VARK Behavior Tracker Extension

This VS Code extension tracks your coding behavior and uses machine learning to predict your VARK learning style. Now with full user authentication!

Features

User Authentication: Secure login/register with email and password
Behavior Tracking: Automatically tracks:
- Keystrokes (characters typed)
- Scroll events
- Compilation/task execution attempts
- Code comments ratio
VARK Prediction: ML-powered learning style prediction
Behavior History: Access and analyze previous behavior data
Secure Token Storage: Uses VS Code's secure secret storage
Periodic Sync: Sends data to backend every 5 minutes

Installation & Setup

Prerequisites

Node.js 14+ and npm
VS Code 1.80.0+
Backend server (included in this repo)

Backend Setup

Install backend dependencies:

cd backend
npm install

Create .env file in backend directory:

PORT=3000
MONGO_URI=mongodb://localhost:27017/vark-tracker
JWT_SECRET=your-secret-key-change-in-production

Start the backend:

npm start

Backend will run on http://localhost:3000

Extension Setup

Install extension dependencies:

cd vscode-extension
npm install

Launch in VS Code:
- Press F5 to start in debug mode, or
- Use vsce package && code --install-extension vark-behavior-tracker-1.0.0.vsix

Usage

First-Time Users

Launch the extension - login screen appears
Click "Register" to create account
Enter name, email, and password
Confirm password and submit
Behavior tracking starts automatically

Returning Users

Click "Login"
Enter email and password
Previous behavior data loads automatically
Coding behavior tracking resumes

Available Commands

Press Ctrl+Shift+P (Windows/Linux) or Cmd+Shift+P (macOS) and search for:

VARK: Login - Login to your account
VARK: Register - Create new account
VARK: Logout - Logout and disable tracking
VARK: Send Behavior Data Now - Manual data submission

How It Works

Data Collection

The extension monitors your VS Code activity:

Keystrokes: Every character you type
Scrolling: Viewport changes while editing
Compilation: Build commands, debug sessions, task runs
Comments: Lines of code that are comments

Data Processing (Every 5 Minutes)

Aggregate collected metrics
Calculate comment ratio
Send to backend with JWT token

VARK Prediction

Backend processes data and calls ML service:

Visual (V) - Tends to scroll more, visual code patterns
Aural (A) - Collaborative patterns, discussions
Reading/Writing (R) - Comment-heavy, documentation
Kinesthetic (K) - Hands-on testing, frequent compilation

Learning Style Update

Your predicted learning style updates after each analysis cycle.

Architecture

VS Code Extension
    ↓
    ├── Authenticate (Login/Register)
    ├── Secure Token Storage
    ├── Behavior Tracking
    │   ├── Keystrokes
    │   ├── Scrolls
    │   ├── Compile Commands
    │   └── Comments
    ↓
Backend Server (Port 3000)
    ├── JWT Authentication
    ├── User Management
    ├── Behavior Storage (MongoDB)
    ├── Behavior History Retrieval
    ↓
ML Service (Port 8001)
    └── VARK Prediction

API Endpoints

Authentication

POST /api/auth/signup - Create account
POST /api/auth/login - Login
GET /api/auth/me - Get current user (requires token)

Behavior

POST /api/behavior - Submit behavior data
GET /api/behavior - Get user's behavior history (requires token)

Tracked Metrics Detail

Metric	Description
keystrokes	Total characters typed in session
scrollEvents	Number of viewport/scroll changes
compileCount	Debug starts or task executions
commentRatio	Comment lines / total lines
sessionDuration	Seconds since last submission
userId	Unique user identifier (from JWT)

Troubleshooting

"Connection refused" error

Problem: Backend not running
Solution: Start backend: cd backend && npm start

"Authentication failed"

Problem: Wrong credentials or server error
Solution:
- Verify email/password
- Check backend is running
- Check backend logs for errors

No behavior data saving

Problem: Network issue or server error
Solution:
- Check internet connection
- Verify backend running on port 3000
- Manually send: VARK: Send Behavior Data Now

VARK prediction shows "unknown"

Problem: ML service not available
Solution: Start ML service or use prediction stub
- Behavior still saves, prediction attempted next cycle

Token expired

Problem: Session expired after 7 days
Solution: Logout and login again with VARK: Logout then VARK: Login

Development

Debug Mode

Press F5 with this folder open
New VS Code window opens with extension loaded
Set breakpoints in extension.js
Use Debug Console for logs

Key Files

extension.js - Main extension logic
package.json - Extension manifest and dependencies

Testing Workflow

Start backend: npm start (in backend/)
Press F5 to launch extension
Register account in test extension window
Type/scroll in files to generate behavior
Check backend logs: [Behavior] ML response:

Security

Passwords: Hashed with bcrypt (never stored plaintext)
Tokens: JWT with 7-day expiration
Secure Storage: VS Code's native secure storage (OS-level encryption)
HTTPS Ready: API URLs can be updated for secure connections

Future Enhancements

Real-time VARK visualization dashboard
Peer comparison and leaderboards
Learning recommendations based on style
Export behavior data reports
Multi-workspace support
IDE plugins for other editors (VS, IntelliJ, etc.)

Support

For issues or questions:

Check troubleshooting section above
Review backend logs: console.log output
Check VS Code extension output: Help → Toggle Developer Tools

License

Final Year Research Project - Learning Style Identifier

Neuromentor

neuromentor