[vlive] fix extraction

youtube_dl/extractor/vlive.py

@@ -1,25 +1,30 @@
 # coding: utf-8
 from __future__ import unicode_literals
 
-import re
-import time
 import itertools
+import json
 
-from .common import InfoExtractor
 from .naver import NaverBaseIE
-from ..compat import compat_str
+from ..compat import (
+    compat_HTTPError,
+    compat_str,
+)
 from ..utils import (
     ExtractorError,
+    int_or_none,
     merge_dicts,
-    remove_start,
     try_get,
     urlencode_postdata,
 )
 
 
-class VLiveIE(NaverBaseIE):
+class VLiveBaseIE(NaverBaseIE):
+    _APP_ID = '8c6cc7b45d2568fb668be6e05b6e5a3b'
+
+
+class VLiveIE(VLiveBaseIE):
     IE_NAME = 'vlive'
-    _VALID_URL = r'https?://(?:(?:www|m)\.)?vlive\.tv/video/(?P<id>[0-9]+)'
+    _VALID_URL = r'https?://(?:(?:www|m)\.)?vlive\.tv/(?:video|embed)/(?P<id>[0-9]+)'
     _NETRC_MACHINE = 'vlive'
     _TESTS = [{
         'url': 'http://www.vlive.tv/video/1326',
@@ -27,7 +32,7 @@ class VLiveIE(NaverBaseIE):
         'info_dict': {
             'id': '1326',
             'ext': 'mp4',
-            'title': "[V LIVE] Girl's Day's Broadcast",
+            'title': "Girl's Day's Broadcast",
             'creator': "Girl's Day",
             'view_count': int,
             'uploader_id': 'muploader_a',
@@ -37,7 +42,7 @@ class VLiveIE(NaverBaseIE):
         'info_dict': {
             'id': '16937',
             'ext': 'mp4',
-            'title': '[V LIVE] 첸백시 걍방',
+            'title': '첸백시 걍방',
             'creator': 'EXO',
             'view_count': int,
             'subtitles': 'mincount:12',
@@ -58,12 +63,11 @@ class VLiveIE(NaverBaseIE):
             'subtitles': 'mincount:10',
         },
         'skip': 'This video is only available for CH+ subscribers',
+    }, {
+        'url': 'https://www.vlive.tv/embed/1326',
+        'only_matching': True,
     }]
 
-    @classmethod
-    def suitable(cls, url):
-        return False if VLivePlaylistIE.suitable(url) else super(VLiveIE, cls).suitable(url)
-
     def _real_initialize(self):
         self._login()
 
@@ -95,173 +99,122 @@ class VLiveIE(NaverBaseIE):
         if not is_logged_in():
             raise ExtractorError('Unable to log in', expected=True)
 
+    def _call_api(self, path_template, video_id, fields=None):
+        query = {'appId': self._APP_ID}
+        if fields:
+            query['fields'] = fields
+        return self._download_json(
+            'https://www.vlive.tv/globalv-web/vam-web/' + path_template % video_id, video_id,
+            'Downloading %s JSON metadata' % path_template.split('/')[-1].split('-')[0],
+            headers={'Referer': 'https://www.vlive.tv/'}, query=query)
+
     def _real_extract(self, url):
         video_id = self._match_id(url)
 
-        webpage = self._download_webpage(
+        try:
-            'https://www.vlive.tv/video/%s' % video_id, video_id)
+            post = self._call_api(
+                'post/v1.0/officialVideoPost-%s', video_id,
+                'author{nickname},channel{channelCode,channelName},officialVideo{commentCount,exposeStatus,likeCount,playCount,playTime,status,title,type,vodId}')
+        except ExtractorError as e:
+            if isinstance(e.cause, compat_HTTPError) and e.cause.code == 403:
+                self.raise_login_required(json.loads(e.cause.read().decode())['message'])
+            raise
 
-        VIDEO_PARAMS_RE = r'\bvlive\.video\.init\(([^)]+)'
+        video = post['officialVideo']
-        VIDEO_PARAMS_FIELD = 'video params'
 
-        params = self._parse_json(self._search_regex(
+        def get_common_fields():
-            VIDEO_PARAMS_RE, webpage, VIDEO_PARAMS_FIELD, default=''), video_id,
+            channel = post.get('channel') or {}
-            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 in ('LIVE_ON_AIR', 'BIG_EVENT_ON_AIR'):
-            return self._live(video_id, webpage)
-        elif status in ('VOD_ON_AIR', 'BIG_EVENT_INTRO'):
-            return self._replay(video_id, webpage, long_video_id, key)
-
-        if status == 'LIVE_END':
-            raise ExtractorError('Uploading for replay. Please wait...',
-                                 expected=True)
-        elif status == 'COMING_SOON':
-            raise ExtractorError('Coming soon!', expected=True)
-        elif status == 'CANCELED':
-            raise ExtractorError('We are sorry, '
-                                 'but the live broadcast has been canceled.',
-                                 expected=True)
-        elif status == 'ONLY_APP':
-            raise ExtractorError('Unsupported video type', expected=True)
-        else:
-            raise ExtractorError('Unknown status %s' % status)
-
-    def _get_common_fields(self, webpage):
-        title = self._og_search_title(webpage)
-        creator = self._html_search_regex(
-            r'<div[^>]+class="info_area"[^>]*>\s*(?:<em[^>]*>.*?</em\s*>\s*)?<a\s+[^>]*>([^<]+)',
-            webpage, 'creator', fatal=False)
-        thumbnail = self._og_search_thumbnail(webpage)
             return {
-            'title': title,
+                'title': video.get('title'),
-            'creator': creator,
+                'creator': post.get('author', {}).get('nickname'),
-            'thumbnail': thumbnail,
+                'channel': channel.get('channelName'),
+                'channel_id': channel.get('channelCode'),
+                'duration': int_or_none(video.get('playTime')),
+                'view_count': int_or_none(video.get('playCount')),
+                'like_count': int_or_none(video.get('likeCount')),
+                'comment_count': int_or_none(video.get('commentCount')),
             }
 
-    def _live(self, video_id, webpage):
+        video_type = video.get('type')
-        init_page = self._download_init_page(video_id)
+        if video_type == 'VOD':
-
+            inkey = self._call_api('video/v1.0/vod/%s/inkey', video_id)['inkey']
-        live_params = self._search_regex(
+            vod_id = video['vodId']
-            r'"liveStreamInfo"\s*:\s*(".*"),',
+            return merge_dicts(
-            init_page, 'live stream info')
+                get_common_fields(),
-        live_params = self._parse_json(live_params, video_id)
+                self._extract_video_info(video_id, vod_id, inkey))
-        live_params = self._parse_json(live_params, video_id)
+        elif video_type == 'LIVE':
-
+            status = video.get('status')
-        formats = []
+            if status == 'ON_AIR':
-        for vid in live_params.get('resolutions', []):
+                stream_url = self._call_api(
-            formats.extend(self._extract_m3u8_formats(
+                    'old/v3/live/%s/playInfo',
-                vid['cdnUrl'], video_id, 'mp4',
+                    video_id)['result']['adaptiveStreamUrl']
-                m3u8_id=vid.get('name'),
+                formats = self._extract_m3u8_formats(stream_url, video_id, 'mp4')
-                fatal=False, live=True))
+                info = get_common_fields()
-        self._sort_formats(formats)
-
-        info = self._get_common_fields(webpage)
                 info.update({
-            'title': self._live_title(info['title']),
+                    'title': self._live_title(video['title']),
                     'id': video_id,
                     'formats': formats,
                     'is_live': True,
                 })
                 return info
-
+            elif status == 'ENDED':
-    def _replay(self, video_id, webpage, long_video_id, key):
+                raise ExtractorError(
-        if '' in (long_video_id, key):
+                    'Uploading for replay. Please wait...', expected=True)
-            init_page = self._download_init_page(video_id)
+            elif status == 'RESERVED':
-            video_info = self._parse_json(self._search_regex(
+                raise ExtractorError('Coming soon!', expected=True)
-                (r'(?s)oVideoStatus\s*=\s*({.+?})\s*</script',
+            elif video.get('exposeStatus') == 'CANCEL':
-                 r'(?s)oVideoStatus\s*=\s*({.+})'), init_page, 'video info'),
+                raise ExtractorError(
-                video_id)
+                    'We are sorry, but the live broadcast has been canceled.',
-            if video_info.get('status') == 'NEED_CHANNEL_PLUS':
+                    expected=True)
-                self.raise_login_required(
+            else:
-                    'This video is only available for CH+ subscribers')
+                raise ExtractorError('Unknown status ' + status)
-            long_video_id, key = video_info['vid'], video_info['inkey']
-
-        return merge_dicts(
-            self._get_common_fields(webpage),
-            self._extract_video_info(video_id, long_video_id, key))
-
-    def _download_init_page(self, video_id):
-        return self._download_webpage(
-            'https://www.vlive.tv/video/init/view',
-            video_id, note='Downloading live webpage',
-            data=urlencode_postdata({'videoSeq': video_id}),
-            headers={
-                'Referer': 'https://www.vlive.tv/video/%s' % video_id,
-                'Content-Type': 'application/x-www-form-urlencoded'
-            })
 
 
-class VLiveChannelIE(InfoExtractor):
+class VLiveChannelIE(VLiveBaseIE):
     IE_NAME = 'vlive:channel'
-    _VALID_URL = r'https?://channels\.vlive\.tv/(?P<id>[0-9A-Z]+)'
+    _VALID_URL = r'https?://(?:channels\.vlive\.tv|(?:(?:www|m)\.)?vlive\.tv/channel)/(?P<id>[0-9A-Z]+)'
-    _TEST = {
+    _TESTS = [{
         'url': 'http://channels.vlive.tv/FCD4B',
         'info_dict': {
             'id': 'FCD4B',
             'title': 'MAMAMOO',
         },
         'playlist_mincount': 110
+    }, {
+        'url': 'https://www.vlive.tv/channel/FCD4B',
+        'only_matching': True,
+    }]
+
+    def _call_api(self, path, channel_key_suffix, channel_value, note, query):
+        q = {
+            'app_id': self._APP_ID,
+            'channel' + channel_key_suffix: channel_value,
         }
-    _APP_ID = '8c6cc7b45d2568fb668be6e05b6e5a3b'
+        q.update(query)
+        return self._download_json(
+            'http://api.vfan.vlive.tv/vproxy/channelplus/' + path,
+            channel_value, note='Downloading ' + note, query=q)['result']
 
     def _real_extract(self, url):
         channel_code = self._match_id(url)
 
-        webpage = self._download_webpage(
+        channel_seq = self._call_api(
-            'http://channels.vlive.tv/%s/video' % channel_code, channel_code)
+            'decodeChannelCode', 'Code', channel_code,
+            'decode channel code', {})['channelSeq']
 
-        app_id = None
-
-        app_js_url = self._search_regex(
-            r'<script[^>]+src=(["\'])(?P<url>http.+?/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(
+            video_list = self._call_api(
-                'http://api.vfan.vlive.tv/vproxy/channelplus/getChannelVideoList',
+                'getChannelVideoList', 'Seq', channel_seq,
-                channel_code, note='Downloading channel list page #%d' % page_num,
+                '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/ytdl-org/youtube-dl/issues/13830
                     # 2. http://channels.vlive.tv/EDBF.
                     'maxNumOfRows': 100,
-                    '_': int(time.time()),
                     'pageNo': page_num
                 }
             )
@@ -269,11 +222,11 @@ class VLiveChannelIE(InfoExtractor):
             if not channel_name:
                 channel_name = try_get(
                     video_list,
-                    lambda x: x['result']['channelInfo']['channelName'],
+                    lambda x: x['channelInfo']['channelName'],
                     compat_str)
 
             videos = try_get(
-                video_list, lambda x: x['result']['videoList'], list)
+                video_list, lambda x: x['videoList'], list)
             if not videos:
                 break
 
@@ -289,79 +242,3 @@ class VLiveChannelIE(InfoExtractor):
 
         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<video_id>[0-9]+)/playlist/(?P<id>[0-9]+)'
-    _VIDEO_URL_TEMPLATE = 'http://www.vlive.tv/video/%s'
-    _TESTS = [{
-        # regular working playlist
-        'url': 'https://www.vlive.tv/video/117956/playlist/117963',
-        'info_dict': {
-            'id': '117963',
-            'title': '아이돌룸(IDOL ROOM) 41회 - (여자)아이들'
-        },
-        'playlist_mincount': 10
-    }, {
-        # playlist with no playlistVideoSeqs
-        'url': 'http://www.vlive.tv/video/22867/playlist/22912',
-        'info_dict': {
-            'id': '22867',
-            'ext': 'mp4',
-            'title': '[V LIVE] Valentine Day Message from MINA',
-            'creator': 'TWICE',
-            'view_count': int
-        },
-        'params': {
-            'skip_download': True,
-        }
-    }]
-
-    def _build_video_result(self, video_id, message):
-        self.to_screen(message)
-        return self.url_result(
-            self._VIDEO_URL_TEMPLATE % video_id,
-            ie=VLiveIE.ie_key(), video_id=video_id)
-
-    def _real_extract(self, url):
-        mobj = re.match(self._VALID_URL, url)
-        video_id, playlist_id = mobj.group('video_id', 'id')
-
-        if self._downloader.params.get('noplaylist'):
-            return self._build_video_result(
-                video_id,
-                'Downloading just video %s because of --no-playlist'
-                % 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)
-
-        raw_item_ids = self._search_regex(
-            r'playlistVideoSeqs\s*=\s*(\[[^]]+\])', webpage,
-            'playlist video seqs', default=None, fatal=False)
-
-        if not raw_item_ids:
-            return self._build_video_result(
-                video_id,
-                'Downloading just video %s because no playlist was found'
-                % video_id)
-
-        item_ids = self._parse_json(raw_item_ids, playlist_id)
-
-        entries = [
-            self.url_result(
-                self._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'<div[^>]+class="[^"]*multicam_playlist[^>]*>\s*<h3[^>]+>([^<]+)',
-            webpage, 'playlist title', fatal=False)
-
-        return self.playlist_result(entries, playlist_id, playlist_name)