HelPT

Generative AI-based QR Code System for Personalized Patient Health Information

Presented at 2023 Korea Knowledge Management Society Fall Conference - Idea Competition

---

Overview

HelPT is a system that generates QR codes containing personalized health information using patient medical data. By combining Generative AI (GPT-4) with the LangChain framework, the system analyzes patient input and converts customized health information into QR codes for easy access.

Key Features

• Real-time Health Information Access: Patients can check their health status anytime, anywhere
• Intuitive Interface: Simple web-based QR code generation
• AI-powered Analysis: Personalized health information using LangChain + GPT-4
• Enhanced Security: Secure data transmission via QR codes

---

Project Structure

HelPT/
├── README.md                 # Project documentation
├── requirements.txt          # Python dependencies
├── .gitignore               # Git ignore file
├── .env.example             # Environment variables template
├── LICENSE                  # MIT License
│
├── docs/                    # Documentation
│   └── presentation.pdf     # Presentation slides
│
├── data/                    # Sample data
│   └── sample_medical_records.csv
│
└── src/                     # Source code
    └── qrsite/
        ├── manage.py
        ├── qrsite/          # Django project settings
        │   ├── settings.py
        │   ├── urls.py
        │   └── wsgi.py
        │
        └── polls/           # Main application
            ├── views.py     # View logic
            ├── models.py    # Data models
            ├── urls.py      # URL routing
            ├── agents.py    # LangChain agents
            ├── agent_config.py  # Agent configuration
            ├── tools.py     # QR code generation tool
            ├── parsers.py   # Output parser
            └── templates/   # HTML templates
                ├── home.html
                └── result.html

---

Tech Stack

Category	Technology	Description
Backend	Django 4.1	Web framework
AI Model	GPT-4 (OpenAI)	Generative AI model
AI Framework	LangChain	LLM application development framework
QR Code	qrcode	QR code generation library
Database	SQLite	Data storage

---

Quick Start

1. Clone the Repository

git clone git@github.com:jinsoo96/HelPT.git
cd HelPT

2. Create and Activate Virtual Environment

python -m venv venv

# Windows
venv\Scripts\activate

# macOS/Linux
source venv/bin/activate

3. Install Dependencies

pip install -r requirements.txt

4. Set Environment Variables

# Copy .env.example to .env
cp .env.example .env

# Open .env file and set your OpenAI API key
OPENAI_API_KEY=your-api-key-here

5. Run Database Migrations

cd src/qrsite
python manage.py migrate

6. Start the Server

python manage.py runserver

7. Access in Browser

http://127.0.0.1:8000/

---

System Architecture

┌─────────────────┐     ┌─────────────────┐     ┌─────────────────┐
│   User Input    │────▶│   LangChain     │────▶│    GPT-4       │
│  (Symptoms)     │     │   Agent         │     │   Analysis      │
└─────────────────┘     └─────────────────┘     └─────────────────┘
                                                        │
                                                        ▼
┌─────────────────┐     ┌─────────────────┐     ┌─────────────────┐
│   QR Code       │◀────│   QR Code       │◀────│   Health Info   │
│   Display       │     │   Generation    │     │   Summary       │
└─────────────────┘     └─────────────────┘     └─────────────────┘

---

Key Features

1. Health Status Input

Users can input their current symptoms or health conditions through the web interface.

2. AI Analysis

The LangChain agent utilizes GPT-4 to analyze the input and generate personalized health information.

3. Medical Data Integration

Combines existing medical records (CSV) with user input to generate comprehensive health information.

4. QR Code Generation

Converts analyzed health information into QR codes for quick access by healthcare providers.

---

Sample Data Format

Date,Visit Type,Diagnosis,Medications,Treatment
2023-01-03,Checkup,Hypertension,Tamiflu,Examination
2023-01-09,Hospital Visit,Cold,"Diabetes medication, Antibiotics",X-ray
...

---

Expected Benefits

Qualitative Effects

• Improved efficiency for healthcare professionals
• Enhanced security through QR code-based data transmission
• Integration of medical data with generative AI

Quantitative Effects

• Reduced consultation time and medical costs
• Increased utilization of medical data
• Improved communication between patients and healthcare providers

---

Contact

For questions or collaborations:

Author: Kim Jin Soo, Kang Minsuk, Kang Seyoung, Jo Yoonju

GitHub: @jinsoo96, @minsuk1003, @strongeryoung, @justlikeazoo

---

References

For presentation materials and references, please refer to the docs/ folder.

---

License

This project is licensed under the MIT License - see the LICENSE file for details.