8 releases (stable)

1.2.2 Nov 29, 2023
1.2.1 Nov 8, 2023
1.1.0 Apr 14, 2023
1.0.1 Mar 10, 2023
0.1.0 Dec 3, 2022

#422 in Web programming

Download history 4/week @ 2024-02-26 3/week @ 2024-03-11 36/week @ 2024-04-01

65 downloads per month

MIT license

25KB
397 lines

hq

Crates.io

jq, but for HTML. Try it in your browser here

hq reads HTML and converts it into a JSON object based on a series of CSS selectors. The selectors are expressed in a similar way to JSON, but where the values are CSS selectors. For example:

{posts: .athing | [ {title: .titleline > a, url: .titleline > a | @(href)} ] }

This will select all .athing elements, and it will create an array (| [{...}]) of objects for each element selected. Then for each element it will select the text of the titleline > a element, and the href attribute (| @(href)).

The end result is the following structure:

{
  "posts": [
    {
      "title": "...",
      "url": "..."
    }
  ]
}

Install

brew install hq, or cargo install html-query

Special query syntax

Text

.foo | @text

This will select the text content from the first element matching .foo.

Selecting attributes

.foo | @(href)

This will select the href attribute from the first element matching .foo.

Parents

.foo | @parent

This will return the parent element from the first element matching .foo.

Siblings

.foo | @sibling(1)

This will return the sibling element from the first element matching .foo.

Examples

Full hacker news story extraction

{posts: .athing | [{href: .titleline > a | @(href), title: .titleline > a, meta: @sibling(1) | {user: .hnuser, posted: .age | @(title) }}]}

This selects each .athing element, extracts the URL from the href attribute as well as the title. It then selects the sibling .athing element, and extracts the user and post time from that:

{
  "posts": [
    {
      "title": "...",
      "url": "...",
      "meta": {
        "posted": "...",
        "user": "..."
      }
    }
  ]
}

Dependencies

~8–14MB
~179K SLoC