Retail Ingestion

Creating a Search Index using Azure AI Search SDK

Prerequisites

Before starting, ensure you have:

An Azure account with an Azure AI Search service created. The service name, admin API key, and endpoint URL.

You can find information on setting up your Azure and all needed requirements here.

Install SDK

Install libraries

Import Libraries

Define Search Index Schema

Put the name of your search index in placeholder below. If you have created search index just update the name in placeholder below and jump to Data Ingestion. Otherwise follow steps below to create your index.

Introduce Vector Search: Define Vector Search Parameters

Introduce Semantic Search: Define Semantic configuration based on Index schema

Create Index in Azure AI Search

Data Ingestion

For ingesting data you will need two components:

1. Images of catalog items: These will be stored in blob storage. It is important that name of each image be exactly articleID of corresponding catalog item. These images are crucial for enhancements during ingestion. Images are provided here
   
2. Listing of product : This listing will be ingested into Azure AI Search. Listing is provided here

Ingest data into Azure AI Search

Search

This part is for testing your ingested data and is not needed for ingesting date directly.

Setup Search Client

Setup Utility Functions

Search Using Vector Similarity