[enh] infobox for wolframalpha

TODO:
    - infobox styles
    - unit tests

ISSUES:
    - no_api version needs to re-call server for additional pods, such
      as plots. therefore, it's even slower than before. comment out the
part that calls get_async_pod if requests reach timeout or increase
timeout in settings.yml.

searx/engines/wolframalpha_api.py

@@ -1,40 +1,56 @@
-# Wolfram Alpha (Maths)
+# Wolfram Alpha (Science)
 #
-# @website     http://www.wolframalpha.com
-# @provide-api yes (http://api.wolframalpha.com/v2/)
+# @website     https://www.wolframalpha.com
+# @provide-api yes (https://api.wolframalpha.com/v2/)
 #
 # @using-api   yes
 # @results     XML
 # @stable      yes
-# @parse       result
+# @parse       url, infobox
 
 from urllib import urlencode
 from lxml import etree
-from re import search
 
 # search-url
-base_url = 'http://api.wolframalpha.com/v2/query'
-search_url = base_url + '?appid={api_key}&{query}&format=plaintext'
-site_url = 'http://www.wolframalpha.com/input/?{query}'
+search_url = 'https://api.wolframalpha.com/v2/query?appid={api_key}&{query}'
+site_url = 'https://www.wolframalpha.com/input/?{query}'
 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::title, "Input")]/subpod/plaintext'
+pods_xpath = '//pod'
+subpods_xpath = './subpod'
+pod_title_xpath = './@title'
+plaintext_xpath = './plaintext'
+image_xpath = './img'
+img_src_xpath = './@src'
+img_alt_xpath = './@alt'
+
+# pods to display as image in infobox
+# this pods do return a plaintext, but they look better and are more useful as images
+image_pods = {'Visual representation',
+              'Manipulatives illustration'}
 
 
 # do search-request
 def request(query, params):
     params['url'] = search_url.format(query=urlencode({'input': query}),
                                       api_key=api_key)
+    params['headers']['Referer'] = site_url.format(query=urlencode({'i': query}))
 
     return params
 
 
 # replace private user area characters to make text legible
 def replace_pua_chars(text):
-    pua_chars = {u'\uf74c': 'd',
+    pua_chars = {u'\uf522': u'\u2192',
+                 u'\uf7b1': u'\u2115',
+                 u'\uf7b4': u'\u211a',
+                 u'\uf7b5': u'\u211d',
+                 u'\uf7bd': u'\u2124',
+                 u'\uf74c': 'd',
                  u'\uf74d': u'\u212f',
                  u'\uf74e': 'i',
                  u'\uf7d9': '='}
@@ -55,23 +71,45 @@ def response(resp):
     if search_results.xpath(failure_xpath):
         return []
 
-    # parse answers
-    answers = search_results.xpath(answer_xpath)
-    if answers:
-        for answer in answers:
-            answer = replace_pua_chars(answer.text)
+    infobox_title = search_results.xpath(input_xpath)
+    if infobox_title:
+        infobox_title = replace_pua_chars(infobox_title[0].text)
+
+    pods = search_results.xpath(pods_xpath)
+    result_chunks = []
+    for pod in pods:
+        pod_title = replace_pua_chars(pod.xpath(pod_title_xpath)[0])
+
+        subpods = pod.xpath(subpods_xpath)
+        if not subpods:
+            continue
+
+        for subpod in subpods:
+            content = subpod.xpath(plaintext_xpath)[0].text
+            image = subpod.xpath(image_xpath)
+            if content and pod_title not in image_pods:
+                content = replace_pua_chars(content)
+                result_chunks.append({'label': pod_title, 'value': content})
 
-            results.append({'answer': answer})
+                # if there's no input pod, infobox_title is content of first pod
+                if not infobox_title:
+                    infobox_title = content
+
+            elif image:
+                result_chunks.append({'label': pod_title,
+                                      'image': {'src': image[0].xpath(img_src_xpath)[0],
+                                                'alt': image[0].xpath(img_alt_xpath)[0]}})
+
+    if not result_chunks:
+        return []
 
-    # if there's no input section in search_results, check if answer has the input embedded (before their "=" sign)
-    try:
-        query_input = search_results.xpath(input_xpath)[0].text
-    except IndexError:
-        query_input = search(u'([^\uf7d9]+)', answers[0].text).group(1)
+    results.append({'infobox': infobox_title,
+                    'attributes': result_chunks,
+                    'urls': [{'title': 'Wolfram|Alpha', 'url': resp.request.headers['Referer']}]})
 
     # append link to site
-    result_url = site_url.format(query=urlencode({'i': query_input.encode('utf-8')}))
-    results.append({'url': result_url,
-                    'title': query_input + " - Wolfram|Alpha"})
+    results.append({'url': resp.request.headers['Referer'],
+                    'title': 'Wolfram|Alpha',
+                    'content': infobox_title})
 
     return results

searx/engines/wolframalpha_noapi.py

@@ -1,23 +1,23 @@
-# WolframAlpha (Maths)
+# Wolfram|Alpha (Science)
 #
-# @website     http://www.wolframalpha.com/
-# @provide-api yes (http://api.wolframalpha.com/v2/)
+# @website     https://www.wolframalpha.com/
+# @provide-api yes (https://api.wolframalpha.com/v2/)
 #
 # @using-api   no
-# @results     HTML
+# @results     JSON
 # @stable      no
-# @parse       answer
+# @parse       url, infobox
 
 from cgi import escape
 from json import loads
 from time import time
 from urllib import urlencode
+from lxml.etree import XML
 
 from searx.poolrequests import get as http_get
 
 # search-url
 url = 'https://www.wolframalpha.com/'
-search_url = url + 'input/?{query}'
 
 search_url = url + 'input/json.jsp'\
     '?async=true'\
@@ -33,13 +33,25 @@ search_url = url + 'input/json.jsp'\
     '&sponsorcategories=true'\
     '&statemethod=deploybutton'
 
-# xpath variables
-scripts_xpath = '//script'
-title_xpath = '//title'
-failure_xpath = '//p[attribute::class="pfail"]'
+referer_url = url + 'input/?{query}'
+
 token = {'value': '',
          'last_updated': None}
 
+# xpath variables
+success_xpath = '/pod[attribute::error="false"]'
+plaintext_xpath = './plaintext'
+title_xpath = './@title'
+image_xpath = './img'
+img_src_xpath = './img/@src'
+img_alt_xpath = './img/@alt'
+
+# pods to display as image in infobox
+# this pods do return a plaintext, but they look better and are more useful as images
+image_pods = {'Visual representation',
+              'Manipulatives illustration',
+              'Symbol'}
+
 
 # seems, wolframalpha resets its token in every hour
 def obtain_token():
@@ -62,13 +74,42 @@ def request(query, params):
     if time() - token['last_updated'] > 3600:
         obtain_token()
     params['url'] = search_url.format(query=urlencode({'input': query}), token=token['value'])
-    params['headers']['Referer'] = 'https://www.wolframalpha.com/input/?i=' + query
+    params['headers']['Referer'] = referer_url.format(query=urlencode({'i': query}))
 
     return params
 
 
+# get additional pod
+# NOTE: this makes an additional requests to server, so the response will take longer and might reach timeout
+def get_async_pod(url):
+    pod = {'subpods': []}
+
+    try:
+        resp = http_get(url, timeout=2.0)
+
+        resp_pod = XML(resp.content)
+        if resp_pod.xpath(success_xpath):
+
+            for subpod in resp_pod:
+                plaintext = subpod.xpath(plaintext_xpath)[0].text
+                if plaintext:
+                    pod['subpods'].append({'title': subpod.xpath(title_xpath)[0],
+                                           'plaintext': plaintext})
+                elif subpod.xpath(image_xpath):
+                    pod['subpods'].append({'title': subpod.xpath(title_xpath)[0],
+                                           'plaintext': '',
+                                           'img': {'src': subpod.xpath(img_src_xpath)[0],
+                                                   'alt': subpod.xpath(img_alt_xpath)[0]}})
+    except:
+        pass
+
+    return pod
+
+
 # get response from search-request
 def response(resp):
+    results = []
+
     resp_json = loads(resp.text)
 
     if not resp_json['queryresult']['success']:
@@ -76,20 +117,45 @@ def response(resp):
 
     # TODO handle resp_json['queryresult']['assumptions']
     result_chunks = []
+    infobox_title = None
     for pod in resp_json['queryresult']['pods']:
         pod_title = pod.get('title', '')
+
         if 'subpods' not in pod:
-            continue
+            # comment this section if your requests always reach timeout
+            if pod['async']:
+                result = get_async_pod(pod['async'])
+                if result:
+                    pod = result
+            else:
+                continue
+
+        # infobox title is input or text content on first pod
+        if pod_title.startswith('Input') or not infobox_title:
+            try:
+                infobox_title = pod['subpods'][0]['plaintext']
+            except:
+                infobox_title = ''
+                pass
+
         for subpod in pod['subpods']:
-            if 'img' in subpod:
-                result_chunks.append(u'<p>{0}<br /><img src="{1}" alt="{2}" /></p>'
-                                     .format(escape(pod_title or subpod['img']['alt']),
-                                             escape(subpod['img']['src']),
-                                             escape(subpod['img']['alt'])))
+            if subpod['plaintext'] != '' and pod_title not in image_pods:
+                # append unless it's not an actual answer
+                if subpod['plaintext'] != '(requires interactivity)':
+                    result_chunks.append({'label': pod_title, 'value': subpod['plaintext']})
+
+            elif 'img' in subpod:
+                result_chunks.append({'label': pod_title, 'image': subpod['img']})
 
     if not result_chunks:
         return []
 
-    return [{'url': resp.request.headers['Referer'].decode('utf-8'),
-             'title': 'Wolframalpha',
-             'content': ''.join(result_chunks)}]
+    results.append({'infobox': infobox_title,
+                    'attributes': result_chunks,
+                    'urls': [{'title': 'Wolfram|Alpha', 'url': resp.request.headers['Referer']}]})
+
+    results.append({'url': resp.request.headers['Referer'],
+                    'title': 'Wolfram|Alpha',
+                    'content': infobox_title})
+
+    return results

searx/settings.yml

@@ -310,10 +310,10 @@ engines:
     shortcut : wa
     # 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: 'apikey' # required!
+    # engine : wolframalpha_api
+    # api_key: '5952JX-X52L3VKWT8' # required!
     engine : wolframalpha_noapi
-    timeout: 6.0
+    timeout: 10.0
     categories : science
 
 #The blekko technology and team have joined IBM Watson! -> https://blekko.com/