Native Rust module for Adblock Plus syntax (e.g. EasyList, EasyPrivacy) filter parsing and matching.
It uses a tokenisation approach for qucikly reducing the potentially matching rule search space against a URL.
The algorithm is inspired by, and closely follows the algorithm of uBlock Origin and Cliqz.
Somewhat graphical explanation of the algorithm:
Demo use in Rust:
extern crate adblock;
use adblock::engine::Engine;
#[test]
fn check_simple_use() {
let rules = vec![
String::from("-advertisement-icon."),
String::from("-advertisement-management/"),
String::from("-advertisement."),
String::from("-advertisement/script."),
];
let blocker = Engine::from_rules(&rules);
let blocker_result = blocker.check_network_urls("http://example.com/-advertisement-icon.", "http://example.com/helloworld", "image");
assert!(blocker_result.matched);
}
Note the Node.js module has overheads inherent to boundary crossing between JS and native code.
const AdBlockClient = require('adblock-rs');
let rules = fs.readFileSync('./data/easylist.to/easylist/easylist.txt', { encoding: 'utf-8' }).split('\n');
const client = new AdBlockClient.Engine(rules);
const serializedArrayBuffer = client.serialize(); // Serialize the engine to an ArrayBuffer
console.log(`Engine size: ${(serializedArrayBuffer.byteLength / 1024 / 1024).toFixed(2)} MB`);
console.log("Matching:", client.check("http://example.com/-advertisement-icon.", "http://example.com/helloworld", "image"))
console.log("Matching:", client.check("https://github.githubassets.com/assets/frameworks-64831a3d.js", "https://github.com/AndriusA", "script"))
- Generate redirect addresses based on provided resources.txt (uBo style)
- Function for extracting CSP directives
- Generate string representation of a rule when debug mode is off (i.e. initial rule is not available)
- Cosmetic filters