Skip to content

Latest commit

 

History

History
318 lines (219 loc) · 15.2 KB

README.md

File metadata and controls

318 lines (219 loc) · 15.2 KB


metascraper

Last version Coverage Status NPM Status

A library to easily get unified metadata from websites using Open Graph, Microdata, RDFa, Twitter Cards, JSON-LD, HTML, and more.

What is it

The metascraper library allows you to easily scrape metadata from an article on the web using Open Graph metadata, regular HTML metadata, and series of fallbacks.

It follows a few principles:

  • Have a high accuracy for online articles by default.
  • Make it simple to add new rules or override existing ones.
  • Don't restrict rules to CSS selectors or text accessors.

Getting started

Let's extract accurate information from the following website:

First, metrascraper expects you provide the HTML markup behind the target URL.

There are multiple ways to get the HTML markup. In our case, we are going to run a programmatic headless browser to simulate real user navigation, so the data obtained will be close to a real-world example.

const getHTML = require('html-get')

/**
 * `browserless` will be passed to `html-get`
 * as driver for getting the rendered HTML.
 */
const browserless = require('browserless')()

const getContent = async url => {
  // create a browser context inside the main Chromium process
  const browserContext = browserless.createContext()
  const promise = getHTML(url, { getBrowserless: () => browserContext })
  // close browser resources before return the result
  promise.then(() => browserContext).then(browser => browser.destroyContext())
  return promise
}

/**
 * `metascraper` is a collection of tiny packages,
 * so you can just use what you actually need.
 */
const metascraper = require('metascraper')([
  require('metascraper-author')(),
  require('metascraper-date')(),
  require('metascraper-description')(),
  require('metascraper-image')(),
  require('metascraper-logo')(),
  require('metascraper-clearbit')(),
  require('metascraper-publisher')(),
  require('metascraper-title')(),
  require('metascraper-url')()
])

/**
 * The main logic
 */
getContent('https://microlink.io')
  .then(metascraper)
  .then(metadata => console.log(metadata))
  .then(browserless.close)
  .then(process.exit)

The output will be something like:

{
  "author": "Microlink HQ",
  "date": "2022-07-10T22:53:04.856Z",
  "description": "Enter a URL, receive information. Normalize metadata. Get HTML markup. Take a screenshot. Identify tech stack. Generate a PDF. Automate web scraping. Run Lighthouse",
  "image": "https://cdn.microlink.io/logo/banner.jpeg",
  "logo": "https://cdn.microlink.io/logo/trim.png",
  "publisher": "Microlink",
  "title": "Turns websites into data — Microlink",
  "url": "https://microlink.io/"
}

What data it detects

Note: Custom metadata detection can be defined using a rule bundle.

Here is an example of the metadata that metascraper can detect:

  • audio — e.g. https://cf-media.sndcdn.com/U78RIfDPV6ok.128.mp3
    A audio URL that best represents the article.

  • author — e.g. Noah Kulwin
    A human-readable representation of the author's name.

  • date — e.g. 2016-05-27T00:00:00.000Z
    An ISO 8601 representation of the date the article was published.

  • description — e.g. Venture capitalists are raising money at the fastest rate...
    The publisher's chosen description of the article.

  • video — e.g. https://assets.entrepreneur.com/content/preview.mp4
    A video URL that best represents the article.

  • image — e.g. https://assets.entrepreneur.com/content/3x2/1300/20160504155601-GettyImages-174457162.jpeg
    An image URL that best represents the article.

  • lang — e.g. en
    An ISO 639-1 representation of the url content language.

  • logo — e.g. https://entrepreneur.com/favicon180x180.png
    An image URL that best represents the publisher brand.

  • publisher — e.g. Fast Company
    A human-readable representation of the publisher's name.

  • title — e.g. Meet Wall Street's New A.I. Sheriffs
    The publisher's chosen title of the article.

  • url — e.g. http://motherboard.vice.com/read/google-wins-trial-against-oracle-saves-9-billion
    The URL of the article.

How it works

metascraper is built out of rules bundles.

It was designed to be easy to adapt. You can compose your own transformation pipeline using existing rules or write your own.

Rules bundles are a collection of HTML selectors around a determinate property. When you load the library, implicitly it is loading core rules.

Each set of rules load a set of selectors in order to get a determinate value.

These rules are sorted with priority: The first rule that resolve the value successfully, stop the rest of rules for get the property. Rules are sorted intentionally from specific to more generic.

Rules work as fallback between them:

  • If the first rule fails, then it fallback in the second rule.
  • If the second rule fails, time to third rule.
  • etc

metascraper do that until finish all the rule or find the first rule that resolves the value.

Importing rules

metascraper exports a constructor that need to be initialized providing a collection of rules to load:

const metascraper = require('metascraper')([
  require('metascraper-author')(),
  require('metascraper-date')(),
  require('metascraper-description')(),
  require('metascraper-image')(),
  require('metascraper-logo')(),
  require('metascraper-clearbit')(),
  require('metascraper-publisher')(),
  require('metascraper-title')(),
  require('metascraper-url')()
])

Again, the order of rules are loaded are important: Just the first rule that resolve the value will be applied.

Use the first parameter to pass custom options specific per each rules bundle:

const metascraper = require('metascraper')([
  require('metascraper-clearbit')({
    size: 256,
    format: 'jpg'
  })
])

Rules bundles

?> Can't find the rules bundle that you want? Let's open an issue to create it.

Official

Rules bundles maintained by metascraper maintainers.

Core essential

Vendor specific

Community

Rules bundles maintained by individual users.

See CONTRIBUTING for adding your own module!

API

constructor(rules)

Create a new metascraper instance declaring the rules bundle to be used explicitly.

rules

Type: Array

The collection of rules bundle to be loaded.

metascraper(options)

Call the instance for extracting content based on rules bundle provided at the constructor.

options

url

Required
Type: String

The URL associated with the HTML markup.

It is used for resolve relative links that can be present in the HTML markup.

it can be used as fallback field for different rules as well.

html

Type: String

The HTML markup for extracting the content.

htmlDom

Type: object

The DOM representation of the HTML markup. When it's not provided, it's get from the html parameter.

rules

Type: Array

You can pass additional rules to add on execution time.

These rules will be merged with your loaded rules at the beginning.

validateUrl

Type: boolean
Default: true

Ensure the URL provided is validated as a WHATWG URL API compliant.

Environment Variables

METASCRAPER_RE2

Type: boolean
Default: true

It attemptt to load re2 to use instead of RegExp.

Benchmark

To give you an idea of how accurate metascraper is, here is a comparison of similar libraries:

Library metascraper html-metadata node-metainspector open-graph-scraper unfluff
Correct 95.54% 74.56% 61.16% 66.52% 70.90%
Incorrect 1.79% 1.79% 0.89% 6.70% 10.27%
Missed 2.68% 23.67% 37.95% 26.34% 8.95%

A big part of the reason for metascraper's higher accuracy is that it relies on a series of fallbacks for each piece of metadata, instead of just looking for the most commonly-used, spec-compliant pieces of metadata, like Open Graph.

metascraper's default settings are targetted specifically at parsing online articles, which is why it's able to be more highly-tuned than the other libraries for that purpose.

If you're interested in the breakdown by individual pieces of metadata, check out the full comparison summary, or dive into the raw result data for each library.

License

metascraper © Microlink, released under the MIT License.
Authored and maintained by Microlink with help from contributors.

microlink.io · GitHub microlinkhq · Twitter @microlinkhq