CURA Framework

CURA (Code Understanding & Reasoning Agent) is a modular framework for programmatic code reasoning, test generation, and code synthesis using LLMs. It is designed for research and practical applications in code generation, evaluation, and dataset creation.

Features

• Modular agent and pipeline design
• Automated prompt engineering for code understanding, test generation, and code synthesis
• Dataset generation and evaluation utilities
• CLI for batch evaluation

Installation

You can install CURA and its dependencies using pip:

pip install .

Or with Poetry:

poetry install

Requirements

• Python 3.9+
• macOS, Linux, or Windows

Usage

Programmatic API

from cura.agent import Agent

agent = Agent(
    name="CodeGenAgent",
    description="Generates code from prompts",
    tools=[],
    model="gpt-4o-mini"
)

result = agent.run("Write a Python function to add two numbers.")
print(result)

CURA Pipeline

from cura.pipeline import CURAPipeline

pipeline = CURAPipeline()
code = pipeline.generate("Write a function to reverse a string.")
print(code)

CLI

You can run dataset evaluation from the command line:

python cli.py evaluate-dataset --subset hard --split complete

Project Structure

cura/
    __init__.py
    agent.py
    base_agent.py
    pipeline.py
    prompts.py
    tasks.py
    utils.py
cli.py
sample.py
setup.py
requirements.txt
README.md

Environment Variables

• OPENROUTER_API_KEY: Your OpenRouter API key for LLM access.

License

MIT License