AI Code Reviewer (Gemini + Python)

CLI project that reviews a single source file using Gemini and prints a structured report in the terminal.

Features

• Reads a code file from disk
• Sends it to Gemini with a strict JSON schema
• Validates response with Pydantic
• Renders results with Rich panels/tables

Project Structure

.
|- analyzer.py      # Gemini request + schema-based parsing
|- schemas.py       # Pydantic response model
|- main.py          # CLI and Rich output
|- examples/        # sample files to analyze

Requirements

• Python 3.10+
• Gemini API key

Installation

python -m venv venv
# Windows:
venv\Scripts\activate

# Linux/macOS:
source venv/bin/activate

pip install google-genai pydantic python-dotenv typer rich

Environment

Create a .env file in the project root:

GEMINI_API_KEY=your_api_key_here

Run

python main.py

When prompted, provide a file path, for example:

examples/linear_search.py

Response Schema

The model returns:

• overall_quality_summary
• general_score (1 to 5)
• estimated_complexity (Big O)
• strengths
• issues
• improvements

Examples Folder

This repository includes 4 code samples:

• examples/linear_search.py
• examples/two_sum_hashmap.py
• examples/search_in_sorted_array.py
• examples/bubble_sort.py

Try them with:

python main.py
# then type one of:
# examples/linear_search.py
# examples/two_sum_hashmap.py
# examples/bubble_sort.py