Rapier Physics Engine

Rapier is a high-performance set of 2D and 3D physics engines written in Rust. It is designed for real-time applications such as video games, animation, and robotics, prioritizing speed, stability, and optional cross-platform determinism.

Key features include:

• Rigid-body collisions and forces
• Joint constraints
• Contact events and sensors
• Snapshotting
• Optional cross-platform determinism
• JavaScript bindings

mjcf-rs is a pure-Rust parser for the MuJoCo XML format (MJCF). It generates a typed Abstract Syntax Tree (AST) that mirrors the MJCF specification 1:1. The parser automatically handles preprocessing tasks such as:

• <include> resolution
• <default> class inheritance
• Angle unit normalization

This allows consumers to use the parsed AST directly without manually implementing MJCF preprocessing logic.

Rapier provides higher-level helpers built on top of the core physics pipeline to simplify complex movement and control tasks. These include:

• Kinematic character controller: For managing character movement, including stepping over obstacles and handling collisions.
• PD / PID controllers: For precise position or velocity control using Proportional-Derivative or Proportional-Integral-Derivative logic.
• Ray-cast vehicle controller (3D-only): For simulating vehicles using ray-casting for wheels.

These controllers are available in the rapier3d module.

The rapier3d-meshloader crate allows you to create shapes compatible with rapier3d and parry3d by loading 3D mesh files. It supports the following file formats:

• .stl (STL files)
• .dae (Collada files)
• .obj (Wavefront OBJ files)

The rapier-compat package uses several tsconfig.json files to manage different compilation targets:

• tsconfig.common.json: Contains shared TypeScript options used across the project.
• tsconfig.pkg2d.json: Configuration specifically for compiling rapier2d-compat.
• tsconfig.pkg3d.json: Configuration specifically for compiling rapier3d-compat.
• tsconfig.json: Used for IDE support (e.g., VSCode) and running unit tests. This file includes Jest types.

Rapier provides 2D and 3D physics engines for Rust. To begin using the library:

1. Consult the User Guides: Visit the official documentation at https://www.rapier.rs/docs/.
2. Run Examples: You can explore the engine's capabilities by running the built-in examples using Cargo:
   • For 2D examples: cargo run --release --bin all_examples2
   • For 3D examples: cargo run --release --bin all_examples3 The source code for these examples is located in the examples2d/ and examples3d/ directories.
3. Community Support: Ask questions or suggest features on the Discord server or by opening a GitHub issue.

To support specific mesh formats referenced within your URDF files, you must enable the corresponding Cargo features in your Cargo.toml:

• stl: Enables loading .STL meshes.
• collada: Enables loading .dae meshes.
• wavefront: Enables loading .obj meshes.

To build the projects, use the scripts defined in the ./package.json file. The build process follows these steps:

1. Build Rust WASM projects into their dedicated folders within ./builds/.
2. Copy common JavaScript and generate dimension-specific JavaScript into a common ./pkg folder.
3. Copy the contents of the ./pkg folder into each individual folder within ./builds.

The Rapier website is built using Docusaurus 3 and consists of markdown files in the rapier.rs repository. To contribute:

1. Fork the rapier.rs repository.
2. Fix or improve the content.
3. Create a pull request targeting the master branch.

Rules for contributions:

• All compilable code used to generate illustrations must be located in the src folder.
• Downloadable examples must be located in the examples directory of the main rapier repository.

To run the tests, first build the package with the --release flag, then install the necessary testing dependencies and execute pytest.

maturin develop --release -m python/rapier-py-3d/Cargo.toml
pip install pytest pytest-timeout hypothesis numpy matplotlib
python -m pytest python/tests