WebPlot AutoDigitizer

Extract data points from scatterplot images as CSV. Upload an image, calibrate the axes, and download the digitized data.

Features

• Drag-and-drop image upload (PNG, JPG, WEBP)
• Auto-detected axis lines with manual fiducial adjustment
• Hybrid digitization: blob detection + template matching
• Point overlay on original image
• Copy/download results as CSV
• CPU-only, no GPU required

Quick Start

Docker (recommended)

docker-compose up
# Open http://localhost:8000

Local Development

Backend:

python3 -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt
uvicorn backend.main:app --reload

Frontend:

cd frontend
npm install
npm run dev
# Opens at http://localhost:5173

Tests:

source .venv/bin/activate
python -m pytest tests/ -v

Deployment

Railway / Render / Fly.io

1. Push to a Git repository
2. Connect the repo to your hosting provider
3. Set the build command: docker build -t app .
4. Set the start command: docker run -p 8000:8000 app
5. The app serves both API and frontend on port 8000

Environment Variables

Variable	Default	Description
PORT	8000	Server port

API

Endpoint	Method	Description
/health	GET	Health check
/api/upload	POST	Upload plot image (multipart)
/api/image/{id}	GET	Retrieve uploaded image
/api/detect-axes	POST	Auto-detect axis lines
/api/digitize	POST	Extract data points

How It Works

1. Upload: Image is validated (max 10MB, max 4000x4000, PNG/JPG/WEBP) and normalized to RGB
2. Axis Detection: Hough line transform finds axis boundaries, optional OCR reads tick labels
3. Calibration: User adjusts fiducial handles and enters axis min/max values
4. Digitization: Hybrid method runs blob detection first, falls back to template matching for non-standard markers
5. Export: Points displayed as table and available as CSV

Tech Stack

• Backend: Python 3.11+, FastAPI, OpenCV, scikit-image, numpy, scipy
• Frontend: React 18, TypeScript, Vite
• Deployment: Docker (CPU only)