根据全局变量设置选择不同的功能

Question

I want to try out different random number generators but not change my code everywhere.

So I define a global variable rG

rG = None

# and provide an intialisation routine

def initrG(randomGenerator):
    global rG
    if (randomGenerator == "secret"):
        rG = randbelow
    elif (randomGenerator == "mple"):
        rG = random.randint
    elif (randomGenerator == "numpyrandint"):
        rG = numpy.random.randint
    else:
        rG = None

Then use rG like in

randomNumber = rg (10) 

calling the different generators dependent on the initial call of the initrG routine.

Two issues:

a) Somehow the assignment of the functions seems not to work.

b) the functions have different number of parameters. How should this be handled.

Answer 1

It generally would be clearer to have rG be a class with state that determines which algorithm to use, but you specifically asked to not have to rewrite code.

Here's a potential way to do it, using functools.partial to create small wrapper functions. Alternatively create the wrapper functions yourself, it'll be a little clearer to see what each is doing.

from functools import partial
from secrets import randbelow
import numpy as np
import random

rG = None

def initrG(randomGenerator):
    global rG
    if (randomGenerator == "secret"):
        rG = randbelow
    elif (randomGenerator == "mple"):
        rG = partial(random.randint, 0)
    elif (randomGenerator == "numpyrandint"):
        rG = partial(np.random.randint, 0)
    else:
        rG = None

initrG("secret")
print(rG(10), rG.__module__)
initrG("mple")
print(rG(10), rG.func.__module__)
initrG("numpyrandint")
print(rG(10), rG.func)