LangGraph Code Generator Module

This module implements a workflow for generating, executing, reviewing, and packaging Python code using LangGraph and E2B.

Features

• Code generation using LangChain and Anthropic's Claude
• Safe code execution in E2B sandbox environment
• Automated code review
• Python module packaging

Installation

git clone https://github.com/awohletz/langgraph_code_generator.git
cd langgraph_code_generator
pip install -e .

Usage

from langgraph_code_generator import CodeGeneratorModule

# Initialize the module
generator = CodeGeneratorModule()

# Generate a module from a prompt
result = generator.generate_module(
    "Create a function that calculates the Fibonacci sequence"
)

# Check the result
if result["success"]:
    print("Generated code:", result["code"])
    print("Execution result:", result["execution_result"])
    print("Review result:", result["review_result"])
else:
    print("Error:", result["error"])

or just update the example.py file with your own prompt and run it.

python langgraph_code_generator/example.py

Workflow

1. Code Generation: Takes a prompt and generates Python code
2. Code Execution: Executes the code in an E2B sandbox
3. Code Review: Reviews the code for quality and issues
4. Code Packaging: Creates a Python module from approved code

Requirements

• Python 3.8+

Environment Variables

• E2B API key (set as environment variable E2B_API_KEY)
• Anthropic API key (set as environment variable ANTHROPIC_API_KEY)
• Anthropic large model (set as environment variable ANTHROPIC_LARGE_MODEL)
• Anthropic small model (set as environment variable ANTHROPIC_SMALL_MODEL) Update the .env.example file with your keys and models and rename it to .env.

Visualization

To visualize the workflow, run python langgraph_code_generator/visualize_graph.py.