Avoid duplicates of special domains when showing full notice version

berkmancenter · Sep 5, 2022 · 83bea79 · 83bea79
1 parent d21e3e5
commit 83bea79
Show file tree

Hide file tree

Showing 17 changed files with 81 additions and 55 deletions.
diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -1,5 +1,9 @@
 The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/). It uses [CalVer](https://calver.org/) as of May 2019.
 
+## [22.09a](https://github.com/berkmancenter/lumendatabase/releases/tag/2022.09a) - 2022-09-05
+### Fixed
+* Fixed duplicates of `special domains` when showing the full notice version.
+
 ## [22.09](https://github.com/berkmancenter/lumendatabase/releases/tag/2022.09) - 2022-09-05
 ### Added
 * Introduced special domains.

diff --git a/app/helpers/notices_helper.rb b/app/helpers/notices_helper.rb
@@ -2,9 +2,9 @@ module NoticesHelper
   # Determines whether 'click to request access' should be an option.
   # Yes, the messy function signature implies there is a lot going on with the
   # business logic.
-  def access_requestable?(notice, show_original, show_infringing)
+  def access_requestable?(notice, show_copyrighted, show_infringing)
     [
-      show_original || show_infringing,
+      show_copyrighted || show_infringing,
       # Additional access cannot be requested for confidential court orders
       # as there is nothing further to display.
       !confidential_order?(notice),
@@ -130,13 +130,6 @@ def placeholder_kind(user, role)
     { selected: user.entity.kind }
   end
 
-  def work_url(url)
-    return url if SpecialDomain.where('? ~~* domain_name', url).where("why_special ? 'full_urls_only_for_researchers'").none? ||
-                  (Current.user && (Current.user.role?(Role.researcher) || Current.user.role?(Role.super_admin)))
-
-    Work.fqdn_from_url(url)
-  end
-
   private
 
   def confidential_order?(notice)

diff --git a/app/models/url.rb b/app/models/url.rb
@@ -6,6 +6,7 @@ class Url
   include ValidatesUrls
 
   attr_reader :url_original
+  attr_accessor :only_fqdn
 
   def [](index)
     self.instance_variable_get("@#{index}")

diff --git a/app/models/work.rb b/app/models/work.rb
@@ -87,6 +87,14 @@ def self.fqdn_from_url(url)
     fqdn
   end
 
+  def infringing_urls_public
+    filter_urls(infringing_urls)
+  end
+
+  def copyrighted_urls_public
+    filter_urls(copyrighted_urls)
+  end
+
   # == Private Methods =========================================================
   private
 
@@ -167,4 +175,28 @@ def valid_url(type, params)
   def not_valid_url_return(type)
     type.new(url: 'invalid')
   end
+
+  def filter_urls(url_instances)
+    filtered_raw_urls = []
+    fqdns = []
+    user_allowed = (Current.user && (Current.user.role?(Role.researcher) || Current.user.role?(Role.super_admin)))
+
+    url_instances.each do |url_instance|
+      just_url = url_instance.url
+
+      filtered_raw_urls << url_instance and next if SpecialDomain.where('? ~~* domain_name', just_url).where("why_special ? 'full_urls_only_for_researchers'").none? ||
+                                                    user_allowed
+
+      fqdn = Work.fqdn_from_url(just_url)
+
+      next if fqdns.include?(fqdn)
+
+      fqdns << fqdn
+      url_instance.url = fqdn
+      url_instance.only_fqdn = true
+      filtered_raw_urls << url_instance
+    end
+
+    filtered_raw_urls.uniq
+  end
 end
diff --git a/app/views/notices/_counterfeit_works.html.erb b/app/views/notices/_counterfeit_works.html.erb
@@ -11,9 +11,9 @@
 
         <%= render 'works_urls',
                     work: work,
-                    original_title: 'Original URLs:',
+                    copyrighted_title: 'Original URLs:',
                     infringing_title: 'Allegedly Infringing Counterfeit URLs:',
-                    show_original: false,
+                    show_copyrighted: false,
                     show_infringing: true
         %>
       <% end %>

diff --git a/app/views/notices/_counternotice_works.html.erb b/app/views/notices/_counternotice_works.html.erb
@@ -16,9 +16,9 @@
 
         <%= render 'works_urls',
                     work: work,
-                    original_title: 'Original URLs:',
+                    copyrighted_title: 'Original URLs:',
                     infringing_title: 'Allegedly Infringing URLs:',
-                    show_original: true,
+                    show_copyrighted: true,
                     show_infringing: true
         %>
       <% end %>

diff --git a/app/views/notices/_court_order_works.html.erb b/app/views/notices/_court_order_works.html.erb
@@ -12,7 +12,7 @@
         <%= render 'works_urls',
                     work: work,
                     infringing_title: 'Targeted URLs:',
-                    show_original: false,
+                    show_copyrighted: false,
                     show_infringing: true
         %>
       <% end %>

diff --git a/app/views/notices/_data_protection_works.html.erb b/app/views/notices/_data_protection_works.html.erb
@@ -7,7 +7,7 @@
         <%= render 'works_urls',
                     work: work,
                     infringing_title: 'Location of Some of the Material Requested for Removal:',
-                    show_original: false,
+                    show_copyrighted: false,
                     show_infringing: true
         %>
       <% end %>

diff --git a/app/views/notices/_defamation_works.html.erb b/app/views/notices/_defamation_works.html.erb
@@ -7,7 +7,7 @@
         <%= render 'works_urls',
                     work: work,
                     infringing_title: 'URLs of Allegedly Defamatory Material:',
-                    show_original: false,
+                    show_copyrighted: false,
                     show_infringing: true
         %>
       <% end %>

diff --git a/app/views/notices/_dmca_works.html.erb b/app/views/notices/_dmca_works.html.erb
@@ -16,9 +16,9 @@
 
         <%= render 'works_urls',
                     work: work,
-                    original_title: 'Original URLs:',
+                    copyrighted_title: 'Original URLs:',
                     infringing_title: 'Allegedly Infringing URLs:',
-                    show_original: true,
+                    show_copyrighted: true,
                     show_infringing: true
         %>
       <% end %>

diff --git a/app/views/notices/_government_request_works.html.erb b/app/views/notices/_government_request_works.html.erb
@@ -11,9 +11,9 @@
 
         <%= render 'works_urls',
                     work: work,
-                    original_title: 'URLs of original work:',
+                    copyrighted_title: 'URLs of original work:',
                     infringing_title: 'URLs mentioned in request:',
-                    show_original: true,
+                    show_copyrighted: true,
                     show_infringing: true
         %>
       <% end %>

diff --git a/app/views/notices/_law_enforcement_request_works.html.erb b/app/views/notices/_law_enforcement_request_works.html.erb
@@ -11,9 +11,9 @@
 
         <%= render 'works_urls',
                     work: work,
-                    original_title: 'URLs of original work:',
+                    copyrighted_title: 'URLs of original work:',
                     infringing_title: 'URLs mentioned in request:',
-                    show_original: true,
+                    show_copyrighted: true,
                     show_infringing: true
         %>
       <% end %>

diff --git a/app/views/notices/_other_works.html.erb b/app/views/notices/_other_works.html.erb
@@ -16,9 +16,9 @@
 
         <%= render 'works_urls',
                     work: work,
-                    original_title: 'URLs of original work:',
+                    copyrighted_title: 'URLs of original work:',
                     infringing_title: 'Problematic URLs:',
-                    show_original: true,
+                    show_copyrighted: true,
                     show_infringing: true
         %>
       <% end %>

diff --git a/app/views/notices/_private_information_works.html.erb b/app/views/notices/_private_information_works.html.erb
@@ -16,9 +16,9 @@
 
         <%= render 'works_urls',
                     work: work,
-                    original_title: 'URLs of original work:',
+                    copyrighted_title: 'URLs of original work:',
                     infringing_title: 'URLs with private information:',
-                    show_original: true,
+                    show_copyrighted: true,
                     show_infringing: true
         %>
       <% end %>

diff --git a/app/views/notices/_trademark_works.html.erb b/app/views/notices/_trademark_works.html.erb
@@ -13,7 +13,7 @@
         <%= render 'works_urls',
                     work: work,
                     infringing_title: 'Allegedly Infringing URLs:',
-                    show_original: false,
+                    show_copyrighted: false,
                     show_infringing: true
         %>
       <% end %>

diff --git a/app/views/notices/_works_urls.html.erb b/app/views/notices/_works_urls.html.erb
@@ -1,33 +1,12 @@
-<% if can_see_full_notice_version?(@notice) %>
-  <% if show_original %>
-    <div class="row">
-      <span class="label original-title"><%= original_title %></span>
-      <ol class="list original-urls">
-        <% if work.copyrighted_urls.each do |url| %>
-          <%= content_tag_for(:li, url) do %><%= work_url(url.url) %><% end %>
-        <% end.empty? %>
-          No copyrighted URLs were submitted.
-        <% end %>
-      </ol>
-    </div>
-  <% end %>
+<% copyrighted_title = copyrighted_title || '' %>
 
-  <% if show_infringing %>
-    <div class="row">
-      <span class="label infringing-title"><%= infringing_title %></span>
-      <ol class="list infringing-urls">
-        <% if work.infringing_urls.each do |url| %>
-          <%= content_tag_for(:li, url) do %><%= work_url(url.url) %><% end %>
-        <% end.empty? %>
-          No infringing URLs were submitted.
-        <% end %>
-      </ol>
-    </div>
-  <% end %>
+<% if can_see_full_notice_version?(@notice) %>
+  <%= render 'works_urls_type_list', type: 'copyrighted', show_copyrighted: show_copyrighted, copyrighted_title: copyrighted_title, work: work %>
+  <%= render 'works_urls_type_list', type: 'infringing', show_infringing: show_infringing, infringing_title: infringing_title, work: work %>
 <% else %>
-  <% if show_original %>
+  <% if show_copyrighted %>
     <div class="row">
-      <span class="label original-title"><%= original_title %></span>
+      <span class="label original-title"><%= copyrighted_title %></span>
       <ol class="list original-urls">
         <% if work.copyrighted_urls_counted_by_fqdn.each do |fqdn| %>
           <li class="copyrighted_url"><%= fqdn[:fqdn] %> - <%= fqdn[:count] %> <%= 'URL'.pluralize(fqdn[:count]) %></li>
@@ -60,7 +39,7 @@
       <% end %>
     </div>
   <% else %>
-    <% if access_requestable?(@notice, show_original, show_infringing) %>
+    <% if access_requestable?(@notice, show_copyrighted, show_infringing) %>
       <div class="row">
         <p><%= link_to 'Click here', request_access_notice_path(@notice) %> to request access and see full URLs.</p>
       </div>

diff --git a/app/views/notices/_works_urls_type_list.html.erb b/app/views/notices/_works_urls_type_list.html.erb
@@ -0,0 +1,17 @@
+<% if binding.local_variable_get("show_#{type}") %>
+  <div class="row">
+    <span class="label <%= type %>-title"><%= binding.local_variable_get("#{type}_title") %></span>
+    <ol class="list <%= type %>-urls">
+      <% if work.send("#{type}_urls_public").each do |url_instance| %>
+        <li class="<%= type %>_url">
+          <%= url_instance.url %>
+          <% if url_instance.only_fqdn %>
+            <i class="tooltipster gg-lumen gg-info" title="You can only see the domain name of this url, the full version of the url is available only for logged-in users with specific permissions."></i>
+          <% end %>
+        </li>
+      <% end.empty? %>
+        No <%= type %> URLs were submitted.
+      <% end %>
+    </ol>
+  </div>
+<% end %>