Add google_search tool (#567)

Gemfile

@@ -87,4 +87,5 @@ group :test do
   gem "selenium-webdriver"
   gem "minitest-stub_any_instance"
   gem "rails-controller-testing"
+  gem "webmock"
 end

Gemfile.lock

@@ -122,6 +122,9 @@ GEM
     coderay (1.1.3)
     concurrent-ruby (1.3.4)
     connection_pool (2.4.1)
+    crack (1.0.0)
+      bigdecimal
+      rexml
     crass (1.0.6)
     date (3.3.4)
     debug (1.8.0)
@@ -167,6 +170,7 @@ GEM
       multi_json (~> 1.11)
       os (>= 0.9, < 2.0)
       signet (>= 0.16, < 2.a)
+    hashdiff (1.1.2)
     hashie (5.0.0)
     i18n (1.14.6)
       concurrent-ruby (~> 1.0)
@@ -463,6 +467,10 @@ GEM
       activemodel (>= 6.0.0)
       bindex (>= 0.4.0)
       railties (>= 6.0.0)
+    webmock (3.24.0)
+      addressable (>= 2.8.0)
+      crack (>= 0.3.2)
+      hashdiff (>= 0.4.0, < 2.0.0)
     webrick (1.8.2)
     websocket (1.2.11)
     websocket-driver (0.7.6)
@@ -526,6 +534,7 @@ DEPENDENCIES
   turbo-rails (~> 2.0.5)
   tzinfo-data
   web-console
+  webmock
 
 RUBY VERSION
    ruby 3.3.6p108

app/helpers/messages_helper.rb

@@ -27,12 +27,23 @@ def format_for_speaking(text)
   end
 
   def format_for_display(message, append_inside_tag: nil)
-    if memory_updated?(message)
-      return link_to JSON.parse(message.content_text)["message_to_user"],
+    if message_to_user_from_tool_call?(message)
+      function_name = message.content_tool_calls.dig(:function, :name)
+      message_to_user = JSON.parse(message.content_text)["message_to_user"]
+
+      case function_name
+      when "memory_remember_detail_about_user"
+        return link_to message_to_user,
         settings_memories_path,
         { data: { turbo_frame: "_top" }, class: "text-gray-400 dark:!text-gray-500 font-normal no-underline" }
-    elsif message_to_user_from_tool_call?(message)
-      return content_tag(:span, JSON.parse(message.content_text)["message_to_user"], class: "text-gray-400 dark:!text-gray-500")
+      when "googlesearch_google_search"
+        query = message_to_user.partition(":").last
+        return link_to message_to_user,
+        "https://www.google.com/search?q=#{URI.encode_www_form_component(query)}",
+        { target: :_blank, data: { turbo_frame: "_top" }, class: "text-gray-400 dark:!text-gray-500 font-normal no-underline" }
+      else
+        return content_tag(:span, message_to_user, class: "text-gray-400 dark:!text-gray-500")
+      end
     else
       escaped_text = html_escape(message.content_text)
 
@@ -74,10 +85,6 @@ def message_to_user_from_tool_call?(message)
     false
   end
 
-  def memory_updated?(message)
-    message.tool? && message.content_tool_calls.dig(:function, :name) == "memory_remember_detail_about_user"
-  end
-
   private
 
   def block_code

app/services/sdk/verb.rb

@@ -31,7 +31,7 @@ def handle(response)
     raise ResponseError.new(response) if !response.status.in? @expected_statuses
 
     if response.status.between?(200, 299)
-      response.body.presence && OpenData.for(JSON.parse(response.body))
+      response.body.presence && OpenData.for(JSON.parse(response.body)) rescue response
     else
       response
     end

app/services/toolbox.rb

@@ -7,6 +7,7 @@ def self.descendants
       test_env && Toolbox::HelloWorld,
       Toolbox::OpenMeteo,
       Toolbox::Memory,
+      Toolbox::GoogleSearch,
       gmail_active && Toolbox::Gmail,
       tasks_active && Toolbox::GoogleTasks,
     ].compact

app/services/toolbox/google_search.rb

@@ -0,0 +1,32 @@
+class Toolbox::GoogleSearch < Toolbox
+
+  describe :google_search, <<~S
+    Search Google for the indicated query.
+    Use this to answer questions about current events, look up information, or find answers to questions.
+    Try to use this sparingly; prefer to use the user's memories and the tools you have available to answer questions.
+    When you do use this, try to use exact queries for which you expect to get a definitive answer.
+    When you respond to the user, try to include an answer to the question rather than just a link.
+  S
+  def google_search(query_s:)
+    encoded_query = URI.encode_www_form_component(query_s)
+    response_body = get("https://www.google.com/search").param(q: encoded_query).body
+    doc = Nokogiri::HTML(response_body)
+
+    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}",
+      query_results: results
+    }
+  end
+end

test/helpers/messages_helper_test.rb

@@ -2,11 +2,6 @@
 
 class MessagesHelperTest < ActionView::TestCase
 
-  test "memory_updated responds properly for tool result" do
-    message = messages(:keep_memory_tool_result)
-    assert memory_updated?(message)
-  end
-
   test "message_to_user_from_tool_call responds properly for tool result" do
     message = messages(:keep_memory_tool_result)
     assert message_to_user_from_tool_call?(message)

test/services/toolbox/google_search_test.rb

@@ -0,0 +1,42 @@
+require "test_helper"
+
+class Toolbox::GoogleSearchTest < ActiveSupport::TestCase
+  setup do
+    @google_search = Toolbox::GoogleSearch.new
+    WebMock.enable!
+  end
+
+  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. Learn more (https://www.poodr.com) for more details."
+    }
+    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(
+        headers: {
+          "Accept"=>"*/*",
+          "Accept-Encoding"=>"gzip;q=1.0,deflate;q=0.6,identity;q=0.3",
+          "User-Agent"=>"Ruby"
+        }
+      )
+      .to_return(status: 200, body: html_content, headers: {})
+
+    result = @google_search.google_search(query_s: "Sandi Metz POODR title")
+    assert_equal expected_result, result
+  end
+end

test/services/toolbox/open_meteo_test.rb

@@ -26,6 +26,8 @@ class Toolbox::OpenMeteoTest < ActiveSupport::TestCase
   end
 
   test "get_current_and_todays_weather hits the API and doesn't fail" do
+    WebMock.disable!
+
     allow_request(:get, :get_location) do
       allow_request(:get, :get_current_and_todays_weather) do
         result = @open_meteo.get_current_and_todays_weather(city_s: "Austin", state_province_or_region_s: "Texas")
@@ -35,6 +37,8 @@ class Toolbox::OpenMeteoTest < ActiveSupport::TestCase
   end
 
   test "get_current_and_todays_weather works as a tool call" do
+    WebMock.disable!
+
     allow_request(:get, :get_location) do
       allow_request(:get, :get_current_and_todays_weather) do
         result = Toolbox.call("openmeteo_get_current_and_todays_weather", city: "Austin", state_province_or_region: "Texas")

test/test_helper.rb

@@ -3,6 +3,7 @@
 require "rails/test_help"
 require "minitest/autorun"
 require "pry"
+require "webmock/minitest"
 
 Dir[Rails.root.join("test/support/**/*.rb")].sort.each { |file| require file }