updates sanitize gem to 5.0 with a bug fix for restoring searchLink elements to proper case

Author: bmckinney-ebsco

ebsco-eds.gemspec

@@ -42,9 +42,10 @@ Gem::Specification.new do |spec|
   spec.add_dependency 'csl', '~> 1.4'
   spec.add_dependency 'citeproc-ruby', '~> 1.0', '>= 1.0.2'
   spec.add_dependency 'csl-styles', '~> 1.0', '>= 1.0.1.5'
-  spec.add_dependency 'activesupport', '~> 5.0'
+  spec.add_dependency 'activesupport', '~> 5.2'
   spec.add_dependency 'net-http-persistent', '~> 3.1'
-  spec.add_dependency 'sanitize', '~> 4.6.6'
+  spec.add_dependency 'sanitize', '~> 5.0'
+  spec.add_dependency 'public_suffix', '~>4.0'
 
   spec.add_development_dependency 'bundler', '~> 1.13'
   spec.add_development_dependency 'rake', '~> 12.3'

lib/ebsco/eds/record.rb

@@ -961,7 +961,12 @@ def html_decode_and_sanitize(data, config = nil)
           html = html.gsub(/\\n /, '')
         end
 
-        Sanitize.fragment(html, sanitize_config)
+        sanitized_html = Sanitize.fragment(html, sanitize_config)
+        # sanitize 5.0 fails to restore element case after doing lowercase
+        sanitized_html = sanitized_html.gsub(/<searchlink/, '<searchLink')
+        sanitized_html = sanitized_html.gsub(/<\/searchlink>/, '</searchLink>')
+        sanitized_html
+
       end
 
       # dynamically add item metadata as 'eds_extra_ItemNameOrLabel'

test/html_test.rb

@@ -81,7 +81,7 @@ def test_sanitize_html_fulltext
     if session_with_sanitize.dbid_in_profile 'bah'
       record = session_with_sanitize.retrieve({dbid: 'bah', an: '116897973'})
       #puts 'HTML AFTER: ' + record.eds_html_fulltext.to_s
-      assert record.eds_html_fulltext.to_s.include? '<h1 id="AN0116897973-3">YOU\'VE GOT TO BE KITTEN ME! </h1>'
+      assert record.eds_html_fulltext.to_s.include? '<h1 id="AN0116897973-3">YOU\'VE GOT TO BE KITTEN ME!&nbsp;</h1>'
     else
       puts "WARNING: skipping test_sanitize_html_fulltext, bah db isn't in the profile."
     end