acausal is a Typescript module that makes it easy to create, edit and generate pseudo random data from Weighted Random Distributions and Markov Chains.
Design Philosophy
- Immutable: all classes are built on top of pure functions which do not mutate state, ensuring that models retain their integrity, and making them easy to use with Redux.
- Portable: all classes are easily serializable and deserializable into data transfer objects, making them easy to store, transfer, and rebuild regardless of whether it's on the client or the server.
- Easy to Use: all APIs are written to prioritize developer usability, making it easy to rapidly prototype and implement new models.
- Minimal Dependencies: acausal only depends on random-js and scalr, (which formerly was part of acausal, but separated out for the 2.0.0 release).
Basic Examples:
import { MarkovChain, Distribution, Random } from 'acausal';
// Random Rarity Distribution
const dist = new Distribution({ seed: 1 });
dist.add('Green', 10); // Common
dist.add('Blue', 5); // Uncommon
dist.add('Purple', 1); // Rare
dist.pick(10);
/* Results in:
[
'Green', 'Green', 'Green', 'Blue', 'Green',
'Blue', 'Purple', 'Green', 'Green', 'Green'
]
*/
// Markov Chain Name Generator
const mc = new MarkovChain({ seed: 1 });
mc.addSequence('alice'.split(''));
mc.addSequence('bob'.split(''));
mc.addSequence('erwin'.split(''));
console.log(mc.generate({ order: 1 }));
/* Results in:
[ 'a', 'l', 'i', 'n' ]
*/
// Random Numbers
const rand = new Random({ seed: 1 });
rand.integer(1, 6); // Roll 1d6
// Results in: 6
Run:
npm install -s acausal
acausal is also implemented in Golang. You can find the module here:
A Random Distribution is a simple model which can simulate picks from a weighted distribution of items.
Distributions can be used to model random draws from a discrete collection of items, where each item has a different probability of appearing.
Example Use Cases:
- Simulating drawing a hand from a standard deck of cards (see below).
- Simulating the outcome of a game of roulette (or nearly any casino game).
- Generating eye or hair color for a fictional person.
- Generating the spectral class of fictional stars.
- Modeling how many McDonalds meals you'd need to buy to win Monopoly.
- Modeling any pseudo-random system through observation.
Distribution Quickstart Example - Card Deck:
import { Distribution } from 'acausal';
// Create a Deck of Cards
const suits = ['♣️', '♦️', '♥️', '♠️'];
const ranks = ['A', '2', '3', '4', '5', '6', '7', '8', '9', '10', 'J', 'Q', 'K'];
// Combine the Suits and Ranks
const cards = suits.reduce((last, suit) => {
return [...last, ...ranks.map(rank => `${rank}${suit}`)];
}, []);
/* Should result in:
[
'A♣️', '2♣️', '3♣️', '4♣️', '5♣️', '6♣️', '7♣️',
'8♣️', '9♣️', '10♣️', 'J♣️', 'Q♣️', 'K♣️', 'A♦️',
'2♦️', '3♦️', '4♦️', '5♦️', '6♦️', '7♦️', '8♦️',
'9♦️', '10♦️', 'J♦️', 'Q♦️', 'K♦️', 'A♥️', '2♥️',
'3♥️', '4♥️', '5♥️', '6♥️', '7♥️', '8♥️', '9♥️',
'10♥️', 'J♥️', 'Q♥️', 'K♥️', 'A♠️', '2♠️', '3♠️',
'4♠️', '5♠️', '6♠️', '7♠️', '8♠️', '9♠️', '10♠️',
'J♠️', 'Q♠️', 'K♠️'
]
*/
// Create weighted source data for the Distribution
const src = cards.reduce((last, card) => ({ ...last, [card]: 1 }), {});
/* Should result in:
{
'A♣️': 1,
'2♣️': 1,
'3♣️': 1,
...
'J♠️': 1,
'Q♠️': 1,
'K♠️': 1,
}
*/
// Create the Distribution from the deck.
const deck = new Distribution({
seed: 23, // Random Seed - if this is empty it will be generated.
source: src, // The weighted source to generate the normalized Distribution from.
});
// Add in 2 Jokers
deck.add('🃏', 2);
// Generate 4 picks from the deck without replacement.
const picks = deck.pick(4, undefined, true);
console.log(picks);
/* Should print:
[ 'J♣️', '10♠️', '3♦️', '9♣️' ]
*/
You can learn more about how to use Random Distributions with acausal in the Random Distribution Quickstart.
A Markov Chain is a mathematical model of a system in which the future state of the system depends only on its present state.
Markov Chains are usually generated by building a statistical model off of sample data, such as a list of names, which can then be used to output sequences which resemble the sampled data. A useful property of this process is that sample data can be "mixed" together like paint to achieve a desired result.
For example, if you wanted to generate names which sounded like a mix of Irish and Japanese, you could generate a Markov Chain from a sample of Irish and Japanese names and the resulting model would be able to output names that mixed the two.
Markov Chain Quickstart Example - Name Generator:
import { MarkovChain } from 'acausal';
// Sample Data
const jpNames = ['honoka', 'akari', 'himari', 'mei', 'ema'];
const ieNames = ['grace', 'fiadh', 'emily', 'sophie', 'ava'];
const names = [...jpNames, ...ieNames];
// Prepare Data Source - the class expects an array of arrays.
const src = names.map(name => name.split(''));
/* Should result in:
[
[ 'h', 'o', 'n', 'o', 'k', 'a' ],
[ 'a', 'k', 'a', 'r', 'i' ],
[ 'h', 'i', 'm', 'a', 'r', 'i' ],
[ 'm', 'e', 'i' ],
[ 'e', 'm', 'a' ],
[ 'g', 'r', 'a', 'c', 'e' ],
[ 'f', 'i', 'a', 'd', 'h' ],
[ 'e', 'm', 'i', 'l', 'y' ],
[ 's', 'o', 'p', 'h', 'i', 'e' ],
[ 'a', 'v', 'a' ]
]
*/
// Create the Markov Chain from the source data.
const chain = new MarkovChain({
seed: 33, // Random Seed - if this is empty it will be generated.
maxOrder: 2, // Maximum Order - Chain will generate orders up to this value.
sequences: src, // Source data, expects an array of arrays.
});
// Generate 5 picks.
for (let i = 0; i < 3; i = 1) {
const pick = chain.generate({
min: 4, // Min Picks - This will force the model to pick at least 4 times.
max: 10, // Max Picks - Stops generation after 10 picks if no end has been reached.
order: 2, // Order - The largest gram size used to calculate the next pick.
strict: false // Strict Order - Dynamically adjusts order up or down each pick if false.
});
console.log(pick.join(''));
}
/* Should print:
sophimari
emari
hie
*/
You can learn more about how to use Markov Chains with acausal in the Markov Chain Quickstart
For documentation of underlying classes and functions, please see the API documentation.