[bilibili] fix info extraction

youtube_dl/extractor/bilibili.py

@@ -1,15 +1,12 @@
 # coding: utf-8
 from __future__ import unicode_literals
 
-import re
-import itertools
 import json
 import xml.etree.ElementTree as ET
 
 from .common import InfoExtractor
 from ..utils import (
     int_or_none,
-    unified_strdate,
     ExtractorError,
 )
 
@@ -21,12 +18,15 @@ class BiliBiliIE(InfoExtractor):
         '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/',
@@ -34,109 +34,72 @@ class BiliBiliIE(InfoExtractor):
             'id': '1041170',
             'title': '【BD1080P】刀语【诸神&异域】',
         },
-        'playlist_count': 9,
+        'playlist_count': 12,
     }]
 
-    def _real_extract(self, url):
+    def _extract_video_info(self, cid, view_data, page_num=1, num_pages=1):
-        video_id = self._match_id(url)
+        title = view_data['title']
-        webpage = self._download_webpage(url, video_id)
 
-        if '(此视频不存在或被删除)' in webpage:
+        page = self._download_webpage(
-            raise ExtractorError(
+            'http://interface.bilibili.com/v_cdn_play?appkey=8e9fc618fbd41e28&cid=%s' % cid,
-                'The video does not exist or was deleted', expected=True)
+            cid,
-
+            'Downloading page %d/%d' % (page_num, num_pages)
-        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'
         )
         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 = ET.fromstring(lq_page)
+        doc = ET.fromstring(page)
-        lq_durls = lq_doc.findall('./durl')
+        durls = doc.findall('./durl')
 
-        hq_doc = self._download_xml(
+        entries = []
-            '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)
-
-        i = 1
-        for lq_durl, hq_durl in zip(lq_durls, hq_durls):
-            formats = [{
-                'format_id': 'lq',
-                'quality': 1,
-                'url': lq_durl.find('./url').text,
-                'filesize': int_or_none(
-                    lq_durl.find('./size'), get_attr='text'),
-            }]
-            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)
 
+        for durl in durls:
             entries.append({
-                'id': '%s_part%d' % (video_id, i),
+                'id': '%s_part%s' % (cid, durl.find('./order').text),
                 'title': title,
-                'formats': formats,
+                'url': durl.find('./url').text,
-                'duration': duration,
+                'filesize': int_or_none(durl.find('./filesize').text),
-                'upload_date': upload_date,
+                'ext': 'flv',
-                'thumbnail': thumbnail,
+                'duration': int_or_none(durl.find('./length').text) // 1000,
             })
 
-            i += 1
+        info = {
-
+            'id': cid,
-        return {
+            'title': title,
-            '_type': 'multi_video',
+            'description': view_data.get('description'),
-            'entries': entries,
+            'thumbnail': view_data.get('pic'),
-            'id': video_id,
+            'uploader': view_data.get('author'),
-            'title': title
+            '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',
+                'entries': entries,
+            })
+            return info
+
+    def _real_extract(self, url):
+        video_id = self._match_id(url)
+        view_data = self._download_json('http://api.bilibili.com/view?type=json&appkey=8e9fc618fbd41e28&id=%s' % video_id, video_id)
+
+        num_pages = int_or_none(view_data['pages'])
+        if num_pages > 1:
+            play_list_title = view_data['title']
+            page_list = self._download_json('http://www.bilibili.com/widget/getPageList?aid=%s' % video_id, video_id, 'Downloading page list metadata')
+            entries = []
+            for page in page_list:
+                view_data['title'] = page['pagename']
+                entries.append(self._extract_video_info(str(page['cid']), view_data, page['page'], num_pages))
+            return self.playlist_result(entries, video_id, play_list_title, view_data.get('description'))
+        else:
+            return self._extract_video_info(str(view_data['cid']), view_data)