ProFormA Import: fix file duplication

app/models/code_ocean/file.rb

@@ -60,6 +60,8 @@ class File < ApplicationRecord
     validates :weight, absence: true, unless: :teacher_defined_assessment?
     validates :file, presence: true if :context.is_a?(Submission)
     validates :context_type, inclusion: {in: ALLOWED_CONTEXT_TYPES}
+    validates :path, uniqueness: {scope: %I[name file_type context_id context_type role]}
+
     # xml_id_path must be unique within the scope of an exercise.
     # Initially, it may match the record’s id (set while exporting),
     # but it can later diverge as long as uniqueness is preserved.

app/services/proforma_service/convert_task_to_exercise.rb

@@ -68,7 +68,7 @@ def string_to_bool(str)
     end
 
     def files
-      model_solution_files + test_files + task_files
+      (model_solution_files + test_files + task_files).uniq {|f| [f.name, f.file_type_id, f.role, f.path] }
     end
 
     def test_files
@@ -106,16 +106,20 @@ def task_files
 
     def codeocean_file_from_task_file(file, parent_object = nil)
       extension = File.extname(file.filename)
+      path = File.dirname(file.filename).in?(['.', '']) ? nil : File.dirname(file.filename)
+      name = File.basename(file.filename, '.*')
+      role = extract_meta_data(@task.meta_data&.dig('meta-data'), 'files', "CO-#{file.id}", 'role')
+
       # checking the last element of xml_id_path array for file.id
       codeocean_file = @exercise.files.detect {|f| f.xml_id_path.last == file.id } || @exercise.files.new
       codeocean_file.assign_attributes(
         context: @exercise,
         file_type: file_type(extension),
         hidden: file.visible != 'yes', # hides 'delayed' and 'no'
-        name: File.basename(file.filename, '.*'),
+        name:,
         read_only: file.usage_by_lms != 'edit',
-        role: extract_meta_data(@task.meta_data&.dig('meta-data'), 'files', "CO-#{file.id}", 'role'),
-        path: File.dirname(file.filename).in?(['.', '']) ? nil : File.dirname(file.filename),
+        role:,
+        path:,
         xml_id_path: (parent_object.nil? ? [file.id] : [parent_object.id, file.id]).map(&:to_s)
       )
       if file.binary

spec/models/code_ocean/file_spec.rb

@@ -137,4 +137,18 @@
       end
     end
   end
+
+  context 'when a file with same attributes (path, name file_type context_id context_type role) already exists' do
+    before do
+      create(:file, name: 'static', context: exercise)
+      file.validate
+    end
+
+    let(:exercise) { create(:dummy) }
+    let(:file) { build(:file, name: 'static', context: exercise) }
+
+    it 'has an error for path' do
+      expect(file.errors[:path]).to be_present
+    end
+  end
 end

spec/services/proforma_service/convert_task_to_exercise_spec.rb

@@ -187,6 +187,27 @@
         expect { convert_to_exercise_service.save! }.to change(Exercise, :count).by(1)
       end
 
+      context 'with two files with similar contents' do
+        let(:file_dup) do
+          ProformaXML::TaskFile.new(
+            id: 'id2',
+            content:,
+            filename:,
+            used_by_grader: 'used_by_grader',
+            visible: 'yes',
+            usage_by_lms:,
+            binary:,
+            mimetype:
+          )
+        end
+
+        let(:files) { [file, file_dup] }
+
+        it 'creates an exercises with only one file' do
+          expect(convert_to_exercise_service.files.length).to be 1
+        end
+      end
+
       context 'when file is a Makefile' do
         let(:filename) { "#{path}Makefile" }
 
@@ -346,7 +367,7 @@
           ProformaXML::TaskFile.new(
             id: 'ms-file-2',
             content: 'content',
-            filename: 'filename.txt',
+            filename: 'filename2.txt',
             used_by_grader: 'used_by_grader',
             visible: 'yes',
             usage_by_lms: 'display',
@@ -456,7 +477,7 @@
           ProformaXML::TaskFile.new(
             id: 'test_file_id2',
             content: 'testfile-content',
-            filename: 'testfile.txt',
+            filename: 'testfile2.txt',
             used_by_grader: 'yes',
             visible: 'no',
             usage_by_lms: 'display',