Skip to content

Seedable random number generator supporting many common distributions.

License

Notifications You must be signed in to change notification settings

transitive-bullshit/random

Repository files navigation

random

Seedable random number generator supporting many common distributions.

NPM Build Status Prettier Code Formatting

Welcome to the most random module on npm! 😜

Highlights

  • Simple TS API with zero dependencies
  • Seedable
  • Plugin support for different pseudo random number generators
  • Includes many common distributions
    • uniform, normal, poisson, bernoulli, etc
  • Replacement for seedrandom which hasn't been updated in over 5 years
  • Supports all modern JS/TS runtimes

Install

npm install random

Usage

import random from 'random'

// quick uniform shortcuts
random.float((min = 0), (max = 1)) // uniform float in [ min, max )
random.int((min = 0), (max = 1)) // uniform integer in [ min, max ]
random.boolean() // true or false

// uniform distribution
random.uniform((min = 0), (max = 1)) // () => [ min, max )
random.uniformInt((min = 0), (max = 1)) // () => [ min, max ]
random.uniformBoolean() // () => [ false, true ]

// normal distribution
random.normal((mu = 0), (sigma = 1))
random.logNormal((mu = 0), (sigma = 1))

// bernoulli distribution
random.bernoulli((p = 0.5))
random.binomial((n = 1), (p = 0.5))
random.geometric((p = 0.5))

// poisson distribution
random.poisson((lambda = 1))
random.exponential((lambda = 1))

// misc distribution
random.irwinHall(n)
random.bates(n)
random.pareto(alpha)

For convenience, several common uniform samplers are exposed directly:

random.float() // 0.2149383367670885
random.int(0, 100) // 72
random.boolean() // true

// random array item
random.choice([1, true, 'foo']) // 'foo'

All distribution methods return a thunk (function with no params), which will return a series of independent, identically distributed random variables from the specified distribution.

// create a normal distribution with default params (mu=1 and sigma=0)
const normal = random.normal()
normal() // 0.4855465422678824
normal() // -0.06696771815439678
normal() // 0.7350852689834705

// create a poisson distribution with default params (lambda=1)
const poisson = random.poisson()
poisson() // 0
poisson() // 4
poisson() // 1

Note that returning a thunk here is more efficient when generating multiple samples from the same distribution.

You can change the underlying PRNG or its seed as follows:

// change the underlying pseudo random number generator seed.
// by default, Math.random is used as the underlying PRNG, but it is not seedable,
// so if a seed is given, we use an ARC4 PRNG (the same one used by `seedrandom`).
random.use('my-seed')

// create a new independent random number generator with a different seed
const rng = random.clone('my-new-seed')

// create a third independent random number generator using a custom PRNG
import seedrandom from 'seedrandom'
const rng2 = random.clone(seedrandom('kitty-seed'))

You can also instantiate a fresh instance of Random:

import { Random } from 'random'

const rng = new Random() // (uses Math.random)
const rng2 = new Random('my-seed-string')
const rng3 = new Random(() => {
  /* custom PRNG */ return Math.random()
})

API

Table of Contents

Seedable random number generator supporting many common distributions.

Defaults to Math.random as its underlying pseudorandom number generator.

Type: function (rng)

  • rng (RNG | function) Underlying pseudorandom number generator. (optional, default Math.random)

Type: function ()


  • See: RNG.clone

Creates a new Random instance, optionally specifying parameters to set a new seed.

Type: function (args, seed, opts): Random

  • args ...any
  • seed string? Optional seed for new RNG.
  • opts object? Optional config for new RNG options.

Sets the underlying pseudorandom number generator used via either an instance of seedrandom, a custom instance of RNG (for PRNG plugins), or a string specifying the PRNG to use along with an optional seed and opts to initialize the RNG.

Type: function (args)

  • args ...any

Example:

import random from 'random'

random.use('example_seedrandom_string')
// or
random.use(seedrandom('kittens'))
// or
random.use(Math.random)

Convenience wrapper around this.rng.next()

Returns a floating point number in [0, 1).

Type: function (): number


Samples a uniform random floating point number, optionally specifying lower and upper bounds.

Convence wrapper around random.uniform()

Type: function (min, max): number

  • min number Lower bound (float, inclusive) (optional, default 0)
  • max number Upper bound (float, exclusive) (optional, default 1)

Samples a uniform random integer, optionally specifying lower and upper bounds.

Convence wrapper around random.uniformInt()

Type: function (min, max): number

  • min number Lower bound (integer, inclusive) (optional, default 0)
  • max number Upper bound (integer, inclusive) (optional, default 1)

Samples a uniform random integer, optionally specifying lower and upper bounds.

Convence wrapper around random.uniformInt()

Type: function (min, max): number

  • min number Lower bound (integer, inclusive) (optional, default 0)
  • max number Upper bound (integer, inclusive) (optional, default 1)

Samples a uniform random boolean value.

Convence wrapper around random.uniformBoolean()

Type: function (): boolean


Samples a uniform random boolean value.

Convence wrapper around random.uniformBoolean()

Type: function (): boolean


Returns an item chosen uniformly at random from the given array.

Convence wrapper around random.uniformInt()

Type: function choice <T> (array: Array<T>): T | undefined

  • array Array Array of items to sample from

Generates a Continuous uniform distribution.

Type: function (min, max): function

  • min number Lower bound (float, inclusive) (optional, default 0)
  • max number Upper bound (float, exclusive) (optional, default 1)

Generates a Discrete uniform distribution.

Type: function (min, max): function

  • min number Lower bound (integer, inclusive) (optional, default 0)
  • max number Upper bound (integer, inclusive) (optional, default 1)

Generates a Discrete uniform distribution, with two possible outcomes, true or `false.

This method is analogous to flipping a coin.

Type: function (): function


Generates a Normal distribution.

Type: function (mu, sigma): function

  • mu number Mean (optional, default 0)
  • sigma number Standard deviation (optional, default 1)

Generates a Log-normal distribution.

Type: function (mu, sigma): function

  • mu number Mean of underlying normal distribution (optional, default 0)
  • sigma number Standard deviation of underlying normal distribution (optional, default 1)

Generates a Bernoulli distribution.

Type: function (p): function

  • p number Success probability of each trial. (optional, default 0.5)

Generates a Binomial distribution.

Type: function (n, p): function

  • n number Number of trials. (optional, default 1)
  • p number Success probability of each trial. (optional, default 0.5)

Generates a Geometric distribution.

Type: function (p): function

  • p number Success probability of each trial. (optional, default 0.5)

Generates a Poisson distribution.

Type: function (lambda): function

  • lambda number Mean (lambda > 0) (optional, default 1)

Generates an Exponential distribution.

Type: function (lambda): function

  • lambda number Inverse mean (lambda > 0) (optional, default 1)

Generates an Irwin Hall distribution.

Type: function (n): function

  • n number Number of uniform samples to sum (n >= 0) (optional, default 1)

Generates a Bates distribution.

Type: function (n): function

  • n number Number of uniform samples to average (n >= 1) (optional, default 1)

Generates a Pareto distribution.

Type: function (alpha): function

  • alpha number Alpha (optional, default 1)

Todo

  • Distributions

    • uniform
    • uniformInt
    • uniformBoolean
    • normal
    • logNormal
    • chiSquared
    • cauchy
    • fischerF
    • studentT
    • bernoulli
    • binomial
    • negativeBinomial
    • geometric
    • poisson
    • exponential
    • gamma
    • hyperExponential
    • weibull
    • beta
    • laplace
    • irwinHall
    • bates
    • pareto
  • Generators

    • pluggable prng
    • port more prng from boost / seedrandom

Related

  • d3-random - D3's excellent random number generation library.
  • seedrandom - Seedable pseudo random number generator.
  • random-int - For the common use case of generating uniform random ints.
  • random-float - For the common use case of generating uniform random floats.
  • randombytes - Random crypto bytes for Node.js and the browser.
  • jshash prngs

Credit

Thanks go to Andrew Moss for the TypeScript port and for helping to maintain this package.

Shoutout to Roger Combs for donating the random npm package for this project!

Lots of inspiration from d3-random (@mbostock and @svanschooten).

Some distributions and PRNGs are ported from C++ boost::random.

License

MIT © Travis Fischer

Support my OSS work by following me on twitter twitter