https://github.com/MeshackMusundi/AzureCognitiveSearch-RAG

Retrieval Augmented Generation (RAG) using Azure Cognitive Search
https://github.com/MeshackMusundi/AzureCognitiveSearch-RAG

azure-cognitive-search azure-openai csharp dotnet-core retrieval-augmented-generation

Last synced: about 1 month ago
JSON representation

Retrieval Augmented Generation (RAG) using Azure Cognitive Search

• Host: GitHub
• URL: https://github.com/MeshackMusundi/AzureCognitiveSearch-RAG
• Owner: MeshackMusundi
• License: mit
• Created: 2023-06-11T09:32:26.000Z (over 1 year ago)
• Default Branch: master
• Last Pushed: 2023-06-29T07:13:05.000Z (over 1 year ago)
• Last Synced: 2024-08-01T13:19:37.746Z (4 months ago)
• Topics: azure-cognitive-search, azure-openai, csharp, dotnet-core, retrieval-augmented-generation
• Language: C#
• Homepage:
• Size: 25.4 KB
• Stars: 1
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE.txt

Awesome Lists containing this project

• awesome-semantickernel - Retrieval Augmented Generation (RAG) using Azure Cognitive Search

README

        
# Introduction

This is a simple C# demo highlighting how to have conversational interactions with private data in Azure using Azure Cognitive Search and Azure OpenAI.

Azure Cognitive Search is used for Retrieval Augmented Generation.

# Requirements

You should have the following resources set up on Azure,

- **Azure Storage Container**: This will host your private documents (PDF, Word, Excel,...) that you want to use as the knowledge base. The storage account is added as a data source of your search service.

- **Azure Cognitive Search Service**: The service should be set up in a standard tier which will ensure you can use Semantic Search. Add the storage container as a data source for the search service and set up the indexer, index, and Semantic Search configuration.

- **Azure OpenAI**: Add a model deployment to your Azure OpenAI service, either GPT 3.5 Turbo or GPT 4.

## User Secrets

Add user secrets to the console project as follows, with the neccessary values specified.

```

{

  "AzureOpenAISettings": {

    "ServiceKey": "",

    "ServiceName": "",

    "DeploymentName": "",

    "Endpoint": "https://.openai.azure.com",

    "MaxCompletionTokens": 500,

    "MaxModelTokens": 8192,

    "Instructions": "Assistant is an intelligent chatbot..."

  },

  "AzureCognitiveSearchSettings": {

    "ServiceKey": "",

    "ServiceName": "",

    "Index": "",

    "SemanticConfig": "",

    "ApiVersion": "",

    "ApiURL": "https://.search.windows.net/indexes//docs?api-version=&search={0}&queryLanguage=en-US&queryType=semantic&captions=extractive&answers=extractive%7Ccount-3&semanticConfiguration="

  }

}

```