epsilonq 0.1.0

Lightweight RL utilities: epsilon-greedy and Q-learning update

epsilonq

epsilonq is a lightweight Python package providing basic Reinforcement Learning utilities.
It includes an Epsilon-Greedy action selection function and a Q-learning update function, making it easy to implement simple RL agents.

---

Features

• Epsilon-Greedy Action Selection for balancing exploration and exploitation.
• Q-learning Update for updating Q-values based on experience.
• Beginner-friendly and lightweight — no heavy dependencies.

---

Installation

Install from PyPI:

pip install epsilonq

# USAGE
from epsilonq import epsilon_greedy, q_learning_update

# Initialize Q-table (dictionary)
Q = {}

# Define current state and Q-values for actions
state = "A"
q_values = [0.5, 0.2, 0.9]

# Select an action using epsilon-greedy policy
action = epsilon_greedy(q_values, epsilon=0.1)

# Simulate reward and next state
reward = 1.0
next_state = "B"

# Perform Q-learning update
q_learning_update(Q, state, action, reward, next_state, alpha=0.5, gamma=0.9)

print("Updated Q-table:", Q)


---

## **Step 4 — Create `setup.py`**
This file tells PyPI how to build your package:

```python
from setuptools import setup, find_packages

setup(
    name="epsilonq",
    version="0.1.0",
    author="Your Name",
    author_email="[e-mail]",
    description="Lightweight RL utilities: epsilon-greedy and Q-learning update",
    long_description=open("README.md").read(),
    long_description_content_type="text/markdown",
    url="https://pypi.org/project/epsilonq/",
    packages=find_packages(),
    install_requires=[],
    classifiers=[
        "Programming Language :: Python :: 3",
        "License :: OSI Approved :: MIT License",
        "Operating System :: OS Independent",
    ],
    python_requires=">=3.6",
)