X-Git-Url: https://git.rapsys.eu/youtubedl/blobdiff_plain/d88adeed4f0a22f14db3a2b3fd3c59b266b97a78..83c8b40add31fb501d587b64fb17a936ffd5dbc5:/youtube_dl/extractor/bilibili.py

diff --git a/youtube_dl/extractor/bilibili.py b/youtube_dl/extractor/bilibili.py
index 0d5889f..2d174e6 100644
--- a/youtube_dl/extractor/bilibili.py
+++ b/youtube_dl/extractor/bilibili.py
@@ -1,106 +1,125 @@
 # coding: utf-8
 from __future__ import unicode_literals
 
+import hashlib
 import re
 
 from .common import InfoExtractor
+from ..compat import compat_parse_qs
 from ..utils import (
-    compat_parse_qs,
-    ExtractorError,
     int_or_none,
-    unified_strdate,
+    float_or_none,
+    unified_timestamp,
+    urlencode_postdata,
 )
 
 
 class BiliBiliIE(InfoExtractor):
-    _VALID_URL = r'http://www\.bilibili\.(?:tv|com)/video/av(?P<id>[0-9]+)/'
+    _VALID_URL = r'https?://(?:www\.|bangumi\.|)bilibili\.(?:tv|com)/(?:video/av|anime/v/)(?P<id>\d+)'
 
     _TEST = {
         'url': 'http://www.bilibili.tv/video/av1074402/',
-        'md5': '2c301e4dab317596e837c3e7633e7d86',
+        'md5': '9fa226fe2b8a9a4d5a69b4c6a183417e',
         'info_dict': {
             'id': '1074402',
-            'ext': 'flv',
+            'ext': 'mp4',
             'title': '【金坷垃】金泡沫',
-            'duration': 308,
+            'description': 'md5:ce18c2a2d2193f0df2917d270f2e5923',
+            'duration': 308.315,
+            'timestamp': 1398012660,
             'upload_date': '20140420',
             'thumbnail': 're:^https?://.+\.jpg',
+            'uploader': '菊子桑',
+            'uploader_id': '156160',
         },
     }
 
-    def _real_extract(self, url):
-        mobj = re.match(self._VALID_URL, url)
-        video_id = mobj.group('id')
+    _APP_KEY = '6f90a59ac58a4123'
+    _BILIBILI_KEY = '0bfd84cc3940035173f35e6777508326'
 
+    def _real_extract(self, url):
+        video_id = self._match_id(url)
         webpage = self._download_webpage(url, video_id)
-        video_code = self._search_regex(
-            r'(?s)<div itemprop="video".*?>(.*?)</div>', webpage, 'video code')
-
-        title = self._html_search_meta(
-            'media:title', video_code, 'title', fatal=True)
-        duration_str = self._html_search_meta(
-            'duration', video_code, 'duration')
-        if duration_str is None:
-            duration = None
+
+        if 'anime/v' not in url:
+            cid = compat_parse_qs(self._search_regex(
+                [r'EmbedPlayer\([^)]+,\s*"([^"]+)"\)',
+                 r'<iframe[^>]+src="https://secure\.bilibili\.com/secure,([^"]+)"'],
+                webpage, 'player parameters'))['cid'][0]
         else:
-            duration_mobj = re.match(
-                r'^T(?:(?P<hours>[0-9]+)H)?(?P<minutes>[0-9]+)M(?P<seconds>[0-9]+)S$',
-                duration_str)
-            duration = (
-                int_or_none(duration_mobj.group('hours'), default=0) * 3600 +
-                int(duration_mobj.group('minutes')) * 60 +
-                int(duration_mobj.group('seconds')))
-        upload_date = unified_strdate(self._html_search_meta(
-            'uploadDate', video_code, fatal=False))
-        thumbnail = self._html_search_meta(
-            'thumbnailUrl', video_code, 'thumbnail', fatal=False)
-
-        player_params = compat_parse_qs(self._html_search_regex(
-            r'<iframe .*?class="player" src="https://secure\.bilibili\.(?:tv|com)/secure,([^"]+)"',
-            webpage, 'player params'))
-
-        if 'cid' in player_params:
-            cid = player_params['cid'][0]
-
-            lq_doc = self._download_xml(
-                'http://interface.bilibili.cn/v_cdn_play?cid=%s' % cid,
-                video_id,
-                note='Downloading LQ video info'
-            )
-            lq_durl = lq_doc.find('.//durl')
+            js = self._download_json(
+                'http://bangumi.bilibili.com/web_api/get_source', video_id,
+                data=urlencode_postdata({'episode_id': video_id}),
+                headers={'Content-Type': 'application/x-www-form-urlencoded; charset=UTF-8'})
+            cid = js['result']['cid']
+
+        payload = 'appkey=%s&cid=%s&otype=json&quality=2&type=mp4' % (self._APP_KEY, cid)
+        sign = hashlib.md5((payload + self._BILIBILI_KEY).encode('utf-8')).hexdigest()
+
+        video_info = self._download_json(
+            'http://interface.bilibili.com/playurl?%s&sign=%s' % (payload, sign),
+            video_id, note='Downloading video info page')
+
+        entries = []
+
+        for idx, durl in enumerate(video_info['durl']):
             formats = [{
-                'format_id': 'lq',
-                'quality': 1,
-                'url': lq_durl.find('./url').text,
-                'filesize': int_or_none(
-                    lq_durl.find('./size'), get_attr='text'),
+                'url': durl['url'],
+                'filesize': int_or_none(durl['size']),
             }]
-
-            hq_doc = self._download_xml(
-                'http://interface.bilibili.cn/playurl?cid=%s' % cid,
-                video_id,
-                note='Downloading HQ video info',
-                fatal=False,
-            )
-            if hq_doc is not False:
-                hq_durl = hq_doc.find('.//durl')
+            for backup_url in durl.get('backup_url', []):
                 formats.append({
-                    'format_id': 'hq',
-                    'quality': 2,
-                    'ext': 'flv',
-                    'url': hq_durl.find('./url').text,
-                    'filesize': int_or_none(
-                        hq_durl.find('./size'), get_attr='text'),
+                    'url': backup_url,
+                    # backup URLs have lower priorities
+                    'preference': -2 if 'hd.mp4' in backup_url else -3,
                 })
-        else:
-            raise ExtractorError('Unsupported player parameters: %r' % (player_params,))
 
-        self._sort_formats(formats)
-        return {
+            self._sort_formats(formats)
+
+            entries.append({
+                'id': '%s_part%s' % (video_id, idx),
+                'duration': float_or_none(durl.get('length'), 1000),
+                'formats': formats,
+            })
+
+        title = self._html_search_regex('<h1[^>]+title="([^"]+)">', webpage, 'title')
+        description = self._html_search_meta('description', webpage)
+        timestamp = unified_timestamp(self._html_search_regex(
+            r'<time[^>]+datetime="([^"]+)"', webpage, 'upload time', fatal=False))
+        thumbnail = self._html_search_meta(['og:image', 'thumbnailUrl'], webpage)
+
+        # TODO 'view_count' requires deobfuscating Javascript
+        info = {
             'id': video_id,
             'title': title,
-            'formats': formats,
-            'duration': duration,
-            'upload_date': upload_date,
+            'description': description,
+            'timestamp': timestamp,
             'thumbnail': thumbnail,
+            'duration': float_or_none(video_info.get('timelength'), scale=1000),
         }
+
+        uploader_mobj = re.search(
+            r'<a[^>]+href="https?://space\.bilibili\.com/(?P<id>\d+)"[^>]+title="(?P<name>[^"]+)"',
+            webpage)
+        if uploader_mobj:
+            info.update({
+                'uploader': uploader_mobj.group('name'),
+                'uploader_id': uploader_mobj.group('id'),
+            })
+
+        for entry in entries:
+            entry.update(info)
+
+        if len(entries) == 1:
+            return entries[0]
+        else:
+            for idx, entry in enumerate(entries):
+                entry['id'] = '%s_part%d' % (video_id, (idx + 1))
+
+            return {
+                '_type': 'multi_video',
+                'id': video_id,
+                'title': title,
+                'description': description,
+                'entries': entries,
+            }