Skip to content

Latest commit

 

History

History
119 lines (83 loc) · 6.27 KB

readme.md

File metadata and controls

119 lines (83 loc) · 6.27 KB
Raw
page_type languages name description products urlFragment
sample
csharp
Vector storage and retrieval in C#
Using Azure.Search.Documents, index and query vectors in a RAG pattern or a traditional search solution.
azure
azure-cognitive-search
csharp-vector-search

Vector search in C# (Azure AI Search)

In this .NET console application for Azure AI Search, DotNetVectorDemo provides raw data for which embeddings are generated externally and then pushed into a search index for queries. First, it calls Azure OpenAI resource and a deployment of the text-embedding-3-small model to create embeddings for text in a local text-sample.json file. Next, it pushes the embeddings and other textual content to a search index. The searchable output is a combination of human-readable text and embeddings that can be queried from your code. Note that a pre-populated set of these embeddings is already provided in vector-sample.json.

Prerequisites

  • An Azure subscription, with access to Azure OpenAI service. You must have the Azure OpenAI service endpoint and an API key if you are not using RBAC authentication.

  • A deployment of the text-embedding-3-small embedding model. For the deployment name, the deployment name is the same as the model, "text-embedding-3-small".

  • Model capacity should be sufficient to handle the load. We successfully tested these samples on a deployment model having a 33K tokens per minute rate limit.

  • An Azure AI Search service with room for a new index. You must have full endpoint and an admin API key if you are not using RBAC authentication.

  • Azure SDK for .NET 5.0 or later. The project specifies 11.7.0-beta.1 to use preview features

You can use Visual Studio or Visual Studio Code with the C# extension for these demos.

Setup

  1. Clone this repository.

  2. Create a local.settings.json file in the same directory as the code for each project and include the following variables:

     {
   "AZURE_SEARCH_SERVICE_ENDPOINT": "YOUR-SEARCH-SERVICE-ENDPOINT",
   "AZURE_SEARCH_INDEX_NAME": "YOUR-SEARCH-INDEX",
   "AZURE_OPENAI_ENDPOINT": "YOUR-AZURE-OPENAI-ENDPOINT"
 }

    Here's an example with fictitious values:

    {  
 "AZURE_SEARCH_SERVICE_ENDPOINT": "https://demo-srch-eastus.search.windows.net",  
 "AZURE_SEARCH_INDEX_NAME": "demo-vector-index",  
 "AZURE_OPENAI_ENDPOINT": "https://demo-openai-southcentralus.openai.azure.com/",

You can find more values to set by checking Configuration.cs

Run the code

Before running the code, ensure you have the .NET SDK installed on your machine.

  1. If you're using Visual Studio Code, select Terminal and New Terminal to get a command line prompt.

  2. For each project, navigate to the project folder (e.g., demo-dotnet/DotNetVectorDemo or demo-dotnet/DotNetIntegratedVectorizationDemo) in your terminal and execute the following command to verify .Net 5.0 or later is installed:

    dotnet run -- -h
    dotnet run -- -h
Description:
  .NET Vector demo

Usage:
  DotNetVector [options]

Options:
  --setup-index      Indexes sample documents. text-embedding-3-small embeddings with a dimension of 1024 are used
  --query <query>    Optional text of the search query. By default no query is run. Unless --textOnly is specified, this query is automatically vectorized. []
  --filter <filter>  Optional filter of the search query. By default no filter is applied []
  -k <k>             How many nearest neighbors to use for vector search. Defaults to 50 [default: 50]
  --top <top>        How nany results to return. Defaults to 3 [default: 3]
  --exhaustive       Optional, specifies if the query skips using the index and computes the true nearest neighbors. Can only be used with vector or hybrid queries. [default:
                     False]
  --text-only        Optional, specifies if the query is vectorized before searching. If true, only the text indexed is used for search. [default: False]
  --hybrid           Optional, specifies if the query combines text and vector results. [default: False]
  --semantic         Optional, specifies if the semantic reranker is used to rerank results from the query. [default: False]
  --debug <debug>    Optional, specifies if debug output is included from the query. Only valid values are disabled (default), semantic, or vector [default: disabled]
  --version          Show version information
  -?, -h, --help     Show help and usage information

  3. Setup a sample index using the following command. It will use a file of pre-populated embeddings created by text-embedding-3-small with a dimension of 1024

    dotnet run -- --setup-index

  4. Run a test query using the following examples:

dotnet run -- --query "What is Azure Search" --hybrid --semantic

The output will be the chunks matched by the query and their relevance score. If semantic is specified, captions and answers will be included in the output if they are available.

Output

Output is a search index. You can use the Azure portal to explore the index definition or delete the index if you no longer need it.

Troubleshoot errors

If you get error 429 from Azure OpenAI service, it means the resource is over capacity:

  • Check the Activity Log of the Azure OpenAI service to see what else might be running.

  • Check the Tokens Per Minute (TPM) on the deployed model. On a system that isn't running other jobs, a TPM of 33K or higher should be sufficient to generate vectors for the sample data. You can try a model with more capacity if 429 errors persist.

  • Review these articles for information on rate limits: Understanding rate limits and A Guide to Azure OpenAI Service's Rate Limits and Monitoring.