Files
lammps/tools/i-pi/ipi/utils/prng.py
2020-03-18 23:09:23 +09:00

130 lines
4.1 KiB
Python

"""Contains the classes used to generate pseudo-random numbers.
Copyright (C) 2013, Joshua More and Michele Ceriotti
This program is free software: you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
the Free Software Foundation, either version 3 of the License, or
(at your option) any later version.
This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU General Public License for more details.
You should have received a copy of the GNU General Public License
along with this program. If not, see <http.//www.gnu.org/licenses/>.
Allows the user to specify a seed for the random number generator.
These are used in initialising the velocities and in stochastic thermostats.
The state of the random number generator is kept track of, so that the if the
simulation is restarted from a checkpoint, we will see the same dynamics as if
it had not been stopped.
Classes:
Random: An interface between the numpy.random module and the user.
"""
__all__ = ['Random']
import numpy as np
import math
class Random(object):
"""Class to interface with the standard pseudo-random number generator.
Initializes the standard numpy pseudo-random number generator from a seed
at the beginning of the simulation, and keeps track of the state so that
it can be output to the checkpoint files throughout the simulation.
Attributes:
rng: The random number generator to be used.
seed: The seed number to start the generator.
state: A tuple of five objects giving the current state of the random
number generator. The first is the type of random number generator,
here 'MT19937', the second is an array of 624 integers, the third
is the current position in the array that is being read from, the
fourth gives whether it has a gaussian random number stored, and
the fifth is this stored Gaussian random number, or else the last
Gaussian random number returned.
"""
def __init__(self, seed=12345, state=None):
"""Initializes Random.
Args:
seed: An optional seed giving an integer to initialize the state with.
state: An optional state tuple to initialize the state with.
"""
self.rng = np.random.mtrand.RandomState(seed=seed)
self.seed = seed
if state is None:
self.rng.seed(seed)
else:
self.state = state
def get_state(self):
"""Interface to the standard get_state() function."""
return self.rng.get_state()
def set_state(self, value):
"""Interface to the standard set_state() function.
Should only be used with states generated from another similar random
number generator, such as one from a previous run.
"""
return self.rng.set_state(value)
state=property(get_state, set_state)
@property
def u(self):
"""Interface to the standard random_sample() function.
Returns:
A pseudo-random number from a uniform distribution from 0-1.
"""
return self.rng.random_sample()
@property
def g(self):
"""Interface to the standard standard_normal() function.
Returns:
A pseudo-random number from a normal Gaussian distribution.
"""
return self.rng.standard_normal()
def gamma(self, k, theta=1.0):
"""Interface to the standard gamma() function.
Args:
k: Shape parameter for the gamma distribution.
theta: Mean of the distribution.
Returns:
A random number from a gamma distribution with a shape k and a
mean value theta.
"""
return self.rng.gamma(k,theta)
def gvec(self, shape):
"""Interface to the standard_normal array function.
Args:
shape: The shape of the array to be returned.
Returns:
An array with the required shape where each element is taken from
a normal Gaussian distribution.
"""
return self.rng.standard_normal(shape)