WavePart Python

WavePart Python is a clean Python package for partitioning directional wave spectra into wind sea and multiple swell components. The repository is organized for direct Python use, includes a compact command-line interface, and ships with a Python-native sample dataset so a new user can run it immediately.

Install

python3 -m venv .venv
source .venv/bin/activate
python -m pip install -U pip
python -m pip install -e .[plot,dev]

Quick Start

Use the bundled sample dataset:

wavepart partition data/sample_spectra.npz --index 1 --output out/partition_case.npz
wavepart wind-limits data/sample_spectra.npz --output out/wind_limits.npz
wavepart demo data/sample_spectra.npz --index 287 --output-dir out/demo

Use the package directly:

import numpy as np
from wavepart import partition_spectrum, compute_partition_params

with np.load("data/sample_spectra.npz") as data:
    freq = data["freq"]
    direction = data["direction"]
    spectrum = data["spectra"][:, :, 0]

partition = partition_spectrum(freq, direction, spectrum)
params = compute_partition_params(
    partition.smoothed_spectrum,
    freq,
    direction,
    partition.labels,
    depth=30.0,
)

Dataset Format

The package expects .npz files with these keys:

• freq
• direction
• spectra
• time (optional)

The bundled sample file follows that format directly.

What Is Included

• src/wavepart: core package
• data/sample_spectra.npz: ready-to-run example dataset
• tests: regression, CLI, and plotting tests
• docs/assets/wavepart-3d.png: package preview image

Verification

Run the full test suite:

pytest

Current tests cover:

• reference regression cases
• wind-limit regression
• flat-spectrum and label invariants
• CLI smoke checks
• plotting smoke checks

Origin

This Python implementation is based on the original WavePart work by Douglas Cahl and George Voulgaris. The original repository remains available here:

• https://github.com/gvoulgaris0/WavePart

License

This repository is distributed under the GNU General Public License v3.0. See LICENSE.