Ed1ezac/bw-setswana-nlp

Botswana Setswana-NLP is a standarised package for Setswana digital translation and eventual NLP and AI use. # bw-setswana-nlp > A developer-facing NLP toolkit and language resource for **Setswana as spoken in Botswana** — building the digital foundation for a language that belongs online. -green.svg) --- ## Why This Exists Setswana is spoken by millions of people in Botswana as a first language and serves as the national language of the country. Yet in digital spaces — operating systems, translation engines, NLP pipelines, keyboards, and developer tooling — it remains largely invisible. This project exists to change that. `bw-setswana-nlp` is an open-source effort to build the foundational language resources that allow Setswana to exist as a first-class language in software: locale support, dictionary datasets, morphological tooling, corpora, and eventually NLP/AI-ready resources — all rooted in **Setswana as natively spoken in Botswana** (`tn-BW`). This is not a translation app. It is the **infrastructure layer** that makes Setswana viable for developers building apps, researchers training models, and platform teams adding locale support. --- ## Project Scope This repository targets **Setswana as spoken in Botswana** (`tn-BW` locale). While Setswana/Tswana is also spoken in South Africa and Zimbabwe, this project's primary reference is the Botswana standard. Contributions that introduce dialect conflicts will be discussed in issues before merging. --- ## Roadmap Development is phased to build a solid foundation before higher-level tooling. ### Phase 1 — Foundation *(current focus)* - [ ] Structured lexicon / dictionary dataset (words, definitions, POS tags) - [ ] ISO locale data (`tn-BW`) — date formats, number formats, currency (BWP) - [ ] Basic tokenizer (whitespace + punctuation aware) - [ ] Character and diacritic reference for Setswana orthography ### Phase 2 — Morphology & Grammar - [ ] Noun class system documentation and data structures - [ ] Verb conjugation tables - [ ] Agglutination rules and morphological analyzer - [ ] Part-of-speech tagger …