A Javascript library for parsing metadata in web pages.
npm install --save page-metadata-parser
This library is meant to be used either in the browser (embedded directly in a website or into a browser addon/extension) or on a server (node.js).
Each function expects to be passed a Document object, which may be created either directly by a browser or on the server using a Document compatible object, such as that provided by jsdom.
You may use each rule individually, or together to parse all of the metadata provided by this library.
This library employs parsers for the following formats:
This library is based on the work of Mozilla Fathom, a framework for using rules to parse content on web pages.
A single rule instructs the parser on a possible DOM node to locate a specific piece of content. For instance, a rule to parse the title of a page might look like
['meta[property="og:title"]', node => node.element.content]
A rule consists of two parts, a querySelector compatible string which is used to look up the target content, and a callable which receives a Node (a wrapper around a DOM element) and returns the desired content from that Node.
This rule would be able to successfully find a page's title from the following HTML sample:
<meta property="og:title" content="A Sample Page" />
This library includes many rules for a single desired piece of metadata which should allow it to consistently find metadata across many types of pages. This library is meant to be a community driven effort, and so if there is no rule to find a piece of information from a particular website, contributors are encouraged to add new rules!
This library provides rules to find the following forms of metadata in a page:
| Field | Description |
|---|---|
| type | The type of content as defined by opengraph. |
| url | A canonical URL for the page. |
| title | A user displayable title for the page. |
| description | A user displayable description for the page. |
| icon_url | A URL which contains an icon for the page. |
| image_url | A URL which contains a preview image for the page. |
To use a single rule to find a particular piece of metadata within a page, simply pass that rule a Document object and it will apply each possible selector for that rule until it finds a matching piece of information and return it.
Example:
const {metadataRules} = require('page-metadata-parser');
const pageTitle = metadataRules.title(document);
To parse all of the available metadata on a page using all of the rules provided in this library, simply call getMetadata on the Document.
const {getMetadata} = require('page-metadata-parser');
const pageMetadata = getMetadata(document);
const pageTitle = pageMetadata.title;