Setup Guide

Prerequisites

• Node.js v20+
• Python 3.11+
• MongoDB (local or Atlas)
• Cloudinary account
• Stripe account (test mode)
• Gmail account or any SMTP provider

---

1. Clone

git clone https://github.com/REZ0AN/NutriNavigator.git
cd NutriNavigator

---

2. Backend environment

Create backend/.env using the same variable names as backend/.env.example:

# Server
PORT=4080
NODE_ENV=development

# Database
MONGODB_URI=mongodb+srv://USERNAME:PASSWORD@cluster.mongodb.net/nutrinavigator

# Auth
JWT_SECRET=generate_with_node_crypto_min_32_chars
JWT_EXPIRE=7d
COOKIE_EXPIRE=7

# Frontend URL (CORS + email verification links)
FRONT_END_URI=http://localhost:3000

# Bcrypt
SALT=10

# Cloudinary
CLOUDINARY_NAME=your_cloud_name
CLOUDINARY_API=your_api_key
CLOUDINARY_API_SECRET=your_api_secret

# Email (SMTP)
SMTP_HOST=smtp.gmail.com
SMTP_PORT=587
SMTP_MAIL=your@gmail.com
SMTP_PASS=your_app_password

# Stripe
STRIPE_API_KEY=pk_test_...
STRIPE_SECRET_KEY=sk_test_...
STRIPE_WEBHOOK_SECRET=whsec_...

# ML Server
ML_SERVER_URL=http://127.0.0.1:5000
ML_SECRET=generate_a_random_secret_here

Generate JWT_SECRET:

node -e "console.log(require('crypto').randomBytes(32).toString('hex'))"

Gmail App Password: Google Account → Security → 2-Step Verification → App Passwords. Use this instead of your real password.

---

3. Install backend dependencies

cd backend
npm install

---

4. Verify environment

cd backend
npm run verify

Checks all required env vars, MongoDB connection, Cloudinary ping, and Stripe account retrieval.

---

5. ML Server

cd mlserver

# Create and activate virtual environment
python -m venv .venv
source .venv/bin/activate      # Windows: .venv\Scripts\activate

# Install runtime dependencies
pip install -r requirements.txt

# Create mlserver/.env
PORT=5000
FLASK_ENV=development
ALLOWED_ORIGIN=http://localhost:4080
ML_SECRET=same_value_as_backend_ML_SECRET

# Train the model — only needed once
python train_model.py

# Start
python server.py

---

6. Frontend

cd frontend
npm install

Create frontend/.env:

# ML recommendations proxy through backend — this is for reference only
REACT_APP_ML_URL=http://127.0.0.1:5000

---

7. Start all three services

Terminal 1 — Backend

cd backend && npm run dev
# → http://localhost:4080

Terminal 2 — ML Server

cd mlserver && source .venv/bin/activate && python server.py
# → http://localhost:5000

Terminal 3 — Frontend

cd frontend && npm start
# → http://localhost:3000

---

8. Create your first admin user

Register through the UI and verify your email, then promote to admin:

// mongosh
use nutrinavigator
db.users.updateOne({ email: "your@email.com" }, { $set: { role: "admin" } })

Or use MongoDB Compass — find your user → edit role → set to "admin" → save.

Log out and back in — the Dashboard link appears in the user menu.

---

9. Test the ML server

curl -X POST http://127.0.0.1:5000/recommend \
  -H "Content-Type: application/json" \
  -H "X-ML-Secret: your_ml_secret" \
  -d '{"age": 28, "height": 1.72, "weight": 68, "gender": 1, "diesease": [1, 7]}'

---

API Documentation

Swagger UI available at http://localhost:4080/api/docs in development.