X-Git-Url: https://git.rapsys.eu/youtubedl/blobdiff_plain/87a0165ca7e39af4dacb7ec637063b2cd35ae40b..1270059fa7cf720bc0533c2cdfe6370923ae4e20:/youtube_dl/extractor/vlive.py?ds=inline diff --git a/youtube_dl/extractor/vlive.py b/youtube_dl/extractor/vlive.py index acf9fda..64d0224 100644 --- a/youtube_dl/extractor/vlive.py +++ b/youtube_dl/extractor/vlive.py @@ -2,16 +2,23 @@ from __future__ import unicode_literals import re +import time +import itertools from .common import InfoExtractor +from ..compat import ( + compat_urllib_parse_urlencode, + compat_str, +) from ..utils import ( dict_get, ExtractorError, float_or_none, int_or_none, remove_start, + try_get, + urlencode_postdata, ) -from ..compat import compat_urllib_parse_urlencode class VLiveIE(InfoExtractor): @@ -42,24 +49,34 @@ class VLiveIE(InfoExtractor): }, }] + @classmethod + def suitable(cls, url): + return False if VLivePlaylistIE.suitable(url) else super(VLiveIE, cls).suitable(url) + def _real_extract(self, url): video_id = self._match_id(url) webpage = self._download_webpage( 'http://www.vlive.tv/video/%s' % video_id, video_id) - video_params = self._search_regex( - r'\bvlive\.video\.init\(([^)]+)\)', - webpage, 'video params') - status, _, _, live_params, long_video_id, key = re.split( - r'"\s*,\s*"', video_params)[2:8] + VIDEO_PARAMS_RE = r'\bvlive\.video\.init\(([^)]+)' + VIDEO_PARAMS_FIELD = 'video params' + + params = self._parse_json(self._search_regex( + VIDEO_PARAMS_RE, webpage, VIDEO_PARAMS_FIELD, default=''), video_id, + transform_source=lambda s: '[' + s + ']', fatal=False) + + if not params or len(params) < 7: + params = self._search_regex( + VIDEO_PARAMS_RE, webpage, VIDEO_PARAMS_FIELD) + params = [p.strip(r'"') for p in re.split(r'\s*,\s*', params)] + + status, long_video_id, key = params[2], params[5], params[6] status = remove_start(status, 'PRODUCT_') - if status == 'LIVE_ON_AIR' or status == 'BIG_EVENT_ON_AIR': - live_params = self._parse_json('"%s"' % live_params, video_id) - live_params = self._parse_json(live_params, video_id) - return self._live(video_id, webpage, live_params) - elif status == 'VOD_ON_AIR' or status == 'BIG_EVENT_INTRO': + if status in ('LIVE_ON_AIR', 'BIG_EVENT_ON_AIR'): + return self._live(video_id, webpage) + elif status in ('VOD_ON_AIR', 'BIG_EVENT_INTRO'): if long_video_id and key: return self._replay(video_id, webpage, long_video_id, key) else: @@ -89,7 +106,22 @@ class VLiveIE(InfoExtractor): 'thumbnail': thumbnail, } - def _live(self, video_id, webpage, live_params): + def _live(self, video_id, webpage): + init_page = self._download_webpage( + 'http://www.vlive.tv/video/init/view', + video_id, note='Downloading live webpage', + data=urlencode_postdata({'videoSeq': video_id}), + headers={ + 'Referer': 'http://www.vlive.tv/video/%s' % video_id, + 'Content-Type': 'application/x-www-form-urlencoded' + }) + + live_params = self._search_regex( + r'"liveStreamInfo"\s*:\s*(".*"),', + init_page, 'live stream info') + live_params = self._parse_json(live_params, video_id) + live_params = self._parse_json(live_params, video_id) + formats = [] for vid in live_params.get('resolutions', []): formats.extend(self._extract_m3u8_formats( @@ -98,10 +130,14 @@ class VLiveIE(InfoExtractor): fatal=False, live=True)) self._sort_formats(formats) - return dict(self._get_common_fields(webpage), - id=video_id, - formats=formats, - is_live=True) + info = self._get_common_fields(webpage) + info.update({ + 'title': self._live_title(info['title']), + 'id': video_id, + 'formats': formats, + 'is_live': True, + }) + return info def _replay(self, video_id, webpage, long_video_id, key): playinfo = self._download_json( @@ -135,8 +171,153 @@ class VLiveIE(InfoExtractor): 'ext': 'vtt', 'url': caption['source']}] - return dict(self._get_common_fields(webpage), - id=video_id, - formats=formats, - view_count=view_count, - subtitles=subtitles) + info = self._get_common_fields(webpage) + info.update({ + 'id': video_id, + 'formats': formats, + 'view_count': view_count, + 'subtitles': subtitles, + }) + return info + + +class VLiveChannelIE(InfoExtractor): + IE_NAME = 'vlive:channel' + _VALID_URL = r'https?://channels\.vlive\.tv/(?P[0-9A-Z]+)' + _TEST = { + 'url': 'http://channels.vlive.tv/FCD4B', + 'info_dict': { + 'id': 'FCD4B', + 'title': 'MAMAMOO', + }, + 'playlist_mincount': 110 + } + _APP_ID = '8c6cc7b45d2568fb668be6e05b6e5a3b' + + def _real_extract(self, url): + channel_code = self._match_id(url) + + webpage = self._download_webpage( + 'http://channels.vlive.tv/%s/video' % channel_code, channel_code) + + app_id = None + + app_js_url = self._search_regex( + r']+src=(["\'])(?Phttp.+?/app\.js.*?)\1', + webpage, 'app js', default=None, group='url') + + if app_js_url: + app_js = self._download_webpage( + app_js_url, channel_code, 'Downloading app JS', fatal=False) + if app_js: + app_id = self._search_regex( + r'Global\.VFAN_APP_ID\s*=\s*[\'"]([^\'"]+)[\'"]', + app_js, 'app id', default=None) + + app_id = app_id or self._APP_ID + + channel_info = self._download_json( + 'http://api.vfan.vlive.tv/vproxy/channelplus/decodeChannelCode', + channel_code, note='Downloading decode channel code', + query={ + 'app_id': app_id, + 'channelCode': channel_code, + '_': int(time.time()) + }) + + channel_seq = channel_info['result']['channelSeq'] + channel_name = None + entries = [] + + for page_num in itertools.count(1): + video_list = self._download_json( + 'http://api.vfan.vlive.tv/vproxy/channelplus/getChannelVideoList', + channel_code, note='Downloading channel list page #%d' % page_num, + query={ + 'app_id': app_id, + 'channelSeq': channel_seq, + # Large values of maxNumOfRows (~300 or above) may cause + # empty responses (see [1]), e.g. this happens for [2] that + # has more than 300 videos. + # 1. https://github.com/rg3/youtube-dl/issues/13830 + # 2. http://channels.vlive.tv/EDBF. + 'maxNumOfRows': 100, + '_': int(time.time()), + 'pageNo': page_num + } + ) + + if not channel_name: + channel_name = try_get( + video_list, + lambda x: x['result']['channelInfo']['channelName'], + compat_str) + + videos = try_get( + video_list, lambda x: x['result']['videoList'], list) + if not videos: + break + + for video in videos: + video_id = video.get('videoSeq') + if not video_id: + continue + video_id = compat_str(video_id) + entries.append( + self.url_result( + 'http://www.vlive.tv/video/%s' % video_id, + ie=VLiveIE.ie_key(), video_id=video_id)) + + return self.playlist_result( + entries, channel_code, channel_name) + + +class VLivePlaylistIE(InfoExtractor): + IE_NAME = 'vlive:playlist' + _VALID_URL = r'https?://(?:(?:www|m)\.)?vlive\.tv/video/(?P[0-9]+)/playlist/(?P[0-9]+)' + _TEST = { + 'url': 'http://www.vlive.tv/video/22867/playlist/22912', + 'info_dict': { + 'id': '22912', + 'title': 'Valentine Day Message from TWICE' + }, + 'playlist_mincount': 9 + } + + def _real_extract(self, url): + mobj = re.match(self._VALID_URL, url) + video_id, playlist_id = mobj.group('video_id', 'id') + + VIDEO_URL_TEMPLATE = 'http://www.vlive.tv/video/%s' + if self._downloader.params.get('noplaylist'): + self.to_screen( + 'Downloading just video %s because of --no-playlist' % video_id) + return self.url_result( + VIDEO_URL_TEMPLATE % video_id, + ie=VLiveIE.ie_key(), video_id=video_id) + + self.to_screen( + 'Downloading playlist %s - add --no-playlist to just download video' + % playlist_id) + + webpage = self._download_webpage( + 'http://www.vlive.tv/video/%s/playlist/%s' + % (video_id, playlist_id), playlist_id) + + item_ids = self._parse_json( + self._search_regex( + r'playlistVideoSeqs\s*=\s*(\[[^]]+\])', webpage, + 'playlist video seqs'), + playlist_id) + + entries = [ + self.url_result( + VIDEO_URL_TEMPLATE % item_id, ie=VLiveIE.ie_key(), + video_id=compat_str(item_id)) + for item_id in item_ids] + + playlist_name = self._html_search_regex( + r']+class="[^"]*multicam_playlist[^>]*>\s*]+>([^<]+)', + webpage, 'playlist title', fatal=False) + + return self.playlist_result(entries, playlist_id, playlist_name)