Include links in google_search results

AllYourBot · Nov 30, 2024 · dd22e1b · dd22e1b
1 parent e30905f
commit dd22e1b
Show file tree

Hide file tree

Showing 2 changed files with 28 additions and 4 deletions.
diff --git a/app/services/toolbox/google_search.rb b/app/services/toolbox/google_search.rb
@@ -11,7 +11,18 @@ def google_search(query_s:)
     encoded_query = URI.encode_www_form_component(query_s)
     response_body = get("https://www.google.com/search?q=#{encoded_query}").get_body
     doc = Nokogiri::HTML(response_body)
-    results = doc.css("div.BNeawe").map(&:text).join("\n")
+
+    results = doc.css("div.BNeawe").map do |div|
+      div.children.map do |node|
+        if node.name == "a"
+          anchor_text = node.text.strip
+          href = node["href"]
+          "#{anchor_text} (#{href})"
+        else
+          node.text.strip
+        end
+      end.join(" ")
+    end.join("\n")
 
     {
       message_to_user: "Web query: #{query_s}",

diff --git a/test/services/toolbox/google_search_test.rb b/test/services/toolbox/google_search_test.rb
@@ -9,9 +9,22 @@ class Toolbox::GoogleSearchTest < ActiveSupport::TestCase
   test "google_search returns the expected result" do
     expected_result = {
       message_to_user: "Web query: Sandi Metz POODR title",
-      query_results: "Practical Object-Oriented Design in Ruby by Sandi Metz"
+      query_results: "Practical Object-Oriented Design in Ruby by Sandi Metz. Learn more (https://www.poodr.com) for more details."
     }
-    body = "<!doctype html><html><body><div class=\"BNeawe vvjwJb AP7Wnd\">#{expected_result[:query_results]}</div></body></html>"
+    html_content = <<-HTML
+    <html>
+      <body>
+        <div class="BNeawe">
+          Practical Object-Oriented Design in Ruby by Sandi Metz.#{' '}
+          <a href="https://www.poodr.com\">Learn more</a> for more details.
+        </div>
+        <div class="OtherClass">
+          Not relevant content.
+          <a href="https://www.notrelevant.com">Ignore this</a>.
+        </div>
+      </body>
+    </html>
+    HTML
 
     stub_request(:get, /www.google.com/)
       .with(
@@ -21,7 +34,7 @@ class Toolbox::GoogleSearchTest < ActiveSupport::TestCase
           "User-Agent"=>"Ruby"
         }
       )
-      .to_return(status: 200, body: body, headers: {})
+      .to_return(status: 200, body: html_content, headers: {})
 
     result = @google_search.google_search(query_s: "Sandi Metz POODR title")
     assert_equal expected_result, result