Trackit – Wireless Telemedicine & ECG/PPG Triage System

Trackit is a real-time wireless emergency telemedicine web application. It streams telemetry waveforms (ECG at 256Hz, PPG at 64Hz) to a Java backend via WebSockets, forwards windows to a Python PyTorch ML inference microservice for triage assessment, and displays real-time health indicators and alert escalations on a premium styled frontend dashboard.

---

System Architecture

• Frontend: Vanilla HTML5, CSS3 (custom HSL theme, CSS grid layouts, micro-animations), and high-performance JavaScript Canvas drawing with requestAnimationFrame.
• Backend Service: Java EE web application deployed on Apache Tomcat 10.x. Uses WebSockets (EcgMonitorEndpoint) for sample ingestion and REST for authentication and patient administration.
• Inference Service: Python FastAPI microservice executing a PyTorch wearable triage classification model on CPU.
• Database: MySQL relational store.

---

1. Database Setup

1. Start MySQL Server: Ensure MySQL server is running locally (default port 3306).
2. Import Database Schema: Create the database and tables using the schema script located at:
   • schema.sql
   • You can import it using your MySQL client CLI or UI tool:

     mysql -u root -p < schema.sql

3. Database Configuration:
   • Connection configurations are loaded from the environment variables or the .env file in the root workspace directory.
   • If not set in the environment, the application falls back to defaults defined in:
     • database.properties
     • Default configuration:
       • URL: jdbc:mysql://localhost:3306/trackit?useSSL=false&allowPublicKeyRetrieval=true&serverTimezone=UTC
       • Username: root
       • Password: (empty)

---

2. Python Inference Service Setup

1. Navigate to Directory:

   cd inference-service

2. Setup Virtual Environment & Install Dependencies:

   python -m venv .venv
   # Windows:
   .venv\Scripts\activate
   # macOS/Linux:
   source .venv/bin/activate
   
   pip install -r requirements.txt

3. Start FastAPI Uvicorn Server:

   uvicorn app.main:app --host 0.0.0.0 --port 8001

   Note: The ML service runs on port 8001 and exposes the /infer and /health endpoints.

---

3. Java Web App Build & Deploy

1. Compile & Package: From the root workspace directory, run Maven packaging to compile all source files and generate the WAR deployment package:

   mvn clean package -DskipTests

2. Start the Application Container (Tomcat): Launch the embedded Tomcat 10.x container using Maven Cargo. Ensure you inject the INFERENCE_SERVICE_URL variable:
   • PowerShell (Windows):

     $env:INFERENCE_SERVICE_URL="http://localhost:8001"; mvn cargo:run

   • Bash (macOS/Linux/Git Bash):

     INFERENCE_SERVICE_URL="http://localhost:8001" mvn cargo:run

   • The application context is deployed at: http://localhost:8080/trackit/

---

4. Default Credentials & Flow Validation

Default Doctor Credentials

At application startup, if the database is clean, a default super-administrator user is seeded with the credentials defined in .env:

• Email: admin@trackit.com
• Password: AdminSecurePassword123!

Telemetry Simulation Flow

1. Open the browser and visit http://localhost:8080/trackit/login.html.
2. Sign in using the doctor credentials.
3. Select Find Patient to search for existing patients, or New Patient to register a new record.
4. Open the Patient Connection dashboard tab.
5. In the left panel:
   • Search/select the patient name (uses autocomplete).
   • Click the file selector and select a simulation CSV file from the root directory.
   • Recommended simulation file: alternating_conditions.csv (combines segments of normal rhythm, tachycardia, afib, and bradycardia).
   • Click Connect to patient.
6. The dual-track canvas will display real-time scrolling ECG (256Hz) and PPG (64Hz) waveforms.
7. Triage results from the ML model will trigger every 5 seconds (zero-padded for the first few cycles until the buffer matches the 30-second model layout), showing rhythm analysis and escalations on the right vital signs board.