Feature/jplag implementation #447

JackSCarroll · 2024-09-20T11:40:20Z

Depends on doubtfire-lms/doubtfire-deploy#28
Depends on doubtfire-lms/doubtfire-web#873
Feasibility document: thoth-tech/documentation#528

Description

This branch adds JPLAG software plagiarism detection to OnTrack. The unit_similarity_module.rb file has been modified to extend the check_similarity function to perform a docker exec command to generate a jplag report for each task with "done" files in a unit.
It extracts the code files from the "done" zip files and temporarily stores them in tmp/jplag/unitcode/task while the jplag report is being generated.

Type of change

Breaking change (fix or feature that would cause existing functionality to not work as expected)
This change requires a documentation update

How Has This Been Tested?

Use the modified docker-compose and added jplag.Dockerfile from Feature/add jplag plagiarism detection #28
Create a new task in a unit and have at least 2 student users upload similar code files for the task
Change task due date to be in the past, so submissions can be accepted in the check_similarity check
Run rake submission:generate_pdfs while cd'ed in doubtfire-api. This will create the "done" files.
Run rake submission:check_plagiarism while cd'ed in doubtfire-api. This will run jplag on all units and tasks.

Checklist:

My code follows the style guidelines of this project
I have performed a self-review of my own code
I have commented my code, particularly in hard-to-understand areas
I have made corresponding changes to the documentation if appropriate
My changes generate no new warnings
I have added tests that prove my fix is effective or that my feature works
I have created or extended unit tests to address my new additions
New and existing unit tests pass locally with my changes
Any dependent changes have been merged and published in downstream modules

macite · 2024-09-20T22:57:15Z

Can you recreate commit 90ef645 with a chore tag at the start of the commit message.

macite

This looks like a good start, but there seems to be quite a bit to do still.

macite · 2024-09-20T23:02:07Z

app/models/similarity/unit_similarity_module.rb

+        # 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?


should this still be moss_language?

I'm using the upload requirement language from the task definition ui task-definition-upload.component.html.
Feel like it makes sense to use existing language definition considering the task creator has already decided if plagiarism checks are necessary from MOSS.
So if no requirement for MOSS, no requirement for JPLAG.
Let me know if you don't think is this appropriate.

With this we should remove the old moss code, so these could be renamed...

app/models/similarity/unit_similarity_module.rb

macite · 2024-09-20T23:04:23Z

app/models/similarity/unit_similarity_module.rb

+        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
+      puts "Starting JPLAG container to run on #{tasks_dir}"
+      tasks_dir_split = tasks_dir.to_s.split('/workspace/doubtfire-api')[1]


hardcoded dev container path... won't work in production

Have changed to Rails.root

app/models/similarity/unit_similarity_module.rb

macite · 2024-09-20T23:13:45Z

app/models/similarity/unit_similarity_module.rb

+
+      # 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'`


Should this be outside the upload requirements loop? What happens if you have multiple files being checked in the same task definition?

Sudo is needed to make the docker exec commands work. Not entirely sure why 🙃.
Have moved the dir checks outside the loop and still works as expected.

docker exec doesn't require sudo in production since it'll run as the root user, but the permission errors in the devcontainer are due to it using the vscode user

You can fix this by setting permissions on /var/run/docker.sock in https://github.com/doubtfire-lms/doubtfire-deploy

macite · 2024-09-20T23:15:05Z

app/models/similarity/unit_similarity_module.rb

+      result_file = "#{results_dir}/#{task_definition.id}-result.zip"
+      `sudo docker exec jplag sh -c 'if [ -f "#{result_file}" ]; then rm "#{result_file}"; fi'`
+
+      case type_data[1]


Is this all the languages available?

Because some of the language names in MOSS and JPLAG differ, i've added a check for c++ and python.
Currently only supporting languages defined from the task definition ui task-definition-upload.component.html
Have added comments to both task-definition-upload.component.html and unit-similarity-module.rb as reminder to check MOSS and JPLAG language defintions

app/models/similarity/unit_similarity_module.rb

macite · 2024-09-20T23:17:19Z

app/models/similarity/unit_similarity_module.rb

+    puts "Files to delete: #{Dir.glob("#{tmp_dir}/*")}"
+    FileUtils.rm_rf(Dir.glob("#{tmp_dir}/*"))
+
+    self


Where are the similarity matches created?

We need to parse the report and link related tasks

I've added a function to create the similarity matches. Here's the commit.
It opens the overview.json file from the report and extracts the max similarity between 2 submissions.
It also finds the task id by looking in /files/student_name/task_id and grabbing the task id folder name; it matches the student names from the two that are being looked at currently in overview.json.

…lag report on those

…p jplag ver

macite

Looks like we need to remove the Moss code, switch the languages in the task definitions, and add in some unit tests.

macite · 2024-09-21T21:28:51Z

app/models/similarity/unit_similarity_module.rb

+        # 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?


With this we should remove the old moss code, so these could be renamed...

macite · 2024-09-21T21:37:43Z

app/models/similarity/unit_similarity_module.rb

+      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


We can be removing MOSS with this - so the languages can be updated to those we want from JPlag. This would be part of the front end change as well.

…ort download endpoint

This was referenced Sep 20, 2024

Feature/add jplag frontend doubtfire-lms/doubtfire-web#873

Open

Feature/add jplag plagiarism detection doubtfire-lms/doubtfire-deploy#28

Open

macite requested changes Sep 20, 2024

View reviewed changes

JackSCarroll added 9 commits September 21, 2024 15:59

chore: Add files that werent committed first time

dd70500

refactor: jplag report generation works but only in student-work/new

469b1ac

rework: add code file extraction from done files zips and generate jp…

28c50d9

…lag report on those

refactor: add jplag function to unit_similarity_module and remove tem…

611bbac

…p jplag ver

fix: make report file generation make more sense and remove tmp files

6fa26f0

fix: make report downloading use new path

c83919c

chore: clean files

4d0abe3

chore: clean files

eb8c0e8

chore: remove hard coded path and update file_lang

bf1f7d6

JackSCarroll force-pushed the feature/jplag-implementation branch from c0bad99 to bf1f7d6 Compare September 21, 2024 06:00

macite requested changes Sep 21, 2024

View reviewed changes

JackSCarroll added 6 commits September 25, 2024 16:55

refactor: rename moss language to jplag language

1caabe4

refactor: adjust similarity match code to be suited for jplag

b95fa31

refactor: add plagiarism link from jplag reports

2a23759

refactor: add jplag plagiarism link support for group tasks

f9b0ae0

refactor: add hasJjplagReport value to task defintion and use for rep…

dab43cb

…ort download endpoint

fix: fix report download

a45eb6b

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Feature/jplag implementation #447

Feature/jplag implementation #447

JackSCarroll commented Sep 20, 2024 •

edited

Loading

macite commented Sep 20, 2024

macite left a comment

macite Sep 20, 2024

JackSCarroll Sep 21, 2024

macite Sep 21, 2024

macite Sep 20, 2024

JackSCarroll Sep 21, 2024

macite Sep 20, 2024

macite Sep 20, 2024

JackSCarroll Sep 21, 2024

b0ink Sep 22, 2024

macite Sep 20, 2024

JackSCarroll Sep 21, 2024

macite Sep 20, 2024

macite Sep 20, 2024

JackSCarroll Oct 3, 2024

macite left a comment

macite Sep 21, 2024

macite Sep 21, 2024

Feature/jplag implementation #447

Are you sure you want to change the base?

Feature/jplag implementation #447

Conversation

JackSCarroll commented Sep 20, 2024 • edited Loading

Description

Type of change

How Has This Been Tested?

Checklist:

macite commented Sep 20, 2024

macite left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

macite left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

JackSCarroll commented Sep 20, 2024 •

edited

Loading