Python-从自定义时间生成随机数

Question

So say that I generate a random number using random.random() on one system without a custom seed, but say that I wanted to also generate a random number ahead of time so that if I were to use random.random() at that time without seeding it would return the same number. In essence, "predict" the random number. How might I customize the time that random uses to seed itself? I thought to do the following based on how the random module seeds:

import random
import time

print random.random()
random.seed(long(time.time()*256))
print random.random()

However, that gave me separate numbers and I don't understand why. Should they not be done at the same instance in time?

tl;dr

How can I customize random.seed(x) where it would seed at a custom point in time.

Assume that random.random() is called a-new each time so it is the first number being generated, and not the second, third, fourth, etc.

Note that I do not want to do this in the past, so not get the state and then restore it, but be able to generate what will be generated in the future.

CLARIFICATION: The system this will be used on does not have urandom implemented.

Answer 1

Actually I was wrong.

It's possible to do it: Here's how:

import time
import random

def predicite_future(delay):
    random.seed(long(time.time() + delay) * 256)
    return random.random()

delay = 10

print "The next value will be:",predicite_future(delay)
print "Waiting for %s seconds"%delay
time.sleep(delay)
random.seed(long(time.time()) * 256)
print "As predicited, the value is:",random.random()

output:

The next value will be: 0.359211550742
Waiting for 10 seconds
As predicited, the value is: 0.359211550742

Answer 2

I think you actually do have a urandom of sorts or at the very least the following does not throw an exception

from binascii import hexlify as _hexlify
from os import urandom as _urandom
a = long(_hexlify(_urandom(16)), 16)

does it?

Now assuming that lets modify the random lib which can be found here .

first replace

    if a is None:
        try:
            a = long(_hexlify(_urandom(16)), 16)
        except NotImplementedError:
            import time
            a = long(time.time() * 256) # use fractional seconds

with just

    if a is None:
        import time
        a = long(time.time() * 256) # use fractional seconds
        print ("bottom - the seed is", a) # debug

Now in bottom of the file in the if __name__ == "__main__" remove the test and instead put

if __name__ == '__main__':
    from time import time
    then1 = long(time() * 256)
    from random import random, seed
    then2 = long(time() * 256)
    assert then1 == then2 # if fails here the time has changed

    print random() # unseeded
    seed(then1)
    print random() # use the seed with think it uses

Now assume the time we want to predict is d in the future then we could have then1 = long((time()+d) * 256) to predict a result of random.random() in the future.