RandNumGenerator.py

import time
import numpy as np
import pandas as pd


def seed_time():
    t = time.localtime(time.time())
    return t.tm_year-2000 + 70*(t.tm_mon+12*(t.tm_mday+31*(t.tm_hour+23*(t.tm_min+59*t.tm_sec))))


class Schrage16807(object):
    def __init__(self, seed=42):
        self.__seed = seed
        self.__M = 2147483647
        self.__a = 16807

    def __iter_rand(self, i_n):
    # Schrage Random Number Generator
        a = self.__a
        m = self.__M
        q = m//a
        r = m % a
        i_next = a * (i_n % q) - r * (i_n // q)
        if i_next < 0:
            return i_next + m
        else:
            return i_next

    def __rand_list(self, dim=1) -> tuple[np.ndarray, float]:
    # Random Number Generator
        if dim == 1:  # A random number
            self.__seed = self.__iter_rand(self.__seed)
            return self.__seed/self.__M
        elif dim > 1:  # A list of random numbers
            i_n = self.__seed
            l = np.zeros(dim)
            for i in range(dim):
                i_n = self.__iter_rand(i_n)
                l[i] = i_n
            self.__seed = i_n
            return l/self.__M

    def rand(self, d0=1, d1=None):
        if isinstance(d0, int):  # If d0 is a positive integer
            if d0 == 1:
                if d1 is None:
                    return self.__rand_list(1)
                elif isinstance(d1, int) and (d1 > 0):
                    temp = self.__rand_list(d1)
                    return temp
            elif d0 > 1:
                if d1 is None:
                    return self.__rand_list(d0)
                elif isinstance(d1, int) and (d1 > 0):
                    if d1 == 1:
                        temp = self.__rand_list(d0)
                        return temp.T
                    else:
                        temp = [self.__rand_list(d1) for item in range(d0)]
                        a = np.array(temp)
                        return a.reshape((d0, d1))
        return print("Error, please input a positive integer!\n")

    def __normal(self, size):
        """
        Generate random numbers that follow the normal distribution
        生成服从正态分布的随机数

        Parameters
        ----------
        loc : float
        scale : float
        size : int or tuple of ints
        """
        i = 0
        tmp = np.zeros(size).flatten()
        l = len(tmp)

        while i < l:
            u = self.__rand_list(1)*2-1
            v = self.__rand_list(1)*2-1
            r = np.sqrt(u**2+v**2)
            if r >= 1:
                continue
            else:
                tmp[i] = (u/r)*np.sqrt(-4*np.log(r))
                if i+1 < l:
                    tmp[i+1] = (v/r)*np.sqrt(-4*np.log(r))
                i += 2
        return tmp.reshape(size)

    def normal(self, loc=0, scale=1,size=None):
        if size is None:
            return self.__normal(1)*scale+loc
        else:
            return self.__normal(size)*scale+loc

    def uniform(self, low=0, high=1,size=None):
        if size is None:
            return self.__rand_list(1)*(high-low)+low
        else:
            return self.__rand_list(size)*(high-low)+low

    def randint(self, low=0, high=1,size=None):
        if size is None:
            return np.floor(self.__rand_list(1)*(high-low)+low)
        else:
            return np.floor(self.__rand_list(size)*(high-low)+low)


def makefile(data, filename):
    df = pd.DataFrame(data=data)
    df.to_csv(filename+'.csv', index=False)


def readfile(filename):
    df = pd.read_csv(filename)
    return np.array(df)