Metadata-Version: 2.4
Name: swarm-drone-gym
Version: 2.0.0.1
Summary: Swarm quadrotor RL environments built on PyBullet with a custom drone setup, 5D yaw control, and benchmark-oriented flight behavior
License: MIT
License-File: LICENSE
Author: Swarm Subnet
Requires-Python: >=3.10,<4.0
Classifier: License :: OSI Approved :: MIT License
Classifier: Programming Language :: Python :: 3
Classifier: Programming Language :: Python :: 3.10
Classifier: Programming Language :: Python :: 3.11
Classifier: Programming Language :: Python :: 3.12
Classifier: Programming Language :: Python :: 3.13
Classifier: Programming Language :: Python :: 3.14
Requires-Dist: gymnasium (>=0.28,<1.0)
Requires-Dist: matplotlib (>=3.7,<4.0)
Requires-Dist: numpy (>=1.24,<3.0)
Requires-Dist: pytest (>=7.3,<9.0)
Requires-Dist: scipy (>=1.10,<2.0)
Requires-Dist: stable-baselines3 (>=2.0.0,<3.0.0)
Requires-Dist: swarm-bullet3 (>=2.0.0.1,<4.0.0)
Requires-Dist: transforms3d (>=0.4.1,<0.5.0)
Project-URL: Homepage, https://github.com/swarm-subnet/swarm-gym-pybullet-drones
Project-URL: Repository, https://github.com/swarm-subnet/swarm-gym-pybullet-drones
Description-Content-Type: text/markdown

# Swarm Gym PyBullet Drones

Swarm's fork of `gym-pybullet-drones` packages the quadrotor simulation environments we use for Swarm benchmark and control workloads.
It depends on the published `swarm-bullet3` distribution for the Swarm-maintained PyBullet fork.
The published distribution name for this package is `swarm-drone-gym`.

This fork keeps the Gymnasium and Stable-Baselines3 2.x compatible environment surface, while adding Swarm-specific improvements:

- custom drone parameters used by Swarm evaluation workloads
- 5D action space with yaw control for richer policy outputs
- yaw-control float32 precision fix for stable behavior at extreme values
- benchmark-oriented environment simplifications used by our training and evaluation loops

Repository: https://github.com/swarm-subnet/swarm-gym-pybullet-drones

