PyAOF

PyAOF is a Python library for computing Average Outward Flux (AOF) on 3D meshes using NVIDIA Warp.

Features

• Compute Average Outward Flux (AOF) volumes
• Efficient tiled computation for large volumes
• Optional 3D visualization with Polyscope and Matplotlib

Installation

Basic Installation

pip install pyaof

Installation with Visualization Support

If you want to run the visualization examples that render 3D meshes and point clouds (requires matplotlib and polyscope):

pip install "pyaof[viz]"

Developer Installation

If you are contributing to the project or want to run the local test suite and coverage tools, install the test dependencies:

pip install "pyaof[test]"

Install Everything

To install all optional dependencies at once:

pip install "pyaof[all]"

Quick Start

import vtk
from pyaof import compute_aof, compute_sdf

# 1. Load mesh
reader = vtk.vtkOBJReader()
reader.SetFileName("data/cuboid.obj")
reader.Update()
mesh = reader.GetOutput()

# 2. Compute Signed Distance Field (SDF)
sdf, normalized_mesh, mesh_vertices, spacing = compute_sdf(
    mesh,
    resolution=1000,
    normalize=True
)

# 3. Compute Average Outward Flux (AOF)
aof_vol = compute_aof(
    sdf,
    tile_size=1024
)

Examples

The examples/ directory contains full usage demonstrations, including:

• Mesh loading and preprocessing
• SDF generation
• AOF computation
• 3D visualization using Polyscope

Run an example:

python examples/demo_polyscope.py

Requirements

• Python 3.12+
• NumPy
• VTK
• SciPy

Optional visualization dependencies:

• Matplotlib
• Polyscope

License

MIT License

Citation

If you use PyAOF in academic work, please cite the repository and related publications.