Feature/jplag implementation

Gemfile.lock

@@ -168,6 +168,7 @@ GEM
     faraday-net_http (3.1.0)
       net-http
     ffi (1.17.0-aarch64-linux-gnu)
+    ffi (1.17.0-x86_64-linux-gnu)
     fugit (1.11.0)
       et-orbi (~> 1, >= 1.2.11)
       raabro (~> 1.4)
@@ -265,6 +266,8 @@ GEM
     nio4r (2.7.3)
     nokogiri (1.16.5-aarch64-linux)
       racc (~> 1.4)
+    nokogiri (1.16.5-x86_64-linux)
+      racc (~> 1.4)
     observer (0.1.2)
     orm_adapter (0.5.0)
     parallel (1.24.0)
@@ -490,6 +493,7 @@ GEM
 
 PLATFORMS
   aarch64-linux
+  x86_64-linux
 
 DEPENDENCIES
   better_errors

app/api/task_definitions_api.rb

@@ -614,4 +614,33 @@ class TaskDefinitionsApi < Grape::API
 
     stream_file path
   end
+
+  desc 'Download the JPLAG report for a given task'
+  params do
+    requires :unit_id, type: Integer, desc: 'The unit to download JPLAG report for'
+    requires :task_def_id, type: Integer, desc: 'The task definition to get the JPLAG report of'
+  end
+  get '/units/:unit_id/task_definitions/:task_def_id/jplag_report' do
+    unit = Unit.find(params[:unit_id])
+    task_def = unit.task_definitions.find(params[:task_def_id])
+
+    unless authorise? current_user, unit, :download_jplag_report
+      error!({ error: 'Not authorised to download JPLAG reports of unit' }, 403)
+    end
+
+    file_loc = FileHelper.task_jplag_report_path(unit, task_def)
+    logger.debug "JPLAG report file location: #{file_loc}"
+
+    if file_loc.nil? || !File.exist?(file_loc)
+      file_loc = Rails.root.join('public', 'resources', 'FileNotFound.pdf')
+      header['Content-Disposition'] = 'attachment; filename=FileNotFound.pdf'
+    else
+      header['Content-Disposition'] = "attachment; filename=#{task_def.abbreviation}-jplag-report.zip"
+    end
+    header['Access-Control-Expose-Headers'] = 'Content-Disposition'
+
+    content_type 'application/octet-stream'
+
+    stream_file file_loc
+  end
 end

app/helpers/file_helper.rb

@@ -249,6 +249,16 @@ def student_portfolio_path(unit, username, create = true)
     File.join(student_portfolio_dir(unit, username, create), FileHelper.sanitized_filename("#{username}-portfolio.pdf"))
   end
 
+  def task_jplag_report_dir(unit)
+    file_server = Doubtfire::Application.config.jplag_report_dir
+    dst = "#{file_server}/#{unit.code}-#{unit.id}/" # trust the server config and passed in type for paths
+    dst
+  end
+
+  def task_jplag_report_path(unit, task)
+    File.join(task_jplag_report_dir(unit), FileHelper.sanitized_filename("#{task.id}-result.zip"))
+  end
+
   def comment_attachment_path(task_comment, attachment_extension)
     "#{File.join(student_work_dir(:comment, task_comment.task), "#{task_comment.id.to_s}#{attachment_extension}")}"
   end
@@ -658,4 +668,6 @@ def line_wrap(path, width: 160)
   module_function :known_extension?
   module_function :pages_in_pdf
   module_function :line_wrap
+  module_function :task_jplag_report_dir
+  module_function :task_jplag_report_path
 end

app/models/similarity/unit_similarity_module.rb

@@ -14,17 +14,26 @@ def last_plagarism_scan
   end
 
   # Pass tasks on to plagarism detection software and setup links between students
-  def check_moss_similarity(force: false)
+  def check_similarity(force: false)
     # Get each task...
     return unless active
 
     # need pwd to restore after cding into submission folder (so the files do not have full path)
     pwd = FileUtils.pwd
 
+    # making temp directory for unit - jplag
+    root_work_dir = Rails.root.join("tmp", "jplag", "#{code}-#{id}")
+    unit_code = "#{code}-#{id}"
+    FileUtils.mkdir_p(root_work_dir)
+
     begin
       logger.info "Checking plagiarsm for unit #{code} - #{name} (id=#{id})"
 
       task_definitions.each do |td|
+        # making temp directory for each task - jplag
+        tasks_dir = root_work_dir.join(td.id.to_s)
+        FileUtils.mkdir_p(tasks_dir)
+
         next if td.moss_language.nil? || td.upload_requirements.nil? || td.upload_requirements.select { |upreq| upreq['type'] == 'code' && upreq['tii_check'] }.empty?
 
         type_data = td.moss_language.split
@@ -35,6 +44,9 @@ def check_moss_similarity(force: false)
         tasks = tasks_for_definition(td)
         tasks_with_files = tasks.select(&:has_pdf)
 
+        # JPLAG
+        run_jplag_on_done_files(td, tasks_dir, tasks_with_files, unit_code)
+
         # Skip if not due yet
         next if td.due_date > Time.zone.now
 
@@ -48,6 +60,7 @@ def check_moss_similarity(force: false)
 
         # There are new tasks, check these
 
+
         logger.debug 'Contacting MOSS for new checks'
 
         # Create the MossRuby object
@@ -222,4 +235,60 @@ def add_done_files_for_plagiarism_check_of(task_definition, tmp_path, to_check,
 
     self
   end
+
+  # JPLAG Function - extracts "done" files for each task and packages them into a directory for JPLAG to run on
+  def run_jplag_on_done_files(task_definition, tasks_dir, tasks_with_files, unit_code)
+    type_data = task_definition.moss_language.split
+    return if type_data.nil? || (type_data.length != 2) || (type_data[0] != 'moss')
+    similarity_pct = task_definition.plagiarism_warn_pct
+    return if similarity_pct.nil?
+
+    # Check if the directory exists and create it if it doesn't
+    results_dir = "/jplag/results/#{unit_code}"
+    `sudo docker exec jplag sh -c 'if [ ! -d "#{results_dir}" ]; then mkdir -p "#{results_dir}"; fi'`
+
+    # Remove existing result file if it exists
+    result_file = "#{results_dir}/#{task_definition.id}-result.zip"
+    `sudo docker exec jplag sh -c 'if [ -f "#{result_file}" ]; then rm "#{result_file}"; fi'`
+
+    # get each code file for each task
+    task_definition.upload_requirements.each_with_index do |upreq, idx|
+      # only check code files marked for similarity checks
+      next unless upreq['type'] == 'code' && upreq['tii_check']
+
+      pattern = task_definition.glob_for_upload_requirement(idx)
+
+      tasks_with_files.each do |t|
+        t.extract_file_from_done(tasks_dir, pattern, ->(_task, to_path, name) { File.join(to_path.to_s, t.student.username.to_s, name.to_s) })
+      end
+
+      logger.info "Starting JPLAG container to run on #{tasks_dir}"
+      root_dir = Rails.root.to_s
+      tasks_dir_split = tasks_dir.to_s.split(root_dir)[1]
+
+      # Set the file language based on the type
+      # Currently only supporting C/C++/C#/Python
+      # MOSS and JPLAG use different names for some languages, need to be converted
+      # If new MOSS languages options are added to task-defintion-upload, this will need to be updated
+      file_lang = case type_data[1]
+                  when 'cc'
+                    'cpp'
+                  when 'python'
+                    'python3'
+                  else
+                    type_data[1]
+                  end
+
+      # Run JPLAG on the extracted files
+      `sudo docker exec jplag java -jar /jplag/myJplag.jar #{tasks_dir_split} -l #{file_lang} --similarity-threshold=#{similarity_pct} -M RUN -r #{results_dir}/#{task_definition.id}-result`
+    end
+
+    # Delete the extracted code files from tmp
+    tmp_dir = Rails.root.join("tmp", "jplag")
+    logger.info "Deleting files in: #{tmp_dir}"
+    logger.info "Files to delete: #{Dir.glob("#{tmp_dir}/*")}"
+    FileUtils.rm_rf(Dir.glob("#{tmp_dir}/*"))
+
+    self
+  end
 end

app/models/unit.rb

@@ -28,6 +28,7 @@ def self.permissions
       :download_stats,
       :download_unit_csv,
       :download_grades,
+      :download_jplag_report,
       :exceed_capacity
     ]
 
@@ -46,6 +47,7 @@ def self.permissions
       :change_project_enrolment,
       :download_stats,
       :download_grades,
+      :download_jplag_report,
       :rollover_unit,
       :exceed_capacity,
       :perform_overseer_assessment_test
@@ -66,6 +68,7 @@ def self.permissions
       :download_stats,
       :download_unit_csv,
       :download_grades,
+      :download_jplag_report,
       :exceed_capacity
     ]
 
@@ -177,6 +180,7 @@ def role_for(user)
   scope :set_inactive,          -> { where('active = ?', false) }
 
   include UnitTiiModule
+
   include UnitSimilarityModule
 
   def detailed_name

config/application.rb

@@ -31,6 +31,12 @@ class Application < Rails::Application
     # variable.
     config.student_work_dir = ENV['DF_STUDENT_WORK_DIR'] || "#{Rails.root}/student_work"
 
+    # ==> JPLAG report directory
+    # File server location for storing JPLAG reports. Defaults to `jplag_results`
+    # directory under root but is overridden using DF_JPLAG_REPORT_DIR environment
+    # variable.
+    config.jplag_report_dir = ENV['DF_JPLAG_REPORT_DIR'] || "#{Rails.root}/jplag_results"
+
     # ==> Load credentials from env
     credentials.secret_key_base = ENV.fetch('DF_SECRET_KEY_BASE', Rails.env.production? ? nil : '9e010ee2f52af762916406fd2ac488c5694a6cc784777136e657511f8bbc7a73f96d59c0a9a778a0d7cf6406f8ecbf77efe4701dfbd63d8248fc7cc7f32dea97')
     credentials.secret_key_attr = ENV.fetch('DF_SECRET_KEY_ATTR', Rails.env.production? ? nil : 'e69fc5960ca0e8700844a3a25fe80373b41c0a265d342eba06950113f3766fd983bad9ec51bf36eb615d9711bfe1dd90b8e35f01841b323f604ffee857e32055')

lib/tasks/checks.rake

@@ -69,7 +69,7 @@ namespace :submission do
           puts ' ------------------------------------------------------------ '
           puts "  Starting Plagiarism Check for #{unit.name}"
           puts ' ------------------------------------------------------------ '
-          unit.check_moss_similarity
+          unit.check_similarity
         end
         puts ' ------------------------------------------------------------ '
         puts ' done.'