lerl 0.5.0

Multi code-snippet generator for notebooks

📘 Machine Learning Snippet Generator (lerl)

A lightweight Python library that generates ready-to-run Machine Learning code snippets for students, beginners, and researchers working with AI algorithms.

This library does not run AI algorithms internally — instead, it returns formatted, complete Python scripts for:

• Monte Carlo Prediction
• Temporal Difference (TD(0))
• Policy Iteration
• Value Iteration (coming soon)
• • more upcoming AI algorithms

You can copy the output, paste it into a Jupyter Notebook, Colab, VS Code, or PyCharm, and run it directly.

---

🚀 Features

✔ Generates complete RL scripts as strings
✔ Code always follows correct RL implementations
✔ Uses format_snippet() to return clean, properly formatted code
✔ Ideal for learning, assignments, and quick experimentation
✔ Zero dependencies for running the library itself
✔ Works in: Jupyter Notebook, Google Colab, VS Code, Spyder, PyCharm

---

📦 Installation

Install the package using pip:

pip install lerl

---

Upgrade:

pip install --upgrade lerl

---

🧠 Available Code Snippet Functions

Function Name	Algorithm	Description
monte_carlo_code()	Monte Carlo Prediction	Returns a full script for MC value estimation on FrozenLake
td_code()	Temporal Difference (TD-0)	Generates complete TD learning code with convergence plots
complete_policy_iteration_code()	Policy Iteration	Returns policy evaluation + improvement + convergence visualization
complete_value_iteration_code()	Value Iteration	(Coming soon — ready function placeholder)

---

📝 Example Usage

▶️ Generate Monte Carlo Snippet

from lerl import monte_carlo_code

print(monte_carlo_code())

▶️ Generate TD(0) Snippet

from lerl import td_code

print(td_code())

▶️ Generate Policy Iteration Code

from lerl import complete_policy_iteration_code

print(complete_policy_iteration_code())

Each function returns a complete runnable script containing:

• Gym environment setup
• Algorithm implementation
• Value function updates
• Convergence plots
• Printed results

---

📚 Example Output (Short Preview)

import numpy as np
import gymnasium as gym
import matplotlib.pyplot as plt

env = gym.make('FrozenLake-v1', is_slippery = 'False')

def monte_carlo(env, policy, episodes = 10000, df = 0.99):
    ...

(Full script provided when calling the function.)

---

🏗 Internal Architecture

lerl/
│
├── __init__.py
├── qrl.py                # Main snippet-generating functions
├── utils.py              # format_snippet() helper
└── README.md

---

📜 License

This project is licensed under the MIT License. Feel free to use it in academic or commercial projects.

---

⭐ Support

If this library helps you learn or build RL assignments, give it a ⭐ on GitHub and share it with your friends!