Randomness and Reproducibility

pypuf strives to make all results fully reproducible.

One key ingredient to this is to only use seeded random number generators. To avoid re-using seeds where different random generators are expected, pypuf implements a convenience function to obtain random generators based on a string description.

pypuf.random.prng(description: str) → Generator

Returns an instance of numpy.random.Generator seeded based on a text description.

>>> from pypuf.random import prng
>>> seed = 5
>>> my_prng = prng(f'my favorite random numbers, seed {seed}')
>>> my_prng.integers(2, 6, 3)
array([4, 3, 2])