Response for code challenge

Gemfile

@@ -0,0 +1,7 @@
+# frozen_string_literal: true
+
+source 'https://rubygems.org'
+
+gem 'nokogiri', '~> 1.18'
+gem 'rspec', '~> 3.13'
+gem 'selenium-webdriver', '~> 4.31'

Gemfile.lock

@@ -0,0 +1,42 @@
+GEM
+  remote: https://rubygems.org/
+  specs:
+    base64 (0.2.0)
+    diff-lcs (1.6.1)
+    logger (1.7.0)
+    nokogiri (1.18.7-x64-mingw-ucrt)
+      racc (~> 1.4)
+    racc (1.8.1)
+    rexml (3.4.1)
+    rspec (3.13.0)
+      rspec-core (~> 3.13.0)
+      rspec-expectations (~> 3.13.0)
+      rspec-mocks (~> 3.13.0)
+    rspec-core (3.13.3)
+      rspec-support (~> 3.13.0)
+    rspec-expectations (3.13.3)
+      diff-lcs (>= 1.2.0, < 2.0)
+      rspec-support (~> 3.13.0)
+    rspec-mocks (3.13.2)
+      diff-lcs (>= 1.2.0, < 2.0)
+      rspec-support (~> 3.13.0)
+    rspec-support (3.13.2)
+    rubyzip (2.4.1)
+    selenium-webdriver (4.31.0)
+      base64 (~> 0.2)
+      logger (~> 1.4)
+      rexml (~> 3.2, >= 3.2.5)
+      rubyzip (>= 1.2.2, < 3.0)
+      websocket (~> 1.0)
+    websocket (1.2.11)
+
+PLATFORMS
+  x64-mingw-ucrt
+
+DEPENDENCIES
+  nokogiri (~> 1.18)
+  rspec (~> 3.13)
+  selenium-webdriver (~> 4.31)
+
+BUNDLED WITH
+   2.5.11

bin/test

@@ -0,0 +1,9 @@
+#!/bin/ruby
+# frozen_string_literal: true
+
+require 'json'
+require_relative '../lib/serpapi'
+
+html_file = File.join(__dir__, '../spec/fixtures/richard-estes/page.html')
+search = SerpAPI::GoogleSearch.new
+puts JSON.dump(search.get_hash(html_file))

lib/serpapi.rb

@@ -0,0 +1,8 @@
+# frozen_string_literal: true
+
+##
+# Module for code challenge classes
+module SerpAPI
+end
+
+require_relative 'serpapi/google_search'

lib/serpapi/google_search.rb

@@ -0,0 +1,118 @@
+# frozen_string_literal: true
+
+require 'json'
+require 'nokogiri'
+require 'selenium-webdriver'
+
+# Represents a blank 1x1 GIF image
+# Returned in initial page load by Google Search before image thumbnails have been populated by JavaScript
+EMPTY_BASE64_GIF = 'data:image/gif;base64,R0lGODlhAQABAIAAAP///////yH5BAEKAAEALAAAAAABAAEAAAICTAEAOw=='
+
+module SerpAPI
+  ##
+  # This class contains Google Search-related methods for the code challenge.
+  class GoogleSearch
+    ##
+    # Creates a new SerpAPI::GoogleSearch instance
+    #
+    # NOTE: Real SerpAPI puts the search request parameters in the args. However,
+    # I put them in `get_hash` for the purpose of this code exercise since it
+    # allows one to reuse a single instance of Selenium, saving on
+    # startup/resources.
+    def initialize
+      @driver = make_web_driver
+    end
+
+    ##
+    # Obtains the result hash for the google search.
+    #
+    # NOTE: Unlike SerpAPI code this requires the HTML file with the results to parse (optimization)
+    # Also only inline images are returned for the code challenge purpose.
+    #
+    # @param [String] html_file Path to an HTML file to parse
+    # @return [Hash] Parsed results
+    def get_hash(html_file)
+      page = fetch(html_file)
+
+      {
+        'knowledge_graph' => knowledge_graph(page)
+      }
+    end
+
+    private
+
+    ##
+    # Creates and configures the Selenium web driver.
+    #
+    # @return [Selenium::WebDriver]
+    def make_web_driver
+      options = Selenium::WebDriver::Chrome::Options.new.tap do |o|
+        o.add_argument('--headless')
+        o.add_argument('--no-sandbox') # initial browser startup times out unless this is added
+      end
+
+      # Increase timeouts to avoid exceptions on Selenium startup
+      Selenium::WebDriver.for(:chrome, options: options).tap do |d|
+        d.manage.timeouts.implicit_wait = 120
+        d.manage.timeouts.script_timeout = 120
+        d.manage.timeouts.page_load = 120
+      end
+    end
+
+    ##
+    # Fetches the webpage for Google Search using Selenium.
+    # We must use Selenium because the page runs some clientside Javascript on load that for example
+    # changes the img elements to have the correct base64-encoded thumbnails.
+    #
+    # @param [String] html_file Path to an HTML file to parse
+    # @return [Nokogiri::XML::Node] Nokogiri parsed page
+    def fetch(html_file)
+      @driver.get("file:///#{html_file}")
+      Nokogiri::HTML5(@driver.page_source)
+    end
+
+    ##
+    # Returns knowledge graph information from the search results page (only artworks, this time)
+    #
+    # @param [Nokogiri::XML::Node] page Nokogiri parsed page
+    # @return [Hash] Parsed results under "knowledge_graph" category
+    def knowledge_graph(page)
+      # This this the HTML for the entries under the "Artworks" tab of the
+      # knowledge graph search widget at top of results.
+      works_container = page.css('[data-attrid="kc:/visual_art/visual_artist:works"] > .xac7Kb > .Cz5hV > .iELo6')
+      artworks = works_container.map(&method(:parse_artwork))
+
+      {
+        'artworks' => artworks
+      }
+    end
+
+    def parse_artwork(work)
+      link = work.at('a')
+      image = link.at('img')
+
+      # When loading the raw page, the image thumbnails `src` attribute in the
+      # knowledge graph section are set to a blank 1x1 GIF. After the inline
+      # JavaScript executes, the thumbnails will be replaced with the full
+      # base64-encoded image. However, the thumbnails not visible on the page
+      # (hidden under "More...") will still be set to 1x1 GIFs. In that case,
+      # use the `data-src` attribute for the `image` property as it gives a URL
+      # (https://encrypted-tbn2.gstatic.com) pointing to the thumbnail.
+      image_src = image.attr('src')
+      image_src = image.attr('data-src') if image_src == EMPTY_BASE64_GIF
+
+      result = {
+        'name' => image.attr('alt'),
+        'link' => "https://www.google.com#{link.attr('href')}",
+        'image' => image_src
+      }
+
+      # Some of the images may not have a year included, in that case do not add
+      # any `extensions` to the result array
+      extensions = work.css('.cxzHyb').map { |ext| ext.inner_html.to_s }.reject(&:empty?)
+      result['extensions'] = extensions unless extensions.empty?
+
+      result
+    end
+  end
+end