🔧 Fixie: AI Agent Debugger

An intelligent Python debugging assistant powered by AI agents using LangChain and Ollama Llama 3.2. Fixie automatically detects bugs, analyzes code logic, and suggests fixes through a coordinated multi-agent workflow.

✨ Features

• 🤖 Multi-Agent Architecture: Specialized agents for syntax checking, logic reasoning, and fix suggestions
• 🔍 Intelligent Bug Detection: Identifies syntax errors, runtime issues, and logic problems
• 📍 Line-by-Line Analysis: Pinpoints exact locations of bugs with line numbers
• 🎯 Confidence Scoring: Provides confidence levels for suggested fixes
• 🔄 LangGraph Workflow: Orchestrated agent coordination using LangGraph
• 🦙 Local AI: Powered by Ollama Llama 3.2 for privacy and offline usage

🚀 Quick Start

Prerequisites

• Python 3.8+ (3.10+ preferred)
• Ollama installed and running
• Git (for cloning)

Installation

1. Clone the repository

   git clone https://github.com/kawish918/Fixie-AI-Agent-Debugger.git
   cd fixie-ai-debugger

2. Create virtual environment

   python -m venv fixie
   fixie\Scripts\activate

3. Install dependencies

   pip install langchain langgraph

4. Install and setup Ollama

   Visit Ollama's official website for installation instructions, or:

   Windows/Mac/Linux:

   # Download from: https://ollama.ai/download
   # Then pull the Llama model:
   ollama pull llama3.2

5. Verify Ollama is running

   ollama list
   # Should show llama3.2 in the list

Usage

1. Run on the example buggy code

   python main.py

2. Add your own buggy Python files

   # Place your .py file in the examples/ directory
   # Update main.py to point to your file:
   code = read_python_file('examples/your_buggy_file.py')

📁 Project Structure

fixie-ai-debugger/
├── agents/
│   ├── fix_suggester.py      # AI agent for generating fixes
│   ├── logic_reasoner.py     # AI agent for understanding code logic
│   └── syntax_checker.py     # AI agent for detecting syntax errors
├── core/
│   ├── agent_runner.py       # Simple agent orchestration
│   ├── input_handler.py      # File reading utilities
│   ├── langgraph_runner.py   # LangGraph workflow management
│   └── llama_interface.py    # Ollama API interface
├── examples/
│   └── buggy_code.py         # Sample buggy code for testing
├── main.py                   # Main application entry point
└── README.md

🔄 How It Works

Fixie uses a coordinated multi-agent workflow:

1. Syntax Checker Agent 🔍

   • Analyzes code for syntax and runtime errors
   • Identifies problematic lines and severity levels

2. Logic Reasoner Agent 🧠

   • Understands the intended purpose of the code
   • Provides context for fix generation

3. Fix Suggester Agent 🛠️

   • Combines bug report and logic analysis
   • Generates complete, executable fix suggestions
   • Provides confidence scores

4. LangGraph Orchestration 🔄

   • Manages agent workflow and data flow
   • Ensures proper sequencing and state management

📊 Example Output

--- Fixie AI Debugger ---

Original Code:
def add_nums(a, b):
    return a + b + c

🔍 Debug Results:
==================================================
🐛 Bug Found: NameError - variable 'c' is not defined
📍 Line Number: 2
⚠️  Severity: HIGH
--------------------------------------------------
📊 Fix Confidence: 0.95
💡 Explanation: Variable 'c' is undefined in the function
🔧 Suggested Fix:
def add_nums(a, b):
    return a + b

⚙️ Configuration

Using Different Models

Update the model in any agent:

# In agents/fix_suggester.py
class FixSuggester:
    def __init__(self, model="llama3.2"):  # Change model here
        self.model = model

Customizing Prompts

Modify prompts in each agent class to suit your debugging needs:

# In agents/syntax_checker.py
def check(self, code: str) -> dict:
    prompt = f"""Your custom prompt here..."""

🤝 Contributing

1. Fork the repository
2. Create a feature branch (git checkout -b feature/amazing-feature)
3. Commit your changes (git commit -m 'Add amazing feature')
4. Push to the branch (git push origin feature/amazing-feature)
5. Open a Pull Request

📋 Future Enhancements

• Support for multiple programming languages
• Integration with popular IDEs
• Web interface for easier usage
• Code execution validation
• Integration with static analysis tools
• Custom rule definitions
• Batch processing of multiple files

🔧 Troubleshooting

Common Issues

Ollama not found:

# Make sure Ollama is installed and in PATH
ollama --version

Model not available:

# Pull the required model
ollama pull llama3.2

LangChain import errors:

# Install missing dependencies
pip install langchain langgraph

Permission errors on Windows:

# Run as administrator or check antivirus settings

📚 Resources

• LangChain Documentation
• LangGraph Documentation
• Ollama Documentation
• Llama 3.2 Model Card
• Get LangSmith API Key
• LangGraph Local Server

📄 License

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

Acknowledgments

• LangChain for the agent framework
• Ollama for local AI model hosting
• Meta for the Llama 3.2 model

---

Star ⭐ this repo if you find it helpful!