ivoire-markov

Markov chains for the Ivoire random number generator framework.

• Installing
• Getting Started
• Reference
• Acknowledgements

Installing

To install, use npm:

npm install ivoire-markov

Alternately, you can find the source on Github.

Getting Started

ivoire-markov extends the ivoire package. You can require it directly:

var Ivoire = require('ivoire-markov');

Or you can require it alongside ivoire:

var Ivoire = require('ivoire');
require('ivoire-markov');

Either way, instantiate and start making chains!

var i = new Ivoire();
// Markov chains must be trained on example data.
var chain = Ivoire.train_markov_chain(['foo', 'bar', 'baz', 'boo', 'blah', 'banana']);

// We use the trained chain to create a generator from an Ivoire instance:
var generator = i.get_markov_generator(trained);

// The generator synthesizes new output from the trained chain, using our
// instance's pseudo-random sequence.
console.log(generator.generate())

Reference

ivoire-markov adds some methods to the Ivoire prototype object, making them available on all Ivoire instances. It also adds methods to the Ivoire "class" object itself, available through the Ivoire namespace.

• .train_markov_chain()—train a markov chain on a corpus of text
• #get_markov_generator()—obtain a generator from trained data
• generator.generate()—generate random strings of text

.train_markov_chain()

Syntax

Ivoire.train_markov_chain(corpus[, lookback])

Usage

Train a markov chain on a corpus of text, that is, an array of strings, providing an optional lookback distance for the chain.

var Ivoire = require('ivoire-markov');

var trained = Ivoire.train_markov_chain(['foo', 'foobar', 'bar', 'baz', 'boo', 'blah', 'boom'], 1);

The lookback controls how much context should be considered when generating text from a chain. lookback defaults to 1 if not provided.

#get_markov_generator()

Syntax

ivoire.get_markov_generator(trained)

Usage

Obtain a markov text generator from an Ivoire instance, based on the provided training data (as returned by .train_markov_chain()).

var Ivoire = require('ivoire-markov');
var trained = Ivoire.train_markov_chain(['foo', 'foobar', 'bar', 'baz', 'boo', 'blah', 'boom'], 1);
var ivoire = new Ivoire();

var generator = ivoire.get_markov_generator(trained);

generator.generate()

Usage

Using the generator object obtained from ivoire.get_markov_generator(), the #generate() method will return randomized strings that look similar to the training corpus.

var Ivoire = require('ivoire-markov');
var trained = Ivoire.train_markov_chain(['foo', 'foobar', 'bar', 'baz', 'boo', 'blah', 'boom'], 1);
var ivoire = new Ivoire();
var generator = ivoire.get_markov_generator(trained);

console.log(generator.generate());

Acknowledgements

Markov chain algorithm based on the implementation in Darmok, and modified to interoperate with the Ivoire framework.