From 450ffb99ef47c0b68f6ad99c77a9ec336dcd9696 Mon Sep 17 00:00:00 2001
From: Alex Kiessling <32677188+ajkiessl@users.noreply.github.com>
Date: Mon, 31 Oct 2022 12:59:59 -0400
Subject: [PATCH] Add view components for rendering google scholar meta tags
 for scraping. (#104)

* Add view components for rendering google scholar meta tags for scraping.

* Adds tests for google scholar meta tags

* Add google_scholar_metadata_component_spec.rb

* Moved let statement to top level in google_scholar_metadata test
---
 Gemfile.lock                                  |  3 +
 .../custom_document_component.html.erb        |  3 +
 ...google_scholar_metadata_component.html.erb |  7 +++
 .../google_scholar_metadata_component.rb      | 24 ++++++++
 .../google_scholar_metadata_component_spec.rb | 61 +++++++++++++++++++
 spec/integration/home_page_spec.rb            |  7 +++
 spec/rails_helper.rb                          |  3 +
 7 files changed, 108 insertions(+)
 create mode 100644 app/components/google_scholar_metadata_component.html.erb
 create mode 100644 app/components/google_scholar_metadata_component.rb
 create mode 100644 spec/component/google_scholar_metadata_component_spec.rb

diff --git a/Gemfile.lock b/Gemfile.lock
index 334c302..71c7b49 100644
--- a/Gemfile.lock
+++ b/Gemfile.lock
@@ -236,6 +236,8 @@ GEM
       rubocop-rspec (~> 2)
       scss_lint (~> 0.55)
     nio4r (2.5.8)
+    nokogiri (1.13.7-arm64-darwin)
+      racc (~> 1.4)
     nokogiri (1.13.7-x86_64-darwin)
       racc (~> 1.4)
     nokogiri (1.13.7-x86_64-linux)
@@ -435,6 +437,7 @@ GEM
     zeitwerk (2.5.4)
 
 PLATFORMS
+  arm64-darwin-21
   x86_64-darwin-20
   x86_64-darwin-21
   x86_64-linux
diff --git a/app/components/custom_document_component.html.erb b/app/components/custom_document_component.html.erb
index 8b8790e..4cea85a 100644
--- a/app/components/custom_document_component.html.erb
+++ b/app/components/custom_document_component.html.erb
@@ -8,6 +8,9 @@
                 itemtype: @document.itemtype,
                 class: classes.flatten.join(' ') do %>
   <%= header %>
+  <% if show? %>
+    <%= render GoogleScholarMetadataComponent.new(document: @document) %>
+  <% end %>
   <% if body.present? %>
     <%= body %>
   <% else %>
diff --git a/app/components/google_scholar_metadata_component.html.erb b/app/components/google_scholar_metadata_component.html.erb
new file mode 100644
index 0000000..925f769
--- /dev/null
+++ b/app/components/google_scholar_metadata_component.html.erb
@@ -0,0 +1,7 @@
+<!-- Google Scholar metadata -->
+<meta name="citation_title" content="<%= @document[:title_tesi] %>">
+<meta name="citation_author" content="<%= @document[:author_name_tesi] %>">
+<meta name="citation_publication_date" content="<%= citation_publication_date %>">
+<%- if citation_pdf_url %>
+  <meta name="citation_pdf_url" content="<%= citation_pdf_url %>">
+<%- end %>
diff --git a/app/components/google_scholar_metadata_component.rb b/app/components/google_scholar_metadata_component.rb
new file mode 100644
index 0000000..56aca07
--- /dev/null
+++ b/app/components/google_scholar_metadata_component.rb
@@ -0,0 +1,24 @@
+# frozen_string_literal: true
+
+class GoogleScholarMetadataComponent < ViewComponent::Base
+  attr_reader :document
+
+  def initialize(document)
+    super
+    @document = document.fetch(:document)
+  end
+
+  def citation_publication_date
+    document[:released_metadata_at_dtsi]&.to_date&.year
+  end
+
+  def citation_pdf_url
+    return nil unless document.access_level.current_access_level == 'open_access'
+
+    document.final_submissions.each do |key, value|
+      return final_submission_file_url(key) if File.extname(value) == '.pdf'
+    end
+
+    nil
+  end
+end
diff --git a/spec/component/google_scholar_metadata_component_spec.rb b/spec/component/google_scholar_metadata_component_spec.rb
new file mode 100644
index 0000000..4c286d4
--- /dev/null
+++ b/spec/component/google_scholar_metadata_component_spec.rb
@@ -0,0 +1,61 @@
+# frozen_string_literal: true
+
+require 'rails_helper'
+
+RSpec.describe GoogleScholarMetadataComponent, type: :component do
+  include Rails.application.routes.url_helpers
+
+  subject(:component) { described_class.new(document: doc) }
+
+  let(:doc) { SolrDocument.new(fake_solr_doc) }
+  let(:html) { render_inline(component) }
+
+  before do
+    Blacklight.default_index.connection.add(doc)
+    Blacklight.default_index.connection.commit
+  end
+
+  context 'with a final submission file pdf' do
+    let(:fake_solr_doc) { FakeSolrDocument.new(access_level: 'open_access').doc }
+
+    it 'renders all the meta tags' do
+      expect(html.search('meta[@name="citation_title"]')
+        .first['content']).to eq(doc[:title_tesi])
+      expect(html.search('meta[@name="citation_author"]')
+        .first['content']).to eq(doc[:author_name_tesi])
+      expect(html.search('meta[@name="citation_publication_date"]')
+        .first['content']).to eq(doc[:released_metadata_at_dtsi].to_date.year.to_s)
+      expect(html.search('meta[@name="citation_pdf_url"]').first['content']).to eq(
+        final_submission_file_url(doc.final_submissions.first.first, host: 'http://test.host')
+      )
+    end
+  end
+
+  context 'when the final submission file is not a pdf' do
+    let(:fake_solr_doc) { FakeSolrDocument.new(access_level: 'open_access', file_names: ['word_doc.docx']).doc }
+
+    it 'renders all the meta tags' do
+      expect(html.search('meta[@name="citation_title"]')
+        .first['content']).to eq(doc[:title_tesi])
+      expect(html.search('meta[@name="citation_author"]')
+        .first['content']).to eq(doc[:author_name_tesi])
+      expect(html.search('meta[@name="citation_publication_date"]')
+        .first['content']).to eq(doc[:released_metadata_at_dtsi].to_date.year.to_s)
+      expect(html.search('meta[@name="citation_pdf_url"]')).to be_empty
+    end
+  end
+
+  context 'when the final submission is not open_access' do
+    let(:fake_solr_doc) { FakeSolrDocument.new(access_level: 'restricted_to_institution').doc }
+
+    it 'renders all the meta tags' do
+      expect(html.search('meta[@name="citation_title"]')
+        .first['content']).to eq(doc[:title_tesi])
+      expect(html.search('meta[@name="citation_author"]')
+        .first['content']).to eq(doc[:author_name_tesi])
+      expect(html.search('meta[@name="citation_publication_date"]')
+        .first['content']).to eq(doc[:released_metadata_at_dtsi].to_date.year.to_s)
+      expect(html.search('meta[@name="citation_pdf_url"]')).to be_empty
+    end
+  end
+end
diff --git a/spec/integration/home_page_spec.rb b/spec/integration/home_page_spec.rb
index f64127b..83eb065 100644
--- a/spec/integration/home_page_spec.rb
+++ b/spec/integration/home_page_spec.rb
@@ -19,6 +19,13 @@
     it 'shows Explore button to search' do
       expect(page).to have_button 'Explore'
     end
+
+    it 'does not display Google Scholar meta tags' do
+      expect(page).not_to have_css 'meta[@name="citation_title"]'
+      expect(page).not_to have_css 'meta[@name="citation_author"]'
+      expect(page).not_to have_css 'meta[@name="citation_publication_date"]'
+      expect(page).not_to have_css 'meta[@name="citation_pdf_url"]'
+    end
   end
 
   context 'when a user is logged in' do
diff --git a/spec/rails_helper.rb b/spec/rails_helper.rb
index 3663542..6964402 100644
--- a/spec/rails_helper.rb
+++ b/spec/rails_helper.rb
@@ -10,6 +10,7 @@
 require 'capybara/rails'
 # Add additional requires below this line. Rails is not loaded until this point!
 require 'devise'
+require 'view_component/test_helpers'
 
 # Requires supporting ruby files with custom matchers and macros, etc, in
 # spec/support/ and its subdirectories. Files matching `spec/**/*_spec.rb` are
@@ -64,6 +65,8 @@
   config.include Devise::Test::ControllerHelpers, type: :controller
   config.include Devise::Test::IntegrationHelpers, type: :feature
 
+  config.include ViewComponent::TestHelpers, type: :component
+
   # Filter lines from Rails gems in backtraces.
   config.filter_rails_from_backtrace!
   # arbitrary gems may also be filtered via: