From e68e39dc216a664dc47b1209cff781cc0487836a Mon Sep 17 00:00:00 2001
From: Gregory Saumier-Finch <gregory@culturecreates.com>
Date: Tue, 19 Nov 2024 11:42:23 -0500
Subject: [PATCH] Fixed create_browser to use positional parameters including
 tests

---
 Gemfile                        |  1 +
 Gemfile.lock                   |  4 ++++
 src/lib/headless_browser.rb    | 26 +++++++++++++++++--------
 tests/headless_browser_test.rb | 35 ++++++++++++++++++++++++++++++++++
 4 files changed, 58 insertions(+), 8 deletions(-)

diff --git a/Gemfile b/Gemfile
index 5d4d2cb..0c2f0c8 100644
--- a/Gemfile
+++ b/Gemfile
@@ -6,5 +6,6 @@ gem 'nokogiri'
 gem 'open-uri'
 gem 'rake'
 gem 'ferrum'
+gem 'mocha'
 
 
diff --git a/Gemfile.lock b/Gemfile.lock
index 4e63ccd..03a296f 100644
--- a/Gemfile.lock
+++ b/Gemfile.lock
@@ -75,6 +75,8 @@ GEM
     logger (1.6.0)
     matrix (0.4.2)
     minitest (5.21.2)
+    mocha (2.5.0)
+      ruby2_keywords (>= 0.0.5)
     multi_json (1.15.0)
     net-http-persistent (4.0.2)
       connection_pool (~> 2.2)
@@ -165,6 +167,7 @@ GEM
       rdf (~> 3.3)
       rexml (~> 3.2)
     rexml (3.2.6)
+    ruby2_keywords (0.0.5)
     scanf (1.0.0)
     shacl (0.4.1)
       json-ld (~> 3.3)
@@ -225,6 +228,7 @@ DEPENDENCIES
   ferrum
   linkeddata
   minitest
+  mocha
   nokogiri
   open-uri
   rake
diff --git a/src/lib/headless_browser.rb b/src/lib/headless_browser.rb
index 9c7d72d..59b8bbe 100644
--- a/src/lib/headless_browser.rb
+++ b/src/lib/headless_browser.rb
@@ -1,6 +1,7 @@
 require 'ferrum'
 require 'json'
 require 'linkeddata'
+require 'rbconfig'
 
 module HeadlessBrowser
   def self.fetch_json_ld_objects(entity_urls, base_url, headers, sparql_paths, browser: nil, graph: nil)
@@ -16,15 +17,18 @@ def self.fetch_json_ld_objects(entity_urls, base_url, headers, sparql_paths, bro
     graph
   end
 
-  def self.create_browser(headers: nil)
-    # MacOS "/Applications/Google Chrome.app/Contents/MacOS/Google Chrome"
-    # Linux "/usr/bin/google-chrome-stable"
-    browser = Ferrum::Browser.new(browser_path: "/usr/bin/google-chrome-stable", headless: true, pending_connection_errors: false, process_timeout: 60, xvfb: true, browser_options: { 'no-sandbox': nil })
+  def self.create_browser(headers = nil)
+    browser_path =  if running_on_macos?
+                      "/Applications/Google Chrome.app/Contents/MacOS/Google Chrome"
+                    else
+                      "/usr/bin/google-chrome-stable"
+                    end
+    browser = Ferrum::Browser.new(browser_path: browser_path, headless: true, pending_connection_errors: false, process_timeout: 60, xvfb: true, browser_options: { 'no-sandbox': nil })
     browser.headers.set(headers) if headers
     browser
   end
 
-  def self.process_entity_url(entity_url, browser, graph, add_url_sparql_file: nil)
+  def self.process_entity_url(entity_url, browser, graph, add_url_sparql_file = nil)
     puts "Processing #{entity_url} in headless mode"
     browser.go_to(entity_url)
     sleep 15
@@ -37,13 +41,15 @@ def self.process_entity_url(entity_url, browser, graph, add_url_sparql_file: nil
     puts "Error processing #{entity_url} in headless mode: #{e.message}"
   end
 
-  def self.process_json_ld_script(script, entity_url, graph, add_url_sparql_file)
+  def self.process_json_ld_script(script, entity_url, graph, add_url_sparql_file = nil)
     # Parse the JSON-LD string into a JSON object
     json_ld = string_to_json(script.text)
     # Convert the JSON-LD object to an RDF graph
     loaded_graph = RDF::Graph.new << JSON::LD::API.toRdf(json_ld)
-    # sparql_file_with_url = add_url_sparql_file.gsub("subject_url", entity_url)
-    # loaded_graph.query(SPARQL.parse(sparql_file_with_url, update: true))
+    if add_url_sparql_file
+      sparql_file_with_url = add_url_sparql_file.gsub("subject_url", entity_url)
+      loaded_graph.query(SPARQL.parse(sparql_file_with_url, update: true))
+    end
     graph << loaded_graph
   rescue JSON::ParserError => e
     puts "Error parsing JSON-LD: #{e.message}"
@@ -61,4 +67,8 @@ def self.fetch_entity_urls_headless(url, headers, browser: nil)
     sleep 15
     browser.body
   end
+
+  def self.running_on_macos?
+    RbConfig::CONFIG['host_os'] =~ /darwin|mac os/
+  end
 end
\ No newline at end of file
diff --git a/tests/headless_browser_test.rb b/tests/headless_browser_test.rb
index 6839054..9cdc53e 100644
--- a/tests/headless_browser_test.rb
+++ b/tests/headless_browser_test.rb
@@ -1,4 +1,5 @@
 require 'minitest/autorun'
+require 'mocha/minitest'
 require 'ferrum'
 require 'linkeddata'
 require_relative '../src/lib/headless_browser'
@@ -18,6 +19,40 @@ def test_string_to_json_with_newlines
     assert_equal expected, actual
   end
 
+  # Test create_browser method
+  def test_create_browser
+    # Stub the Ferrum::Browser.new method
+    Ferrum::Browser.stubs(:new).returns(mock_browser)
 
+    # Call the method
+    browser = HeadlessBrowser.create_browser
 
+    # Assertions
+    assert_instance_of Mocha::Mock, browser
+  end
+
+  def test_create_browser_with_headers    
+    # Stub the Ferrum::Browser.new method
+    Ferrum::Browser.stubs(:new).returns(mock_browser)
+    
+    # Call the method with headers
+    browser = HeadlessBrowser.create_browser({"User-Agent" => "Chrome"})
+
+    # Assertions
+    assert_instance_of Mocha::Mock, browser
+    assert_equal "Chrome", browser.headers["User-Agent"]
+  end
+
+  private
+
+  def mock_browser
+    headers_mock = mock('headers')
+    headers_mock.stubs(:set)
+    headers_mock.stubs(:[]).returns("Chrome")
+
+    browser = mock('browser')
+    browser.stubs(:headers).returns(headers_mock)
+    
+    browser
+  end
 end
\ No newline at end of file