Welcome to the Zomi‑Syl Documentation¶

zomi-syl is a modular, dialect‑aware Zomi syllabification library with support for multiple backends, dialect profiles, benchmarking tools, and a clean developer workflow. This documentation provides an overview of the system, how to use it, and how to extend it with new models.

🚀 What is Zomi‑Syl?¶

Zomi‑Syl provides:

• A unified syllabification API
• Multiple backends (rule‑based, CRF, transformer‑ready)
• Dialect‑aware syllabification
• A full CLI for analysis, benchmarking, and diagnostics
• Developer tooling for adding new backends
• Training and evaluation scripts

It is designed for linguists, developers, and researchers working with Zomi language processing.

📦 Installation¶

pip install zomi-syl

🧠 Quick Start¶

Syllabify a word¶

zomi-syl syllabify itna

Analyze a word¶

zomi-syl analyze itna --json

Python API¶

import zomi_syl as zs

zs.syllabify("itna")
zs.analyze("itna")

🧩 Key Components¶

Backends¶

Zomi‑Syl supports multiple syllabification backends:

• Rule‑based
• CRF
• Transformer‑ready

See: Backend Overview

Dialect Profiles¶

Profiles define onsets, nuclei, codas, and rules for each dialect.

See: Dialect Profiles

CLI Tools¶

The CLI provides:

• Syllabification
• Analysis
• Batch processing
• Benchmarking
• Backend comparison
• Diagnostics

See: Top‑Level Commands

Developer Guides¶

Zomi‑Syl includes extensive developer documentation:

• Adding new backends
• Unified Metadata Schema (UMS)
• CRF training
• Test templates
• Recommended folder structure

See: Developer Documentation

📊 Benchmarking¶

Benchmark a backend:

zomi-syl models benchmark crf

Compare multiple backends:

zomi-syl models compare rule crf

See: Benchmarking Guide

🛠 Adding New Models¶

Zomi‑Syl is designed for extensibility.
New backends can be added without modifying the core engine.

See: Adding New Backends

📚 Documentation Structure¶

This documentation includes:

• User Guide
• Installation
• CLI usage
• Python API

• Dialect profiles

• Developer Guide

• Backend architecture
• Unified Metadata Schema
• Adding new backends
• Testing

• Training CRF models

• Reference

• CLI command reference
• Folder structure
• Release checklist

🧪 Testing¶

Run all tests:

pytest

Golden CRF regression data:

tests/golden/crf_golden.tsv

📦 Release Notes¶

Changelog is generated automatically:

make changelog

See: [Looks like the result wasn't safe to show. Let's switch things up and try something else!]

📜 License¶

MIT License — see LICENSE.

🙌 Contributing¶

See: [Looks like the result wasn't safe to show. Let's switch things up and try something else!]