Procedural noise generation for Python.
Uses noise functions from the Rust noise library at https://github.com/brendanzab/noise-rs. Currently, noisily links with a forked version of the library which adds periodicity (for tileable noise).
import noisily as ns
import numpy as np
# Create an array of points. Noisily accepts any array of points so long
# as the shape of the last axis matches the dimension.
points = np.indices((1024, 1024)).T
# Scale for a closer range
points *= 1 / 64
# Create a generator for 2D cell noise using the default seed and period
generator = ns.generator(ns.cell2D_range_inv)
# Call the generator to get the values at each point. This has the same shape
# as the array, sans the last axis. E.g. (1024, 1024, 2) -> (1024, 1024)
values = generator(points)pip install noisily will install a binary distribution from PyPI.
python setup.py install if you need to build from source. Make sure to have Rust installed and cargo in your path.