Controlling Random Number Generation

The generation of random numbers is done through a numpy.random.Generator(). The default is numpy.random.default_rng().

import numpy as np
import pyvista as pv
from pyvista import examples

import pyransame

antarctica = examples.download_antarctica_velocity()
antarctica.points /= 1000.0  # convert to kilometers

pyransame stores the generator being used at pyransame.rng

pyransame.rng

Generator(PCG64) at 0x7F7B777C26C0

Setup common plotting routine for use later.

def plot_points(points):
    pl = pv.Plotter()
    pl.add_mesh(antarctica, color="tan")
    spheres = points.glyph(geom=pv.Sphere(radius=50), scale=False, orient=False)
    pl.add_mesh(spheres, scalars="ssavelocity", clim=[0, 750])
    pl.view_xy()
    pl.show()

Sampling twice in succession will lead to different results.

points = pyransame.random_surface_dataset(antarctica, 500)
plot_points(points)

Static Scene

Interactive Scene

Second sampling with slightly different results

points = pyransame.random_surface_dataset(antarctica, 500)
plot_points(points)

Static Scene

Interactive Scene

This time, we will control the random number generation by using a seed to ensure that the same result is obtained.

pyransame.rng = np.random.default_rng(seed=42)
points = pyransame.random_surface_dataset(antarctica, 500)
plot_points(points)

Static Scene

Interactive Scene

Second sampling with identical results.

pyransame.rng = np.random.default_rng(seed=42)
points = pyransame.random_surface_dataset(antarctica, 500)
plot_points(points)

Static Scene

Interactive Scene