Welcome to the Quera's MATLAB (Octave) Judge — a lightweight, containerized, and automated grading system for evaluating MATLAB-style .m code using @gnu-octave. It's designed for integration with Quera's Paas Judge, enabling hands-free code testing and evaluation, especially in educational and competition settings. This MATLAB (Octave) judge is about 2 to 4 times faster than the current MATLAB judge!

This judge uses GNU Octave under the hood to execute and validate MATLAB-like code. Octave is a powerful open-source tool that closely mirrors MATLAB's syntax and capabilities, making it a great backend for judging MATLAB submissions without licensing restrictions.

• Difference in Compatibility: Around 10-20% (with MATLAB having some proprietary features not in Octave).

The system is optimized for containerized environments using Docker, making deployment simple, portable, and scalable.

The judging process includes:

1. Put quera's input files in in and output files in out folder. Pay attention to names! It's not important to put them in order, because test discoverer will find all of them without problem, but there should be a bijection between input and output files, if not they will be skipped when using test discoverer!

2. Run test dicoverer and generator. You should run generate_tests.py file and give --solution flag and solution file name as an argument to it. It will discover tests you had put in test_cases folder before, generate code for test.py section which is main entry point, generate tester_config.json file which is judge configs and valid_files which is valid files that can be recived from users. This will also let you to customize each test differently!

python3 generate_tests.py --solution=solution.m

3. Thats it! Now just zip all the content and upload as a Quera's Paas (Devops) judge tester file!

4. Receiving some solution .m files submitted by a user.

5. Running the solution files using Octave Container via a python judge script (test.py) that defines how the solution is tested.

6. Comparing outputs with expected results defined in test cases.

This system can be extended or adapted for a wide range of evaluation scenarios, including:

• Autograding in online courses
• Competitive programming contests
• In-browser coding challenges
• Interactive educational tools

• ✅ Automated judging of MATLAB/Octave code
• 🧱 Built on Quera’s Paas Judge
• 🐳 Containerized with Docker for seamless deployment
• 🎓 Open-source and customizable
• 🧪 Testcase discoverer included
• 🧠 Minimal dependencies using standard Octave
• 🔧 Easy integration with Quera

To integrate with Quera's Paas Judge:

• Define a judge config that uses this judge as a tester.
• Format your problem input/output expectations accordingly.

🐳 docker-compose.yml

version: "3.3"
services:
  octave-container:
    image: registry.gitlab.com/qio/standard/gnuoctave/octave:8.2.0
    container_name: "octave-container"
    working_dir: /mnt
    volumes:
      - ./solution:/mnt
    stdin_open: true
    tty: true
    command: ["tail", "-f", "/dev/null"]

📌 What It Does

• Uses the official Octave image from Quera's GitLab Container Registery.

• Mounts your local ./solution directory into the container's /mnt. It will make users solution file accesible form octave's container.

• Starts a persistent container (tail -f /dev/null) so you can run commands inside interactively.

• Also useful for debugging or executing MATLAB/Octave .m files manually.

quera_matlab_judge
├── README.md
├── assets
├── docker-compose.yml      # Octave docker-compose file 
├── generate_tests.py       # Tests generator and discoverer
├── requirements.txt
├── scripts                 # Octave Judge helpers (docker handler and octave runner)
│   ├── docker_handler.py
│   └── octave_runner.py
├── solution                # Solutions will come here (like solution.m)
│   ├── Dockerfile
│   └── solution.m
├── test.py                 # Main Tester, Entry point
├── test_cases              # Test cases for I/O judge
│   ├── in
│   │   └── ...
│   └── out
│       └── ...
├── tester_config.json      # Quera's tester config
└── valid_files             # Quera's tester valid files

Though Octave tries to match MATLAB syntax closely, there are subtle differences to be aware of when writing or judging code: