X-Git-Url: http://git.bitcoin.ninja/index.cgi?a=blobdiff_plain;ds=sidebyside;f=youtube_dl%2Fextractor%2Fyoutube.py;h=2e71795e780b5dbf80aae3ccfc41a605aa9748a3;hb=485047854376465f95309daad4966971f56728ef;hp=d66693c0c3079aaebfb3dcf832f47a548f866171;hpb=dc4e4f90a285dfebd060da55973ad3d1bc88ab7c;p=youtube-dl diff --git a/youtube_dl/extractor/youtube.py b/youtube_dl/extractor/youtube.py index d66693c0c..2e71795e7 100644 --- a/youtube_dl/extractor/youtube.py +++ b/youtube_dl/extractor/youtube.py @@ -673,6 +673,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor): }, }, # video_info is None (https://github.com/rg3/youtube-dl/issues/4421) + # YouTube Red ad is not captured for creator { 'url': '__2ABJjxzNo', 'info_dict': { @@ -1269,37 +1270,57 @@ class YoutubeIE(YoutubeBaseInfoExtractor): sub_lang_list[sub_lang] = sub_formats return sub_lang_list + def make_captions(sub_url, sub_langs): + parsed_sub_url = compat_urllib_parse_urlparse(sub_url) + caption_qs = compat_parse_qs(parsed_sub_url.query) + captions = {} + for sub_lang in sub_langs: + sub_formats = [] + for ext in self._SUBTITLE_FORMATS: + caption_qs.update({ + 'tlang': [sub_lang], + 'fmt': [ext], + }) + sub_url = compat_urlparse.urlunparse(parsed_sub_url._replace( + query=compat_urllib_parse_urlencode(caption_qs, True))) + sub_formats.append({ + 'url': sub_url, + 'ext': ext, + }) + captions[sub_lang] = sub_formats + return captions + + # New captions format as of 22.06.2017 + player_response = args.get('player_response') + if player_response and isinstance(player_response, compat_str): + player_response = self._parse_json( + player_response, video_id, fatal=False) + if player_response: + renderer = player_response['captions']['playerCaptionsTracklistRenderer'] + base_url = renderer['captionTracks'][0]['baseUrl'] + sub_lang_list = [] + for lang in renderer['translationLanguages']: + lang_code = lang.get('languageCode') + if lang_code: + sub_lang_list.append(lang_code) + return make_captions(base_url, sub_lang_list) + # Some videos don't provide ttsurl but rather caption_tracks and # caption_translation_languages (e.g. 20LmZk1hakA) + # Does not used anymore as of 22.06.2017 caption_tracks = args['caption_tracks'] caption_translation_languages = args['caption_translation_languages'] caption_url = compat_parse_qs(caption_tracks.split(',')[0])['u'][0] - parsed_caption_url = compat_urllib_parse_urlparse(caption_url) - caption_qs = compat_parse_qs(parsed_caption_url.query) - - sub_lang_list = {} + sub_lang_list = [] for lang in caption_translation_languages.split(','): lang_qs = compat_parse_qs(compat_urllib_parse_unquote_plus(lang)) sub_lang = lang_qs.get('lc', [None])[0] - if not sub_lang: - continue - sub_formats = [] - for ext in self._SUBTITLE_FORMATS: - caption_qs.update({ - 'tlang': [sub_lang], - 'fmt': [ext], - }) - sub_url = compat_urlparse.urlunparse(parsed_caption_url._replace( - query=compat_urllib_parse_urlencode(caption_qs, True))) - sub_formats.append({ - 'url': sub_url, - 'ext': ext, - }) - sub_lang_list[sub_lang] = sub_formats - return sub_lang_list + if sub_lang: + sub_lang_list.append(sub_lang) + return make_captions(caption_url, sub_lang_list) # An extractor error can be raise by the download process if there are # no automatic captions but there are subtitles - except (KeyError, ExtractorError): + except (KeyError, IndexError, ExtractorError): self._downloader.report_warning(err_msg) return {} @@ -1353,10 +1374,16 @@ class YoutubeIE(YoutubeBaseInfoExtractor): start_time = parse_duration(time_point) if start_time is None: continue + if start_time > duration: + break end_time = (duration if next_num == len(chapter_lines) else parse_duration(chapter_lines[next_num][1])) if end_time is None: continue + if end_time > duration: + end_time = duration + if start_time > end_time: + break chapter_title = re.sub( r']+>[^<]+', '', chapter_line).strip(' \t-') chapter_title = re.sub(r'\s+', ' ', chapter_title) @@ -1481,8 +1508,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor): if not video_info_webpage: continue get_video_info = compat_parse_qs(video_info_webpage) - if get_video_info.get('use_cipher_signature') != ['True']: - add_dash_mpd(get_video_info) + add_dash_mpd(get_video_info) if not video_info: video_info = get_video_info if 'token' in get_video_info: @@ -1624,7 +1650,21 @@ class YoutubeIE(YoutubeBaseInfoExtractor): video_webpage, 'license', default=None) m_music = re.search( - r']+class="title"[^>]*>\s*Music\s*\s*]*>\s*
  • (?P.+?) by (?P<creator>.+?)(?:\(.+?\))?</li', + r'''(?x) + <h4[^>]+class="title"[^>]*>\s*Music\s*</h4>\s* + <ul[^>]*>\s* + <li>(?P<title>.+?) + by (?P<creator>.+?) + (?: + \(.+?\)| + <a[^>]* + (?: + \bhref=["\']/red[^>]*>| # drop possible + >\s*Listen ad-free with YouTube Red # YouTube Red ad + ) + .*? + )?</li + ''', video_webpage) if m_music: video_alt_title = remove_quotes(unescapeHTML(m_music.group('title'))) @@ -1716,12 +1756,8 @@ class YoutubeIE(YoutubeBaseInfoExtractor): format_id = url_data['itag'][0] url = url_data['url'][0] - if 'sig' in url_data: - url += '&signature=' + url_data['sig'][0] - elif 's' in url_data: - encrypted_sig = url_data['s'][0] + if 's' in url_data or self._downloader.params.get('youtube_include_dash_manifest', True): ASSETS_RE = r'"assets":.+?"js":\s*("[^"]+")' - jsplayer_url_json = self._search_regex( ASSETS_RE, embed_webpage if age_gate else video_webpage, @@ -1742,6 +1778,11 @@ class YoutubeIE(YoutubeBaseInfoExtractor): video_webpage, 'age gate player URL') player_url = json.loads(player_url_json) + if 'sig' in url_data: + url += '&signature=' + url_data['sig'][0] + elif 's' in url_data: + encrypted_sig = url_data['s'][0] + if self._downloader.params.get('verbose'): if player_url is None: player_version = 'unknown'