Skip to content

🏎 The fastest HTML tag and attributes parser for JavaScript

License

Notifications You must be signed in to change notification settings

AndreasPizsa/hypertag

Repository files navigation

</​hypertag> npm-version-badge npm-license-badge

The fastest HTML tag and attributes parser.

hypertag is an HTML tag parser built for speed. Use it to find specific HTML tags and their attributes in HTML documents. It’s like a superfast getElementsByTagName without the DOM.

✨ Features

  • ✅  Hyperfast. 50 × faster than cheerio, 30 × parse5, 10 × htmlparser2.
  • ✅  Tiny. < 500 bytes gzipped.
  • ✅  Complete Zero dependencies.
  • ✅  Robust. 100% Code Coverage. coveralls-badge travis-build-badge

💻 Use

const html = `
  <html><head>
    <meta name="hello" content="world">
    <meta name="hello" content="moon">
  </head><body>
    <div><h1>Hello, world!</h1></div>
  </body></html>
`

const result = parseHtmlTags(html, 'meta')
console.log(result)

[
  {
    '<' : 'meta',
    name: 'hello',
    content: 'world'
  },
  {
    '<' : 'meta',
    name: 'hello',
    content: 'moon'
  }
]

Examples

Getting Favicons

const result = parseHtmlTags(html, 'link')
  .filter(({rel}) => /^(shortcut\s+)?icon/i.test(rel))

[
  {
    '<': 'link',
    rel: 'icon',
    href: 'favicon.png',
    sizes: '16x16'
    type: 'image/png'
  }
]

Getting OpenGraph Images

const result = parseHtmlTags(html, 'meta')
  .filter(({property}) => property.toLowerCase() === 'og:image')

[
  {
    '<': 'meta',
    property: 'og:image',
    content: 'http://static01.nyt.com/images/2015/02/19/arts/international/19iht-btnumbers19A/19iht-btnumbers19A-facebookJumbo-v2.jpg'
  }
]

Benchmarks 🍏🍊

Run benchmarks with

$ ./benchmark.js

Benchmark Design

The tested packages all do different things and have their strengths in different areas, so the benchmark by design compares apples to oranges.

The question this benchmark aims to answer is

How fast can I find tags of interest in an HTML string?

Most of the tested parsers come with many more features and allow you to do more complex queries than hypertag; for example, parse5 and cheerio create a whole DOM, and similarly html-parse-stringify2 creates an AST. html-tag-parser parses tags but not attributes.

One objection could be that this is an unfair test, since the parsers are just too different. This can be rebutted by the fact that one ought to pick the right tool for the job: a sports car is faster than a truck, but the truck can load more freight. Do you need a fast and simple parser to find a few tags or do you want to manipulate a DOM?

For this benchmark, we load a pretty "standard" web page (specifically, apple.com) and the let each of the parsers parse the HTML.

Results

hypertag x 10,248 ops/sec ±0.78% (88 runs sampled)
fast-html x 980 ops/sec ±1.36% (87 runs sampled)
parse5 x 323 ops/sec ±1.68% (83 runs sampled)
htmlparser2 x 1,079 ops/sec ±0.87% (88 runs sampled)
html-tag-parser x 1,482 ops/sec ±0.71% (91 runs sampled)
cheerio x 182 ops/sec ±5.20% (70 runs sampled)
html-parse-stringify2 x 499 ops/sec ±1.07% (87 runs sampled)
Fastest is hypertag

About

🏎 The fastest HTML tag and attributes parser for JavaScript

Topics

Resources

License

Stars

Watchers

Forks

Packages

No packages published