Sunbird ALL - Offline ASR Integration

This project implements offline Automated Speech Recognition (ASR) capabilities for the Sunbird ALL platform, enhancing language learning through interactive exercises and real-time feedback.

Features

• Offline ASR with ≥ 90% accuracy in multiple languages
• Real-time speech recognition and feedback
• Interactive language learning exercises
• Progress tracking dashboard
• Cross-platform support (Web and Mobile)

Project Structure

.
├── backend/                 # Python backend services
│   ├── app/                # Main application code
│   ├── models/             # ASR models and ML components
│   ├── tests/              # Backend tests
│   └── Dockerfile          # Backend container configuration
├── frontend/               # React/React Native frontend
│   ├── web/               # React web application
│   ├── mobile/            # React Native mobile app
│   └── shared/            # Shared components and utilities
└── docker-compose.yml     # Container orchestration

Prerequisites

• Python 3.8+
• Node.js 16+
• Docker and Docker Compose
• React Native development environment (for mobile)

Setup Instructions

Backend Setup

1. Navigate to the backend directory:

   cd backend

2. Create and activate a virtual environment:

   python -m venv venv
   source venv/bin/activate  # On Windows: venv\Scripts\activate

3. Install dependencies:

   pip install -r requirements.txt

Frontend Setup

1. Navigate to the frontend directory:

   cd frontend/web  # or frontend/mobile for mobile app

2. Install dependencies:

   npm install

Running the Application

1. Start the backend services:

   docker-compose up -d

2. Start the frontend development server:

   # For web
   cd frontend/web
   npm start
   
   # For mobile
   cd frontend/mobile
   npm run android  # or npm run ios

Technology Stack

• Backend:

  • Python 3.8+
  • FastAPI
  • Vosk (Offline ASR Engine)
  • SQLAlchemy
  • Docker

• Frontend:

  • React (Web)
  • React Native (Mobile)
  • TypeScript
  • Material-UI
  • Redux Toolkit