From 576d37f256649b570a9c8591a795acd85ac499bc Mon Sep 17 00:00:00 2001 From: a01200356 Date: Sun, 3 Jan 2016 15:58:01 -0600 Subject: [PATCH] [fix] unescape htmlentities in wolframalpha_noapi's answer --- searx/engines/wolframalpha_noapi.py | 7 ++++++- searx/settings.yml | 3 ++- 2 files changed, 8 insertions(+), 2 deletions(-) diff --git a/searx/engines/wolframalpha_noapi.py b/searx/engines/wolframalpha_noapi.py index a730ed60b..0f0315630 100644 --- a/searx/engines/wolframalpha_noapi.py +++ b/searx/engines/wolframalpha_noapi.py @@ -12,6 +12,7 @@ from re import search, sub from json import loads from urllib import urlencode from lxml import html +import HTMLParser # search-url url = 'http://www.wolframalpha.com/' @@ -62,7 +63,11 @@ def response(resp): # extract answer from json answer = line[line.find('{'):line.rfind('}')+1] answer = loads(answer.encode('unicode-escape')) - answer = answer['stringified'].decode('unicode-escape') + answer = answer['stringified'] + + # clean plaintext answer + h = HTMLParser.HTMLParser() + answer = h.unescape(answer.decode('unicode-escape')) answer = sub(r'\\', '', answer) results.append({'answer': answer}) diff --git a/searx/settings.yml b/searx/settings.yml index 600c20c0d..63e944060 100644 --- a/searx/settings.yml +++ b/searx/settings.yml @@ -305,9 +305,10 @@ engines: # You can use the engine using the official stable API, but you need an API key # See : http://products.wolframalpha.com/api/ # engine : wolframalpha_api - # api_key: '5952JX-X52L3VKWT8' # required! + # api_key: '' # required! engine : wolframalpha_noapi timeout: 6.0 + disabled : True #The blekko technology and team have joined IBM Watson! -> https://blekko.com/ # - name : blekko images