shredword 0.1.2

fast bpe, unigram & word tokenizer for experimental models

Shredword

A fast and efficient tokenizer library for natural language processing tasks, built with Python and optimized C backend.

Features

• High Performance: Fast tokenization powered by optimized C libraries
• Multiple Encodings: Support for various tokenization models and vocabularies
• Flexible API: Easy-to-use Python interface with comprehensive functionality
• Special Tokens: Built-in support for special tokens and custom vocabularies
• Fallback Mechanisms: Robust error handling with fallback tokenization
• BPE Support: Byte Pair Encoding implementation for subword tokenization
• Word Tokenization: Fast word-level tokenization with contraction handling
• TF-IDF Embeddings: Built-in TF-IDF vectorization with dense and sparse representations

Installation

pip install shredword

Quick Start

BPE Tokenization

from shred import load_encoding

tokenizer = load_encoding("pre_16k")

tokens = tokenizer.encode("Hello, world!")
print(tokens)

text = tokenizer.decode(tokens)
print(text)

print(f"Vocabulary size: {tokenizer.vocab_size}")
print(f"Special tokens: {tokenizer.special_tokens}")

Word Tokenization & TF-IDF Embeddings

from shred import WordTokenizer, TfidfEmbedding

tokenizer = WordTokenizer()
tokens = tokenizer.tokenize("Hello, world! This is a test.")
print(tokens)

embedding = TfidfEmbedding()
embedding.add_documents([
  "The quick brown fox jumps over the lazy dog",
  "Python programming is fun and exciting"
])

ids = embedding.encode_ids("The lazy fox")
dense_vec = embedding.encode_tfidf_dense("The lazy fox")
indices, values = embedding.encode_tfidf_sparse("The lazy fox")

embedding.save("vocab.txt")
loaded = TfidfEmbedding.load("vocab.txt")

Documentation

For detailed usage instructions, API reference, and examples, please see our User Documentation.

Supported Encodings

Shredword supports various pre-trained tokenization models. The library automatically downloads vocabulary files from the official repository when needed.

Contributing

We welcome contributions! Please feel free to submit issues, feature requests, or pull requests.

Development Setup

1. Clone the repository
2. Install development dependencies: pip install -r requirements.txt (there are none!)
3. Run tests: python -m pytest

Guidelines

• Follow PEP 8 style guidelines
• Add tests for new features
• Update documentation as needed
• Ensure all tests pass before submitting PRs

License

This project is licensed under the Apache 2.0 License - see the LICENSE file for details.

Support

• Issues: Report bugs or request features on GitHub Issues
• Discussions: Join community discussions on GitHub Discussions