dayleetow/kikuyu-correction-hub

Digitizing Gikuyu Language. In this project, Karicho is building a digital version of the Gikuyu language and Culture to make it global and easily accessible ensuring its survival # Kikuyu Correction Hub — Streamlit MVP Review, correct, and approve **Gĩkũyũ** prompts for Common Voice and language datasets. ## 🚀 Deploy (Streamlit Community Cloud) 1. Create a public GitHub repo and add these files. 2. Go to streamlit.io, click **New app**, select your repo and `app.py`. 3. Deploy — share the URL with reviewers. ## 📂 Files - `app.py` — Streamlit app (inline editing, approval, export) - `requirements.txt` — minimal deps - `data/kikuyu_spontaneous_prompts_seed.tsv` — starter prompts - `data/kikuyu_spontaneous_prompts_review.csv` — review queue (with QA columns) - `.streamlit/config.toml` — theming ## 🧭 Workflow - **Review Queue**: editors fix orthography (e.g., `Warereirwo ku?`), add `dialect_note` - Set `status=APPROVED` when 2 reviewers agree - Export **APPROVED.tsv** for Common Voice or training ## ✍️ Notes - Prefer modern, natural Kikuyu forms - Keep prompts 5–12 words; avoid PII - Use `dialect_note` to tag Nyeri/Murang'a/Kirinyaga variants ## 📜 License This template is provided for your project. Prompts you add remain your content.