Skip to content

DEPLOYMENT.md

Latest commit

 

History

History
175 lines (118 loc) · 3.68 KB

DEPLOYMENT.md

File metadata and controls

175 lines (118 loc) · 3.68 KB

NeuroScan — Deployment Guide

This guide covers the current production deployment and how to redeploy or run locally.

Current Production Deployment

Service Platform URL
Frontend Vercel neuroscan.vercel.app
Backend API Hugging Face Spaces yashnaiduu-neurosacn.hf.space

Prerequisites

  • Python 3.9+
  • Docker (for containerized deployment)
  • Git
  • Hugging Face account (for backend)
  • Vercel account (for frontend)

Environment Variables

Variable Required Description
PORT Optional Backend port (default: 5050, HF Spaces: 7860)
FLASK_ENV Optional Set to production in deployment
MODEL_URL Optional Remote URL to download the .h5 model file

No external API keys are required. MRI validation uses CLIP (runs locally).

Backend — Hugging Face Spaces

The backend runs as a Docker container on Hugging Face Spaces.

Redeploy

# Clone the HF Space
git clone https://yashnaiduu:YOUR_HF_TOKEN@huggingface.co/spaces/yashnaiduu/neurosacn hf_space
cd hf_space

# Copy updated files
cp ../server1.py ../requirements.txt ../Dockerfile ../entrypoint.sh .

# Commit and push (triggers rebuild)
git add . && git commit -m "Update deployment" && git push

Key Files

File Purpose
Dockerfile Container definition
entrypoint.sh Startup script (downloads model, starts gunicorn)
server1.py Flask application
requirements.txt Python dependencies

Health Check

curl https://yashnaiduu-neurosacn.hf.space/health

Expected response:

{
  "status": "healthy",
  "model_loaded": true,
  "clip_available": true,
  "uptime": 3600.5
}

Frontend — Vercel

The frontend is a static HTML/CSS/JS site deployed on Vercel.

Redeploy

Push to the main branch on GitHub — Vercel auto-deploys on every push.

git add client/
git commit -m "Update frontend"
git push

Configuration

vercel.json handles routing — all requests are served from client/index.html.

Local Development

# Clone the repository
git clone https://github.com/yashnaiduu/NeuroScan-Brain-Tumor-Classification.git
cd NeuroScan-Brain-Tumor-Classification

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

# Install dependencies
pip install -r requirements.txt

# Run backend
python server1.py               # http://localhost:5050

# Run frontend (separate terminal)
cd client && python3 -m http.server 8000   # http://localhost:8000

Local Docker

# Build
docker build -t neuroscan:latest .

# Run
docker run -d \
  --name neuroscan \
  -p 5050:5050 \
  -e FLASK_ENV=production \
  neuroscan:latest

# Check health
curl http://localhost:5050/health

Troubleshooting

Model Not Loading

Symptom: /health returns model_loaded: false

  1. Check MODEL_URL is accessible and points to a valid .h5 file
  2. Check container logs for download errors
  3. Ensure sufficient disk space (model is ~15MB)

High Memory Usage

Symptom: Container crashes or restarts

  1. Increase container memory to 2GB minimum
  2. Reduce WEB_CONCURRENCY in entrypoint.sh
  3. Monitor with /stats endpoint

Slow Predictions

Symptom: Requests take longer than expected

  1. CLIP model loads in the background on startup — first request may be slower
  2. Increase GUNICORN_TIMEOUT if needed
  3. CPU inference is expected; GPU instances will be significantly faster

Support

Yash Naidu — yashnnaidu@gmail.com