commit: a8907224a1c433b1227fd707e9bb2524dd405109
parent 09ee2aa69dbd4815e0e1e1de53f3571972e04903
Author: Lorenzo J. Lucchini <LuccoJ@users.noreply.github.com>
Date: Thu, 7 Jul 2016 00:33:03 +0200
Improving Wolfram Alpha search hit content
Making WA search hits contain
- the (parsed) input inside the "title" instead of just "Wolfram|Alpha", to better match other hit titles and to confirm correct parsing of input to the user
- the first output field that contains any text (skipping ones that are only pictures; this is usually the most meaningful "result" field) instead of the raw input as the "content", making it additionally possible to obtain WA computations from JSON API calls
Diffstat:
1 file changed, 9 insertions(+), 7 deletions(-)
diff --git a/searx/engines/wolframalpha_api.py b/searx/engines/wolframalpha_api.py
@@ -18,7 +18,6 @@ api_key = '' # defined in settings.yml
# xpath variables
failure_xpath = '/queryresult[attribute::success="false"]'
-answer_xpath = '//pod[attribute::primary="true"]/subpod/plaintext'
input_xpath = '//pod[starts-with(attribute::id, "Input")]/subpod/plaintext'
pods_xpath = '//pod'
subpods_xpath = './subpod'
@@ -76,11 +75,11 @@ def response(resp):
try:
infobox_title = search_results.xpath(input_xpath)[0].text
except:
- infobox_title = None
+ infobox_title = ""
pods = search_results.xpath(pods_xpath)
- result = ""
result_chunks = []
+ result_content = ""
for pod in pods:
pod_id = pod.xpath(pod_id_xpath)[0]
pod_title = pod.xpath(pod_title_xpath)[0]
@@ -97,8 +96,9 @@ def response(resp):
if content and pod_id not in image_pods:
- if pod_is_result:
- result = content
+ if pod_is_result or not result_content:
+ if pod_id != "Input":
+ result_content = "%s: %s" % (pod_title, content)
# if no input pod was found, title is first plaintext pod
if not infobox_title:
@@ -115,6 +115,8 @@ def response(resp):
if not result_chunks:
return []
+ title = "Wolfram|Alpha (%s)" % infobox_title
+
# append infobox
results.append({'infobox': infobox_title,
'attributes': result_chunks,
@@ -122,7 +124,7 @@ def response(resp):
# append link to site
results.append({'url': resp.request.headers['Referer'].decode('utf8'),
- 'title': infobox_title + ' - Wolfram|Alpha',
- 'content': result})
+ 'title': title,
+ 'content': result_content})
return results