Code Debugging Agent with LangGraph

This project implements an AI-powered code debugging assistant using the LangGraph framework. The agent is designed to analyze code, identify bugs, suggest solutions, and provide detailed explanations to streamline the debugging process for the developers.

Features

• Bug Identification: Automatically detects syntax errors, runtime issues, and potential logical bugs.

• In-depth Analysis: Utilizes tools to assess code complexity and inspect the Abstract Syntax Tree (AST) for static issues.

• Stateful Workflow: Employs a state machine to move through a logical sequence of debugging operations.

• Graceful Error Handling: Manages internal errors to provide partial analysis and helpful tips even when a full solution can't be generated.

How It Works

The agent is built as a StateGraph where each node represents a step in the debugging process:

1. identify_bugs: The entry point, which performs an initial analysis.

2. analyze_code: Conducts a deeper inspection of the code's structure and complexity.

3. generate_fixes: Proposes solutions to the identified problems.

4. create_explanation: Summarizes the findings in a comprehensive report.

5. handle_error: A terminal node for managing any failures in the workflow.

The agent's flow is controlled by a DebuggingStatus enum, which directs the graph from one node to the next based on the outcome of the current step.

Getting Started

To run the agent, you need to define your code snippet and any known error messages. The main function will then process the input and print the results.

Usage Example:

# 1. Define the code to be debugged
test_code = """
def calculate_average(numbers):
    total = 0
    for num in numbers
        total += num
    return total / len(numbers)
"""

# 2. Provide any known error message
error_message = "SyntaxError: invalid syntax"


# 3. Run the debugging assistant
# (Assuming the run_debugging_assistant function is defined)
run_debugging_assistant(test_code, error_message)