The pyudorandom module lets you iterate over a list in a non-succsessive, yet deterministic way. It comes in handy when you want to mix up the items, but don't need any guarantees of randomness. Also, it makes sure that it only gives you the elements once.
If you have a iterable of length n, pyudorandom will first find a random number m between 0 and n-1 such that the gcd(m, n) == 1. That number will then be used to generate 0 through n-1 by using integers modulo n.
As such, it might be slow on small data, but shall be significantly faster than random.shuffle for longer lists.
Given a list ls = [1, 5, 7, 3, ..., 321, 994] and the imported module import pyudorandom.
##pyudorandom.items(ls)
Draw 'random' items from ls.
>>> for i in pyudorandom.items(ls):
... print(i)
...
7
321
...
...
5Get a new list with the elements of ls in a new order.
>>> new_order = pyudorandom.shuffle(ls)
>>> new_order == ls
False
>>> set(new_order) == set(ls)
TrueGet the indices of the list in a 'random' order.
See source in perf.py.
| methods | n=10 | n=100 | n=1000 | n=10000 | n=100000 | n=1000000 | n=10000000 |
|---|---|---|---|---|---|---|---|
| random.shuffle | 9.4878e-05 | 0.000697057 | 0.0079944 | 0.0994185 | 1.15775 | 13.1515 | 158.261 |
| pyudorandom.shuffle | 9.2293e-05 | 0.000397103 | 0.00311404 | 0.0421352 | 0.577736 | 6.40441 | 88.2447 |
| pyudorandom.items | 7.2781e-05 | 0.000374695 | 0.00349281 | 0.0377044 | 0.551456 | 6.10636 | 84.6845 |
Pyudorandom can be twice as fast as random.