Bantaba - West African Language Translation App

A React Native app for building word datasets for Mandinka, Wolof, Jola, and Fula languages through AI-powered translation generation and user verification.

🌍 Supported Languages

• 🇬🇧 English (source language)
• 🇬🇲 Mandinka - Gambia, Senegal, Guinea
• 🇸🇳 Wolof - Senegal, Gambia, Mauritania
• 🇸🇳 Jola (Joola) - Senegal, Gambia, Guinea-Bissau
• 🇬🇳 Fula (Fulfulde) - West Africa

✨ Features

• Translation Generation: Generate translations using AI models (OpenAI, custom APIs)
• User Verification: Verify, correct, and submit translations with notes
• Batch Processing: Generate multiple translations at once
• Dataset Management: Track statistics and export verified data
• Category Support: Organize words by categories (greetings, food, etc.)

🚀 Setup

1. Install dependencies:

npm install

2. Set up Supabase (Required for backend):

   • Create a new project at supabase.com
   • Go to SQL Editor and run the migration script from supabase_migration.sql
   • Get your Supabase URL and anon key from Settings > API
   • Create a .env file in the root directory:

     EXPO_PUBLIC_SUPABASE_URL=your_supabase_project_url
     EXPO_PUBLIC_SUPABASE_ANON_KEY=your_supabase_anon_key
     EXPO_PUBLIC_USE_BACKEND=true

   • See SUPABASE_SCHEMA.md for detailed setup instructions

3. Add your translation data (Optional - for local fallback):

   • Add JSON files to /data/ directory:
     • mandinka_lang.json ✅ (already added)
     • wolof_lang.json (empty - add your data)
     • fula_lang.json (empty - add your data)
     • jola_lang.json (empty - add your data)

4. Start the app:

npm start

Then press a for Android, i for iOS, or w for Web.

⚙️ Configuration

Supabase Backend

The app now uses Supabase as the primary backend database. All translations are stored in Supabase and synced across devices.

• Primary Storage: Supabase PostgreSQL database
• Local Cache: AsyncStorage (for offline support and fallback)
• Setup: See SUPABASE_SCHEMA.md for database schema and setup instructions

JSON Data Format

Your JSON files should follow this structure:

[
  {
    "english": "hello",
    "mandinka": "salaam",
    "wolof": "nanga def",
    "fula": "jam",
    "jola": "kasumay"
  },
  {
    "english": "goodbye", 
    "mandinka": "fo dii",
    "wolof": "ba beneen yoon",
    "fula": "sellam",
    "jola": "kasumay"
  }
]

Current Status:

• ✅ Mandinka: ~28,000+ translations loaded
• ⚠️ Wolof: No data (add to data/wolof_lang.json)
• ⚠️ Fula: No data (add to data/fula_lang.json)
• ⚠️ Jola: No data (add to data/jola_lang.json)

📁 Project Structure

src/
├── config/
│   └── languages.ts      # Language definitions
├── screens/
│   ├── HomeScreen.tsx
│   ├── TranslationVerificationScreen.tsx
│   └── DatasetManagementScreen.tsx
├── services/
│   ├── TranslationService.ts
│   └── StorageService.ts
└── types/
    └── index.ts          # TypeScript interfaces

📊 How to Use

1. Select Language: Choose Mandinka, Wolof, Jola, or Fula from home screen
2. Review Translation: App shows random English word with AI translation
3. Edit if Needed: Modify the translation if it's incorrect
4. Submit or Skip: Save verified translation or skip to next word
5. Track Progress: View statistics and export dataset anytime

Workflow per Language

Each language has its own screen where you:

• Get a random English word
• See OpenAI-generated translation
• Edit the translation if needed
• Submit (saves to dataset) or Skip (moves to next word)
• Track your progress in real-time

🗄️ Database

The app uses Supabase (PostgreSQL) for data storage. See SUPABASE_SCHEMA.md for:

• Database schema documentation
• SQL migration scripts
• Security policies (RLS)
• Setup instructions

🔜 Next Steps

1. ✅ Set up backend sync for collaborative dataset building (Done - using Supabase)
2. Add audio recording/playback for pronunciation verification
3. Implement CSV/JSON batch import
4. Add user authentication
5. Integrate speech-to-text for audio transcription

📱 Built with Expo

This is an Expo project using file-based routing.