Merge pull request #738 from microlinkhq/happydom

perf(readability): use happy-dom

Author: Kikobeats

packages/metascraper-readability/README.md

@@ -14,6 +14,19 @@
 $ npm install metascraper-readability --save
 ```
 
+## API
+
+### metascraper-readability([options])
+
+#### options
+
+##### getDocument
+
+Type: `function`<br>
+Default: [source code](https://github.com/microlinkhq/metascraper/blob/master/packages/metascraper-readability/src/index.js#L14-L20)
+
+The function to be called to serialized html into a DOM document.
+
 ## License
 
 **metascraper-readability** © [Microlink](https://microlink.io), released under the [MIT](https://github.com/microlinkhq/metascraper/blob/master/LICENSE.md) License.<br>

packages/metascraper-readability/benchmark/fixture.html

@@ -0,0 +1,42 @@
+'use strict'
+
+const { readFileSync } = require('fs')
+
+const url = 'https://arxiv.org/pdf/2412.06592'
+const html = readFileSync('./fixture.html', 'utf8')
+
+const jsdom = () => {
+  const { JSDOM, VirtualConsole } = require('jsdom')
+  const dom = new JSDOM(html, { url, virtualConsole: new VirtualConsole() })
+  return dom.window.document
+}
+
+const happydom = () => {
+  const { Window } = require('happy-dom')
+  const window = new Window({ url })
+  const document = window.document
+  document.documentElement.innerHTML = html
+  return document
+}
+
+const { Readability } = require('@mozilla/readability')
+
+const measure = fn => {
+  const now = Date.now()
+  const parsed = new Readability(fn()).parse()
+  return { parsed, duration: Date.now() - now }
+}
+
+const jsdomResult = measure(jsdom)
+const happydomResult = measure(happydom)
+
+const isEqual = (value1, value2) =>
+  JSON.stringify(value1) === JSON.stringify(value2)
+
+if (!isEqual(jsdomResult.parsed, happydomResult.parsed)) {
+  console.error('Results are different')
+  process.exit(1)
+}
+
+console.log(`   jsdom: ${jsdomResult.duration}ms`)
+console.log(`happydom: ${happydomResult.duration}ms`)

packages/metascraper-readability/benchmark/index.js

@@ -0,0 +1,9 @@
+{
+  "name": "@metascraper-readability/benchmark",
+  "private": true,
+  "version": "1.0.0",
+  "devDependencies": {
+    "dom-parser": "latest",
+    "happy-dom": "latest"
+  }
+}

packages/metascraper-readability/benchmark/package.json

@@ -25,7 +25,7 @@
   "dependencies": {
     "@metascraper/helpers": "workspace:*",
     "@mozilla/readability": "~0.5.0",
-    "jsdom": "~25.0.1"
+    "happy-dom": "~16.5.3"
   },
   "devDependencies": {
     "ava": "5",

packages/metascraper-readability/package.json

@@ -1,9 +1,7 @@
 'use strict'
 
 const { memoizeOne, composeRule } = require('@metascraper/helpers')
-
 const { Readability } = require('@mozilla/readability')
-const { JSDOM, VirtualConsole } = require('jsdom')
 
 const parseReader = reader => {
   try {
@@ -13,15 +11,25 @@ const parseReader = reader => {
   }
 }
 
-const readability = memoizeOne((url, html) => {
-  const dom = new JSDOM(html, { url, virtualConsole: new VirtualConsole() })
-  const reader = new Readability(dom.window.document)
-  return parseReader(reader)
-}, memoizeOne.EqualityFirstArgument)
+const defaultGetDocument = ({ url, html }) => {
+  const { Window } = require('happy-dom')
+  const window = new Window({ url })
+  const document = window.document
+  document.documentElement.innerHTML = html
+  return document
+}
+
+module.exports = ({ getDocument = defaultGetDocument } = {}) => {
+  const readability = memoizeOne((url, html, getDocument) => {
+    const document = getDocument({ url, html })
+    const reader = new Readability(document)
+    return parseReader(reader)
+  }, memoizeOne.EqualityFirstArgument)
 
-const getReadbility = composeRule(($, url) => readability(url, $.html()))
+  const getReadbility = composeRule(($, url) =>
+    readability(url, $.html(), getDocument)
+  )
 
-module.exports = () => {
   return {
     author: getReadbility({ from: 'byline', to: 'author' }),
     description: getReadbility({ from: 'excerpt', to: 'description' }),

packages/metascraper-readability/src/index.js

@@ -46,8 +46,8 @@ Generated by [AVA](https://avajs.dev).
 
     {
       author: null,
-      description: null,
+      description: 'Virtual Tour of 219 Shale Rd.',
       lang: null,
       publisher: null,
-      title: null,
+      title: '219 Shale Rd - Virtual Tour',
     }