🏥 Doctor-Patient Translation Service

A real-time translation service for doctor-patient communication using Sarvam AI, supporting English, Hindi, Marathi, and Kannada languages.

✨ Features

• Real-time Translation: Instant translation between doctor (English) and patient (Hindi/Marathi/Kannada)
• Voice-to-Voice Agent: Live intermediary translator that recognizes speakers and translates conversations
• Web Interface: Modern browser-based UI with real-time capabilities
• Sarvam AI Integration: Uses state-of-the-art Indic language translation
• Interactive Interface: User-friendly command-line interface with Rich formatting
• Speaker Recognition: Automatically detects doctor vs patient speech
• Text Fallbacks: Works with text input when voice features unavailable
• Real-time Communication: WebSocket-based live updates
• Error Handling: Robust error handling with helpful messages
• Extensible: Easy to add more languages and features

🚀 Quick Start

1. Install Dependencies

pip install -r requirements.txt

2. Get Sarvam AI API Key

1. Visit Sarvam AI Dashboard
2. Sign up and get your API key
3. Set your API key in config.py:

   SARVAM_API_KEY = "your_actual_api_key_here"

3. Run the Application

python main.py

🌐 Web Interface

The service includes a modern web-based interface with real-time capabilities:

Features

• Live Voice Translation: Real-time voice-to-voice translation in the browser
• Language Selection: Easy language switching with visual indicators
• Text Translation: Manual text translation with instant results
• Conversation Log: Complete conversation history with timestamps
• Status Indicators: Real-time status updates and system messages
• Responsive Design: Works on desktop and mobile devices
• WebSocket Communication: Instant updates without page refresh

Accessing the Web Interface

python main.py
# Choose option 3: "🌐 Web Interface (Browser-based UI)"

The web server will start at http://localhost:5000. Open this URL in your browser.

Web Demo

If you can't install Flask dependencies, you can view a static demo of the web interface:

open web_demo.html  # On macOS
# or
# Open web_demo.html in your browser

The demo shows the full interface design and functionality preview.

Web Interface Sections

Language Selection

• Choose between Hindi, Marathi, and Kannada for patient communication
• Visual language buttons with native script display

Voice Translation Agent

• Start/stop voice conversations
• Real-time status indicators (listening/speaking/ready)
• Automatic speaker detection and turn management

Text Translation

• Tabbed interface for Doctor→Patient and Patient→Doctor translation
• Manual text input with instant translation
• Text-to-speech capabilities for both languages

Conversation Log

• Complete history of all translations and interactions
• Color-coded entries (Doctor/Patient/System)
• Clear log functionality

🎤 Voice-to-Voice Translation Agent

The service includes an advanced Voice Translation Agent that acts as a live intermediary translator:

How It Works

1. Speaker Recognition: Automatically detects whether the doctor or patient is speaking
2. Speech-to-Text: Converts speech to text (with Google Speech Recognition)
3. Translation: Translates using Sarvam AI's high-quality Indic language models
4. Text-to-Speech: Speaks the translation aloud (with pyttsx3)
5. Conversation Flow: Manages turn-taking between doctor and patient

Example Conversation Flow

Doctor speaks: "Hello, how are you feeling today?"
→ Agent recognizes doctor, translates to Hindi
→ Agent speaks: "नमस्ते, आज आप कैसी महसूस कर रहे हैं?"

Patient responds: "मैं ठीक हूं, लेकिन सिरदर्द है"
→ Agent recognizes patient, translates to English
→ Agent speaks: "I am fine but yes, I have a headache"

Doctor asks: "When did the headache start?"
→ Agent recognizes doctor, translates to Hindi
→ Agent speaks: "सिरदर्द कब शुरू हुआ?"

Voice Mode Features

• Real-time Processing: Instant translation with minimal delay
• Fallback Support: Works with text input when microphone unavailable
• State Management: Tracks conversation flow and speaker turns
• Multi-language: Supports Hindi, Marathi, and Kannada patients
• Error Recovery: Graceful handling of recognition/translation failures

📋 Supported Languages

Language	Code	Usage
English	en-IN	Doctor's language
Hindi	hi-IN	Patient language
Marathi	mr-IN	Patient language
Kannada	kn-IN	Patient language

🎯 Usage Examples

Interactive Mode

🏥 Doctor-Patient Translation Service

Select Patient's Language:
┌─────────┬──────────┬─────────────────┐
│ Option  │ Language │ Native Name     │
├─────────┼──────────┼─────────────────┤
│ 1       │ Hindi    │ हिंदी (Hindi)    │
│ 2       │ Marathi  │ मराठी (Marathi)  │
│ 3       │ Kannada  │ ಕನ್ನಡ (Kannada)  │
└─────────┴──────────┴─────────────────┘

Enter language number (1-3): 1

📝 Translation Demo:
Doctor: Hello, how are you feeling today?
Patient (Hindi): नमस्ते, आज आप कैसे महसूस कर रहे हैं?

🎯 Interactive Translation Mode
Doctor (English): Please describe your symptoms.
→ Patient (Hindi): कृपया अपने लक्षणों का वर्णन करें।

Commands

• quit - Exit the application
• switch - Change patient language
• demo - Show translation examples

🛠️ API Integration

Basic Usage

from sarvam_service import SarvamService, DoctorPatientTranslator

# Initialize service
service = SarvamService()
translator = DoctorPatientTranslator(service)

# Translate doctor to patient
patient_text = translator.doctor_to_patient("Hello, how are you?", "hindi")
print(patient_text)  # नमस्ते, आप कैसे हैं?

# Translate patient to doctor
doctor_text = translator.patient_to_doctor("मैं ठीक हूं", "hindi")
print(doctor_text)  # I am fine

Direct API Usage

from sarvam_service import SarvamService

service = SarvamService()

# Translate between any supported languages
result = service.translate_text("Hello world", "en-IN", "hi-IN")
print(result)  # हैलो वर्ल्ड

📁 Project Structure

├── main.py                 # Main application interface
├── sarvam_service.py       # Sarvam AI integration and translation logic
├── voice_agent.py          # Voice translation agent with speaker recognition
├── web_app.py             # Flask web application with SocketIO
├── config.py              # Configuration and settings
├── requirements.txt       # Python dependencies
├── test_sarvam.py         # Unit tests
├── setup.py               # Setup automation script
├── templates/
│   └── index.html         # Web interface template
├── static/
│   ├── css/
│   │   └── styles.css     # Web interface styles
│   └── js/
│       └── app.js         # Frontend JavaScript
└── README.md              # This file

🔧 Configuration

Edit config.py to customize:

• API key
• Supported languages
• Translation pairs
• API endpoints

🚨 Troubleshooting

Common Issues

1. API Key Error

   • Ensure your Sarvam AI API key is correctly set in config.py
   • Verify the key is active and has credits

2. Network Errors

   • Check your internet connection
   • Ensure firewall allows HTTPS requests

3. Translation Errors

   • Verify the language pair is supported
   • Check that input text is not empty

Error Messages

• "API request failed" - Check API key and network connection
• "Translation from X to Y is not supported" - Verify language codes
• "Network error" - Check internet connection

🤝 Contributing

1. Fork the repository
2. Create a feature branch
3. Add tests for new features
4. Submit a pull request

📄 License

This project is open source and available under the MIT License.

🙏 Acknowledgments

• Sarvam AI for providing excellent Indic language translation
• Built for improving healthcare communication in multilingual India

---

Made with ❤️ for better doctor-patient communication