/cognitive-search-vector-pr

A repository of code samples for Vector search capabilities in Azure Cognitive Search.

Primary LanguageJupyter NotebookMIT LicenseMIT

Vector search (public preview) - Azure Cognitive Search

DISCLAIMER Preview functionality is provided under Supplemental Terms of Use, without a service level agreement, and isn't recommended for production workloads.

This repository provides code samples for the vector search (preview) in Azure Cognitive Search.

The demos currently use beta versions of the client libraries for the Azure SDKs.

Sample Description
.NET A .NET Console App that calls Azure OpenAI to create vectorized data. It then calls Azure Cognitive Search to create, load, and query the data.
Python Five notebooks that demonstrate aspects of vector search, including data chunking and vectorization of image content.
JavaScript There are three code samples. One is an end-to-end code sample that calls Azure OpenAI for embeddings and Azure Cognitive Seach to create, load, and query an index that contains vectors. Another sample calls just Azure OpenAI and is used to generate embeddings for fields in an index. The last one also calls just Azure OpenAI and is used to generate an embedding for a vector query.
Postman collection A collection of REST API calls to an Azure Cognitive Search instance to create, load, and query a search index that contains text and vector fields. The requests in this collection include an index schema, sample documents, and sample queries. The collection is reference in Quickstart: Vector search. Each query demonstrates key scenarios.

Use the Postman app and import the collection.

Set collection variables to provide your search service URI and admin key

If you're unfamiliar with Postman, see this Postman/REST quickstart for Cognitive Search.

Related resources