README.md

---
title: Docker JEE
emoji: 📝
colorFrom: blue
colorTo: green
sdk: docker
pinned: false
---

# JEE PDF Extraction API

FastAPI service for deterministic Linux PDF extraction. It uses Poppler and Pillow to crop question images, then returns a ZIP containing:

- `metadata.json`
- `q-1.png`, `q-2.png`, ...

## Local Run

```bash
docker build -t jee-pdf-extractor .
docker run --rm -p 7860:7860 -e ALLOWED_ORIGINS=http://localhost:5173 jee-pdf-extractor
```

## Hugging Face Space

Create a new Docker Space and upload the contents of this folder.

Set Space variables:

```env
ALLOWED_ORIGINS=https://your-frontend-domain.vercel.app,http://localhost:5173
EXTRACTION_TIMEOUT_SECONDS=900
```

Health check:

```text
https://YOUR-SPACE.hf.space/health
```

Frontend env:

```env
VITE_EXTRACTOR_API_URL=https://YOUR-SPACE.hf.space
```