TreeGrammar

Simple context-free grammar text generator.

This module creates random texts based on a set of rules.

const Tg = require('treegrammar');

const rules = {
    '<start>': ['<template>'],
    '<template>': [
        'the <noun> is <adj>', 'A <adj> <noun>'
    ],
    '<noun>': [
        'house', 'cat'
    ],
    '<adj>': [
        'green', 'big'
    ]
};

const generator = new Tg(rules);

generator.generate(); //the house is big
generator.generate(); //A big cat
generator.generate(); //A green house
generator.generate(); //the cat is green

Install

npm install treegrammar

Usage

First, you need to create a grammar based on a set of rules.

The rule name may be any name surrounded by angle brackets, letters, numbers an underscore are allowed. The rule content must be an Array of strings, those strings should be any combination of text and/or references to other rules. See example above.

Then create a new generator object with that set of rules or you can add the rules after using the API. The generator object can be used to .generate() random strings and it also has some other utility methods documented below.

API

Method | Description | ---------|---------------| generate(startRule) | Generates a new random string. By default the text will be generated starting from the rule with the name <start> but you can also define your starting point using the optional parameter startRule. addRule(ruleName, ruleContent, overwrite) | Adds a new rule. The <ruleName> and [ruleContent] are required. If the rule exists it will throw an Error unless the optional overwrite param is set to true. deleteRule(ruleName) | Deletes the rule <ruleName>.

Building

• git clone https://github.com/bertez/treegrammar
• npm install
• gulp

This will lint the code and run the tests.

License

MIT.

(c) 2016, Berto Yáñez.

The rules syntax is based on the awesome RitaJS module