Deploy the sample labeling tool

The Form Recognizer sample labeling tool is an application that provides a simple user interface (UI), which you can use to manually label forms (documents) for the purpose of supervised learning. In this article, we'll provide links and instructions that teach you how to:

Run the sample labeling tool locally

The fastest way to start labeling data is to run the sample labeling tool locally. The following quickstart uses the Form Recognizer REST API and the sample labeling tool to train a custom model with manually labeled data.

Deploy with Azure Container Instances (ACI)

Before we get started, it's important to note that there are two ways to deploy the sample labeling tool to an Azure Container Instance (ACI). Both options are used to run the sample labeling tool with ACI:

Azure portal

Follow these steps to create a new resource using the Azure portal:

  1. Sign in to the Azure portal.

  2. Select Create a resource.

  3. Next, select Web App.

    Select web app

  4. First, make sure that the Basics tab is selected. Now, you're going to need to provide some information:

    Select Basics

    • Subscription - Select an existing Azure subscription
    • Resource Group - You can reuse an existing resource group or create a new one for this project. Creating a new resource group is recommended.
    • Name - Give your web app a name.
    • Publish - Select Docker Container
    • Operating System - Select Linux
    • Region - Choose a region that makes sense for you.
    • Linux Plan - Select a pricing tier/plan for your app service.

    Configure your web app

  5. Next, select the Docker tab.

    Select Docker

  6. Now let's configure your Docker container. All fields are required unless otherwise noted:

    • Options - Select Single Container
    • Image Source - Select Private Registry
    • Server URL - Set this to https://mcr.microsoft.com
    • Username (Optional) - Create a username.
    • Password (Optional) - Create a secure password that you'll remember.
    • Image and tag - Set this to mcr.microsoft.com/azure-cognitive-services/custom-form/labeltool:latest
    • Continuous Deployment - Set this to On if you want to receive automatic updates when the development team makes changes to the sample labeling tool.
    • Startup command - Set this to ./run.sh eula=accept

    Configure Docker

  7. That's it. Next, select Review + Create, then Create to deploy your web app. When complete, you can access your web app at the URL provided in the Overview for your resource.

Note

When creating your web app, you can also configure authorization/authentication. This is not necessary to get started.

Azure CLI

As an alternative to using the Azure portal, you can create a resource using the Azure CLI. Before you continue, you'll need to install the Azure CLI. You can skip this step if you're already working with the Azure CLI.

There's a few things you need know about this command:

  • DNS_NAME_LABEL=aci-demo-$RANDOM generates a random DNS name.
  • This sample assumes that you have a resource group that you can use to create a resource. Replace <resource_group_name> with a valid resource group associated with your subscription.
  • You'll need to specify where you want to create the resource. Replace <region name> with your desired region for the web app.
  • This command automatically accepts EULA.

From the Azure CLI, run this command to create a web app resource for the sample labeling tool:

DNS_NAME_LABEL=aci-demo-$RANDOM

az container create \
  --resource-group <resource_group_name> \
  --name <name> \
  --image mcr.microsoft.com/azure-cognitive-services/custom-form/labeltool \
  --ports 3000 \
  --dns-name-label $DNS_NAME_LABEL \
  --location <region name> \
  --cpu 2 \
  --memory 8 \
  --command-line "./run.sh eula=accept"

Connect to Azure AD for authorization

It's recommended that you connect your web app to Azure Active Directory. This ensures that only users with valid credentials can sign in and use your web app. Follow the instructions in Configure your App Service app to connect to Azure Active Directory.

Open source on GitHub

The OCR Form Labeling Tool is also available as an open-source project on GitHub. The tool is a web application built using React + Redux, and is written in TypeScript. To learn more or contribute, see OCR Form Labeling Tool.

Next steps

Use the Train with labels quickstart to learn how to use the tool to manually label training data and perform supervised learning.