Mental Health Support App Using LLM

This project is an innovative mental health support application that leverages Gemini (GPT-based LLM) and FastAPI to provide empathetic responses, coping strategies, and resources for professional help. The application enables users to anonymously discuss their mental health concerns, offering support in real-time.

Table of Contents

• Features
• Tech Stack
• Installation
• Configuration
• Usage
• API Documentation
• Frontend (Streamlit)
• Backend (FastAPI)
• Best Practices
• Future Enhancements

---

Features

• Anonymous Mental Health Support: Users can submit concerns without revealing their identity.
• Empathetic Responses: Powered by Gemini LLM, the app provides empathetic and non-judgmental responses.
• Coping Strategies: Users receive personalized coping strategies based on their input.
• Professional Resources: The app suggests professional mental health resources if needed.
• Interactive Client Interface: Simple and intuitive user interface using Streamlit.

---

Tech Stack

• Backend: FastAPI
• Frontend: Streamlit
• LLM: Gemini (via OpenAI GPT API)
• Other Libraries:
  • requests for API calls
  • pydantic for data validation in FastAPI
  • uvicorn to run the FastAPI server

---

Installation

1. Clone the Repository

git clone https://github.com/mr-rakesh-ranjan/mental-health-support-app.git
cd mental-health-support-app

2. Set up Virtual Environment

It’s recommended to use a virtual environment to avoid dependency conflicts.

python3 -m venv venv
source venv/bin/activate  # For Linux/MacOS
venv\Scripts\activate     # For Windows

3. Install Dependencies

pip install -r requirements.txt

---

Configuration

You need an GEMINI API Key to interact with the Gemini LLM.

1. Create a .env file or set environment variables for your API key.

GEMINI_API_KEY=your_gemini_api_key

2. You can also update the config.py file to hardcode your API key (not recommended for production).

---

Usage

1. Run the FastAPI Backend

First, start the backend API service using FastAPI and Uvicorn.

cd backend
uvicorn backend.main:app --reload

This will start the backend on http://localhost:8000.

2. Run the Streamlit Frontend

Next, start the Streamlit client interface where users can submit their concerns.

cd frontend
streamlit run frontend/app.py

This will launch the frontend in your browser at http://localhost:8501.

---

API Documentation

The FastAPI backend exposes the following endpoints:

1. Analyze Input (POST /analyze/)

• Description: Analyzes user input to provide empathetic responses and coping strategies.
• Request:
  • Body (JSON):

    {
      "text": "I feel very stressed and anxious about work."
    }
    

• Response (JSON):

  {
    "llm_response": "It's completely normal to feel anxious during tough times at work. Try taking breaks and practicing mindfulness.",
    "coping_strategies": [
      "Take deep breaths and focus on the present moment.",
      "Engage in a hobby or activity that brings you joy.",
      "Reach out to a trusted friend or family member."
    ]
  }
  

2. Get Professional Resources (GET /resources/)

• Description: Provides a list of professional mental health resources.
• Response (JSON):

  [
    {"name": "National Suicide Prevention Lifeline", "contact": "1-800-273-8255"},
    {"name": "BetterHelp", "website": "https://www.betterhelp.com"},
    {"name": "Talkspace", "website": "https://www.talkspace.com"}
  ]
  

---

Frontend (Streamlit)

The Streamlit client provides an easy-to-use interface for users to input their mental health concerns.

Key Features:

• User Input: Users can submit their text, which is sent to the FastAPI backend for analysis.
• Real-Time Response: The app shows LLM-generated empathetic responses and coping strategies directly on the interface.

How to Use:

1. Enter your mental health concerns in the provided text area.
2. Click Submit to receive suggestions and feedback.
3. View the Empathetic Response and a list of Coping Strategies.

---

Backend (FastAPI)

The FastAPI backend provides the main logic for processing input, calling the Gemini LLM API, and returning coping strategies.

Key Components:

1. LLM Service: This handles communication with the Gemini API (through OpenAI) to generate responses based on user input.
2. Coping Strategies Service: This module returns helpful strategies and professional resources.
3. Data Validation: The input is validated using Pydantic models.

FastAPI Server

You can run the FastAPI server locally using uvicorn.

uvicorn backend.main:app --reload

---

Best Practices

1. Security

• API Key Management: Store your OpenAI API key in environment variables, not in code.
• Input Sanitization: Ensure proper input sanitization to avoid malicious content being processed.

2. Modularity

• Separation of Concerns: The project is divided into backend and frontend components for better modularity and maintainability.
• Service-Oriented: The backend services (LLM, coping strategies) are isolated into separate modules for future scalability.

3. Scalability

• Containerization: Consider using Docker to containerize the FastAPI and Streamlit services for easy deployment.
• Caching: Implement caching for frequently used resources (e.g., coping strategies).

---

Future Enhancements

1. User Authentication: Implement user authentication for personalized experiences and mental health tracking over time.
2. Analytics Dashboard: Use Streamlit or another frontend to provide users with data on their progress and insights.
3. Advanced NLP Models: Incorporate more advanced NLP techniques or custom-trained models for better text understanding.
4. Mental Health Assessment: Expand the service to offer assessments for various mental health conditions.
5. Support for Multiple Languages: Implement support for users in different languages.

---

License

This project is licensed under the MIT License. See the LICENSE file for more information.

---

Feel free to modify the README as per your specific project requirements and branding! Let me know if you need further adjustments or additions.