Skip to content

Latest commit

 

History

History
 
 

postman-collection

README.md

page_type languages name description products urlFragment
sample
rest
Vector REST APIs for Azure AI Search
Demonstrates generally available and preview REST APIs for indexing and querying vectors in Azure AI Search.
azure
azure-cognitive-search
rest-api-vectors

Vector REST APIs for Azure AI Search

Flask sample MIT license badge

Use Postman and the Azure AI Search REST APIs to work with vector and hybrid queries.

There are two collections:

Collection Description
AzureSearchVectors_2023-11-01 Demonstrates vector indexing and queries.
AzureSearchVectors_2023-10-01-preview Adds integrated vectorization, with data chunking and embedded calls to Azure OpenAI. This workflow adds an indexer, data source, and skillset. It uses PDFs for the sample data.

Prerequisites

  • Postman Desktop app
  • Azure AI Search service
  • (optional) Semantic ranking for queries that use it.
  • Azure OpenAI with a deployment of text-embedding-ada-002 is required for the preview collection and optional for the GA version. It's optional for the GA version because requests in the collection include pre-vectorized content for queries and indexing.
  • (optional) Azure Storage for the preview API workflow. Azure Storage provides the PDFs used in data chunking and embedded vectorization.

Setup

  1. Clone or download this sample repository.
  2. Extract contents if the download is a zip file. Make sure the files are read-write.

Import and set collection variables

  1. For the preview collection only, upload the sample health plan PDFs to a blob container on Azure Storage. Make a note of the container name and a connection string so that you can provide these values as variables.
  2. Start Postman and import either one of the collections.
  3. Select the collection, open the actions menu, select Edit.
  4. Enter variables:
    • Index or prefix name to apply a naming convention to the created objects.
    • Search service name and an admin API key, which you can obtain from the Azure portal.
    • Azure OpenAI service name, key, REST API version, and model deployment name.
    • For the preview collection, provide an Azure Storage blob container name and connection string.
  5. Select Save.
  6. Send each request to the service.

Next steps

You can learn more about Azure AI Search on the official documentation site.