From 5695ec58df14615e154072e97afc4bd6ed505812 Mon Sep 17 00:00:00 2001 From: asciimoo Date: Tue, 4 Feb 2014 19:42:32 +0100 Subject: [PATCH] [fix] whitespace duplication fix in nonhtml outputs --- searx/webapp.py | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) diff --git a/searx/webapp.py b/searx/webapp.py index ec8fe8fb3..c52fae9d9 100644 --- a/searx/webapp.py +++ b/searx/webapp.py @@ -190,7 +190,9 @@ def index(): else: if 'content' in result: result['content'] = html_to_text(result['content']).strip() - result['title'] = html_to_text(result['title']).strip() + # removing html content and whitespace duplications + result['title'] = ' '.join(html_to_text(result['title'])\ + .strip().split()) if len(result['url']) > 74: url_parts = result['url'][:35], result['url'][-35:] result['pretty_url'] = '{0}[...]{1}'.format(*url_parts)