censorship-id

A robust, feature-rich, and lightweight Indonesian profanity filter. Fully redesigned for v2.0.0, it supports structured metadata detection, severity levels, categories, whitelisting, leetspeak evasion handling, and custom dictionary management.

Features

• Multi-purpose API: Censor text, detect profane words with metadata (severity/category), or perform simple boolean checks.
• Smart Mode: Handles advanced evasion techniques (leetspeak, repeated characters, punctuation-in-word).
• Categorization & Severity: Dictionary entries are classified into categories (vulgar, slur, sexual, offensive, slang) and severity levels (low, medium, high).
• Whitelisting: Easily allow specific words to bypass filtering.
• Dictionary Management: Dynamically add or remove words globally at runtime.
• Lightweight: Zero production dependencies. Built for CJS and ESM with full TypeScript support.

Installation

npm install censorship-id

Reorganized Structure

censorship-id/
├── src/
│   ├── index.js          # Core filter implementation (detect, censorId, isProfane)
│   ├── dictionary.js     # Structured Indonesian profanity database
│   └── index.d.ts        # TypeScript declarations
├── test/
│   └── index.test.js     # Jest unit tests
├── dist/                 # Build outputs (CJS, ESM, Types)
└── package.json

API Reference

1. censorId(text, options)

Replaces profane words with a mask character.

const { censorId } = require("censorship-id");

// Basic
censorId("anjing kau!");
// => "****** kau!"

// Keep first and last letter
censorId("anjing kau!", { keepFirstAndLast: true });
// => "a****g kau!"

// Custom mask character
censorId("anjing kau!", { mask: "#" });
// => "###### kau!"

Options:

• mask (string, default: *): Mask character.
• keepFirstAndLast (boolean, default: false): Keep first and last letters visible.
• allowedWords (string[]): Words to bypass filtering.
• customWords (string[] | object[]): Additional words to censor.
• smartMode (boolean, default: false): Handle obfuscation (e.g., 4njing, baaaaabiii).
• minSeverity ('low' | 'medium' | 'high'): Minimum severity level to censor.
• categories (string[]): Only censor specific categories (e.g., ['slur', 'sexual']).

2. detect(text, options)

Scans text and returns structured metadata about all matches.

const { detect } = require("censorship-id");

const results = detect("dasar anjing kau!");
console.log(results);
/*
Output:
[
  {
    match: "anjing",
    word: "anjing",
    index: 6,
    length: 6,
    severity: "high",
    category: "offensive"
  }
]
*/

// Using smartMode
detect("dasar 4njing!", { smartMode: true });
/*
Output:
[
  {
    match: "4njing",
    word: "anjing",
    index: 6,
    length: 6,
    severity: "high",
    category: "offensive"
  }
]
*/

Options: Same as censorId (minus mask and keepFirstAndLast) plus:

• unique (boolean, default: false): If true, only returns the first occurrence of each unique word.

3. isProfane(text, options)

Returns a boolean indicating if the text contains any profanity.

const { isProfane } = require("censorship-id");

isProfane("Halo apa kabar?"); // => false
isProfane("Dasar anjing!"); // => true
isProfane("Dasar anjing!", { minSeverity: "high" }); // => true

4. Dictionary Management

Modify the default database at runtime.

const { addWords, removeWords, getDictionary } = require("censorship-id");

// Add string words (assigns medium severity & custom category)
addWords(["perkosa", "bodoh"]);

// Add words with metadata
addWords([{ word: "jelek", severity: "low", category: "offensive" }]);

// Remove words
removeWords(["jelek", "bodoh"]);

// Retrieve copy of active dictionary
const activeDictionary = getDictionary();

Severity & Categories

The package classifies Indonesian words into the following categories and levels:

| Category | Level | Description | Examples | | ----------- | ------ | ---------------------------------------- | --------------------------------------- | | sexual | High | Highly vulgar/anatomical terms | kontol, memek, ngentot, pepek | | slur | High | Offensive slurs targeting identity | banci, bencong, lonte, perek | | offensive | High | Harsh expressions, insults | bangsat, bajingan, anjing, babi | | vulgar | Medium | Inappropriate bodily or vulgar slang | berak, tai, sempak | | offensive | Medium | Insulting terms for intellect | bego, goblok, tolol, dungu | | slang | Medium | Inappropriate regional slang/swear words | anjir, cuk, dancuk, jancuk | | offensive | Low | Mild insults or slang | buaya, najis, ndeso, kere | | vulgar | Low | Mild sexual references, innuendo | bokep, sange, lendir |

License

MIT © rheatkhs