X-Git-Url: http://git.bitcoin.ninja/index.cgi?a=blobdiff_plain;f=youtube_dl%2Fextractor%2Fyoutube.py;h=87a5a452e011ba5270b39dd815eb6abc611781c3;hb=0f2999fe2b352795d54e6fcc4027e6a64ce5bc1d;hp=57b8fdff7204b1fbb5dc6d88742704638995b0eb;hpb=65697b3bf3bf6eaeb91a34e5308a6d2239118071;p=youtube-dl diff --git a/youtube_dl/extractor/youtube.py b/youtube_dl/extractor/youtube.py index 57b8fdff7..87a5a452e 100644 --- a/youtube_dl/extractor/youtube.py +++ b/youtube_dl/extractor/youtube.py @@ -40,7 +40,7 @@ class YoutubeBaseInfoExtractor(InfoExtractor): """Provide base functions for Youtube extractors""" _LOGIN_URL = 'https://accounts.google.com/ServiceLogin' _LANG_URL = r'https://www.youtube.com/?hl=en&persist_hl=1&gl=US&persist_gl=1&opt_out_ackd=1' - _AGE_URL = 'http://www.youtube.com/verify_age?next_url=/&gl=US&hl=en' + _AGE_URL = 'https://www.youtube.com/verify_age?next_url=/&gl=US&hl=en' _NETRC_MACHINE = 'youtube' # If True it will raise an error if no login info is provided _LOGIN_REQUIRED = False @@ -111,7 +111,8 @@ class YoutubeBaseInfoExtractor(InfoExtractor): 'next_url': '/', 'action_confirm': 'Confirm', } - req = compat_urllib_request.Request(self._AGE_URL, compat_urllib_parse.urlencode(age_form)) + req = compat_urllib_request.Request(self._AGE_URL, + compat_urllib_parse.urlencode(age_form).encode('ascii')) self._download_webpage( req, None, @@ -207,6 +208,12 @@ class YoutubeIE(YoutubeBaseInfoExtractor, SubtitlesInfoExtractor): '141': {'ext': 'm4a', 'format_note': 'DASH audio', 'vcodec': 'none', 'abr': 256, 'preference': -50}, # Dash webm + '167': {'ext': 'webm', 'height': 360, 'width': 640, 'format_note': 'DASH video', 'container': 'webm', 'vcodec': 'VP8', 'acodec': 'none', 'preference': -40}, + '168': {'ext': 'webm', 'height': 480, 'width': 854, 'format_note': 'DASH video', 'container': 'webm', 'vcodec': 'VP8', 'acodec': 'none', 'preference': -40}, + '169': {'ext': 'webm', 'height': 720, 'width': 1280, 'format_note': 'DASH video', 'container': 'webm', 'vcodec': 'VP8', 'acodec': 'none', 'preference': -40}, + '170': {'ext': 'webm', 'height': 1080, 'width': 1920, 'format_note': 'DASH video', 'container': 'webm', 'vcodec': 'VP8', 'acodec': 'none', 'preference': -40}, + '218': {'ext': 'webm', 'height': 480, 'width': 854, 'format_note': 'DASH video', 'container': 'webm', 'vcodec': 'VP8', 'acodec': 'none', 'preference': -40}, + '219': {'ext': 'webm', 'height': 480, 'width': 854, 'format_note': 'DASH video', 'container': 'webm', 'vcodec': 'VP8', 'acodec': 'none', 'preference': -40}, '242': {'ext': 'webm', 'height': 240, 'resolution': '240p', 'format_note': 'DASH webm', 'preference': -40}, '243': {'ext': 'webm', 'height': 360, 'resolution': '360p', 'format_note': 'DASH webm', 'preference': -40}, '244': {'ext': 'webm', 'height': 480, 'resolution': '480p', 'format_note': 'DASH webm', 'preference': -40}, @@ -276,16 +283,17 @@ class YoutubeIE(YoutubeBaseInfoExtractor, SubtitlesInfoExtractor): u"url": u"http://www.youtube.com/watch?v=a9LDPn-MO4I", u"file": u"a9LDPn-MO4I.m4a", u"note": u"256k DASH audio (format 141) via DASH manifest", - u"params": { - u"format": "141" - }, u"info_dict": { u"upload_date": "20121002", u"uploader_id": "8KVIDEO", u"description": "No description available.", u"uploader": "8KVIDEO", u"title": "UHDTV TEST 8K VIDEO.mp4" - } + }, + u"params": { + u"youtube_include_dash_manifest": True, + u"format": "141", + }, }, ] @@ -1007,7 +1015,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor, SubtitlesInfoExtractor): def _get_available_subtitles(self, video_id, webpage): try: sub_list = self._download_webpage( - 'http://video.google.com/timedtext?hl=en&type=list&v=%s' % video_id, + 'https://video.google.com/timedtext?hl=en&type=list&v=%s' % video_id, video_id, note=False) except ExtractorError as err: self._downloader.report_warning(u'unable to download video subtitles: %s' % compat_str(err)) @@ -1023,7 +1031,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor, SubtitlesInfoExtractor): 'fmt': self._downloader.params.get('subtitlesformat', 'srt'), 'name': unescapeHTML(l[0]).encode('utf-8'), }) - url = u'http://www.youtube.com/api/timedtext?' + params + url = u'https://www.youtube.com/api/timedtext?' + params sub_lang_list[lang] = url if not sub_lang_list: self._downloader.report_warning(u'video doesn\'t have subtitles') @@ -1289,7 +1297,8 @@ class YoutubeIE(YoutubeBaseInfoExtractor, SubtitlesInfoExtractor): 'url': video_real_url, 'player_url': player_url, } - dct.update(self._formats[itag]) + if itag in self._formats: + dct.update(self._formats[itag]) formats.append(dct) return formats @@ -1355,7 +1364,8 @@ class YoutubeIE(YoutubeBaseInfoExtractor, SubtitlesInfoExtractor): # Look for the DASH manifest dash_manifest_url_lst = video_info.get('dashmpd') - if dash_manifest_url_lst and dash_manifest_url_lst[0]: + if (dash_manifest_url_lst and dash_manifest_url_lst[0] and + self._downloader.params.get('youtube_include_dash_manifest', False)): try: dash_doc = self._download_xml( dash_manifest_url_lst[0], video_id, @@ -1520,7 +1530,7 @@ class YoutubeTopListIE(YoutubePlaylistIE): channel = mobj.group('chann') title = mobj.group('title') query = compat_urllib_parse.urlencode({'title': title}) - playlist_re = 'href="([^"]+?%s[^"]+?)"' % re.escape(query) + playlist_re = 'href="([^"]+?%s.*?)"' % re.escape(query) channel_page = self._download_webpage('https://www.youtube.com/%s' % channel, title) link = self._html_search_regex(playlist_re, channel_page, u'list') url = compat_urlparse.urljoin('https://www.youtube.com/', link) @@ -1545,7 +1555,7 @@ class YoutubeChannelIE(InfoExtractor): IE_DESC = u'YouTube.com channels' _VALID_URL = r"^(?:https?://)?(?:youtu\.be|(?:\w+\.)?youtube(?:-nocookie)?\.com)/channel/([0-9A-Za-z_-]+)" _MORE_PAGES_INDICATOR = 'yt-uix-load-more' - _MORE_PAGES_URL = 'http://www.youtube.com/c4_browse_ajax?action_load_more_videos=1&flow=list&paging=%s&view=0&sort=da&channel_id=%s' + _MORE_PAGES_URL = 'https://www.youtube.com/c4_browse_ajax?action_load_more_videos=1&flow=list&paging=%s&view=0&sort=da&channel_id=%s' IE_NAME = u'youtube:channel' def extract_videos_from_page(self, page): @@ -1601,9 +1611,9 @@ class YoutubeChannelIE(InfoExtractor): class YoutubeUserIE(InfoExtractor): IE_DESC = u'YouTube.com user videos (URL or "ytuser" keyword)' _VALID_URL = r'(?:(?:(?:https?://)?(?:\w+\.)?youtube\.com/(?:user/)?(?!(?:attribution_link|watch)(?:$|[^a-z_A-Z0-9-])))|ytuser:)(?!feed/)([A-Za-z0-9_-]+)' - _TEMPLATE_URL = 'http://gdata.youtube.com/feeds/api/users/%s' + _TEMPLATE_URL = 'https://gdata.youtube.com/feeds/api/users/%s' _GDATA_PAGE_SIZE = 50 - _GDATA_URL = 'http://gdata.youtube.com/feeds/api/users/%s/uploads?max-results=%d&start-index=%d&alt=json' + _GDATA_URL = 'https://gdata.youtube.com/feeds/api/users/%s/uploads?max-results=%d&start-index=%d&alt=json' IE_NAME = u'youtube:user' @classmethod @@ -1734,7 +1744,7 @@ class YoutubeFeedsInfoExtractor(YoutubeBaseInfoExtractor): action = 'action_load_system_feed' if self._PERSONAL_FEED: action = 'action_load_personal_feed' - return 'http://www.youtube.com/feed_ajax?%s=1&feed_name=%s&paging=%%s' % (action, self._FEED_NAME) + return 'https://www.youtube.com/feed_ajax?%s=1&feed_name=%s&paging=%%s' % (action, self._FEED_NAME) @property def IE_NAME(self): @@ -1803,7 +1813,10 @@ class YoutubeFavouritesIE(YoutubeBaseInfoExtractor): class YoutubeTruncatedURLIE(InfoExtractor): IE_NAME = 'youtube:truncated_url' IE_DESC = False # Do not list - _VALID_URL = r'(?:https?://)?[^/]+/watch\?feature=[a-z_]+$' + _VALID_URL = r'''(?x) + (?:https?://)?[^/]+/watch\?feature=[a-z_]+$| + (?:https?://)?(?:www\.)?youtube\.com/attribution_link\?a=[^&]+$ + ''' def _real_extract(self, url): raise ExtractorError(