aamraz 0.0.5

This project is a collection of Natural Language Processing tools for Kurdish Language.

Aamraz - Kurdish NLP collection

• Overview
• Features
• Installation
• Pre-trained Models
• Usage

Overview

Aamraz which is written "ئامراز" in kurdish script means "instrument". This project is a collection of Natural Language Processing tools for Kurdish Language.

Base Features

• Normalization
• Tokenization
• Word Embedding: Creates vector representations of words.
• Sentences Embedding: Creates vector representations of sentences.

Tools

Installation

pip install aamraz

PretrainedModels

some useful pre-trained Models:

Model	Description	Size
FastText WordEmbedding	Model trained using FastText method on our own Corpus. This is bot the fasttext & skip-gram model itself (fasttext model.	~ 2.3 GB
FastText WordEmbedding - Lite	Model trained using FastText method on our own Corpus. This is bot the fasttext & skip-gram model itself (fasttext model.	~ 800 MB

Usage

import aamraz

# Normalization
normalizer= aamraz.Normalizer()
sample_sentence="قڵبە‌کە‌م‌ بە‌  کوردی‌  قسە‌ دە‌کات‌."
normalized_sentence=normalizer.normalize(sample_sentence)
print(normalized_sentence)

# Tokenization
tokenizer = aamraz.WordTokenizer()
sample_sentence="زوانی له دربره"
tokens = tokenizer.tokenize(sample_sentence)
print(tokens)

# Embedding
model_path = 'kurdish_fasttext_skipgram_dim300_v1.bin'
embedding_model = aamraz.EmbeddingModel(model_path, dim=50)

sample_word="ئامراز"
sample_sentence="زوانی له دربره"

word_vector = embedding_model.word_embedding(sample_word)
sentence_vector = embedding_model.sentence_embedding(sample_sentence)

print(word_vector)
print(sentence_vector)

License

This project is licensed under the MIT License. You are free to use, distribute, modify, and build upon this work, even for commercial purposes, as long as you include a copy of the original MIT License and provide proper attribution.

To view a copy of this license, visit: https://opensource.org/licenses/MIT