random-seedable

Fully-fledged random number generator library offering both 32 bit and 64 bit high quality implementations of Xorshift, Xorwow, Mersenne Twister, PCG, and LCG. Each implements a standard API producing number distributions that exactly match the original implementations. Typescript compatible.

Highlights

• Avoids the state overflow problems that plague other javascript implemented random number generators.
• Matches the output of original authored C/C++ implementations for all algorithms.
• 32 bit and 64 bit generators.
• Simple, common API to all generators.
• Light footprint.
• Browser support.
• ES Style module.

Install

npm install random-seedable --save

Getting Started

Just want to use a random generator with no hassle? All you've got to do is import random, and use it as you would a generator you initialised yourself.

Simply import random and call whatever method you please,

import random from 'random-seedable';

random.bool(); // false
random.int(); // 2346
random.float(); // 0.34527
random.shuffle([1, 2, 3, 1, 4]); // [2, 4, 3, 1, 1]
random.choice([1, 2, 3, 1, 4]); // 2

Want a random generator with a seed you can set? Just import the generator you wish and initialise it yourself.

import { XORShift } from 'random-seedable';

const random = new XORShift(123456789);
random.int(); // 123312
random.bool(); // true
random.shuffle([1, 2, 3, 1, 4]); // [2, 4, 3, 1, 1]
random.choice([1, 2, 3, 1, 4]); // 2

Once the generator is initialised, the seed can be changed using the seed setter function.

import { XORShift } from 'random-seedable';

const random = new XORShift(123456789);

console.log(random.seed); // 123456789

random.seed = 987654321;

console.log(random.seed); // 987654321

Condensed Documentation

PRNGs

Supported PRNGs and their default initialisations.

| Class | Default Initialisation | Integer output | | -----------| ------ | ------ | | LCG |new LCG(Date.now(), 1664525, 1013904223, 4294967296); | 32 bit | | PCG | new PCG(Date.now(), 6364136223846793005n, 1442695040888963407n); | 32 bit | | MersenneTwister | new MersenneTwister(Date.now(), 624, 397); | 32 bit | | XORShift | new XORShift(Date.now(), 13, 17, 5); | 32 bit | | XORShift64 | new XORShift64(Date.now(), 13, 7, 17); | 64 bit | | XORShift128 | new XORShift128(Date.now(), 362436069, 521288629, 88675123); | 32 bit | | XORShift128Plus | new XORShift128Plus(Date.now(), 362436069); | 64 bit | | XORWow | new XORWow(Date.now(), 362436069, 521288629, 88675123, 5783321, 6615241, 362437); | 32 bit | | random | default PRNG, same as XORShift64 | 64 bit |

PRNG methods.

Each PRNG has the following methods.

| Method | Parameters | Return | | ----------- | -------- | ------ | | .bool() | None | Boolean. | | .coin(pTrue) | pTrue:Number | Boolean. | | .int() | None | Number. | | .bigInt() | None | BigInt. | | .float() | None | Float. | | .float53() | None | Float spread over full range. | | .randRange(min, max) | min:Number, max:Number | min <= Number <= max | | .randBelow(max) | max:Number | Number <= Max | | .choice(array) | array:[?] | Item from array of type ? | | .shuffle(array, inPlace = true) | array:[?], inPlace:Boolean | Shuffled Array[?] | | .boolArray(size) | size:Number |Array[Boolean] of length size. | | .coinArray(size, pTrue) | size:Number, pTrue:Number |Array[Boolean] of length size. | | .intArray(size) | size:Number | Array[Number] of length size. | | .bigIntArray(size) | size:Number | Array[BigInt] of length size. | | .randRangeArray(size, min, max) | size:Number, min:Number, max:Number | Array[Number] of length size filled w/ min <= num <= max. | | .floatArray(size) | size:Number | Array[Number] between 0.0 - 1.0 of length size. | | .float53Array(size) | size:Number | Array[Number] between 0.0 - 1.0 of length size. |

Full Documentation

LCG

Linear Congruential Generator (LCG) is a simple generator originally devised in 1951, if you need something quick with minimal memory usage and not the best quality randomness, this is for you. 32 bits of output.

Parameters

• seed -> Initial seed.
• a -> Multiplier parameter.
• c -> Increment parameter.
• m -> Modulus parameter.

Example

const random = new LCG(1234, 1664525, 1013904223, 4294967296);

PCG

Permuted Congruential Generator (PCG) is again, a relatively simple generator that improves on the qualites of LCG by improving its randomness quality by increasing its state size and using only the most significant bits to produce the output. 32 bits of output.

Parameters

• seed -> Initial seed.
• mul -> Multiplier parameter.
• inc -> Increment parameter.

Example

const random = new PCG(0x4d595df4d0f33173n, 6364136223846793005n, 1442695040888963407n);

MersenneTwister

Mersenne Twister is a widely used PRNG, most well known for being the Python and Excel default with an extremely large state. 32 bits of output.

Parameters

• seed -> Initial seed.
• n -> Degree of recurrence.
• m -> Middle word, offset used during recurrence.

Example

const random = new MersenneTwister(5489, 624, 397);

XORShift

XorShift generators are fast, efficient generators with good randomness quality. This generator has 32 bit output with 32 bits of internal state.

Parameters

• seed -> Initial seed.
• a -> First bit shift parameter.
• b -> Second bit shift parameter.
• c -> Third bit shift parameter.

Example

const random = new XORShift(11234, 13, 17, 5);

XORShift64

XorShift generators are fast, efficient generators with good randomness quality. This implementation has 64 bit output with 64 bits of internal state.

Parameters

• seed -> Initial seed.
• a -> First bit shift parameter.
• b -> Second bit shift parameter.
• c -> Third bit shift parameter.

Example

const random = new XORShift64(11234, 13, 7, 17);

XORShift128

XorShift generators are fast, efficient generators with good randomness quality. This implementation has 32 bit output with 128 bits of internal state.

Parameters

• seed -> Initial seed.
• y -> First bit shift parameter.
• z -> Second bit shift parameter.
• w -> Third bit shift parameter.

Example

const random = new XORShift128(Date.now(), 362436069, 521288629, 88675123);

XORShift128Plus

XorShift generators are fast, efficient generators with good randomness quality. 64 bits of output with 128 internal state.

Parameters

• seed -> Initial seed.
• y -> Second seed.

Example

const random = new XORShift128Plus(Date.now(), 362436069);

XORWow

XorWow is an improved version of XorShift and default generator of Nvidia CUDA. 32 bits of output.

Parameters

• seed -> Initial seed.
• y -> First state initial value.
• z -> Second state initial value.
• w -> Third state initial value.
• v -> Fourth state initial value.
• d -> Fifth state initial value.
• weyl -> Additive counter.

Example

const random = new XORWow(123456789, 362436069, 521288629, 88675123, 5783321, 6615241, 362437);

bool

random.bool()

Generates a boolean with the formula random.float() >= 0.5

Parameters

None.

Returns

Boolean True/False.

Example

random.bool(); // true

coin

random.coin(pTrue)

Generates a random boolean with probability of it being true denoted by the pTrue parameter. For example, when pTrue=0.8, 80% of the numbers generated with this method will be true.

Parameters

• pTrue -> Probability of generating a true value.

Returns

Boolean True/False.

Example

random.coin(0.8); // true

int

random.int()

Generates and returns the next number in the PRNGs sequence.

Parameters

None.

Returns

Number less than 2 ** 32 for 32 bit generators.

Example

random.int(); // 85424123

bigInt

random.bigInt()

Generates and returns the next number in the PRNGs sequence and returns it as a Bigint.

Parameters

None.

Returns

Number less than 2 ** 32 for 32 bit generators represented as a BigInt class. Further reading on Big Integers

Example

random.bigInt(); // 85424123n

float

random.float()

Generates a random floating point number.

Parameters

None.

Returns

Float between 0.0 - 1.0.

Example

random.float(); // 0.234242

float53

random.float53()

Generates a random floating point number.

Parameters

None.

Returns

Float between 0.0 - 1.0.

Example

random.float53(); // 0.2342422341231

randRange

random.randRange(min, max)

Generates a number within the given range.

Parameters

• min -> Lower bound of the numbers to generate (inclusive).
• max -> Upper bound of the numbers to generate (inclusive).

Returns

Number min <= Number <= max.

Example

const lowerBound = 4;
const upperBound = 2432;
random.randRange(lowerBound, upperBound); // 36.

randBelow

random.randBelow(max)

Generates a number below the given maximum.

Parameters

• max -> Upper bound of the numbers to generate (inclusive).

Returns

Number <= max

Example

const upperBound = 2432;
random.randBelow(upperBound);  // 285.

choice

random.choice(array)

Picks a random element from the array.

Parameters

• array -> Array of any type from which we randomly select one item.

Returns

A singular item from the array of type ?.

Example

const arr = [1, 4, 2, 3];
random.choice(arr); // 4

shuffle

random.shuffle(array, inPlace = false)

Randomly shuffles the given array using the fisher-yates algorithm.

Parameters

• array -> Array of any type to be shuffled.
• inPlace -> Whether to shuffle the reference input array or return a new, shuffled array.

Returns

Array shuffled (inPlace === false), shuffled copy of array (inPlace === true).

Examples

const arr = [1, 4, 2, 3];
const shuffled = random.shuffle(arr, false);
console.log(arr); // [1, 4, 2, 3]
console.log(shuffled); // [4, 2, 3, 1]

const arr = [1, 4, 2, 3];
const shuffled = random.shuffle(arr, true);
console.log(arr); // [4, 2, 3, 1]
console.log(shuffled); // [4, 2, 3, 1]

boolArray

random.boolArray(size)

Generates an n size array populated with booleans.

Parameters

• size -> Size of the array to generate.

Returns

Array[Boolean] of length size.

Example

const size = 256;
random.boolArray(size);

coinArray

random.coinArray(size, pTrue)

Generates an n size array of random booleans with probability of it being true denoted by the pTrue parameter. For example, when pTrue=0.8, 80% of the numbers in the generated array will be true.

Parameters

• size -> Size of the array to generate.
• pTrue -> Probability of generating a true value.

Returns

Array[Boolean] of length size.

Example

const size = 256;
const pTrue = 0.8;
random.coinArray(size, pTrue);

intArray

random.intArray(size)

Generates an n size array populated with integers.

Parameters

• size -> Size of the array to generate.

Returns

Array[Number] of length size.

Example

const size = 256;
random.intArray(size);

bigIntArray

random.bigIntArray(size)

Generates an n size array populated with Big Integers.

Parameters

• size -> Size of the array to generate.

Returns

Array[BigInt] of length size.

Example

const size = 256;
random.bigIntArray(size);

rangeRangeArray

random.randRangeArray(size, min, max)

Generates an n size array populated within the given range.

Parameters

• size -> Size of the array to generate.
• min -> Lower bound of the numbers to generate (inclusive).
• max -> Upper bound of the numbers to generate (inclusive).

Returns

Array[Number] of length size filled w/ min <= num <= max.

Example

const size = 256;
const lowerBound = 4;
const upperBound = 2432;
random.randRangeArray(size, lowerBound, upperBound);

floatArray

random.floatArray(size)

Generates an n size array populated with floats.

Parameters

• size -> Size of the array to generate.

Returns

Array[Number] between 0.0 - 1.0 of length size.

Example

const size = 256;
random.floatArray(size);

float53Array

random.float53Array(size)

Generates an n size array populated with floats.

Parameters

• size -> Size of the array to generate.

Returns

Array[Number] between 0.0 - 1.0 of length size.

Example

const size = 256;
random.float53Array(size);

License

See LICENSE file.

Resources

• Mersenne Twister
• Linear Congruential Generator
• Piecewise Congruential Generator
• XorShift/XorWow
• Javascript Big Integers
• Ranged random number generation