Unittest AI Agent is a Python tool that automatically generates comprehensive unit tests for your Python functions and classes using OpenAI's GPT models. It analyzes your source code, prepares context-rich prompts, and writes robust pytest-based test suites.

• Automatic Source Analysis: Extracts imports, functions, and class context from your Python files.
• Prompt Engineering: Uses customizable prompt templates for both standalone functions and class methods.
• LLM Integration: Sends code context to OpenAI GPT-4o to generate high-quality unit tests.
• Test Postprocessing: Cleans and adapts generated code for your project structure.
• Test Writing: Saves generated tests to the appropriate directory.

0. Requirements:

   • Python 3.9 or higher
   • OpenAI API key
   • Poetry curl -sSL https://install.python-poetry.org | python3

1. Install dependencies:

   poetry install

2. Set your OpenAI API key:

   • Rename the .env.example to .env:

     OPENAI_API_KEY=your_openai_api_key_here
     

3. Run the test generator:

   poetry run ut generate example/converter.py

   This will analyze ut/example/converter.py and generate tests in ut/example/tests/.

• Prompt Templates: Edit files in ut/prompts/ to change how prompts are constructed for the LLM.

Given a function like:

def convert_date_to_iso(date_str: str, format: str = "%d/%m/%Y") -> str:
    ...

The agent will generate a suite of pytest tests covering various edge cases and save them to ut_output/test_convert.py.

This project is under active development. Below is a summary of our progress and a look at what's ahead.

Contributions are highly encouraged!

👉 Roadmap: https://focusmap.pro/roadmap/45a1b599-aead-4c11-b749-032e5ea168e1

Apache 2.0