MedComp Assistant

MedComp Assistant is a health-focused, AI-powered assistant designed to provide accurate, reliable, and empathetic support to users of all ages, especially teenagers. With over 10 years of medical knowledge integrated into its system, MedComp aims to simplify complex health information into easy-to-understand advice while offering personalized care.

A web application and chatbot backend built using Vite for the frontend and backend and FastAPI for the chatbot backend.

---

Table of Contents

• Features
• Bot in Action
• System Architecture Overview
• Future Scope
• Tech Stack
• Installation
• Setting Up Environment Variables
• Running the Project
  • Frontend
  • Backend
• Build and Deployment
• Contributing
• License

---

Features

System Features

• Users can share articles, tips, and experiences about
• A blog platform for people to exchange ideas and learn from others' experiences.
• Users can create and manage schedules for:
  • Medication reminders.
  • Workout plans.
  • Health checkups or doctor appointments.
• An intuitive interface to help users align their daily health plans

Bot Features

• Users can ask the chatbot any questions related to:
  • Diseases and symptoms.
  • Psychology and mental health.
  • Nutrition advice for healthy living.
  • Tips for improving lifestyle.
• The chatbot analyzes user schedules and provides personalized recommendations to improve health and wellness.
• Suggests optimized plans for fitness, meals, and mental well-being.
• Provides information about diseases, their symptoms, prevention, and treatment options.
• Answers questions in easy-to-understand language, avoiding medical jargon.
• Offers personalized advice to help users maintain a healthy lifestyle.
• Answers questions about food choices, fitness routines, and mental relaxation techniques.

---

Tech Stack

• Frontend: Vite, React
• Backend: Express, FastAPI, Python 3.11
• Package Manager: npm (for frontend and backend dependencies), pip (for chatbot backend dependencies)
• Database: MongoDB

---

Bot in Action

Watch the below video to get a better idea on how bot works

---

System Architecture Overview

Common System Chatbot System

---

Future Scope

• Optimized Prompt and Tool Descriptions

  • Design more specific and effective prompts to enhance chatbot performance.
  • Refine descriptions of tools to ensure clarity and usability.

• Admin Dashboard

  • Develop a dedicated dashboard for administrators to manage users, content, and system settings efficiently.

• Search for Missing Information

  • Enable the chatbot to search the Internet for missing information or provide guidance when it lacks data to respond accurately.

• Web Deployment

  • Deploy the web application for public access on platforms like Vercel, Netlify, or a dedicated server.

• Image and Voice Processing

  • Add support for users to upload images for analysis or interpretation.
  • Implement voice interaction capabilities for queries and responses.

• Schedule Database Management

  • Introduce features for users to modify their schedules directly, such as deleting or updating entries in the database.

---

Installation

Prerequisites

• Node.js (v16 or newer)
• Python 3.11
• pip for Python dependency management

Steps

1. Clone the repository:

   git clone https://github.com/dngan0365/HealthcareBlogChat
   cd HealthcareBlogChat
   

2. Install dependencies:

   • Frontend:

     cd client
     npm install
     

   • Backend:

     cd backend
     npm install
     

   • Chatbot Backend:

     cd chatbot-backend
     pip install -r requirements.txt
     

---

Setting Up Environment Variables

1. Frontend:

   • Create a .env file in the client directory.
   • Add your environment variables:

     VITE_API_URL=http://localhost:3000
     VITE_IK_URL_ENDPOINT=your-vite-ik-url-endpoint
     VITE_IK_PUBLIC_KEY=your-vite-ik-public-key 
     VITE_IK_SECRET_KEY=your-vite-ik-secret-key
     VITE_CLERK_PUBLISHABLE_KEY=your-vite-clerk-publishable
     VITE_API_URL=your-vite-api-url
     VITE_GEMINI_PUBLIC_KEY=your-gemini-public
     LICENSE_KEY_SYNCFUSION=your-license-key-syncfusion

   • For IK (support images): ImageKit
   • For Syncfusion (support schedule): Syncfusion

2. Backend:

   • Create a .env file in the backend directory.
   • Add your environment variables:

     MONGO=your-mongo-url
     CLIENT_URL="http://localhost:5173"
     CLERK_WEBHOOK_SECRET=your-webhook-secret
     CLERK_PUBLISHABLE_KEY=your-publishable-key
     CLERK_SECRET_KEY=your-clerk-secret-key
     IK_URL_ENDPOINT=your-ik-url-endpoint
     IK_PUBLIC_KEY=your-ik-public-key
     IK_PRIVATE_KEY=your-ik-private-key 

   • Mongo for database: Mongo
   • Clerk for Authentication: Cleark
   • IK fro images: ImageKit

3. Chatbot Backend:

   • Create a api_keys.yaml file in configs folder in the chatbot-backend directory.

   • Add your environment variables:

     api_keys.yaml

     openai:
         api_key: your-key
     weaviate:
         url: your-weaviate-url
         api_key: your-weaviate-api-key
     mongodb:
         url: your-mongo-url
     

4. Ensure the api_keys.yaml files are listed in .gitignore to prevent them from being pushed to the repository.

---

Running the Project

Frontend

1. Start the development server:

   cd client
   npm run dev
   

2. Open your browser and visit: http://localhost:5173

Backend

1. Start the FastAPI server:

   cd backend
   npm start
   

2. Open your browser and visit: http://localhost:3000

Chatbot Backend

1. Start the FastAPI server:

   cd backend
   cd src
   uvicorn backend.main:app --reload --host 0.0.0.0 --port 8000
   

2. API documentation will be available at:

   • Swagger UI
   • ReDoc

---

Build and Deployment

Frontend

1. Build for production:

   cd frontend
   npm run build
   

2. The production-ready files will be in the dist folder.

Backend

1. Use a production-grade server like gunicorn or uvicorn with multiple workers:

   python -m backend.uvicorn main:app --host 0.0.0.0 --port 8000