🤖 Code Generation Agent

A sophisticated AI-powered code generation system that uses a unique "thought-first" approach with large language models to generate high-quality code implementations and test cases.

🌟 Key Features

• 🧠 Two-stage thinking process using specialized LLMs
• 🔄 Automated test case generation
• 💻 Intelligent code implementation
• 🔍 Comprehensive codebase analysis
• 📝 Detailed logging and documentation

🎯 Core Components

1. Test Case Generator

• Clones target repository
• Analyzes existing codebase
• Generates comprehensive test cases using pytest
• Implements test fixtures and edge cases

2. Feature Implementer

• Implements new features based on test cases
• Ensures code quality and best practices
• Validates implementation against tests
• Maintains existing codebase integrity

🧠 Unique "Think-First" Approach

This project implements a novel two-stage thinking process:

1. Thought Generation 🤔

# First, generate deep analysis using DeepSeek-V3
thought = together_client.chat.completions.create(
    model="deepseek-ai/DeepSeek-R1",
    messages=[...],
    stop=['</think>']
)

2. Code Implementation ⌨️

# Then, use Qwen Coder with the generated thoughts
prompt_with_thought = prompt + f"""
Consider this analysis when implementing:
<think>
{thought_result}
</think>
"""

Why This Approach Works

• 🎯 Better problem understanding
• 🔍 More thorough solution analysis
• 💡 Improved code quality
• 🐛 Fewer bugs and edge cases missed

🚀 Getting Started

Prerequisites

• Python 3.8+
• Together AI API key
• Git

Installation

git clone https://github.com/yourusername/code_generation_agent
cd code_generation_agent
pip install -r requirements.txt

Usage

from code_generation_agent import TestCaseGenerator, FeatureImplementer

# Initialize and run test generation
generator = TestCaseGenerator(
    repo_url="your_repo_url",
    feature_description="your_feature_description",
    together_api_key="your_api_key"
)
generator.generate_and_run_tests()

# Implement features
implementer = FeatureImplementer(
    repo_url="your_repo_url",
    feature_description="your_feature_description",
    together_api_key="your_api_key"
)
implementer.implement_features()

📊 Model Configuration

The system uses specific LLM models for different tasks:

• 🤔 Thought Generation: DeepSeek-V3
• 💻 Code Generation: Qwen2.5-Coder-32B-Instruct

🔐 Security

• Never commit API keys to version control
• Use environment variables for sensitive data
• Regularly rotate API keys

📝 License

This project is licensed under the MIT License - see the LICENSE file for details.

🤝 Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

📧 Contact

For questions and support, please open an issue in the repository.