(self, html): @@ -718,7 +801,7 @@ class InfoExtractor(object): 'mature': 17, 'restricted': 19, } - return RATING_TABLE.get(rating.lower(), None) + return RATING_TABLE.get(rating.lower()) def _family_friendly_search(self, html): # See http://schema.org/VideoObject @@ -733,12 +816,74 @@ class InfoExtractor(object): '0': 18, 'false': 18, } - return RATING_TABLE.get(family_friendly.lower(), None) + return RATING_TABLE.get(family_friendly.lower()) def _twitter_search_player(self, html): return self._html_search_meta('twitter:player', html, 'twitter card player') + def _search_json_ld(self, html, video_id, expected_type=None, **kwargs): + json_ld = self._search_regex( + r'(?s)]+type=(["\'])application/ld\+json\1[^>]*>(?P.+?)', + html, 'JSON-LD', group='json_ld', **kwargs) + default = kwargs.get('default', NO_DEFAULT) + if not json_ld: + return default if default is not NO_DEFAULT else {} + # JSON-LD may be malformed and thus `fatal` should be respected. + # At the same time `default` may be passed that assumes `fatal=False` + # for _search_regex. Let's simulate the same behavior here as well. + fatal = kwargs.get('fatal', True) if default == NO_DEFAULT else False + return self._json_ld(json_ld, video_id, fatal=fatal, expected_type=expected_type) + + def _json_ld(self, json_ld, video_id, fatal=True, expected_type=None): + if isinstance(json_ld, compat_str): + json_ld = self._parse_json(json_ld, video_id, fatal=fatal) + if not json_ld: + return {} + info = {} + if not isinstance(json_ld, (list, tuple, dict)): + return info + if isinstance(json_ld, dict): + json_ld = [json_ld] + for e in json_ld: + if e.get('@context') == 'http://schema.org': + item_type = e.get('@type') + if expected_type is not None and expected_type != item_type: + return info + if item_type == 'TVEpisode': + info.update({ + 'episode': unescapeHTML(e.get('name')), + 'episode_number': int_or_none(e.get('episodeNumber')), + 'description': unescapeHTML(e.get('description')), + }) + part_of_season = e.get('partOfSeason') + if isinstance(part_of_season, dict) and part_of_season.get('@type') == 'TVSeason': + info['season_number'] = int_or_none(part_of_season.get('seasonNumber')) + part_of_series = e.get('partOfSeries') or e.get('partOfTVSeries') + if isinstance(part_of_series, dict) and part_of_series.get('@type') == 'TVSeries': + info['series'] = unescapeHTML(part_of_series.get('name')) + elif item_type == 'Article': + info.update({ + 'timestamp': parse_iso8601(e.get('datePublished')), + 'title': unescapeHTML(e.get('headline')), + 'description': unescapeHTML(e.get('articleBody')), + }) + elif item_type == 'VideoObject': + info.update({ + 'url': e.get('contentUrl'), + 'title': unescapeHTML(e.get('name')), + 'description': unescapeHTML(e.get('description')), + 'thumbnail': e.get('thumbnailUrl'), + 'duration': parse_duration(e.get('duration')), + 'timestamp': unified_timestamp(e.get('uploadDate')), + 'filesize': float_or_none(e.get('contentSize')), + 'tbr': int_or_none(e.get('bitrate')), + 'width': int_or_none(e.get('width')), + 'height': int_or_none(e.get('height')), + }) + break + return dict((k, v) for k, v in info.items() if v is not None) + @staticmethod def _hidden_inputs(html): html = re.sub(r'', '', html) @@ -746,7 +891,7 @@ class InfoExtractor(object): for input in re.findall(r'(?i)]+)>', html): if not re.search(r'type=(["\'])(?:hidden|submit)\1', input): continue - name = re.search(r'name=(["\'])(?P.+?)\1', input) + name = re.search(r'(?:name|id)=(["\'])(?P.+?)\1', input) if not name: continue value = re.search(r'value=(["\'])(?P.*?)\1', input) @@ -765,6 +910,12 @@ class InfoExtractor(object): if not formats: raise ExtractorError('No video formats found') + for f in formats: + # Automatically determine tbr when missing based on abr and vbr (improves + # formats sorting in some cases) + if 'tbr' not in f and f.get('abr') is not None and f.get('vbr') is not None: + f['tbr'] = f['abr'] + f['vbr'] + def _formats_key(f): # TODO remove the following workaround from ..utils import determine_ext @@ -772,19 +923,23 @@ class InfoExtractor(object): f['ext'] = determine_ext(f['url']) if isinstance(field_preference, (list, tuple)): - return tuple(f.get(field) if f.get(field) is not None else -1 for field in field_preference) + return tuple( + f.get(field) + if f.get(field) is not None + else ('' if field == 'format_id' else -1) + for field in field_preference) preference = f.get('preference') if preference is None: - proto = f.get('protocol') - if proto is None: - proto = compat_urllib_parse_urlparse(f.get('url', '')).scheme - - preference = 0 if proto in ['http', 'https'] else -0.1 + preference = 0 if f.get('ext') in ['f4f', 'f4m']: # Not yet supported preference -= 0.5 + protocol = f.get('protocol') or determine_protocol(f) + proto_preference = 0 if protocol in ['http', 'https'] else (-0.5 if protocol == 'rtsp' else -0.1) + if f.get('vcodec') == 'none': # audio only + preference -= 50 if self._downloader.params.get('prefer_free_formats'): ORDER = ['aac', 'mp3', 'm4a', 'webm', 'ogg', 'opus'] else: @@ -795,6 +950,8 @@ class InfoExtractor(object): except ValueError: audio_ext_preference = -1 else: + if f.get('acodec') == 'none': # video only + preference -= 40 if self._downloader.params.get('prefer_free_formats'): ORDER = ['flv', 'mp4', 'webm'] else: @@ -814,6 +971,7 @@ class InfoExtractor(object): f.get('vbr') if f.get('vbr') is not None else -1, f.get('height') if f.get('height') is not None else -1, f.get('width') if f.get('width') is not None else -1, + proto_preference, ext_preference, f.get('abr') if f.get('abr') is not None else -1, audio_ext_preference, @@ -832,6 +990,16 @@ class InfoExtractor(object): item='%s video format' % f.get('format_id') if f.get('format_id') else 'video'), formats) + @staticmethod + def _remove_duplicate_formats(formats): + format_urls = set() + unique_formats = [] + for f in formats: + if f['url'] not in format_urls: + format_urls.add(f['url']) + unique_formats.append(f) + formats[:] = unique_formats + def _is_valid_url(self, url, video_id, item='video'): url = self._proto_relative_url(url, scheme='http:') # For now assume non HTTP(S) URLs always valid @@ -873,7 +1041,7 @@ class InfoExtractor(object): def _extract_f4m_formats(self, manifest_url, video_id, preference=None, f4m_id=None, transform_source=lambda s: fix_xml_ampersands(s).strip(), - fatal=True): + fatal=True, m3u8_id=None): manifest = self._download_xml( manifest_url, video_id, 'Downloading f4m manifest', 'Unable to download f4m manifest', @@ -883,7 +1051,21 @@ class InfoExtractor(object): fatal=fatal) if manifest is False: - return manifest + return [] + + return self._parse_f4m_formats( + manifest, manifest_url, video_id, preference=preference, f4m_id=f4m_id, + transform_source=transform_source, fatal=fatal, m3u8_id=m3u8_id) + + def _parse_f4m_formats(self, manifest, manifest_url, video_id, preference=None, f4m_id=None, + transform_source=lambda s: fix_xml_ampersands(s).strip(), + fatal=True, m3u8_id=None): + # currently youtube-dl cannot decode the playerVerificationChallenge as Akamai uses Adobe Alchemy + akamai_pv = manifest.find('{http://ns.adobe.com/f4m/1.0}pv-2.0') + if akamai_pv is not None and ';' in akamai_pv.text: + playerVerificationChallenge = akamai_pv.text.split(';')[0] + if playerVerificationChallenge.strip() != '': + return [] formats = [] manifest_version = '1.0' @@ -891,14 +1073,36 @@ class InfoExtractor(object): if not media_nodes: manifest_version = '2.0' media_nodes = manifest.findall('{http://ns.adobe.com/f4m/2.0}media') + # Remove unsupported DRM protected media from final formats + # rendition (see https://github.com/rg3/youtube-dl/issues/8573). + media_nodes = remove_encrypted_media(media_nodes) + if not media_nodes: + return formats base_url = xpath_text( manifest, ['{http://ns.adobe.com/f4m/1.0}baseURL', '{http://ns.adobe.com/f4m/2.0}baseURL'], 'base URL', default=None) if base_url: base_url = base_url.strip() + + bootstrap_info = xpath_element( + manifest, ['{http://ns.adobe.com/f4m/1.0}bootstrapInfo', '{http://ns.adobe.com/f4m/2.0}bootstrapInfo'], + 'bootstrap info', default=None) + for i, media_el in enumerate(media_nodes): - if manifest_version == '2.0': - media_url = media_el.attrib.get('href') or media_el.attrib.get('url') + tbr = int_or_none(media_el.attrib.get('bitrate')) + width = int_or_none(media_el.attrib.get('width')) + height = int_or_none(media_el.attrib.get('height')) + format_id = '-'.join(filter(None, [f4m_id, compat_str(i if tbr is None else tbr)])) + # If is present, the specified f4m is a + # stream-level manifest, and only set-level manifests may refer to + # external resources. See section 11.4 and section 4 of F4M spec + if bootstrap_info is None: + media_url = None + # @href is introduced in 2.0, see section 11.6 of F4M spec + if manifest_version == '2.0': + media_url = media_el.attrib.get('href') + if media_url is None: + media_url = media_el.attrib.get('url') if not media_url: continue manifest_url = ( @@ -908,40 +1112,58 @@ class InfoExtractor(object): # since bitrates in parent manifest (this one) and media_url manifest # may differ leading to inability to resolve the format by requested # bitrate in f4m downloader - if determine_ext(manifest_url) == 'f4m': + ext = determine_ext(manifest_url) + if ext == 'f4m': f4m_formats = self._extract_f4m_formats( - manifest_url, video_id, preference, f4m_id, fatal=fatal) - if f4m_formats: - formats.extend(f4m_formats) + manifest_url, video_id, preference=preference, f4m_id=f4m_id, + transform_source=transform_source, fatal=fatal) + # Sometimes stream-level manifest contains single media entry that + # does not contain any quality metadata (e.g. http://matchtv.ru/#live-player). + # At the same time parent's media entry in set-level manifest may + # contain it. We will copy it from parent in such cases. + if len(f4m_formats) == 1: + f = f4m_formats[0] + f.update({ + 'tbr': f.get('tbr') or tbr, + 'width': f.get('width') or width, + 'height': f.get('height') or height, + 'format_id': f.get('format_id') if not tbr else format_id, + }) + formats.extend(f4m_formats) + continue + elif ext == 'm3u8': + formats.extend(self._extract_m3u8_formats( + manifest_url, video_id, 'mp4', preference=preference, + m3u8_id=m3u8_id, fatal=fatal)) continue - tbr = int_or_none(media_el.attrib.get('bitrate')) formats.append({ - 'format_id': '-'.join(filter(None, [f4m_id, compat_str(i if tbr is None else tbr)])), + 'format_id': format_id, 'url': manifest_url, - 'ext': 'flv', + 'ext': 'flv' if bootstrap_info is not None else None, 'tbr': tbr, - 'width': int_or_none(media_el.attrib.get('width')), - 'height': int_or_none(media_el.attrib.get('height')), + 'width': width, + 'height': height, 'preference': preference, }) - self._sort_formats(formats) - return formats - def _extract_m3u8_formats(self, m3u8_url, video_id, ext=None, - entry_protocol='m3u8', preference=None, - m3u8_id=None, note=None, errnote=None, - fatal=True): - - formats = [{ + def _m3u8_meta_format(self, m3u8_url, ext=None, preference=None, m3u8_id=None): + return { 'format_id': '-'.join(filter(None, [m3u8_id, 'meta'])), 'url': m3u8_url, 'ext': ext, 'protocol': 'm3u8', - 'preference': preference - 1 if preference else -1, + 'preference': preference - 100 if preference else -100, 'resolution': 'multiple', 'format_note': 'Quality selection URL', - }] + } + + def _extract_m3u8_formats(self, m3u8_url, video_id, ext=None, + entry_protocol='m3u8', preference=None, + m3u8_id=None, note=None, errnote=None, + fatal=True, live=False): + + formats = [self._m3u8_meta_format(m3u8_url, ext, preference, m3u8_id)] format_url = lambda u: ( u @@ -954,28 +1176,38 @@ class InfoExtractor(object): errnote=errnote or 'Failed to download m3u8 information', fatal=fatal) if res is False: - return res + return [] m3u8_doc, urlh = res m3u8_url = urlh.geturl() + + # We should try extracting formats only from master playlists [1], i.e. + # playlists that describe available qualities. On the other hand media + # playlists [2] should be returned as is since they contain just the media + # without qualities renditions. + # Fortunately, master playlist can be easily distinguished from media + # playlist based on particular tags availability. As of [1, 2] master + # playlist tags MUST NOT appear in a media playist and vice versa. + # As of [3] #EXT-X-TARGETDURATION tag is REQUIRED for every media playlist + # and MUST NOT appear in master playlist thus we can clearly detect media + # playlist with this criterion. + # 1. https://tools.ietf.org/html/draft-pantos-http-live-streaming-17#section-4.3.4 + # 2. https://tools.ietf.org/html/draft-pantos-http-live-streaming-17#section-4.3.3 + # 3. https://tools.ietf.org/html/draft-pantos-http-live-streaming-17#section-4.3.3.1 + if '#EXT-X-TARGETDURATION' in m3u8_doc: # media playlist, return as is + return [{ + 'url': m3u8_url, + 'format_id': m3u8_id, + 'ext': ext, + 'protocol': entry_protocol, + 'preference': preference, + }] last_info = None last_media = None - kv_rex = re.compile( - r'(?P[a-zA-Z_-]+)=(?P"[^"]+"|[^",]+)(?:,|$)') for line in m3u8_doc.splitlines(): if line.startswith('#EXT-X-STREAM-INF:'): - last_info = {} - for m in kv_rex.finditer(line): - v = m.group('val') - if v.startswith('"'): - v = v[1:-1] - last_info[m.group('key')] = v + last_info = parse_m3u8_attributes(line) elif line.startswith('#EXT-X-MEDIA:'): - last_media = {} - for m in kv_rex.finditer(line): - v = m.group('val') - if v.startswith('"'): - v = v[1:-1] - last_media[m.group('key')] = v + last_media = parse_m3u8_attributes(line) elif line.startswith('#') or not line.strip(): continue else: @@ -986,35 +1218,45 @@ class InfoExtractor(object): format_id = [] if m3u8_id: format_id.append(m3u8_id) - last_media_name = last_media.get('NAME') if last_media and last_media.get('TYPE') != 'SUBTITLES' else None - format_id.append(last_media_name if last_media_name else '%d' % (tbr if tbr else len(formats))) + last_media_name = last_media.get('NAME') if last_media and last_media.get('TYPE') not in ('SUBTITLES', 'CLOSED-CAPTIONS') else None + # Despite specification does not mention NAME attribute for + # EXT-X-STREAM-INF it still sometimes may be present + stream_name = last_info.get('NAME') or last_media_name + # Bandwidth of live streams may differ over time thus making + # format_id unpredictable. So it's better to keep provided + # format_id intact. + if not live: + format_id.append(stream_name if stream_name else '%d' % (tbr if tbr else len(formats))) f = { 'format_id': '-'.join(format_id), 'url': format_url(line.strip()), 'tbr': tbr, 'ext': ext, + 'fps': float_or_none(last_info.get('FRAME-RATE')), 'protocol': entry_protocol, 'preference': preference, } - codecs = last_info.get('CODECS') - if codecs: - # TODO: looks like video codec is not always necessarily goes first - va_codecs = codecs.split(',') - if va_codecs[0]: - f['vcodec'] = va_codecs[0].partition('.')[0] - if len(va_codecs) > 1 and va_codecs[1]: - f['acodec'] = va_codecs[1].partition('.')[0] resolution = last_info.get('RESOLUTION') if resolution: width_str, height_str = resolution.split('x') f['width'] = int(width_str) f['height'] = int(height_str) + # Unified Streaming Platform + mobj = re.search( + r'audio.*?(?:%3D|=)(\d+)(?:-video.*?(?:%3D|=)(\d+))?', f['url']) + if mobj: + abr, vbr = mobj.groups() + abr, vbr = float_or_none(abr, 1000), float_or_none(vbr, 1000) + f.update({ + 'vbr': vbr, + 'abr': abr, + }) + f.update(parse_codecs(last_info.get('CODECS'))) if last_media is not None: f['m3u8_media'] = last_media last_media = None formats.append(f) last_info = {} - self._sort_formats(formats) return formats @staticmethod @@ -1029,8 +1271,8 @@ class InfoExtractor(object): out.append('{%s}%s' % (namespace, c)) return '/'.join(out) - def _extract_smil_formats(self, smil_url, video_id, fatal=True, f4m_params=None): - smil = self._download_smil(smil_url, video_id, fatal=fatal) + def _extract_smil_formats(self, smil_url, video_id, fatal=True, f4m_params=None, transform_source=None): + smil = self._download_smil(smil_url, video_id, fatal=fatal, transform_source=transform_source) if smil is False: assert not fatal @@ -1047,10 +1289,10 @@ class InfoExtractor(object): return {} return self._parse_smil(smil, smil_url, video_id, f4m_params=f4m_params) - def _download_smil(self, smil_url, video_id, fatal=True): + def _download_smil(self, smil_url, video_id, fatal=True, transform_source=None): return self._download_xml( smil_url, video_id, 'Downloading SMIL file', - 'Unable to download SMIL file', fatal=fatal) + 'Unable to download SMIL file', fatal=fatal, transform_source=transform_source) def _parse_smil(self, smil, smil_url, video_id, f4m_params=None): namespace = self._parse_smil_namespace(smil) @@ -1107,21 +1349,24 @@ class InfoExtractor(object): formats = [] rtmp_count = 0 http_count = 0 + m3u8_count = 0 - videos = smil.findall(self._xpath_ns('.//video', namespace)) - for video in videos: - src = video.get('src') - if not src: + srcs = [] + media = smil.findall(self._xpath_ns('.//video', namespace)) + smil.findall(self._xpath_ns('.//audio', namespace)) + for medium in media: + src = medium.get('src') + if not src or src in srcs: continue - - bitrate = float_or_none(video.get('system-bitrate') or video.get('systemBitrate'), 1000) - filesize = int_or_none(video.get('size') or video.get('fileSize')) - width = int_or_none(video.get('width')) - height = int_or_none(video.get('height')) - proto = video.get('proto') - ext = video.get('ext') + srcs.append(src) + + bitrate = float_or_none(medium.get('system-bitrate') or medium.get('systemBitrate'), 1000) + filesize = int_or_none(medium.get('size') or medium.get('fileSize')) + width = int_or_none(medium.get('width')) + height = int_or_none(medium.get('height')) + proto = medium.get('proto') + ext = medium.get('ext') src_ext = determine_ext(src) - streamer = video.get('streamer') or base + streamer = medium.get('streamer') or base if proto == 'rtmp' or streamer.startswith('rtmp'): rtmp_count += 1 @@ -1144,12 +1389,20 @@ class InfoExtractor(object): continue src_url = src if src.startswith('http') else compat_urlparse.urljoin(base, src) + src_url = src_url.strip() if proto == 'm3u8' or src_ext == 'm3u8': m3u8_formats = self._extract_m3u8_formats( src_url, video_id, ext or 'mp4', m3u8_id='hls', fatal=False) - if m3u8_formats: - formats.extend(m3u8_formats) + if len(m3u8_formats) == 1: + m3u8_count += 1 + m3u8_formats[0].update({ + 'format_id': 'hls-%d' % (m3u8_count if bitrate is None else bitrate), + 'tbr': bitrate, + 'width': width, + 'height': height, + }) + formats.extend(m3u8_formats) continue if src_ext == 'f4m': @@ -1160,10 +1413,8 @@ class InfoExtractor(object): 'plugin': 'flowplayer-3.2.0.1', } f4m_url += '&' if '?' in f4m_url else '?' - f4m_url += compat_urllib_parse.urlencode(f4m_params) - f4m_formats = self._extract_f4m_formats(f4m_url, video_id, f4m_id='hds', fatal=False) - if f4m_formats: - formats.extend(f4m_formats) + f4m_url += compat_urllib_parse_urlencode(f4m_params) + formats.extend(self._extract_f4m_formats(f4m_url, video_id, f4m_id='hds', fatal=False)) continue if src_url.startswith('http') and self._is_valid_url(src, video_id): @@ -1179,26 +1430,17 @@ class InfoExtractor(object): }) continue - self._sort_formats(formats) - return formats def _parse_smil_subtitles(self, smil, namespace=None, subtitles_lang='en'): + urls = [] subtitles = {} for num, textstream in enumerate(smil.findall(self._xpath_ns('.//textstream', namespace))): src = textstream.get('src') - if not src: + if not src or src in urls: continue - ext = textstream.get('ext') or determine_ext(src) - if not ext: - type_ = textstream.get('type') - SUBTITLES_TYPES = { - 'text/vtt': 'vtt', - 'text/srt': 'srt', - 'application/smptett+xml': 'tt', - } - if type_ in SUBTITLES_TYPES: - ext = SUBTITLES_TYPES[type_] + urls.append(src) + ext = textstream.get('ext') or mimetype2ext(textstream.get('type')) or determine_ext(src) lang = textstream.get('systemLanguage') or textstream.get('systemLanguageName') or textstream.get('lang') or subtitles_lang subtitles.setdefault(lang, []).append({ 'url': src, @@ -1249,10 +1491,270 @@ class InfoExtractor(object): }) return entries + def _extract_mpd_formats(self, mpd_url, video_id, mpd_id=None, note=None, errnote=None, fatal=True, formats_dict={}): + res = self._download_webpage_handle( + mpd_url, video_id, + note=note or 'Downloading MPD manifest', + errnote=errnote or 'Failed to download MPD manifest', + fatal=fatal) + if res is False: + return [] + mpd, urlh = res + mpd_base_url = re.match(r'https?://.+/', urlh.geturl()).group() + + return self._parse_mpd_formats( + compat_etree_fromstring(mpd.encode('utf-8')), mpd_id, mpd_base_url, formats_dict=formats_dict) + + def _parse_mpd_formats(self, mpd_doc, mpd_id=None, mpd_base_url='', formats_dict={}): + """ + Parse formats from MPD manifest. + References: + 1. MPEG-DASH Standard, ISO/IEC 23009-1:2014(E), + http://standards.iso.org/ittf/PubliclyAvailableStandards/c065274_ISO_IEC_23009-1_2014.zip + 2. https://en.wikipedia.org/wiki/Dynamic_Adaptive_Streaming_over_HTTP + """ + if mpd_doc.get('type') == 'dynamic': + return [] + + namespace = self._search_regex(r'(?i)^{([^}]+)?}MPD$', mpd_doc.tag, 'namespace', default=None) + + def _add_ns(path): + return self._xpath_ns(path, namespace) + + def is_drm_protected(element): + return element.find(_add_ns('ContentProtection')) is not None + + def extract_multisegment_info(element, ms_parent_info): + ms_info = ms_parent_info.copy() + segment_list = element.find(_add_ns('SegmentList')) + if segment_list is not None: + segment_urls_e = segment_list.findall(_add_ns('SegmentURL')) + if segment_urls_e: + ms_info['segment_urls'] = [segment.attrib['media'] for segment in segment_urls_e] + initialization = segment_list.find(_add_ns('Initialization')) + if initialization is not None: + ms_info['initialization_url'] = initialization.attrib['sourceURL'] + else: + segment_template = element.find(_add_ns('SegmentTemplate')) + if segment_template is not None: + start_number = segment_template.get('startNumber') + if start_number: + ms_info['start_number'] = int(start_number) + segment_timeline = segment_template.find(_add_ns('SegmentTimeline')) + if segment_timeline is not None: + s_e = segment_timeline.findall(_add_ns('S')) + if s_e: + ms_info['total_number'] = 0 + ms_info['s'] = [] + for s in s_e: + r = int(s.get('r', 0)) + ms_info['total_number'] += 1 + r + ms_info['s'].append({ + 't': int(s.get('t', 0)), + # @d is mandatory (see [1, 5.3.9.6.2, Table 17, page 60]) + 'd': int(s.attrib['d']), + 'r': r, + }) + else: + timescale = segment_template.get('timescale') + if timescale: + ms_info['timescale'] = int(timescale) + segment_duration = segment_template.get('duration') + if segment_duration: + ms_info['segment_duration'] = int(segment_duration) + media_template = segment_template.get('media') + if media_template: + ms_info['media_template'] = media_template + initialization = segment_template.get('initialization') + if initialization: + ms_info['initialization_url'] = initialization + else: + initialization = segment_template.find(_add_ns('Initialization')) + if initialization is not None: + ms_info['initialization_url'] = initialization.attrib['sourceURL'] + return ms_info + + mpd_duration = parse_duration(mpd_doc.get('mediaPresentationDuration')) + formats = [] + for period in mpd_doc.findall(_add_ns('Period')): + period_duration = parse_duration(period.get('duration')) or mpd_duration + period_ms_info = extract_multisegment_info(period, { + 'start_number': 1, + 'timescale': 1, + }) + for adaptation_set in period.findall(_add_ns('AdaptationSet')): + if is_drm_protected(adaptation_set): + continue + adaption_set_ms_info = extract_multisegment_info(adaptation_set, period_ms_info) + for representation in adaptation_set.findall(_add_ns('Representation')): + if is_drm_protected(representation): + continue + representation_attrib = adaptation_set.attrib.copy() + representation_attrib.update(representation.attrib) + # According to [1, 5.3.7.2, Table 9, page 41], @mimeType is mandatory + mime_type = representation_attrib['mimeType'] + content_type = mime_type.split('/')[0] + if content_type == 'text': + # TODO implement WebVTT downloading + pass + elif content_type == 'video' or content_type == 'audio': + base_url = '' + for element in (representation, adaptation_set, period, mpd_doc): + base_url_e = element.find(_add_ns('BaseURL')) + if base_url_e is not None: + base_url = base_url_e.text + base_url + if re.match(r'^https?://', base_url): + break + if mpd_base_url and not re.match(r'^https?://', base_url): + if not mpd_base_url.endswith('/') and not base_url.startswith('/'): + mpd_base_url += '/' + base_url = mpd_base_url + base_url + representation_id = representation_attrib.get('id') + lang = representation_attrib.get('lang') + url_el = representation.find(_add_ns('BaseURL')) + filesize = int_or_none(url_el.attrib.get('{http://youtube.com/yt/2012/10/10}contentLength') if url_el is not None else None) + f = { + 'format_id': '%s-%s' % (mpd_id, representation_id) if mpd_id else representation_id, + 'url': base_url, + 'ext': mimetype2ext(mime_type), + 'width': int_or_none(representation_attrib.get('width')), + 'height': int_or_none(representation_attrib.get('height')), + 'tbr': int_or_none(representation_attrib.get('bandwidth'), 1000), + 'asr': int_or_none(representation_attrib.get('audioSamplingRate')), + 'fps': int_or_none(representation_attrib.get('frameRate')), + 'vcodec': 'none' if content_type == 'audio' else representation_attrib.get('codecs'), + 'acodec': 'none' if content_type == 'video' else representation_attrib.get('codecs'), + 'language': lang if lang not in ('mul', 'und', 'zxx', 'mis') else None, + 'format_note': 'DASH %s' % content_type, + 'filesize': filesize, + } + representation_ms_info = extract_multisegment_info(representation, adaption_set_ms_info) + if 'segment_urls' not in representation_ms_info and 'media_template' in representation_ms_info: + if 'total_number' not in representation_ms_info and 'segment_duration': + segment_duration = float(representation_ms_info['segment_duration']) / float(representation_ms_info['timescale']) + representation_ms_info['total_number'] = int(math.ceil(float(period_duration) / segment_duration)) + media_template = representation_ms_info['media_template'] + media_template = media_template.replace('$RepresentationID$', representation_id) + media_template = re.sub(r'\$(Number|Bandwidth|Time)\$', r'%(\1)d', media_template) + media_template = re.sub(r'\$(Number|Bandwidth|Time)%([^$]+)\$', r'%(\1)\2', media_template) + media_template.replace('$$', '$') + + # As per [1, 5.3.9.4.4, Table 16, page 55] $Number$ and $Time$ + # can't be used at the same time + if '%(Number' in media_template: + representation_ms_info['segment_urls'] = [ + media_template % { + 'Number': segment_number, + 'Bandwidth': representation_attrib.get('bandwidth'), + } + for segment_number in range( + representation_ms_info['start_number'], + representation_ms_info['total_number'] + representation_ms_info['start_number'])] + else: + representation_ms_info['segment_urls'] = [] + segment_time = 0 + + def add_segment_url(): + representation_ms_info['segment_urls'].append( + media_template % { + 'Time': segment_time, + 'Bandwidth': representation_attrib.get('bandwidth'), + } + ) + + for num, s in enumerate(representation_ms_info['s']): + segment_time = s.get('t') or segment_time + add_segment_url() + for r in range(s.get('r', 0)): + segment_time += s['d'] + add_segment_url() + segment_time += s['d'] + if 'segment_urls' in representation_ms_info: + f.update({ + 'segment_urls': representation_ms_info['segment_urls'], + 'protocol': 'http_dash_segments', + }) + if 'initialization_url' in representation_ms_info: + initialization_url = representation_ms_info['initialization_url'].replace('$RepresentationID$', representation_id) + f.update({ + 'initialization_url': initialization_url, + }) + if not f.get('url'): + f['url'] = initialization_url + try: + existing_format = next( + fo for fo in formats + if fo['format_id'] == representation_id) + except StopIteration: + full_info = formats_dict.get(representation_id, {}).copy() + full_info.update(f) + formats.append(full_info) + else: + existing_format.update(f) + else: + self.report_warning('Unknown MIME type %s in DASH manifest' % mime_type) + return formats + + def _parse_html5_media_entries(self, base_url, webpage): + def absolute_url(video_url): + return compat_urlparse.urljoin(base_url, video_url) + + def parse_content_type(content_type): + if not content_type: + return {} + ctr = re.search(r'(?P[^/]+/[^;]+)(?:;\s*codecs="?(?P[^"]+))?', content_type) + if ctr: + mimetype, codecs = ctr.groups() + f = parse_codecs(codecs) + f['ext'] = mimetype2ext(mimetype) + return f + return {} + + entries = [] + for media_tag, media_type, media_content in re.findall(r'(?s)(<(?Pvideo|audio)[^>]*>)(.*?)', webpage): + media_info = { + 'formats': [], + 'subtitles': {}, + } + media_attributes = extract_attributes(media_tag) + src = media_attributes.get('src') + if src: + media_info['formats'].append({ + 'url': absolute_url(src), + 'vcodec': 'none' if media_type == 'audio' else None, + }) + media_info['thumbnail'] = media_attributes.get('poster') + if media_content: + for source_tag in re.findall(r']+>', media_content): + source_attributes = extract_attributes(source_tag) + src = source_attributes.get('src') + if not src: + continue + f = parse_content_type(source_attributes.get('type')) + f.update({ + 'url': absolute_url(src), + 'vcodec': 'none' if media_type == 'audio' else None, + }) + media_info['formats'].append(f) + for track_tag in re.findall(r']+>', media_content): + track_attributes = extract_attributes(track_tag) + kind = track_attributes.get('kind') + if not kind or kind == 'subtitles': + src = track_attributes.get('src') + if not src: + continue + lang = track_attributes.get('srclang') or track_attributes.get('lang') or track_attributes.get('label') + media_info['subtitles'].setdefault(lang, []).append({ + 'url': absolute_url(src), + }) + if media_info['formats']: + entries.append(media_info) + return entries + def _live_title(self, name): """ Generate the title for a live video """ now = datetime.datetime.now() - now_str = now.strftime("%Y-%m-%d %H:%M") + now_str = now.strftime('%Y-%m-%d %H:%M') return name + ' ' + now_str def _int(self, v, name, fatal=False, **kwargs): @@ -1309,7 +1811,7 @@ class InfoExtractor(object): any_restricted = False for tc in self.get_testcases(include_onlymatching=False): - if 'playlist' in tc: + if tc.get('playlist', []): tc = tc['playlist'][0] is_restricted = age_restricted( tc.get('info_dict', {}).get('age_limit'), age_limit) @@ -1325,7 +1827,7 @@ class InfoExtractor(object): return {} def _get_subtitles(self, *args, **kwargs): - raise NotImplementedError("This method must be implemented by subclasses") + raise NotImplementedError('This method must be implemented by subclasses') @staticmethod def _merge_subtitle_items(subtitle_list1, subtitle_list2): @@ -1351,7 +1853,23 @@ class InfoExtractor(object): return {} def _get_automatic_captions(self, *args, **kwargs): - raise NotImplementedError("This method must be implemented by subclasses") + raise NotImplementedError('This method must be implemented by subclasses') + + def mark_watched(self, *args, **kwargs): + if (self._downloader.params.get('mark_watched', False) and + (self._get_login_info()[0] is not None or + self._downloader.params.get('cookiefile') is not None)): + self._mark_watched(*args, **kwargs) + + def _mark_watched(self, *args, **kwargs): + raise NotImplementedError('This method must be implemented by subclasses') + + def geo_verification_headers(self): + headers = {} + geo_verification_proxy = self._downloader.params.get('geo_verification_proxy') + if geo_verification_proxy: + headers['Ytdl-request-proxy'] = geo_verification_proxy + return headers class SearchInfoExtractor(InfoExtractor): @@ -1391,7 +1909,7 @@ class SearchInfoExtractor(InfoExtractor): def _get_n_results(self, query, n): """Get a specified number of results for a query""" - raise NotImplementedError("This method must be implemented by subclasses") + raise NotImplementedError('This method must be implemented by subclasses') @property def SEARCH_KEY(self):