Skip to content

Azure AI Document Intelligence Custom Extraction Connected Containers Helm Chart

License

Notifications You must be signed in to change notification settings

jamesmcroft/azure-ai-document-intelligence-helm-chart

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

6 Commits
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Azure AI Document Intelligence Custom Extraction Connected Container Helm Chart

This repository contains the Helm chart for deploying the Azure AI Document Intelligence Connected Containers for use with custom extraction scenarios.

The templates in this Helm chart will deploy the following components to a Kubernetes cluster:

This Helm chart also includes a custom StorageClass used for provisioning persistent storage as Azure File Shares with nobrl mount option enabled to prevent file locking issues with the Azure AI Document Intelligence Studio sqlite database file.

Getting Started

To deploy the Azure AI Document Intelligence Connected Containers, you will need:

Pre-requisites

Deploy the Helm Chart

The following steps will guide you through deploying the Azure AI Document Intelligence Connected Containers to your Kubernetes cluster.

Note

The templates have been pre-configured with default values except for the required documentIntelligence.env.billing and documentIntelligence.env.apikey values. You can override these values using the --set option or a values.yaml file as described below. To customize the deployment further, you can modify the chart's default values.yaml file in the ai-document-intelligence directory.

kubectl create namespace di
helm install di-extraction ai-document-intelligence --namespace di --set documentIntelligence.env.billing.value=your-document-intelligence-endpoint-value --set documentIntelligence.env.apikey.value=your-document-intelligence-apikey-value

When using secret values, you can configure the billing endpoint and API key with the --set documentIntelligence.env.billing.valueFrom.secretKeyRef and --set documentIntelligence.env.apikey.valueFrom.secretKeyRef options with both the required name and key values.

Alternatively, you can use a values.yaml file to configure the deployment. See more on Helm values files.

Access the Azure AI Document Intelligence Studio

By default, this Helm chart deploys only the containers required and does not expose any services to the public internet. To access the Azure AI Document Intelligence Studio and nginx proxy, you can use kubectl port-forward to forward the service ports to your local machine.

kubectl port-forward svc/di-extraction-ai-document-intelligence-nginx 5000:5000 --namespace di
kubectl port-forward svc/di-extraction-ai-document-intelligence-studio 5001:5001 --namespace di

You can then access the Studio UI at http://localhost:5001. When creating a new custom extraction project, the Form Recognizer Service Endpoint value will be set to the nginx proxy URL http://localhost:5000.

Note

In a real-world scenario, you would expose the services using an ingress controller, service mesh, or other application gateway to control access to the services and provide a secure connection. This is not covered in this Helm chart as you may already be implementing this. Please refer to the Kubernetes documentation for more information.

FAQ

How do I configure tolerations, affinity, and node selectors?

Although not included in the default values.yaml file, you can configure these settings using the Helm --set option or a values.yaml file for each of the deployment configurations. The values include tolerations, affinity, and nodeSelector for each of the deployments.

Here is an example values.yaml for how to configure the layout deployment with tolerations, affinity, and node selectors:

layout:
  tolerations:
    - key: "key"
      operator: "Equal"
      value: "value"
      effect: "NoSchedule"
  affinity:
    nodeAffinity:
      requiredDuringSchedulingIgnoredDuringExecution:
        nodeSelectorTerms:
          - matchExpressions:
              - key: "key"
                operator: "In"
                values:
                  - "value"
  nodeSelector:
    key: "value"

About

Azure AI Document Intelligence Custom Extraction Connected Containers Helm Chart

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages