From 14d05d3b2088fa405a92b6839332416c1d4287ed Mon Sep 17 00:00:00 2001
From: pl <0x00-pl@gmail.com>
Date: Sat, 11 Oct 2014 00:15:54 +0800
Subject: [PATCH] fix douyutv title regex
---
src/you_get/extractors/douyutv.py | 6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)
diff --git a/src/you_get/extractors/douyutv.py b/src/you_get/extractors/douyutv.py
index 0292c94e..1dde4141 100644
--- a/src/you_get/extractors/douyutv.py
+++ b/src/you_get/extractors/douyutv.py
@@ -8,11 +8,11 @@ import json
def douyutv_download(url, output_dir = '.', merge = True, info_only = False):
html = get_html(url)
- room_id_patt = '"room_id":(\d{1,99}),'
- title_patt = '([^<]{0,1000})'
+ room_id_patt = r'"room_id":(\d{1,99}),'
+ title_patt = r'\s*
([^<]{1,9999})
\s*'
roomid = re.findall(room_id_patt,html)[0]
- title = re.findall(title_patt,html)[0]
+ title = unescape_html(re.findall(title_patt,html)[0])
conf = get_html("http://www.douyutv.com/api/client/room/"+roomid)
metadata = json.loads(conf)