From b5752c2132b7568218a17232f4ca51d60c9b7929 Mon Sep 17 00:00:00 2001
From: Caio <117518+caiosba@users.noreply.github.com>
Date: Fri, 22 Nov 2024 22:02:58 -0300
Subject: [PATCH] Ticket CV2-5653 [WIP]: Adding migration and rake task to
 delete duplicate relationships

---
 ...add_unique_index_to_relationships_table.rb | 33 +++++++++++++++++++
 db/schema.rb                                  |  4 +--
 ...030351_delete_duplicate_relationships.rake | 10 +++---
 3 files changed, 41 insertions(+), 6 deletions(-)
 create mode 100644 db/migrate/20241123001206_add_unique_index_to_relationships_table.rb

diff --git a/db/migrate/20241123001206_add_unique_index_to_relationships_table.rb b/db/migrate/20241123001206_add_unique_index_to_relationships_table.rb
new file mode 100644
index 0000000000..28e1700050
--- /dev/null
+++ b/db/migrate/20241123001206_add_unique_index_to_relationships_table.rb
@@ -0,0 +1,33 @@
+class AddUniqueIndexToRelationshipsTable < ActiveRecord::Migration[6.1]
+  def change
+    # The code below is a copy-paste from the rake task lib/tasks/migrate/20230216030351_delete_duplicate_relationships.rake
+    # and it's responsible for deleting any remaining duplicate relationship, but first, before running this migration, be 
+    # sure to run the rake task above, with: rake check:migrate:delete_duplicate_relationships
+    duplicates = Relationship.group(:target_id).having('COUNT(id) > 1').count
+    n = duplicates.size
+    i = 0
+    duplicates.each do |pm_id, count|
+      i += 1
+      puts "[#{Time.now}] #{i}/#{n}"
+      if count > 1
+        relationships = Relationship.where(target_id: pm_id).order('id ASC').to_a
+        # Keep the confirmed relationship, or the one whose model is image, video or audio... if none, keep the first one
+        keep = relationships.find{ |r| r.relationship_type == Relationship.confirmed_type } || relationships.find{ |r| ['image', 'video', 'audio'].include?(r.model) } || relationships.first
+        raise "No relationship to keep for target_id #{pm_id}!" if keep.nil?
+        relationships.each do |relationship|
+          if relationship.id == keep.id
+            puts "  Keeping relationship ##{r.id}"
+          else
+            puts "  Deleting relationship ##{r.id}"
+            relationship.destroy!
+          end
+          relationship.source.clear_cached_fields
+          relationship.target.clear_cached_fields
+        end
+      end
+    end
+
+    remove_index :relationships, name: 'index_relationships_on_target_id'
+    add_index :relationships, :target_id, unique: true
+  end
+end
diff --git a/db/schema.rb b/db/schema.rb
index dfb106d588..bc95642b32 100644
--- a/db/schema.rb
+++ b/db/schema.rb
@@ -10,7 +10,7 @@
 #
 # It's strongly recommended that you check this file into your version control system.
 
-ActiveRecord::Schema.define(version: 2024_10_15_223059) do
+ActiveRecord::Schema.define(version: 2024_11_23_001206) do
 
   # These are extensions that must be enabled in order to support this database
   enable_extension "plpgsql"
@@ -596,7 +596,7 @@
     t.index ["source_id", "target_id", "relationship_type"], name: "relationship_index", unique: true
     t.index ["source_id"], name: "index_relationships_on_source_id"
     t.index ["target_id", "relationship_type"], name: "index_relationships_on_target_id_and_relationship_type"
-    t.index ["target_id"], name: "index_relationships_on_target_id"
+    t.index ["target_id"], name: "index_relationships_on_target_id", unique: true
     t.check_constraint "source_id <> target_id", name: "source_target_must_be_different"
   end
 
diff --git a/lib/tasks/migrate/20230216030351_delete_duplicate_relationships.rake b/lib/tasks/migrate/20230216030351_delete_duplicate_relationships.rake
index 77c7aff648..593a2ff2d7 100644
--- a/lib/tasks/migrate/20230216030351_delete_duplicate_relationships.rake
+++ b/lib/tasks/migrate/20230216030351_delete_duplicate_relationships.rake
@@ -8,10 +8,10 @@ namespace :check do
         i += 1
         puts "[#{Time.now}] #{i}/#{n}"
         if count > 1
-          relationships = Relationship.where(target_id: pm_id).to_a
-          # Keep the relationship whose model is image, video or audio... if none, keep the first one
-          keep = relationships.find{ |r| ['image', 'video', 'audio'].include?(r.model) } || relationships.first
-          raise "No relationship to keeo for target_id #{pm_id}!" if keep.nil?
+          relationships = Relationship.where(target_id: pm_id).order('id ASC').to_a
+          # Keep the confirmed relationship, or the one whose model is image, video or audio... if none, keep the first one
+          keep = relationships.find{ |r| r.relationship_type == Relationship.confirmed_type } || relationships.find{ |r| ['image', 'video', 'audio'].include?(r.model) } || relationships.first
+          raise "No relationship to keep for target_id #{pm_id}!" if keep.nil?
           relationships.each do |relationship|
             if relationship.id == keep.id
               puts "  Keeping relationship ##{r.id}"
@@ -19,6 +19,8 @@ namespace :check do
               puts "  Deleting relationship ##{r.id}"
               relationship.destroy!
             end
+            relationship.source.clear_cached_fields
+            relationship.target.clear_cached_fields
           end
         end
       end