v0.1.2

🎯 Research Purpose

SpaceMining serves as a testbed for evaluating large language models' ability to design reward functions for completely unfamiliar environments, addressing concerns about pretraining bias in standard RL benchmarks.

🔧 Key Features

• 2D physics-based space mining simulation
• Energy management and resource collection mechanics
• Partial observability with limited observation radius
• Moving obstacles and collision detection
• Continuous action space with thrust and mining controls
• Customizable reward structure

📦 Installation

pip install space-mining

🚀 Quick Start

import gymnasium as gym
import space_mining

env = gym.make("SpaceMining")

📚 Documentation

• Getting Started Guide
• Examples

Version Info

• Python: 3.10+
• License: MIT
• Dependencies: Gymnasium, Stable-Baselines3, Pygame, NumPy

---

Authors: Xinning Zhu & Lunde Chen
Repository: https://github.com/reveurmichael/space_mining