Introduction

  • 3 minutes

Azure AI Language is a cloud-based Natural Language Processing (NLP) service. It allows users to understand and analyze text with features such as key phrase extraction, entity recognition, personally identifiable information (PII) detection, and more. Learn more here: What is Azure AI Language?.

The azure_ai extension for Azure Database for PostgreSQL flexible server integrates the database with the Azure AI Language services. It provides user-defined functions to access the language APIs within SQL. This access allows you to extract insights from text directly from the database without writing natural language processing code in client applications.

Example scenario

Consider a company that provides vacation listings. The company's marketing team wants to better understand customer preferences by identifying the key phrases in the most popular listings and highlighting which businesses, attractions, and places are highlighted. Also, the company's information security team wants to scan listings for personally identifiable information (PII) to ensure safety and privacy.

Learning objectives

To accomplish our three tasks, you use the azure_ai extension to integrate an Azure Database for PostgreSQL flexible server with Azure AI Language. First, you extract the key phrases from listing descriptions and store them in the database, enabling the data science team to analyze which appear the most in popular listings. Then, you extract & store named entities for the same purpose. Lastly, you flag all PII listings and store the redacted text.

The main goal is to understand how to use the azure_ai PostgreSQL extension to access these Azure AI Language services: key phrase extraction, entity recognition, and PII detection. You learn how to access these APIs in SQL and store the result data in columns.

Setup: enable and authorize azure_ai

This learning path uses the azure_ai extension. There are a few steps to install and configure it. You need to add it to your allowlist as described in how to use PostgreSQL extensions. Then, install the extension by running this SQL query:

CREATE EXTENSION azure_ai;

If you want to remove the extension, use:

DROP EXTENSION azure_ai;

Installing the extension creates these three schemas:

  • azure_ai: the main schema storing configuration data & functions.
  • azure_openai: functions and composite types related to OpenAI.
  • azure_cognitive: functions and composite types related to Cognitive Services.

Our tasks use Azure Cognitive Services. After enabling and configuring the azure_ai extension, you can integrate it with Azure Cognitive Services to access key phrase extraction from SQL.

If you don't already have one, create a Language resource in the Azure portal. Once you have a Language resource, go to Resource Management > Keys and Endpoint to get your key and endpoint for Azure Cognitive Services.

Next, authorize the Azure Database for PostgreSQL flexible server's azure_ai extension by running this SQL:

select azure_ai.set_setting('azure_cognitive.endpoint','https://<endpoint>.cognitiveservices.azure.com');
select azure_ai.set_setting('azure_cognitive.subscription_key', '<API Key>');
-- the region setting is only required for the translate function
select azure_ai.set_setting('azure_cognitive.region', '<API Key>');

More information is available in the Azure Cognitive Services documentation.