mereview 1.0.0

A CLI tool for reviewing coding assignments and Jupyter notebooks with AI-powered analysis

mereview

A CLI tool for reviewing coding assignments and Jupyter notebooks with AI-powered analysis.

Features

• AI Detection: Checks if plans.txt is AI-generated using GPTZero API (optional)
• Notebook Analysis: Converts Jupyter notebooks to Python and performs diff analysis
• Diff Reports: Generates both statistical and HTML diff reports
• Test Validation: Uses Gemini AI to validate test assertions against plans and initial code

Installation

pip install -e .

Or install dependencies directly:

pip install -r requirements.txt

Usage

mereview [OPTIONS]

Options

• --dir, -d: Directory containing plans.txt and zip file (default: current directory)
• --output, -o: Output directory for results (default: review_results)
• --help: Show help message

Example

# Run in current directory
mereview

# Specify directory and output location
mereview --dir ./submissions --output ./reviews

Required Files

The tool expects the following files in the specified directory:

1. plans.txt - The project plans document
2. [name].zip - A zip file containing:
   • A folder starting with TASK_
   • Inside the folder:
     • initial_notebook.ipynb
     • final_notebook.ipynb
     • test_notebook.py

Environment Variables

Set these environment variables before running the tool (optional):

export GPTZERO_API_KEY="your_gptzero_api_key"
export GEMINI_API_KEY="your_gemini_api_key"

• GPTZERO_API_KEY (optional): API key for GPTZero AI detection service - Get from https://gptzero.me/
• GEMINI_API_KEY (optional): API key for Google Gemini AI analysis - Get from https://ai.google.dev/

If these keys are not provided, the respective steps will be skipped with warnings.

Output

The tool generates the following files in the output directory:

• initial_notebook.py - Converted initial notebook
• final_notebook.py - Converted final notebook
• diff_report.html - HTML visualization of code differences
• gemini_analysis.txt - Gemini AI analysis of test assertions
• results_summary.json - Complete results summary in JSON format

How It Works

1. Step 1: Checks if plans.txt is AI-generated (optional with GPTZero)
2. Step 2: Extracts the zip file and locates the TASK_ folder
3. Step 3: Converts Jupyter notebooks to Python files
4. Step 4: Calculates diff statistics with docstrings included
5. Step 5: Removes docstrings and recalculates diff statistics
6. Step 6: Generates an HTML diff report
7. Step 7: Analyzes test file with Gemini AI to check for undocumented assertions
8. Step 8: Saves all results to the output directory

Requirements

• Python 3.8+
• Dependencies listed in requirements.txt