gobot 0.1.5

Python bindings for the Gobot robotics scene, simulation, and rendering engine.

gobot

Gobot is a Linux robotics simulation package with a Python-first workflow. Use it from Python to create scenes, step simulation, inspect robot state, and drive reinforcement-learning experiments.

Install

Gobot currently publishes Linux wheels.

pip install gobot -i https://pypi.org/simple

Check the install:

import gobot

print(gobot.__file__)
print(gobot.__version__)
print(gobot.backend_infos())

Python Usage

Create and inspect a small test scene:

import gobot

scene = gobot.create_test_scene()
root = scene.root

print(root.name)
print(root.type)
print([child.name for child in root.children])

Run a minimal reinforcement-learning environment:

import gobot

gobot.app.context().set_project_path("/tmp/gobot-demo")
gobot.scene.save_cartpole_scene("res://cartpole.jscn")

env = gobot.rl.ManagerBasedEnv(
    {"scene": "res://cartpole.jscn", "backend": "null", "controlled_joints": ["slider"]}
)
obs, info = env.reset(seed=1)

for _ in range(10):
    obs, reward, terminated, truncated, info = env.step([[0.0]])
    if terminated[0] or truncated[0]:
        obs, info = env.reset()

Use the Gymnasium-style adapter:

import gobot

env = gobot.rl.GymWrapper(
    gobot.rl.ManagerBasedEnv(
        {"scene": "res://cartpole.jscn", "backend": "null", "controlled_joints": ["slider"]}
    )
)
obs, info = env.reset(seed=1)
obs, reward, terminated, truncated, info = env.step([0.0])

Set a project root when working with res:// paths:

import gobot

gobot.set_project_path("/path/to/project")
scene = gobot.load_scene("res://world.jscn")

Editor

The wheel also installs the editor command:

gobot_editor

The executable and its local shared libraries are installed inside the Python package, so a normal pip install gobot is enough for the packaged runtime. Check the packaged editor version with gobot_editor --version.

Packaged examples are available from the editor start screen under Examples. The same files are installed inside the Python package at gobot/examples/. The first bundled example is cartpole, a MuJoCo-backed inverted pendulum scene with a Python NodeScript controller. The Python launcher records packaged example roots in ~/.gobot/projects.json under example_roots, and the GUI reads that list instead of guessing from the editor executable path. If both a source checkout and a pip install are present, same-name examples are de-duplicated and the pip-installed root is preferred when launched through the Python package. See doc/examples.md for the example project layout and packaging rules.

To locate examples from Python:

from importlib import resources

print(resources.files("gobot").joinpath("examples"))

Local Wheel Build

From a source checkout:

git clone https://github.com/RobSimulatorGroup/gobot.git
cd gobot
git submodule update --init --recursive
python -m pip install -U build scikit-build-core
python -m build --wheel
python -m pip install --force-reinstall dist/gobot-0.1.5-*.whl

For a faster local build without MuJoCo:

python -m build --wheel -Ccmake.define.GOB_BUILD_MUJOCO=OFF

Notes

• Supported platform: Linux.
• Python package name: gobot.
• Main RL entry point: gobot.rl.ManagerBasedEnv.
• Gym-style helpers: gobot.rl.GymWrapper.
• MuJoCo support is included in release wheels when available in the build.
• Packaged examples: gobot/examples/ in wheels and examples/ in source.
• MuJoCo RL roadmap: doc/mujoco_rl_plan.md.