decancer

A library that removes common unicode confusables/homoglyphs from strings.

• Its core is written in Rust and utilizes a form of Binary Search to ensure speed!
• By default, it's capable of filtering 221,522 (19.88%) different unicode codepoints like:
  • All whitespace characters
  • All diacritics, this also eliminates all forms of Zalgo text
  • Most leetspeak characters
  • Most homoglyphs
  • Several emojis
• Unlike other packages, this package is unicode bidi-aware where it also interprets right-to-left characters in the same way as it were to be rendered by an application!
• Its behavior is also highly customizable to your liking!
• And it's available in the following languages:
  • Rust
  • JavaScript (Node.js/Browser)
  • C/C++
  • Java
  • Python (unofficial)

Installation

In your Cargo.toml:

decancer = "3.2.0"

In your shell:

$ npm install decancer

In your code (CommonJS):

const decancer = require('decancer')

In your code (ESM):

import decancer from 'decancer'

In your code:

<script type="module">
  import init from 'https://cdn.jsdelivr.net/gh/null8626/[email protected]/bindings/wasm/bin/decancer.min.js'

  const decancer = await init()
</script>

As a dependency

In your build.gradle:

repositories {
  mavenCentral()
  maven { url 'https://jitpack.io' }
}

dependencies {
  implementation 'com.github.null8626:decancer:v3.2.0'
}

In your pom.xml:

<repositories>
  <repository>
    <id>central</id>
    <url>https://repo.maven.apache.org/maven2</url>
  </repository>
  <repository>
    <id>jitpack.io</id>
    <url>https://jitpack.io</url>
  </repository>
</repositories>

<dependencies>
  <dependency>
    <groupId>com.github.null8626</groupId>
    <artifactId>decancer</artifactId>
    <version>v3.2.0</version>
  </dependency>
</dependencies>

Building from source

$ git clone https://github.com/null8626/decancer.git --depth 1
$ cd ./decancer/bindings/java
$ unzip ./bin/bindings.zip -d ./bin
$ chmod +x ./gradlew
$ ./gradlew build --warning-mode all

Download

• Download for ARM64 macOS (11.0+, Big Sur+)
• Download for ARM64 iOS
• Download for Apple iOS Simulator on ARM6
• Download for ARM64 Android
• Download for ARM64 Windows MSVC
• Download for ARM64 Linux (kernel 4.1, glibc 2.17+)
• Download for ARM64 Linux with MUSL
• Download for ARMv6 Linux (kernel 3.2, glibc 2.17)
• Download for ARMv5TE Linux (kernel 4.4, glibc 2.23)
• Download for ARMv7-A Android
• Download for ARMv7-A Linux (kernel 4.15, glibc 2.27)
• Download for ARMv7-A Linux, hardfloat (kernel 3.2, glibc 2.17)
• Download for 32-bit Linux w/o SSE (kernel 3.2, glibc 2.17)
• Download for 32-bit MSVC (Windows 7+)
• Download for 32-bit FreeBSD
• Download for 32-bit Linux (kernel 3.2+, glibc 2.17+)
• Download for PPC64LE Linux (kernel 3.10, glibc 2.17)
• Download for RISC-V Linux (kernel 4.20, glibc 2.29)
• Download for S390x Linux (kernel 3.2, glibc 2.17)
• Download for SPARC Solaris 11, illumos
• Download for Thumb2-mode ARMv7-A Linux with NEON (kernel 4.4, glibc 2.23)
• Download for 64-bit macOS (10.12+, Sierra+)
• Download for 64-bit iOS
• Download for 64-bit MSVC (Windows 7+)
• Download for 64-bit FreeBSD
• Download for 64-bit illumos
• Download for 64-bit Linux (kernel 3.2+, glibc 2.17+)
• Download for 64-bit Linux with MUSL

Building from source

Building from source requires Rust v1.65 or later.

$ git clone https://github.com/null8626/decancer.git --depth 1
$ cd decancer/bindings/native
$ cargo build --release

And the binary files should be generated in the target/release directory.

Examples

For more information, please read the documentation.

let mut cured = decancer::cure!(r"vＥⓡ𝔂 𝔽𝕌Ňℕｙ ţ乇𝕏𝓣 wWiIiIIttHh l133t5p3/-\|<").unwrap();

assert_eq!(cured, "very funny text with leetspeak");

// WARNING: it's NOT recommended to coerce this output to a Rust string
//          and process it manually from there, as decancer has its own
//          custom comparison measures, including leetspeak matching!
assert_ne!(cured.as_str(), "very funny text with leetspeak");

assert!(cured.contains("funny"));

cured.censor("funny", '*');
assert_eq!(cured, "very ***** text with leetspeak");

cured.censor_multiple(["very", "text"], '-');
assert_eq!(cured, "---- ***** ---- with leetspeak");

const assert = require('assert')
const cured = decancer('vＥⓡ𝔂 𝔽𝕌Ňℕｙ ţ乇𝕏𝓣 wWiIiIIttHh l133t5p3/-\\|<')

assert(cured.equals('very funny text with leetspeak'))

// WARNING: it's NOT recommended to coerce this output to a JavaScript string
//          and process it manually from there, as decancer has its own
//          custom comparison measures, including leetspeak matching!
assert(cured.toString() !== 'very funny text with leetspeak')
console.log(cured.toString())
// => very funny text wwiiiiitthh l133t5p3/-\|<

assert(cured.contains('funny'))

cured.censor('funny', '*')
console.log(cured.toString())
// => very ***** text wwiiiiitthh l133t5p3/-\|<

cured.censorMultiple(['very', 'text'], '-')
console.log(cured.toString())
// => ---- ***** ---- wwiiiiitthh l133t5p3/-\|<

<!DOCTYPE html>
<html lang="en">
  <head>
    <meta charset="utf-8" />
    <title>Decancerer!!! (tm)</title>
    <style>
      textarea {
        font-size: 30px;
      }

      #cure {
        font-size: 20px;
        padding: 5px 30px;
      }
    </style>
  </head>
  <body>
    <h3>Input cancerous text here:</h3>
    <textarea rows="10" cols="30"></textarea>
    <br />
    <button id="cure" onclick="cure()">cure!</button>
    <script type="module">
      import init from 'https://cdn.jsdelivr.net/gh/null8626/[email protected]/bindings/wasm/bin/decancer.min.js'

      const decancer = await init()

      window.cure = function () {
        const textarea = document.querySelector('textarea')

        if (!textarea.value.length) {
          return alert("There's no text!!!")
        }

        textarea.value = decancer(textarea.value).toString()
      }
    </script>
  </body>
</html>

See this in action here.

import com.github.null8626.decancer.CuredString;

public class Program {
  public static void main(String[] args) {
    CuredString cured = new CuredString("vＥⓡ𝔂 𝔽𝕌Ňℕｙ ţ乇𝕏𝓣 wWiIiIIttHh l133t5p3/-\\|<");
    
    assert cured.equals("very funny text with leetspeak");
    
    // WARNING: it's NOT recommended to coerce this output to a Java String
    //          and process it manually from there, as decancer has its own
    //          custom comparison measures, including leetspeak matching!
    assert !cured.toString().equals("very funny text with leetspeak");
    System.out.println(cured.toString());
    // => very funny text wwiiiiitthh l133t5p3/-\|<
    
    assert cured.contains("funny");
    
    cured.censor("funny", '*');
    System.out.println(cured.toString());
    // => very ***** text wwiiiiitthh l133t5p3/-\|<
    
    String[] keywords = { "very", "text" };
    cured.censorMultiple(keywords, '-');
    System.out.println(cured.toString());
    // => ---- ***** ---- wwiiiiitthh l133t5p3/-\|<
    
    cured.destroy();
  }
}

UTF-8 example:

#include <decancer.h>

#include <string.h>
#include <stdlib.h>
#include <stdio.h>

// global variable for assertion purposes only
decancer_cured_t cured;

static void assert(const bool expr, const char *message)
{
    if (!expr)
    {
        fprintf(stderr, "assertion failed (%s)\n", message);
        decancer_cured_free(cured);
        
        exit(1);
    }
}

static void print_error(decancer_error_t *error)
{
    char message[90];
    uint8_t message_size;
    
    memcpy(message, error->message, error->message_size);
   
    // rust strings are NOT null-terminated
    message[error->message_size] = '\0';
    
    fprintf(stderr, "error: %s", message);
}

int main(void) {
    decancer_error_t error;

    // UTF-8 bytes for "vＥⓡ𝔂 𝔽𝕌Ňℕｙ ţ乇𝕏𝓣"
    uint8_t string[] = {0x76, 0xef, 0xbc, 0xa5, 0xe2, 0x93, 0xa1, 0xf0, 0x9d, 0x94, 0x82, 0x20, 0xf0, 0x9d,
                        0x94, 0xbd, 0xf0, 0x9d, 0x95, 0x8c, 0xc5, 0x87, 0xe2, 0x84, 0x95, 0xef, 0xbd, 0x99,
                        0x20, 0xc5, 0xa3, 0xe4, 0xb9, 0x87, 0xf0, 0x9d, 0x95, 0x8f, 0xf0, 0x9d, 0x93, 0xa3};

    cured = decancer_cure(string, sizeof(string), DECANCER_OPTION_DEFAULT, &error);

    if (cured == NULL)
    {
        print_error(&error);
        return 1;
    }

    assert(decancer_equals(cured, (uint8_t *)("very funny text"), 15), "equals");
    assert(decancer_contains(cured, (uint8_t *)("funny"), 5), "contains");

    // coerce output as a raw UTF-8 pointer and retrieve its size (in bytes)
    size_t output_size;
    const uint8_t *output_raw = decancer_cured_raw(cured, &output_size);

    assert(output_size == 15, "raw output size");

    // UTF-8 bytes for "very funny text"
    const uint8_t expected_raw[] = {0x76, 0x65, 0x72, 0x79, 0x20, 0x66, 0x75, 0x6e,
                                    0x6e, 0x79, 0x20, 0x74, 0x65, 0x78, 0x74};

    char assert_message[38];
    for (uint32_t i = 0; i < sizeof(expected_raw); i++)
    {
        sprintf(assert_message, "mismatched utf-8 contents at index %u", i);
        assert(output_raw[i] == expected_raw[i], assert_message);
    }

    decancer_cured_free(cured);    
    return 0;
}

UTF-16 example:

#include <decancer.h>

#include <string.h>
#include <stdlib.h>
#include <stdio.h>

// global variable for assertion purposes only
decancer_cured_t cured;
decancer_cured_raw_wide_t wide = NULL;

static void assert(const bool expr, const char *message)
{
    if (!expr)
    {
        fprintf(stderr, "assertion failed (%s)\n", message);
        
        if (wide != NULL)
        {
            decancer_cured_raw_wide_free(wide);
        }
        
        decancer_cured_free(cured);
        
        exit(1);
    }
}

static void print_error(decancer_error_t *error)
{
    char message[90];
    uint8_t message_size;
    
    memcpy(message, error->message, error->message_size);
   
    // rust strings are NOT null-terminated
    message[error->message_size] = '\0';
    
    fprintf(stderr, "error: %s", message);
}

int main(void) {
    decancer_error_t error;

    // UTF-16 bytes for "vＥⓡ𝔂 𝔽𝕌Ňℕｙ ţ乇𝕏𝓣"
    uint16_t string[] = {
        0x0076, 0xff25, 0x24e1,
        0xd835, 0xdd02, 0x0020,
        0xd835, 0xdd3d, 0xd835,
        0xdd4c, 0x0147, 0x2115,
        0xff59, 0x0020, 0x0163,
        0x4e47, 0xd835, 0xdd4f,
        0xd835, 0xdce3
    };

    cured = decancer_cure_wide(string, sizeof(string), DECANCER_OPTION_DEFAULT, &error);

    if (cured == NULL)
    {
        print_error(&error);
        return 1;
    }

    assert(decancer_equals(cured, (uint8_t *)("very funny text"), 15), "equals");
    assert(decancer_contains(cured, (uint8_t *)("funny"), 5), "contains");

    // coerce output as a raw UTF-16 pointer and retrieve its size (in bytes)
    uint16_t *output_ptr;
    size_t utf16_output_size;
    wide = decancer_cured_raw_wide(cured, &output_ptr, &utf16_output_size);

    assert(utf16_output_size == (15 * sizeof(uint16_t)), "raw output size");

    // UTF-16 bytes for "very funny text"
    const uint16_t expected_raw[] = {0x76, 0x65, 0x72, 0x79, 0x20, 0x66, 0x75, 0x6e,
                                     0x6e, 0x79, 0x20, 0x74, 0x65, 0x78, 0x74};

    char assert_message[39];
    for (uint32_t i = 0; i < sizeof(expected_raw) / sizeof(uint16_t); i++)
    {
        sprintf(assert_message, "mismatched utf-16 contents at index %u", i);
        assert(output_raw[i] == expected_raw[i], assert_message);
    }

    decancer_cured_raw_wide_free(wide);
    decancer_cured_free(cured);    
    return 0;
}

Donations

If you want to support my eyes for manually looking at thousands of unicode characters, consider donating! ❤

Contributing

Please read CONTRIBUTING.md for newbie contributors who want to contribute!