[bilibili] Extract multipart videos (closes #3250)

youtube_dl/extractor/bilibili.py

@@ -2,6 +2,7 @@
 from __future__ import unicode_literals
 
 import re
+import itertools
 
 from .common import InfoExtractor
 from ..utils import (
@@ -14,18 +15,25 @@ from ..utils import (
 class BiliBiliIE(InfoExtractor):
     _VALID_URL = r'http://www\.bilibili\.(?:tv|com)/video/av(?P<id>[0-9]+)/'
 
-    _TEST = {
+    _TESTS = [{
         'url': 'http://www.bilibili.tv/video/av1074402/',
         'md5': '2c301e4dab317596e837c3e7633e7d86',
         'info_dict': {
-            'id': '1074402',
+            'id': '1074402_part1',
             'ext': 'flv',
             'title': '【金坷垃】金泡沫',
             'duration': 308,
             'upload_date': '20140420',
             'thumbnail': 're:^https?://.+\.jpg',
         },
-    }
+    }, {
+        'url': 'http://www.bilibili.com/video/av1041170/',
+        'info_dict': {
+            'id': '1041170',
+            'title': '【BD1080P】刀语【诸神&异域】',
+        },
+        'playlist_count': 9,
+    }]
 
     def _real_extract(self, url):
         video_id = self._match_id(url)
@@ -57,19 +65,14 @@ class BiliBiliIE(InfoExtractor):
 
         cid = self._search_regex(r'cid=(\d+)', webpage, 'cid')
 
+        entries = []
+
         lq_doc = self._download_xml(
             'http://interface.bilibili.com/v_cdn_play?appkey=1&cid=%s' % cid,
             video_id,
             note='Downloading LQ video info'
         )
-        lq_durl = lq_doc.find('./durl')
+        lq_durls = lq_doc.findall('./durl')
-        formats = [{
-            'format_id': 'lq',
-            'quality': 1,
-            'url': lq_durl.find('./url').text,
-            'filesize': int_or_none(
-                lq_durl.find('./size'), get_attr='text'),
-        }]
 
         hq_doc = self._download_xml(
             'http://interface.bilibili.com/playurl?appkey=1&cid=%s' % cid,
@@ -77,8 +80,20 @@ class BiliBiliIE(InfoExtractor):
             note='Downloading HQ video info',
             fatal=False,
         )
-        if hq_doc is not False:
+        hq_durls = hq_doc.findall('./durl') if hq_doc is not False else itertools.repeat(None)
-            hq_durl = hq_doc.find('./durl')
+
+        assert len(lq_durls) == len(hq_durls)
+
+        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:
                 formats.append({
                     'format_id': 'hq',
                     'quality': 2,
@@ -87,13 +102,22 @@ class BiliBiliIE(InfoExtractor):
                     'filesize': int_or_none(
                         hq_durl.find('./size'), get_attr='text'),
                 })
-
             self._sort_formats(formats)
-        return {
+
-            'id': video_id,
+            entries.append({
+                'id': '%s_part%d' % (video_id, i),
                 'title': title,
                 'formats': formats,
                 'duration': duration,
                 'upload_date': upload_date,
                 'thumbnail': thumbnail,
+            })
+
+            i += 1
+
+        return {
+            '_type': 'multi_video',
+            'entries': entries,
+            'id': video_id,
+            'title': title
         }