node-markovify

A Markov Chain module implemented in Node.js. Included functionality supports plain text files as well as arrays of Tweets. For more information on Markov Chains, please visit http://setosa.io/ev/markov-chains/ for an excellent visual explanation.

Functionality is in it's earliest stages and will be upgraded frequently in the coming days.

Installation

node-markovify is available on NPM:

npm install --save node-markovify

Usage

Markov Chain Only

var markovChain = require('node-markovify').markovChain;
var thisMarkovChain = new markovChain(options);

Markov Text Only

var markovText = require('node-markovify').markovText;
var thisMarkovText = new markovText(options);

Markov Twitter Only

var markovTwitter = require('node-markovify').markovTwitter;
var thisMarkovTwitter = new markovTwitter(options);

Options Examples

Markov Text Options

var options = {
    state_size : 2
    text: 'Hello World',
    corpus : [
        'Sentence 1',
        'Sentence 2'
    ], // Optional, corpus will be built from options.text if no corpus is passed
    DEFAULT_MAX_OVERLAP_RATIO : .6, // Optional, default .6
    DEFAULT_MAX_OVERLAP_TOTAL : 15, // Optional, default 15
    DEFAULT_TRIES : 100, // Optional, default 100
    BEGIN : '___BEGIN___', // Optional, default ___BEGIN___
    END : '___END___', // Optional, dafault ___END___
}

Markov Twitter Options

var options = {
    tweets : [
        'Tweet 1',
        'Tweet 2',
        '...'
    ],
    state_size : 2,
    numTweetsToPredict : 10
}

Examples

Markov Chain using input from a text file

Markov Chain using input from Twitter