rust-simulation-tools 0.1.11

Fast algorithms for MD trajectories

Rust Simulation Tools

Fast, numerically stable MD trajectory processing implemented in Rust with a clean python API. Installable via pip, integrates smoothly with MDAnalysis or mdtraj, and ships with tests.

Installation

pip install rust-simulation-tools

Features

• ⚡ Fast: Rust implementation with SIMD optimizations
• 🎯 Accurate: Numerically stable Kabsch alignment and fragment-based unwrapping
• 🧪 Well tested: Comprehensive test suite with >80% coverage
• 🧩 Easy integration: Works directly with MDAnalysis selections/indices

Usage

import MDAnalysis as mda
from rust_simulation_tools import kabsch_align

# Load trajectory
u = mda.Universe("topology.pdb", "trajectory.dcd")

# Select alignment atoms
align_selection = u.select_atoms("backbone")
align_indices = align_selection.indices.astype(np.int32)

# Get coordinates
reference = u.atoms.positions.copy().astype(np.float32)
trajectory = np.array([ts.positions for ts in u.trajectory], dtype=np.float32)

# Align
aligned = kabsch_align(trajectory, reference, align_indices)

API Reference

kabsch_align(
    trajectory: np.ndarray,      # float, shape [n_frames, n_atoms, 3]
    reference: np.ndarray,       # float, shape [n_atoms, 3]
    align_idx: np.ndarray        # int,   shape [n_alignment_atoms]
) -> np.ndarray                  # float, shape [n_frames, n_atoms, 3]

unwrap_system(
    trajectory: np.ndarray,      # float, shape [n_frames, n_atoms, 3]
    box_dimensions: np.ndarray,  # float, shape [n_frames, 3]
    fragment_idx: np.ndarray     # int,   shape [n_atoms]
) -> np.ndarray                  # float, shape [n_frames, n_atoms, 3]

calculate_sasa_trajectory(
    coordinates: np.ndarray,     # np.float64, shape [n_frames, n_atoms, 3]
    radii: list,                 # int, shape [n_atoms]
    residue_indices: np.ndarray  # np.int64, shape [n_atoms]
    probe_radius: float=1.4      # float, default 1.4
    n_sphere_points: int=960     # int, default 960
) -> dict[str, np.ndarray]       # np.float64, keys ['total', 'per_atom', 'per_residue']

Development

# Clone repository
git clone https://github.com/msinclair-py/rust-simulation-tools.git
cd rust-simulation-tools

# Install development dependencies
pip install maturin pytest pytest-cov numpy

# Build and install in development mode
maturin develop --release

# Install with pip
pip install target/wheels/name-of-package.whl

# Run tests
pytest tests/ -v --cov

License

MIT License