Merge pull request #1830 from sul-dlss/1761

Add rake task to remediate empty anchors fixes #1761

.rubocop_todo.yml

@@ -1,19 +1,12 @@
 # This configuration was generated by
 # `rubocop --auto-gen-config`
-# on 2020-02-27 21:19:26 -0800 using RuboCop version 0.80.0.
+# on 2020-03-04 10:41:06 -0700 using RuboCop version 0.80.1.
 # The point is for the user to remove these configuration records
 # one by one as the offenses are removed from the code base.
 # Note that changes in the inspected code, or installation of new
 # versions of RuboCop, may require this file to be generated again.
 
-# Offense count: 1
-# Cop supports --auto-correct.
-# Configuration parameters: AutoCorrect, AllowHeredoc, AllowURI, URISchemes, IgnoreCopDirectives, IgnoredPatterns.
-# URISchemes: http, https
-Layout/LineLength:
-  Max: 122
-
-# Offense count: 12
+# Offense count: 13
 Metrics/AbcSize:
   Max: 42
 
@@ -56,7 +49,7 @@ RSpec/ExpectInHook:
     - 'spec/views/catalog/_exhibits_document_header_default.html.erb_spec.rb'
     - 'spec/views/viewers/edit.html.erb_spec.rb'
 
-# Offense count: 113
+# Offense count: 114
 # Configuration parameters: AggregateFailuresByDefault.
 RSpec/MultipleExpectations:
   Max: 6
@@ -96,6 +89,11 @@ Style/MixinUsage:
     - 'lib/traject/dor_config.rb'
     - 'spec/features/viewers_spec.rb'
 
+# Offense count: 1
+Style/MultilineBlockChain:
+  Exclude:
+    - 'lib/tasks/spotlight.rake'
+
 # Offense count: 1
 # Cop supports --auto-correct.
 # Configuration parameters: EnforcedStyle.

lib/tasks/spotlight.rake

@@ -10,4 +10,18 @@ namespace :spotlight do
       system 'bundle exec rake db:migrate'
     end
   end
+  desc 'Remove empty anchors'
+  task remediate_anchors: :environment do
+    url_regex = %r{<a href="(https?|mailto:)[\S]+">(<br>)?<\/a>}
+    Spotlight::Page.find_each.select do |p|
+      p.content.any? { |c| c.text.to_s.match? url_regex }
+    end.each do |p|
+      dup = p.content
+      p.content.each_with_index do |c, i|
+        dup[i].text = c.text.to_s.gsub(url_regex, '') if c.text.to_s.match? url_regex
+      end
+      p.content = dup
+      p.save
+    end
+  end
 end