From 640f3168bc6928f8661208f1abdbf03d65bd772e Mon Sep 17 00:00:00 2001
From: Gregor Lasnibat <gregor@scinote.net>
Date: Tue, 12 Mar 2024 17:11:45 +0100
Subject: [PATCH] (dev) Detect duplicated item IDs when parsing a spreadsheet
 [SCI-10399]

---
 app/controllers/repositories_controller.rb      |  6 ++++++
 app/services/spreadsheet_parser.rb              |  8 ++++++++
 .../repository_import_parser/importer.rb        | 17 +++++++++++++----
 .../repositories/_parse_records_modal.html.erb  |  5 +++++
 config/locales/en.yml                           |  1 +
 5 files changed, 33 insertions(+), 4 deletions(-)

diff --git a/app/controllers/repositories_controller.rb b/app/controllers/repositories_controller.rb
index 98fe0a5a2..249089ccf 100644
--- a/app/controllers/repositories_controller.rb
+++ b/app/controllers/repositories_controller.rb
@@ -278,6 +278,12 @@ class RepositoriesController < ApplicationController
             items_size: Constants::IMPORT_REPOSITORY_ITEMS_LIMIT)
         )
       else
+        sheet = SpreadsheetParser.open_spreadsheet(import_params[:file])
+        duplicate_ids = SpreadsheetParser.duplicate_ids(sheet)
+        if duplicate_ids.any?
+          @importing_duplicates_warning = t('repositories.import_records.error_message.importing_duplicates', duplicate_ids: duplicate_ids)
+        end
+
         @import_data = parsed_file.data
 
         if @import_data.header.blank? || @import_data.columns.blank?
diff --git a/app/services/spreadsheet_parser.rb b/app/services/spreadsheet_parser.rb
index fbdbe6ebd..c12b438a9 100644
--- a/app/services/spreadsheet_parser.rb
+++ b/app/services/spreadsheet_parser.rb
@@ -66,4 +66,12 @@ class SpreadsheetParser
       row.map(&:to_s)
     end
   end
+
+  def self.duplicate_ids(sheet)
+    # Extracting IDs from sheet and removing header row
+    ids = sheet.drop(1).map(&:first)
+
+    # Selecting duplicate IDs
+    ids.group_by { |id| id }.select { |_, group| group.size > 1 }.keys
+  end
 end
diff --git a/app/utilities/repository_import_parser/importer.rb b/app/utilities/repository_import_parser/importer.rb
index 1e97b3f21..05e2786d4 100644
--- a/app/utilities/repository_import_parser/importer.rb
+++ b/app/utilities/repository_import_parser/importer.rb
@@ -57,6 +57,7 @@ module RepositoryImportParser
 
     def import_rows!(can_edit_existing_items, should_overwrite_with_empty_cells)
       errors = false
+      duplicate_ids = SpreadsheetParser.duplicate_ids(@sheet)
 
       @repository.transaction do
         batch_counter = 0
@@ -66,6 +67,9 @@ module RepositoryImportParser
           # Skip empty rows
           next if row.blank?
 
+          # Skip duplicates
+          next if duplicate_ids.include?(row.first)
+
           unless @header_skipped
             @header_skipped = true
             next
@@ -82,11 +86,8 @@ module RepositoryImportParser
           incoming_row.each_with_index do |value, index|
             if index == @name_index
 
-              # extract row_id
-              row_id = try_decimal_to_string(value)
-
               # check if row (inventory) already exists
-              existing_row = RepositoryRow.find_by(name: row_id, repository: @repository)
+              existing_row = RepositoryRow.find_by(id: incoming_row[0])
 
               # if it doesn't exist create it
               unless existing_row
@@ -145,7 +146,15 @@ module RepositoryImportParser
     end
 
     def import_batch_to_database(full_row_import_batch, can_edit_existing_items, should_overwrite_with_empty_cells)
+      skipped_rows = []
+
       full_row_import_batch.each do |full_row|
+        # skip archived rows and rows that belong to other repositories
+        if full_row[:repository_row].archived || full_row[:repository_row].repository_id != @repository.id
+          skipped_rows << full_row[:repository_row]
+          next
+        end
+
         full_row[:repository_row].save!(validate: false)
         @new_rows_added += 1
 
diff --git a/app/views/repositories/_parse_records_modal.html.erb b/app/views/repositories/_parse_records_modal.html.erb
index f1072461e..df2a0d056 100644
--- a/app/views/repositories/_parse_records_modal.html.erb
+++ b/app/views/repositories/_parse_records_modal.html.erb
@@ -84,6 +84,11 @@
           </ul>
         </div>
       </div>
+      <% if @importing_duplicates_warning %>
+        <div class="alert alert-warning">
+          <%= @importing_duplicates_warning %>
+        </div>
+      <% end %>
       <div class="modal-footer">
         <button type="button" class="btn btn-secondary" data-dismiss="modal"><%= t('general.cancel')%></button>
         <input type="submit" class="btn btn-success" value="<%= t('repositories.modal_parse.import') %>">
diff --git a/config/locales/en.yml b/config/locales/en.yml
index c3558b1e3..3b0768652 100644
--- a/config/locales/en.yml
+++ b/config/locales/en.yml
@@ -2140,6 +2140,7 @@ en:
       partial_success_flash: "%{nr} of %{total_nr} successfully imported. Other rows contained errors."
       error_message:
         items_limit: "The imported file contains too many rows. Max %{items_size} items allowed to upload at once."
+        importing_duplicates: "Items with duplicates detected: %{duplicate_ids}. These will be ignored on import."
         temp_file_not_found: "This file could not be found. Your session might expire."
         session_expired: "Your session expired. Please try again."
         no_data_to_parse: "There's nothing to be parsed."