pyfastnoiselite-ledfx

Note: This is a fork of the original PyFastNoiseLite maintained by the LedFx team.

Why this fork exists:

• The original project is sporadically active, and its 0.0.7 release has no Python 3.14 wheels and no sdist on PyPI
• We build against the CPython stable ABI, so one wheel per platform covers Python 3.11 and every later release
• We ship armv7l wheels for 32-bit Raspberry Pi OS
• LedFx depends on pyfastnoiselite and needs a reliable, up-to-date release

All credit for pyfastnoiselite goes to Tiziano Bettio, and for FastNoise Lite to Jordan Peck (Auburn). This fork exists solely to provide maintained releases for projects that depend on it. The build changes are offered upstream in tizilogic/PyFastNoiseLite#3.

Original project: https://github.com/tizilogic/PyFastNoiseLite\ This fork: https://github.com/LedFx/pyfastnoiselite-ledfx

A Cython wrapper for Auburn's FastNoise Lite noise generation library.

Installation

pip install pyfastnoiselite-ledfx

Binary wheels are published for CPython 3.11+ on Windows (x86_64), macOS (x86_64, arm64) and Linux glibc/musl (x86_64, aarch64, armv7l). Building from the sdist needs a C++11 compiler.

The abi3 wheels target standard (GIL-enabled) CPython. Free-threaded Python requires a separate source build. On 32-bit Raspberry Pi OS, NumPy may also need a distribution-provided package or a source build because PyPI does not provide NumPy ARMv7 wheels.

The distribution is renamed but the import name is unchanged, so it is a drop-in replacement for pyfastnoiselite. Don't install both: they provide the same module.

Note: This wrapper lacks the domain warping functionality.

Usage

from pyfastnoiselite.pyfastnoiselite import FastNoiseLite, NoiseType

# Initializing with seed
noise = FastNoiseLite(seed=1337)
# Set noise type (optional, defaults to OpenSimplex2)
noise.noise_type = NoiseType.NoiseType_OpenSimplex2S

# Get 2D noise
print(noise.get_noise(34, 22))  # 0.7130074501037598
print(noise.get_noise(100, 110))  # -0.3495847284793854

# Get 3D noise
print(noise.get_noise(95, 100, 30))  # -0.4522402286529541


import numpy as np

Xs = [3, 57, 95]
Ys = [4, 13, 100]
Zs = [0, -4, 30]
coords = np.array([Xs, Ys, Zs], dtype=np.float32)
# Generate noise for each coordinate
print(noise.gen_from_coords(coords))

For many points, gen_from_coords is much faster than calling get_noise in a Python loop. It takes a float32 array of shape (2, N) or (3, N) (read-only arrays and views are fine) and returns a float32 array of shape (N,).

For batches of 1024 points or more, gen_from_coords releases the GIL, so separate FastNoiseLite instances can generate in parallel threads. Don't change an instance's settings from another thread while it is generating.

License

This project is licensed under the MIT license. FastNoise Lite is also MIT licensed.

Release publication

The existing build.yml/pypi Trusted Publisher identity and native/ABI3 build matrix are retained. After all CI gates pass, one queued job downloads the same run's cibw-* wheels/sdist and calls the SHA-pinned shared release transaction with the generated wheel plan and existing pyproject.toml metadata. Coverage includes compressed glibc aliases, musllinux and ARMv7 with ABI3 reuse.

The scoped App token and caller OIDC/attestation steps verify archive metadata, source/tag identity, SHA-256 and provenance before publishing the existing draft last. Matching partial uploads can resume from the original run. Missing drafts or conflicting assets now fail instead of creating fallback releases or using --clobber; release-please must create the draft and preserve its notes. Higher stable drafts/releases veto latest promotion, so an abandoned newer draft may delay latest without blocking immutable version publication. Keep the retained snapshot/bundles and follow the shared recovery guide; do not rebuild the same version to replace published bytes. PR/manual runs do not publish.

Wheel support is maintained in [tool.cibuildwheel]; platform rows live in [[tool.release-ci.targets]] in pyproject.toml. Update the single wheel-build cibuildwheel dependency pin and uv.lock together. Planning and builds use that locked tool and configuration; publication rejects missing platform coverage.