Arabic Cleaner

A lightweight Python library for cleaning and normalizing Arabic text, designed for natural language processing (NLP) tasks such as Quranic text preprocessing.

Overview

Arabic text often contains diacritics, inconsistent character forms, and non-standard symbols that can negatively affect downstream NLP models. This library provides a simple and efficient pipeline to normalize and clean Arabic text before further processing.

Features

• Remove Arabic diacritics (harakat): Efficiently strips all Arabic harakat.
• Normalize common character variations: Standardizes variations of Alif, Hamza, etc.
• Filter non-Arabic characters: Removes unwanted symbols to ensure data purity.
• Simple and modular cleaning pipeline: Designed to be lightweight and easily integrated.
• Command-line interface (CLI) support: Quick processing directly from the terminal.

Installation

Install from PyPI:

pip install arabic-cleaner

For development (editable mode):

pip install -e .

Usage

Python

from arabic_cleaner import clean_text

text = "إِنَّ ٱللَّهَ غَفُورٌ رَّحِيمٌ"
result = clean_text(text)

print(result)

Output:

ان الله غفور رحيم

Command Line Interface

arabic-clean "إِنَّ ٱللَّهَ غَفُورٌ رَّحِيمٌ"

API

clean_text(text, remove_diacritics=True)

Example:

Input: إِنَّ ٱللَّهَ غَفُورٌ رَّحِيمٌ

Output: ان الله غفور رحيم

Project Structure

arabic_cleaner/
│── __init__.py
│── cleaner.py
│── normalizer.py
│── cli.py

Use Cases

• Arabic NLP preprocessing
• Quranic text normalization
• Speech-to-text postprocessing
• Machine learning data cleaning pipelines

License

MIT License