Autograder for Web Projects

Demonstration of the Autograder to automatically evaluate student code with GitHub Actions and criteria-oriented feedback.

Autograder WebTech | Criteria | Feedback | Workflow

---

Table of Contents

• Autograder for Web Projects
  • Table of Contents
  • Quick Start
  • About the Project
  • Repository Architecture
  • Evaluation Flow
  • Detailed Documentation
  • Practical Usage
    • When you want to change what is evaluated
    • When you want to change how the result is shown
    • When you want to change when/how it runs
  • Quick Troubleshooting
  • Summary

---

Quick Start

1. Edit the student files in submission/.
2. Adjust criteria in .github/autograder/criteria.json when needed.
3. Configure feedback display in .github/autograder/feedback.json.
4. Push or open a PR to main.
5. Check the Autograder workflow execution in GitHub Actions.

Important: test paths must follow the pattern submission/file.ext.

---

About the Project

This repository is a demonstration of automated evaluation for web projects. The goal is to clearly separate:

• student-submitted code;
• evaluation rules;
• feedback policy;
• automated execution pipeline.

With this separation, you can evolve the rubric without directly modifying submission code.

Repository Architecture

Component	Function
submission/	Files submitted by the student (index.html, styles.css, app.js, etc.)
.github/autograder/criteria.json	Defines weights, blocks, and evaluation tests
.github/autograder/feedback.json	Defines how the result will be displayed
.github/workflows/classroom.yml	Triggers and runs the autograder on GitHub Actions
.github/autograder/setup.json	Additional setup file (not used in the current flow)

---

Evaluation Flow

1. A CI event happens (push, pull_request, or workflow_dispatch).
2. The workflow in .github/workflows/classroom.yml starts the grading job.
3. The webtech-network/autograder@main action runs the tests defined in criteria.json.
4. The final result is presented according to feedback.json.

Attention: the workflow uses openai-key: ${{ secrets.ENGINE }}. Make sure this secret exists in the repository.

---

Detailed Documentation

To go deeper into the configuration of each file:

• docs/criteria.md: rubric structure, weights, test types, and safe maintenance.
• docs/feedback.md: result display strategy and feedback configurations.
• docs/workflow-classroom.md: triggers, jobs, steps, and execution parameters.

Practical Usage

When you want to change what is evaluated

Edit .github/autograder/criteria.json and validate in a PR whether the tests represent the pedagogical objective.

When you want to change how the result is shown

Edit .github/autograder/feedback.json (for example, to show or hide passed tests).

When you want to change when/how it runs

Edit .github/workflows/classroom.yml to adjust triggers, permissions, or action parameters.

---

Quick Troubleshooting

• Workflow did not generate feedback: confirm that the ENGINE secret is configured.
• Tests fail for no apparent reason: review paths in criteria.json for submission/....
• Rule change was not reflected: verify that the executed workflow is the latest one from the branch.

---

Summary

This repository exemplifies an ideal scenario for running the Autograder for web activities:

• submission/ contains the submission;
• criteria.json defines the grading;
• feedback.json defines the presentation;
• classroom.yml runs everything in CI.

For day-to-day operation, use this README as a quick entry point and the files in docs/ as in-depth reference.