Simple Text Parser

This is a very simple text parser written in TypeScript. It's based around strings and regular expressions so it's highly customizable, synchronous, and relatively fast.

Install

Install via NPM/Yarn and use it in your package of choice. This package is compatible with the browser.

npm i simple-text-parser

yarn add simple-text-parser

Usage

The simple-text-parser package exports a Parser class. Create a new instance from it.

import { Parser } from "simple-text-parser";

const parser = new Parser();

Example

This library works by taking a plain text string and searching it for substrings and regular expressions. When a match is found, it is parsed out into a tree and replaced.

Let's start by defining a parsing rule. Say we want to parse some text for hash tags (#iamahashtag) and replace it with some custom html:

// Define a rule using a regular expression
parser.addRule(/\#[\S]+/gi, function (tag) {
  // Return the tag minus the `#` and surrond with html tags
  return `<span class="tag">${tag.substr(1)}</span>`;
});

Now let's render some text using our rule and output the resulting string:

parser.render("Some text #iamahashtag foo bar.");

becomes...

Some text <span class="tag">iamahashtag</span> foo bar.

Of course we can also parse some text into an array of nodes for more custom handling and to retrieve the parsed data:

parser.toTree("Some text #iamahashtag foo bar.");

outputs...

[
  { type: "text", text: "Some text " },
  { type: "text", text: '<span class="tag">iamahashtag</span>' },
  { type: "text", text: " foo bar." },
];

Of course a type of text on a tag isn't helpful when specifically trying to parse out tags. Let's modify our parsing rule to be more specific:

// Define a rule using a regular expression
// RegExp capture groups are passed as extra arguments
parser.addRule(/#([\S]+)/gi, function (tag, clean_tag) {
  // create the replacement text with surrounding html tags
  const html = `<span class="tag">${clean_tag}</span>`;

  // return a node describing this tag
  return { type: "tag", text: html, value: clean_tag };
});

Now lets rerun render() and toTree() on the original text. Notice that render() outputs the same thing as before, but toTree() includes the custom meta data.

Some text <span class="tag">iamahashtag</span> foo bar.

[
  { type: "text", text: "Some text " },
  {
    type: "tag",
    text: '<span class="tag">iamahashtag</span>',
    value: "iamahashtag",
  },
  { type: "text", text: " foo bar." },
];

Now the rule we've been using is actually already included as a preset. Presets are easy to use, they include the match side, you need to set a replace value.

// Define a rule using a preset
parser.addPreset("tag", function (tag, clean_tag) {
  const html = `<span class="tag">${clean_tag}</span>`;
  return { type: "tag", text: html, value: clean_tag };
});

There are actually 3 included presets: tag, url, and email. You can also add your own presets to extend the parser globally by using Parser.registerPreset().

API Documentation

Instance Methods

These methods can be called on objects returned from new Parser().

parser.addRule()

Add a rule to this parser. A rule consists of a match and optionally a replace and type.

addRule(match: Match, replace?: Replace, type?: string): this
addRule(rule: Rule): this

• match - The search to perform. If a string, it is searched for exactly. If a regular expression, a simple match is performed and any capture groups are passed to replace. If a function, it is called with a single argument, the full string passed to render(), and should return an array with an index and length of the match.
• replace - Replaces the match when found. If a string, it replaces exactly. Functions are called with matched substrings and possibly any regular expression capture groups. The function should return a string to replace with or an object representing a tree node. This argument is optional and when not provided the matched content is preserved.
• type - The type of the rule, which will also be the default type used in parsed tree nodes.
• rule - The above arguments as an object.

parser.addPreset()

Add a registered global preset rule within this parser and give it a replace. The preset must first be registered using Parser.registerPreset() before it can be used with this method.

addPreset(type: string, replace?: Replace): this

• type - The string id of the preset as declared by Parser.registerPreset(). This will be the node's type when returned by toTree().
• replace - Replaces the match when found. Same as the replace in addRule().

parser.toTree()

Returns the parsed string as an array of nodes. Every node includes at least type and text properties. type defaults to "text" but could be any value as returned by replace. The text key is used to replaced the matched string by render().

toTree(str: string): Node[]

• str - A plain text string to parse.

parser.render()

Returns a parsed string with all matches replaced.

render(str: string): string

• str - A plain text string to parse and replace.

Class Methods

These methods can be called from the Parser class.

Parser.registerPreset()

Register a new global preset rule. Presets don't handle the replacing, only the matching. There are three pre-included presets: tag, url, and email.

static registerPreset(type: string, match: Match): void

• name - The string id of the preset. This will become the node's type when returned by toTree().
• match - The search to perform. Same as the replace in addRule().

Parser.renderTree()

Rasterize an array of nodes into a string by concatenating all their text properties. Used internally by render().

static renderTree(tree: Node[]): string

• tree - Array of node objects, usually what is returned by toTree().