HTML parser for any Javascript runtime environment. Small, Fast, Easy to use, and highly customizable
Most HTML parsers will force you to learn their Javascript API after the parse result. They won't allow you to tap into the processing to access the nodes as they are parsed or let you create your own API for the final result that adapts to your project instead of the other way around.
This parser
- Is one of the fastest HTML parser out there averaging 1ms per HTML page of different sizes. Check benchmark.
- Uses a DOM like API which is a custom Lite DOM built for performance
- Can use browser DOM API or JsDOM to give you the parsed HTML allowing it to be used in any js runtime environment
- You can use your own custom DOM API like to gain absolute control
- Accepts a callback, so you can access the nodes as they are being parsed
- Super simple to use. No need for extensive options list. Parses everything in a performant way
- Handles SVG and HTML easily including comments and script tags with HTML inside
npm install @beforesemicolon/html-parser
<!DOCTYPE html>
<html lang="en">
<head>
<!-- Grab the latest version -->
<script src="https://unpkg.com/@beforesemicolon/html-parser/dist/parser-client.min.js"></script>
<!-- Or a specific version -->
<script src="https://unpkg.com/@beforesemicolon/html-parser@1.0.0/dist/parser-client.min.js"></script>
</head>
<body></body>
</html>- Only works with HTML and SVG tags. Duh!
- Handles custom tags, style, script tags and comments by default without differences in the performance
-
<!Doctype>tag is ignored - Honor the format by keeping all white spaces which are returned as text nodes
By default, it will return a document fragment as root. The API is DOM-like, meaning, if you know the DOM API you already know this. The DOM-like API is minimal and built for performance allowing you to easily use the same code in the browser, Node, Deno or any other javascript runtime environment.
See custom handler section to understand what this Document-like API looks like.
import {parse} from "@beforesemicolon/html-parser";
const frag = parse('<h1>site title</h1>'); // return DocumentFragment-like object
frag.children[0] // h1 ElementThis parser works with the DOM API by default so if you want to use it in Node, Deno or any Javascript runtime environment, make sure to import jsDom or similar and provide the Document object.
import * as jsdom from "jsdom";
const {JSDOM} = jsdom;
const document = new JSDOM('').window.document;
// import the parser
import {parse} from "@beforesemicolon/html-parser";
const frag = parse('<h1>site title</h1>', document); // return DocumentFragment
frag.children[0] // h1 Element<script>
const {parse, Doc} = window.BFS;
// uses a like Document-like object by default
const frag1 = parse('<h1>site title</h1>'); // returns DocumentFragment-like
// use the native DOM Document object
const frag2 = parse('<h1>site title</h1>', document); // returns DocumentFragment object
frag1.children[0] // h1 Element
frag2.children[0] // h1 Element
</script>You may also pass a callback function as second parameter which will get called as the nodes are being parsed and created. This will use the document as default so the callback will be get called with DOM Nodes and Element.
const frag = parse('<h1>site title</h1>', (node) => {
// handle node here
});The parser itself if fast but depending on the API you use for the final parsed result the performance will varies on their algorithm. Here are two examples using htmlparser-benchmark.
import {parse} from "@beforesemicolon/html-parser";
parse(aReallyMassimeHTMLString);
// avg duration: 1.86113 ms/file ± 1.09698This is up to 30 times faster than the DOM Document API
This is using the custom jsDom in NodeJs:
import * as jsdom from "jsdom";
import {parse} from "@beforesemicolon/html-parser";
const {JSDOM} = jsdom;
const document = new JSDOM('').window.document;
parse(aReallyMassimeHTMLString, document);
// avg duration: 27.3563 ms/file ± 19.1060`The best thing about this parser is the ability to crate your own handler to transform HTML into anything you like.
Here is an example of a simple implementation you can start from.
const MyCustomDoc = {
createComment: (value: string) => ({type: 'comment', value}),
createTextNode: (value: string) => ({type: 'text', value}),
createDocumentFragment: () => {
const children: unknown[] = []
return {
type: 'fragment',
children,
appendChild: (node: unknown) => {
children.push(node)
}
}
},
createElementNS: (namespace: string, tagName: string) => {
const children: unknown[] = []
const attributes: Record<string, unknown> = {}
return {
namespace,
tagName,
children,
attributes,
type: 'node',
appendChild(node: unknown) {
children.push(node)
},
setAttribute(name: string, value: string) {
attributes[name] = value;
}
}
},
}
const result = parse<typeof MyCustomDoc>(`...`, MyCustomDoc);