Azure Text Analytics Integration

Introduction

Azure Text Analytics is a cloud-based service that provides advanced natural language processing over raw text. One of its main functions includes Named Entity Recognition (NER), which has the ability to identify different entities in text and categorize them into pre-defined classes or types. This document will demonstrate Presidio integration with Azure Text Analytics.

Supported entity categories in the Text Analytics API

Text Analytics supports multiple PII entity categories. The Text Analytics service runs a predictive model to identify and categorize named entities from an input document. The service's latest version includes the ability to detect personal (PII) and health (PHI) information. A list of all supported entities can be found in the official documentation.

Prerequisites

To use Text Analytics with Preisido, an Azure Text Analytics resource should first be created under an Azure subscription. Follow the official documentation for instructions. The key and endpoint, generated once the resource is created, will be used when integrating with Text Analytics, using a Presidio Text Analytics recognizer.

Text Analytics Recognizer

Sample implementation of a TextAnalyticsRecognizer. The sample suggests an implementation of a Remote Recognizer calling the Text Analytics service REST API. There is also an alternative solution of using azure-ai-textanalytics python package, replacing the client implementation in the example. The sample reads from a yaml file, defining the entities that should be recognized by Text Analytics, and their corresponding Preisdio entity types. Use the supported entity categories reference above to extend the list of entities defined in the file, with the entities that you want Text Analytics to recognize.