@webalytics/metadata
Given some HTML website, extract common metadata fields. Part of the Webalytics Toolbox.
Installation
npm install --save @webalytics/metadata
Extracted Properties (if possible)
interface Metadata {
title: string
description: string
url: string
image: string
feeds: string[]
favicon: string
keywords: string[]
author: string
}Usage (convenience)
This package works out-of-the-box with any HTML document without further configuration:
import metadata from '@webalytics/metadata'
const html = '<html><head><title>abc</title></head><body></body></html>'
const data = metadata(html) // { title: 'abc' }Usage (convenience, with url aid)
When given an additional hint with the base url of the HTML document, relative urls can be resolved correctly:
import metadata from '@webalytics/metadata'
const url = 'http://example.com'
const html = '<html><head><link rel="icon" href="/fav" /></head><body /></html>'
const data = metadata(html, url) // { favicon: 'http://example.com/fav' }Usage (selective)
When the HTML document or snippet shall not be processed completely, use the underlying parser class directly and select just the fields you want:
import { Parser } from '@webalytics/metadata'
const url = 'http://example.com'
const html = '<html><head><title>abc</title></head><body></body></html>'
const document = new Parser(html, url) // same signature as default method
const title = document.selectTitle() // 'abc'Usage (fully customized)
If the pre-chosen selectors not suit you completely, you can also hook directly into the underlying cheerio DOM selector engine. It's like jQuery, but in node:
import { Parser } from '@webalytics/metadata'
const url = 'http://example.com'
const html = '<html><head><title>abc</title></head><body></body></html>'
const document = new Parser(html, url) // same signature as default method
const title = document.$('title').text() // 'abc'License
LGPL v3. You can use this code any way you want without restrictions, but I want bugfixes and improvements to flow back to this repository to benefit everyone.