---
title: document-ocr-demo
app_file: src/app.py
sdk: gradio
sdk_version: 4.26.0
---
# Document OCR Demo

This repository contains a web demo for extracting key-value pairs from documents using Azure Document Intelligence.

## Getting Started

### Prerequisites

- Git
- Docker and Docker Compose

### Installation

1. **Clone the repository:**

   ```bash
   git clone git@github.com:vincent-lo-greenaitech/document-ocr-demo.git
   ```

2. **Configure environment variables:**

   Locate the `.env.template` file in the repository, replace the placeholders with the actual Azure endpoint and key, and rename the file to `.env`.

   For example, convert:

   **.env.template**

   ```plaintext
   AZURE_ENDPOINT=<YOUR_ENDPOINT>
   AZURE_KEY=<YOUR_KEY>
   ```

   To:

   **.env**

   ```plaintext
   AZURE_ENDPOINT="https://123.com"
   AZURE_KEY="abc123kljfdkkvvs"
   ```

3. **Start the server with Docker Compose:**

   Ensure `docker compose` is installed and available, then run the following command:

   ```bash
   make run
   ```

   Or:

   ```bash
   docker compose up --detach
   ```

4. **Access the web demo:**

   The web demo should now be accessible at `http://localhost:7860` (or the configured port).

### Shutting Down

To shut down the server and clean up Docker containers, use the following command:

```bash
make stop
```

Or:

```bash
docker compose down
```

## Update image

To update the image and run the server again, use the following command:

```bash
make update-image-and-run
```

## Development

### Managing Python Dependencies

This project uses `poetry` for managing Python dependencies. It is recommended to install `poetry` through `pipx` for an isolated setup. Refer to the [official Poetry documentation](https://python-poetry.org/docs/) for more detailed instructions.