htmlgrep is a suite of command-line tools for searching HTML documents, that allows selecting elements by various types of selectors.

The suite consists of the following programs:

The tools are built with the HTML tree manipulation library Kuchiki (朽木), which uses the same HTML parser as the Servo browser engine.

Installation

Using the cargo package manager:

% cargo install htmlgrep

Usage

As a CLI tool, given the followig HTML document, blog.html_:

<!doctype html>
<meta charset=utf>
<title>My first blog post</title>
<meta name=keywords content=blog,first,hello>
<meta name=description content="First entry to blog.">

To find all occurrences of <meta> elements:

% css blog.html meta
first.html  <meta content="blog,first,hello" name="keywords">
first.html  <meta content="First entry to blog." name="description">

And to only look for <meta> elements with a name attribute equal to keywords and a content attribute containing blog in a space-separated list:

% css meta[name=keywords][content~=blog] blog.html
first.html  <meta content="blog,first,hello" name="keywords">

It can also receive streaming content from stdin:

% curl -L https://sny.no/ | css title
/dev/stdin  <title>Andreas Tolfsen</title>

Library

Programmatically, with the htmlgrep crate:

extern crate htmlgrep;

fn main() {
    let input = r#"
        <!doctype html>
        <meta charset=utf>
        <title>My first blog post</title>
        <meta name=keywords content=blog,first,hello>
        <meta name=description content="First entry to blog.">
    "#;

    let matches = htmlgrep::select("meta[name=keywords]", input.as_bytes()).unwrap();

    for node in matches {
        println!("{}", node.source);
    }
}