#neural-network #arabic #restoration #text #cross-platform #diacritics #diacritic

libtashkeel_base

Arabic-text diacritic restoration using neural networks

3 stable releases

new 1.5.0 Apr 20, 2024
1.2.0 Oct 4, 2023
1.1.0 Oct 3, 2023

#129 in Machine learning

Download history 9/week @ 2024-02-19 29/week @ 2024-02-26 38/week @ 2024-04-01 109/week @ 2024-04-15

147 downloads per month

MIT/Apache

4.5MB
421 lines

Libtashkeel

Libtashkeelis a cross-platform library for diacritic restoration of Arabic text.

Libtashkeel is written in Rust, and provides both a standalone linkable library and a command line tool.

The library uses models trained mainly on MSA data, from Hareef.

Getting Libtashkeel

You need to build the project yourself, see the Building section for a step-by-step guide.

Usage

Using the library

To use Libtashkeel from your C/C++ project, just include libtashkeel.h and you are good to go.

The API consists of a single entry point for diacritizing a **utf-8 ** encoded string. Please take a look at ffi_usage_example.py for sample usage.

From Python

Python bindings are also provided.

After building the wheels (see the Building section), install the wheel using pip:

pip install ./target/wheels/pylibtashkeel*.whl

and then:

>>> from pylibtashkeel import tashkeel
>>> tashkeel("إن روعة اللغة العربية لا تتبدى إلا لعشاقها")
'إِنَّ رَوْعَةَ اللُّغَةِ الْعَرَبِيَّةِ لَا تَتَبَدَّى إِلَّا لِعُشَّاقِهَا'

Command line tool

Libtashkeel provides a standalone executable called tashkeel for diacritizing text from the command line.

$ tashkeel --help
Arabic-text diacritic restoration using neural networks

Usage: tashkeel [OPTIONS]

Options:
  -f, --input-file <INPUT_FILE>    Input file (default `stdin`)
  -o, --output-file <OUTPUT_FILE>  Output file (default `stdout`)
  -i, --interactive                Use interactive mode (useful for testing)
  -t, --taskeen                    Use sukoon for case-ending diacritic if the model is uncertain
  -p, --prob <PROB>                Taskeen threshold probability [default: 0.95]
  -x, --onnx <ONNX_MODEL>          ONNX model (default: use bundled model if available)
  -h, --help                       Print help
  -V, --version                    Print version

Building

Libtashkeel is written in Rust, you need to install Rust first

To build the linkable library libtashkeel, and the command line tool tashkeel, run the following command from the root of the repository:

$ cargo build --release

Then, the built library and executable is found under target directory.

To build Python bindings as a wheel, you need to install maturin

Run the following to build the wheel:

$ cd pylibtashkeel
$ python3 -m venv .venv
$ source .venv/bin/activate
$ pip install maturin
$ maturin build --release --strip -i .venv/bin/python

Then, the built wheel is found under target/wheels directory.

Licence

Copyright (c) Musharraf Omer. This project is licenced under the terms of The MIT License

Dependencies

~8–17MB
~191K SLoC