Merge pull request #7210 from remitamine/bilibili

[bilibili] fix info extraction(fixes #7182)

youtube_dl/extractor/bilibili.py

@@ -2,7 +2,6 @@
 from __future__ import unicode_literals
 
 import re
-import itertools
 import json
 
 from .common import InfoExtractor
@@ -11,134 +10,108 @@ from ..compat import (
 )
 from ..utils import (
     int_or_none,
-    unified_strdate,
+    unescapeHTML,
     ExtractorError,
 )
 
 
 class BiliBiliIE(InfoExtractor):
-    _VALID_URL = r'http://www\.bilibili\.(?:tv|com)/video/av(?P<id>[0-9]+)/'
+    _VALID_URL = r'http://www\.bilibili\.(?:tv|com)/video/av(?P<id>\d+)(?:/index_(?P<page_num>\d+).html)?'
 
     _TESTS = [{
         'url': 'http://www.bilibili.tv/video/av1074402/',
         'md5': '2c301e4dab317596e837c3e7633e7d86',
         'info_dict': {
-            'id': '1074402_part1',
+            'id': '1554319',
             'ext': 'flv',
             'title': '【金坷垃】金泡沫',
-            'duration': 308,
+            'duration': 308313,
             'upload_date': '20140420',
             'thumbnail': 're:^https?://.+\.jpg',
+            'description': 'md5:ce18c2a2d2193f0df2917d270f2e5923',
+            'timestamp': 1397983878,
+            'uploader': '菊子桑',
         },
     }, {
         'url': 'http://www.bilibili.com/video/av1041170/',
         'info_dict': {
             'id': '1041170',
             'title': '【BD1080P】刀语【诸神&异域】',
+            'description': '这是个神奇的故事~每个人不留弹幕不给走哦~切利哦！~',
+            'uploader': '枫叶逝去',
+            'timestamp': 1396501299,
         },
         'playlist_count': 9,
     }]
 
     def _real_extract(self, url):
-        video_id = self._match_id(url)
-        webpage = self._download_webpage(url, video_id)
-
-        if '(此视频不存在或被删除)' in webpage:
-            raise ExtractorError(
-                'The video does not exist or was deleted', expected=True)
-
-        if '>你没有权限浏览！ 由于版权相关问题 我们不对您所在的地区提供服务<' in webpage:
-            raise ExtractorError(
-                'The video is not available in your region due to copyright reasons',
-                expected=True)
-
-        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
-        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)
-
-        cid = self._search_regex(r'cid=(\d+)', webpage, 'cid')
-
-        entries = []
-
-        lq_page = self._download_webpage(
-            'http://interface.bilibili.com/v_cdn_play?appkey=1&cid=%s' % cid,
-            video_id,
-            note='Downloading LQ video info'
+        mobj = re.match(self._VALID_URL, url)
+        video_id = mobj.group('id')
+        page_num = mobj.group('page_num') or '1'
+
+        view_data = self._download_json(
+            'http://api.bilibili.com/view?type=json&appkey=8e9fc618fbd41e28&id=%s&page=%s' % (video_id, page_num),
+            video_id)
+        if 'error' in view_data:
+            raise ExtractorError('%s said: %s' % (self.IE_NAME, view_data['error']), expected=True)
+
+        cid = view_data['cid']
+        title = unescapeHTML(view_data['title'])
+
+        page = self._download_webpage(
+            'http://interface.bilibili.com/v_cdn_play?appkey=8e9fc618fbd41e28&cid=%s' % cid,
+            cid,
+            'Downloading page %s/%s' % (page_num, view_data['pages'])
         )
         try:
-            err_info = json.loads(lq_page)
+            err_info = json.loads(page)
             raise ExtractorError(
                 'BiliBili said: ' + err_info['error_text'], expected=True)
         except ValueError:
             pass
 
-        lq_doc = compat_etree_fromstring(lq_page)
-        lq_durls = lq_doc.findall('./durl')
+        doc = compat_etree_fromstring(page)
 
-        hq_doc = self._download_xml(
-            'http://interface.bilibili.com/playurl?appkey=1&cid=%s' % cid,
-            video_id,
-            note='Downloading HQ video info',
-            fatal=False,
-        )
-        if hq_doc is not False:
-            hq_durls = hq_doc.findall('./durl')
-            assert len(lq_durls) == len(hq_durls)
-        else:
-            hq_durls = itertools.repeat(None)
+        entries = []
 
-        i = 1
-        for lq_durl, hq_durl in zip(lq_durls, hq_durls):
+        for durl in doc.findall('./durl'):
+            size = durl.find('./filesize|./size')
             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.find('./url').text,
+                'filesize': int_or_none(size.text) if size else None,
+                'ext': 'flv',
             }]
-            if hq_durl is not None:
-                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'),
-                })
-            self._sort_formats(formats)
+            backup_urls = durl.find('./backup_url')
+            if backup_urls is not None:
+                for backup_url in backup_urls.findall('./url'):
+                    formats.append({'url': backup_url.text})
+            formats.reverse()
 
             entries.append({
-                'id': '%s_part%d' % (video_id, i),
+                'id': '%s_part%s' % (cid, durl.find('./order').text),
                 'title': title,
+                'duration': int_or_none(durl.find('./length').text) // 1000,
                 'formats': formats,
-                'duration': duration,
-                'upload_date': upload_date,
-                'thumbnail': thumbnail,
             })
 
-            i += 1
-
-        return {
-            '_type': 'multi_video',
-            'entries': entries,
-            'id': video_id,
-            'title': title
+        info = {
+            'id': str(cid),
+            'title': title,
+            'description': view_data.get('description'),
+            'thumbnail': view_data.get('pic'),
+            'uploader': view_data.get('author'),
+            'timestamp': int_or_none(view_data.get('created')),
+            'view_count': view_data.get('play'),
+            'duration': int_or_none(doc.find('./timelength').text),
         }
+
+        if len(entries) == 1:
+            entries[0].update(info)
+            return entries[0]
+        else:
+            info.update({
+                '_type': 'multi_video',
+                'id': video_id,
+                'entries': entries,
+            })
+            return info