kwant.digest – Random-access random numbers¶Random-access random numbers
This module provides routines that given some input compute a “random” output that depends on the input in a (cryptographically) intractable way.
This turns out to be very useful when one needs to map some irregular objects to random numbers in a deterministic and reproducible way.
Internally, the md5 hash algorithm is used. The randomness thus generated is
good enough to pass the “dieharder” battery of tests: see the function test of
this module.
kwant.digest.uniform(input, salt='')¶md5-hash input and salt and map the result to the [0,1) interval.
input must be some object that supports the buffer protocol (i.e. a string or a numpy/tinyarray array). salt must be a string or a bytes object.
kwant.digest.gauss(input, salt='')¶md5-hash input and salt and return the result as a standard normal distributed variable.
input must be some object that supports the buffer protocol (i.e. a string or a numpy/tinyarray array). salt must be a string or a bytes object.
kwant.digest.test(n=20000)¶Test the generator with the dieharder suite generating n**2 samples.
Executing this function may take a very long time.