@cliqz/adblocker

Adblocker

Efficient · Minimal · JavaScript · TypeScript · uBlock Origin- and Easylist-compatible
Node.js · Puppeteer · Electron · WebExtension

---

Getting Started

Install: npm install --save @cliqz/adblocker.

Usage

There are multiple ways you can create an instance of the blocking engine to start blocking ads:

If you already have filters locally:

import { FiltersEngine } from '@cliqz/adblocker';
const engine = FiltersEngine.parse(fs.readFileSync('easylist.txt', 'utf-8'));

Fetching lists from URLs:

import { FiltersEngine } from '@cliqz/adblocker';
engine = await FiltersEngine.fromLists(fetch, [
 'https://easylist.to/easylist/easylist.txt'
]);

Use ready-made configs to block ads and optionally trackers:

import { FiltersEngine } from '@cliqz/adblocker';
engine = await FiltersEngine.fromPrebuiltAdsOnly(fetch); // ads only
engine = await FiltersEngine.fromPrebuiltAdsAndTracking(fetch); // ads and tracking

Once you have your engine, start matching requests and block ads:

import { Request } from '@cliqz/adblocker';

const { match } = engine.match(Request.fromRawDetails({
  type: 'script',
  url: 'https://domain.com/ads.js',
}));

Request Abstraction

To abstract over network requests independently from platforms (Node.js, WebExtension, etc.), the Request provides a unified APIs and helpers functions for initialization on different platforms:

import { Request } from '@cliqz/adblocker';

const request = Request.fromRawDetails({
  url: 'https://sub.example.com',
  type: 'main_frame',
});

console.log(request.isMainFrame()); // true
console.log(request.url); // https://sub.example.com
console.log(request.hostname); // sub.example.com
console.log(request.domain); // example.com

Manipulating Individual Filters

Content blockers usually manipulate two kinds of filters: network and cosmetics. The former allows to specify which network requests should be blocked (or redirected), usually from the WebRequest API of extensions. The later allows to alter the DOM of pages directly, hiding elements or injecting scripts.

Network Filters

Here is how one can parse and match individual network filters using the NetworkFilter class. It offers multiple accessors and helpers to parse, match and manipulate network filters.

import { NetworkFilter } from '@cliqz/adblocker';

// Parse filter from string
const filter = NetworkFilter.parse('||domain.com/ads.js$script');

// Filter attributes
console.log(filter.isHostnameAnchor()); // true
console.log(filter.getHostname()); // 'domain.com'
console.log(filter.getFilter()); // '/ads.js'

// Request options
console.log(filter.fromScript()); // true = can match 'script' requests
console.log(filter.fromImage()); // false = cannot match 'image' requests

Matching network filter against requests:

import { Request } from '@cliqz/adblocker';

const request = Request.fromRawDetails({
  type: 'script',
  url: 'https://sub.domain.com/ads.js?param=42',
  sourceUrl: 'https://frame-domain.com',
});

console.log(filter.match(request)); // true

Cosmetic Filters

Similarly, one can parse cosmetic filters using the CosmeticFilter class.

const { CosmeticFilter } = require('@cliqz/adblocker');

// Parsing filter from string
const filter = CosmeticFilter.parse('domain.*,domain2.com###selector');

// Properties
console.log(filter.hasHostnameConstraint()); // true
console.log(filter.getSelector()); // '#selector'
console.log(filter.isUnhide()); // false

// Matching a cosmetic filter requires both a hostname and domain
filter.match('sub.domain.com', 'domain.com'); // true

Filters Engine

Manipulating filters at a low level is useful to build tooling or debugging, but they are not appropriate for efficient blocking of requests (it would require iterating on all the filters to know if a request needs to be blocked). Instead, we can make use of the FiltersEngine class which can be seen as a "container" for both network and cosmetic filters. The filters are organized in a very compact way which also enables fast matching.

import { FiltersEngine, NetworkFilter, CosmeticFilter, Request } from '@cliqz/adblocker';

// Parse multiple filters at once
let engine = FiltersEngine.parse(`
! This is a custom list
||domain.com/ads.js$script

###selector
domain.com,entity.*##+js(script,args1)
`);

Updating an existing engine with new filters:

// Update with individual filters
engine.update({
  newNetworkFilters: [NetworkFilter.parse('/ads.js')]
  newCosmeticFilters: [CosmeticFilter.parse('###selector')],
});

Serializing an engine to Uint8Array and reloading it to its original form:

// Serialize the full engine to a Uint8Array for caching
const serialized = engine.serialize();
engine = FiltersEngine.deserialize(serialized);

Matching requests:

// Matching network filters
const {
  match, // `true` if there is a match
  redirect, // data url to redirect to if any
  exception, // instance of NetworkFilter exception if any
  filter, // instance of NetworkFilter which matched
} = engine.match(Request.fromRawDetails({
  type: 'script',
  url: 'https://sub.domain.com/ads.js',
}));

Checking for CSP injection rules for a given frame:

// Matching CSP (content security policy) filters.
const directives = engine.getCSPDirectives(Request.fromRawDetails({
  type: 'main_frame',
  url: 'https://sub.domain.com/',
}));

Checking for cosmetics injection:

// Matching cosmetic filters
const {
  styles, // stylesheet to inject in the page
  scripts, // Array of scriptlets to inject in the page
} = engine.getCosmeticsFilters({
  url: 'https://sub.domain.com/path',
  hostname: 'sub.domain.com',
  domain: 'domain.com',
});

Changelog

v1.31.1 (Thu Aug 01 2024)

:bug: Bug Fix

• @cliqz/adblocker-content
  • fix(adblocker-electron): An object could not be cloned #4157 (@seia-soto)

:nut_and_bolt: Dependencies

• Build(deps-dev): Bump @eslint/js from 9.7.0 to 9.8.0 #4151 (@dependabot[bot])
• Build(deps): Bump electron from 31.2.1 to 31.3.1 #4154 (@dependabot[bot])
• Build(deps-dev): Bump tsx from 4.16.2 to 4.16.4 #4161 (@dependabot[bot])
• Build(deps-dev): Bump @types/chai from 4.3.16 to 4.3.17 #4162 (@dependabot[bot])

Authors: 2

• @dependabot[bot]
• HoJeong Go (@seia-soto)

---