PublicMirrors/you-get
1
0
Fork 0
mirror of https://github.com/soimort/you-get.git synced 2025-02-03 08:43:58 +03:00
Code Issues Actions Packages Projects Releases 1 Wiki Activity

fix douyutv title regex

Browse Source
This commit is contained in:
pl 2014-10-11 00:15:54 +08:00
parent c3b16cfbda
commit 14d05d3b20
1 changed files with 3 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

6
src/you_get/extractors/douyutv.py
View File

@ -8,11 +8,11 @@ import json
def douyutv_download(url, output_dir = '.', merge = True, info_only = False): def douyutv_download(url, output_dir = '.', merge = True, info_only = False):
html = get_html(url) html = get_html(url)
room_id_patt = '"room_id":(\d{1,99}),' room_id_patt = r'"room_id":(\d{1,99}),'
title_patt = '<title>([^<]{0,1000})</title>' title_patt = r'<div class="headline clearfix">\s*<h1>([^<]{1,9999})</h1>\s*</div>'
roomid = re.findall(room_id_patt,html)[0] 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) conf = get_html("http://www.douyutv.com/api/client/room/"+roomid)
metadata = json.loads(conf) metadata = json.loads(conf)