C-Grader Pro - AI-Powered C Programming Exam Evaluation for Professors

C-Grader Pro is a modern, single-page web application designed to automate the evaluation of C programming exams using the power of Google's Gemini 2.0 Flash AI. It provides a streamlined workflow for professors to grade multiple student submissions with consistent, high-quality feedback and detailed scoring.

🚀 Key Features

• AI-Driven Evaluation: Utilizes Google's cutting-edge Gemini 2.0 Flash Exp API for superior code understanding and reasoning.
• Three-Phase Workflow:
  1. Criteria Generation: Automatically generates a strict grading rubric based on your exercise context.
  2. Automated Grading: Evaluates student code against the rubric, checking for correctness, style, and efficiency.
  3. Report Generation: Produces downloadable detailed text reports and CSV summary files.
• Smart Criteria Editor: Review and modify the AI-generated grading criteria before applied to students.
• Evaluation Viewer: interactive side-by-side viewer to inspect the AI's annotated feedback vs. the student's original code.
• Deterministic Grading: Configured with strict model parameters (Temperature 0.0) to ensure consistent, fair grading across attempts.
• Local & Secure: Runs entirely in your browser. API keys are stored locally and files are processed on your machine.
• Resilient Architecture: Built-in rate limiting and auto-retry logic to handle API quotas smoothly.
• Multi-Language Support: Fully localized interface and AI feedback in English, Spanish, Galician, and Romanian. Select your preferred language directly from the header.
• Most importantly: academic responsibility always stays with the professor — aligning with the European Union’s Explainable AI in education principles, which highlight the need for human oversight and shared responsibility in AI-assisted educational tools (see Publications Office of the EU report: https://op.europa.eu/en/publication-detail/-/publication/0020c74c-7402-11f0-9af8-01aa75ed71a1/language-en).

📋 Prerequisites

• A modern web browser (Chrome, Edge, or Firefox).
• A Google Gemini API Key. You can get one for free at Google AI Studio.
• Student submissions organized in a directory (one folder per student containing their .c file).

🛠️ How to Use

1. Launch the App: Simply open c-grader-pro.html in your web browser.
2. Configuration:
   • API Key
     • Paste your Google Gemini API key
     • It's stored locally in your browser for convenience
     • Never transmitted anywhere except to Google's API
3. Load Files:
   • Root Directory
     • Enter the path to your submissions folder
     • OR click "Browse for Folder" to select it graphically
     • Expected structure:

       Submissions/
       ├── Student_1/
       │   └── main.c
       ├── Student_2/
       │   └── solution.c
       └── ...
       

   • Exercise Description: Paste or load the exam problem statement/requirements.
   • Professor's Solution
     • Paste or load your reference solution code.
     • Use "Auto-Generate Solution" to obtain a C code proposal from Gemini AI.
4. Start Evaluation:
   • The system will first generate a grading rubric.
   • Review Phase: Edit the rubric if necessary and confirm.
   • Grading Phase: The app will iterate through all students.
5. View Results:
   • Download the [FolderName]_grading_details.md (detailed grading) and [FolderName]__only_grades.csv (gradebook).
   • Use the Evaluation Viewer at the bottom to inspect specific student results.

📂 Output Files

Reports are saved to your Download folder with localized filenames corresponding to your selected language:

• [FolderName]_grading_details.md (or _detalles_calificacion.md, etc.): A comprehensive report containing grades, deductions, and fully annotated code for every student.
• [FolderName]_only_grades.csv (or _solo_notas.csv, etc.): A spreadsheet-friendly format listing Student Name, Final Grade, and Deduction Summaries.
• [FolderName]_grading_criteria.md (or _criterios_calificacion.md): The finalized rubric used for the session.

🏗️ Technical Details

• Stack: Vanilla HTML5, CSS3, JSON, and JavaScript (ES6+).
• Styling: Modern, responsive dark UI using CSS Variables and Flexbox/Grid.
• API: Direct integration with Google Generative Language API (REST).

👥 Credits & Affiliations

• Research Groups: RNASA, CITIC, UDC
• Author: Cristian R Munteanu

---

Disclaimer: This tool is an assistant for grading. While strictly prompt-engineered for accuracy, final grades should always be reviewed by a human instructor.