NatanimS/wolaytta-english-nlp

A Bilingual Wolaytta-English Parallel Corpus and Interactive Translation Engine Dashboard. # 🌍 Wolaytta-English NLP Translation Engine & Parallel Corpus Matrix A foundational, highly interactive Natural Language Processing (NLP) framework designed for Wolaytta, a low-resource Omotic language spoken in southwestern Ethiopia. This repository bridges digital data scarcity by aligning scriptural corpora, injecting everyday vocabulary, and demonstrating structural translations via a live web sandboxing interface. ## 🚀 Live Demonstration The deployment phase is fully solved and hosted globally on decentralized CDNs via Netlify. 👉 **Explore the Live App:** bejewelled-taffy-c865dd.net… --- ## 👥 Research Team & Contributors * **Betselot Mekonen** (ID: 17,034/25) * **Bontu Tesfaye** (ID: 17,026/25) * **Hildana Getachew** (ID: 17,044/25) * **Medmim Teferi** (ID: 17,007/25) * **Natanim Seifu** (ID: 17,240/25) **Submitted to:** Masresha Asegede (UNIX Course Project) **Submission Date:** July 2026 --- ## 🛠️ Project Core & Features ### 1. Structured Parallel Corpus Frame * **Ecclesiastical Ingestion:** Cleaned and verse-aligned scriptural tokens sourced from verified textual repositories (Genesis, Exodus, Matthew). * **Deterministic Alignment:** Utilizes cross-linguistic verse keys to guarantee ground-truth structural sentence pairs, completely avoiding model-based misalignment errors. ### 2. Multi-Category Vocabulary Augmentation ($N = 4,235+$ Terms) To resolve the out-of-vocabulary (OOV) bottlenecks inherent to classical scriptural text domains, we engineered an injection algorithm overlaying everyday conversational tokens: * **Grammatical Connectors:** *shin* (but), *woykko* (or), *yaatidi* (then). * **Kinship Structures:** *aawa* (father), *aaye* (mother), *naata* (kids). * **Modern Logistics & Infrastructure:** *bilbiliya* (phone), *makiina* (car), *kompiuteriya* (computer). * **Action Verbs & Colors:** *ma* (eat), *uy* (drink), *zo'o* (red), *bootta* (white). ### 3. Production Web Sandbox Application * **Glassmorphic UI Engine:** Styled us …