AI Search Solution

A comprehensive AI-powered search solution based on Azure AI Search multimodal capabilities, featuring a .NET 9 backend API and a .NET 9 Razor Pages frontend with Azure AD authentication.

📝 Attribution: This .NET implementation is inspired by and based on the Azure AI Search Multimodal Sample Python repository. We've adapted the concepts and functionality to the .NET ecosystem while adding enterprise-grade features like Azure AD authentication and a modern Razor Pages frontend.

🤖 AI-Assisted Development: This project was developed with significant assistance from AI tools, demonstrating effective human-AI collaboration. Primary AI partners: Claude Sonnet 4 (Anthropic) for architecture and code generation, and GPT-5 Preview (OpenAI) for refinement and optimization. The result showcases how AI can accelerate development while maintaining high code quality and comprehensive documentation.

🏗️ Project Structure

AISearch/
├── dotnet/
│   ├── AISearch.Web/              # .NET 9 Razor Pages/MVC frontend
│   ├── AISearch.Api/              # .NET 9 Web API backend
│   └── AISearch.Core/             # Core business logic and models
├── docs/                          # Documentation
└── scripts/                       # Build and deployment scripts

📐 System Architecture

Architecture Overview:

Frontend Layer: .NET 9 Razor Pages with Azure AD authentication
API Layer: .NET 9 Web API with comprehensive REST endpoints
Business Logic: Clean architecture with dependency injection
Azure Services: AI Search, OpenAI, Document Intelligence, Storage
Authentication: Azure AD with token management and refresh
Data Flow: Document processing → Embedding generation → Vector search → AI chat

✨ Features

Frontend (.NET 9 Razor Pages/MVC)

Document Management: Upload, index, and manage multimodal documents
AI Search Interface: Semantic search with configurable parameters
Chat Interface: Interactive chat with document grounding
Index Management: Visual index creation and configuration
Azure AD Authentication: Secure login with Microsoft Identity
Material Design UI: Bootstrap-based responsive interface
Real-time Updates: AJAX-powered interactions with jQuery

Backend (.NET 9 Web API)

Document Management: Upload, index, and manage multimodal documents
AI Search: Semantic search with Azure AI Search
Vector Search: Embedding-based similarity search
Multimodal RAG: Chat interface with document grounding
Index Management: Create and manage search indexes
Swagger Documentation: Interactive API testing
Azure Integration: AI Search, OpenAI, Document Intelligence, and Storage services

🚀 Getting Started

Prerequisites

.NET 9 SDK
Visual Studio 2022 or VS Code
Azure subscription with:
- Azure AI Search service
- Azure OpenAI service
- Azure Storage account
- Azure Document Intelligence service
- Azure AD app registration

Frontend Setup (AISearch.Web)

Navigate to frontend directory:
```
cd dotnet/AISearch.Web
```

Configure application settings: Update appsettings.json with your Azure service credentials:

{
  "Logging": {
    "LogLevel": {
      "Default": "Information",
      "Microsoft.AspNetCore": "Warning",
      "Microsoft.AspNetCore.Authentication": "Information",
      "Microsoft.Identity.Web": "Information"
    }
  },
  "AllowedHosts": "*",
  "UseSimpleAuth": true,
  "ApiSettings": {
    "BaseUrl": "https://localhost:7001"
  },
  "AzureAd": {
    "Instance": "https://login.microsoftonline.com/",
    "Domain": "your-domain.onmicrosoft.com",
    "ClientId": "your-client-id",
    "TenantId": "your-tenant-id",
    "CallbackPath": "/signin-oidc",
    "SignedOutCallbackPath": "/signout-callback-oidc",
    "ClientSecret": "your-client-secret"
  },
  "DownstreamApi": {
    "BaseUrl": "https://localhost:7001",
    "Scopes": "User.Read"
  }
}

Restore dependencies:
```
dotnet restore
```
Build and run:
```
dotnet build
dotnet run
```
Access application: Open https://localhost:7002

Backend Setup (AISearch.Api)

Navigate to backend directory:
```
cd dotnet/AISearch.Api
```

Configure application settings: Update appsettings.json with your Azure service credentials:

{
  "AzureSearch": {
    "ServiceEndpoint": "https://your-search-service.search.windows.net",
    "IndexName": "knowledge-index",
    "SearchAdminKey": "your-search-admin-key",
    "OpenAIEndpoint": "https://your-openai-service.openai.azure.com",
    "OpenAIModelName": "gpt-4",
    "OpenAIDeploymentName": "gpt-4",
    "OpenAIEmbeddingModel": "text-embedding-3-large",
    "OpenAIEmbeddingDeploymentName": "text-embedding-3-large",
    "VectorDimensions": 3072,
    "OpenAIApiKey": "your-openai-api-key",
    "StorageAccountUrl": "https://your-storage.blob.core.windows.net",
    "StorageAccountKey": "your-storage-key",
    "ArtifactsContainer": "artifacts",
    "SamplesContainer": "samples",
    "KnowledgeAgentName": "knowledge-agent",
    "DocumentIntelligenceEndpoint": "https://your-doc-intelligence.cognitiveservices.azure.com/",
    "DocumentIntelligenceApiKey": "your-doc-intelligence-key"
  },
  "AzureAd": {
    "Instance": "https://login.microsoftonline.com/",
    "Domain": "your-domain.onmicrosoft.com",
    "ClientId": "your-client-id",
    "TenantId": "your-tenant-id",
    "Audience": "your-client-id"
  }
}

Build and run:
```
dotnet build
dotnet run
```
Access Swagger UI: Open https://localhost:7001/swagger

📊 Data Flow Architecture

Document processing pipeline from upload to search and chat interactions

Processing Pipeline:

Document Upload → Azure Storage + Document Intelligence
Content Extraction → Text, tables, images processing
Intelligent Chunking → Optimal content segmentation
Embedding Generation → OpenAI text-embedding-3-large
Vector Indexing → Azure AI Search with hybrid capabilities
Search & Chat → Semantic search + GPT-4 chat completions

📚 API Endpoints

Documents

POST /api/documents/upload - Upload and index documents
GET /api/documents - List all documents
GET /api/documents/{id} - Get document details
DELETE /api/documents/{id} - Delete document
POST /api/documents/{id}/reindex - Re-index document

Search

POST /api/search - Perform text/semantic search
POST /api/search/vector - Vector similarity search
POST /api/search/similar/{documentId} - Find similar documents

Chat

POST /api/chat - Chat with document grounding
POST /api/chat/stream - Streaming chat responses

Index Management

POST /api/indexes - Create new search index
GET /api/indexes - List all indexes
GET /api/indexes/{name} - Get index details
DELETE /api/indexes/{name} - Delete index

🛠️ Technologies Used

Frontend (AISearch.Web)

.NET 9: Latest .NET framework
ASP.NET Core Razor Pages/MVC: Web framework
Azure AD Authentication: Microsoft Identity integration
Bootstrap 5: Responsive CSS framework
jQuery: JavaScript library for DOM manipulation
Material Icons: Google Material Design icons
AJAX: For seamless API integration

Backend (AISearch.Api)

.NET 9: Latest .NET framework
Azure AI Search SDK: For search and indexing capabilities
Azure.AI.OpenAI: For embeddings and chat completions
Azure Storage Blobs: For document storage
Azure Document Intelligence: For document processing
Swagger/OpenAPI: API documentation

Core (AISearch.Core)

Azure.Search.Documents: Search operations
Azure.AI.DocumentIntelligence: Document processing
Azure.Identity: Authentication
Microsoft.Extensions: Dependency injection and configuration

🔧 Configuration

Development Environment

Create appsettings.Development.json for local development:

{
  "Logging": {
    "LogLevel": {
      "Default": "Information",
      "Microsoft.AspNetCore": "Warning"
    }
  },
  "UseSimpleAuth": true,
  "ApiSettings": {
    "BaseUrl": "https://localhost:7001"
  },
  "AzureAd": {
    "Instance": "https://login.microsoftonline.com/",
    "Domain": "your-dev-domain.onmicrosoft.com",
    "ClientId": "your-dev-client-id",
    "TenantId": "your-dev-tenant-id"
  }
}

Authentication Configuration

The application supports multiple authentication modes:

Enhanced Auth: Full Azure AD integration with token refresh
Simple Auth: Basic Azure AD authentication
Standard Auth: Default authentication mode

Configure in appsettings.json:

{
  "UseSimpleAuth": true,
  "UseEnhancedAuth": false
}

🚀 Running the Application

Development Mode

Start the API backend:
```
cd dotnet/AISearch.Api
dotnet run
```
API will be available at: https://localhost:7001
Start the Razor Pages frontend:
```
cd dotnet/AISearch.Web
dotnet run
```
Web app will be available at: https://localhost:7002
Access the application:
- Frontend: https://localhost:7002
- API: https://localhost:7001
- Swagger UI: https://localhost:7001/swagger

Production Deployment

Build for production:
```
dotnet publish -c Release -o ./publish
```
Configure production settings in appsettings.Production.json
Deploy to Azure App Service or your preferred hosting platform

📖 Usage Examples

Access the Web Interface

Navigate to https://localhost:7002
Sign in with your Azure AD credentials
Use the navigation menu to access different features

Upload Documents via API

curl -X POST "https://localhost:7001/api/documents/upload" \
  -H "Content-Type: multipart/form-data" \
  -F "file=@document.pdf" \
  -F "title=My Document" \
  -F "description=Document description"

Search Documents via API

curl -X POST "https://localhost:7001/api/search" \
  -H "Content-Type: application/json" \
  -d '{
    "query": "artificial intelligence",
    "top": 10,
    "threshold": 0.7
  }'

Chat with Documents via API

curl -X POST "https://localhost:7001/api/chat" \
  -H "Content-Type: application/json" \
  -d '{
    "messages": [
      {
        "role": "user",
        "content": [{"type": "text", "text": "What is machine learning?"}]
      }
    ],
    "useGrounding": true
  }'

🖥️ User Interface Features

Main Pages

Home: Dashboard with authentication status
Search: Advanced search interface with filters and configuration
Chat: Interactive chat with document grounding
Document Management: Upload and manage documents
Index Management: Create and configure search indexes

UI Components

Material Design: Consistent with modern web standards
Responsive Layout: Works on desktop and mobile devices
Real-time Feedback: Loading states and progress indicators
Error Handling: User-friendly error messages
Authentication Flow: Seamless Azure AD integration

🔒 Security Features

Azure AD Authentication: Secure user authentication
CSRF Protection: Built-in request forgery protection
XSS Prevention: Input sanitization and encoding
Token Management: Automatic token refresh
Secure Headers: Security headers configuration
Input Validation: Server-side validation
CORS Configuration: Proper cross-origin resource sharing

🏗️ Build Status

✅ Frontend (.NET 9 Razor Pages) - Build Successful
✅ Backend (.NET 9 API) - Build Successful
✅ Core Library (.NET 9) - Build Successful
⚠️ Configuration Required - Azure service credentials needed

🔗 Related Projects

Original Python Implementation

This .NET implementation is based on the excellent work from the Azure Samples team:

Azure AI Search Multimodal Sample

Original Python/FastAPI implementation
Comprehensive multimodal search capabilities
Streamlit frontend for quick prototyping
Excellent documentation and examples

Key Differences in .NET Version

Enterprise Authentication: Full Azure AD integration with token management
Razor Pages Frontend: Server-side rendered UI with modern responsive design
Type Safety: Strong typing with C# throughout the application
Performance: Optimized for .NET runtime performance characteristics
Security: Enhanced security features built into ASP.NET Core
Deployment: Ready for enterprise deployment scenarios
AI-Assisted Development: Built using modern AI development tools and practices

Development Methodology

This project demonstrates AI-Assisted Development best practices:

Human Vision + AI Implementation: Strategic decisions by humans, rapid implementation with AI
Iterative Refinement: Continuous improvement through human-AI collaboration
Quality Assurance: Human oversight ensuring production-ready code
Documentation: Comprehensive documentation generated with AI assistance

🤝 Contributing

Fork the repository
Create a feature branch
Follow the existing code style and patterns
Ensure all features maintain UI consistency
Add appropriate error handling and loading states
Test across different browsers and device sizes
Submit a pull request

📝 License

This project is licensed under the MIT License.

🆘 Support

For issues and questions:

Check the Swagger documentation at https://localhost:7001/swagger
Review the configuration section above
Ensure all Azure services are properly configured
Check the application logs for detailed error information
Verify Azure AD app registration settings

📋 Prerequisites Checklist

.NET 9 SDK installed
Azure subscription active
Azure AI Search service created
Azure OpenAI service deployed
Azure Document Intelligence service configured
Azure Storage account configured
Azure AD app registration completed
Application settings configured
Dependencies restored

🔄 .NET 9 Migration Notes

What's New in .NET 9

Performance Improvements: Enhanced runtime performance and memory management
C# 13 Features: New language features and syntax improvements
ASP.NET Core Enhancements: Better performance and new features
Enhanced Security: Improved security features and vulnerability mitigation

Migration Steps Completed

✅ Updated all project files to target net9.0
✅ Upgraded Microsoft packages to version 9.0.0
✅ Maintained Azure SDK versions (they are .NET 9 compatible)
✅ Updated documentation to reflect .NET 9

Post-Migration Steps

Install .NET 9 SDK: Download from Microsoft .NET Downloads
Restore packages: Run dotnet restore in each project directory
Build and test: Verify all projects build successfully
Update CI/CD: Update build pipelines to use .NET 9

Note: Make sure to configure your Azure services and update both appsettings.json files (Web and API projects) before running the application. The frontend requires the backend API to be running for full functionality.

AI Development Note: This project showcases the potential of AI-assisted development, combining human expertise with AI capabilities to create production-ready software efficiently and effectively.

Name		Name	Last commit message	Last commit date
Latest commit History 17 Commits
docs/img		docs/img
dotnet		dotnet
scripts		scripts
.gitignore		.gitignore
AISearch.sln		AISearch.sln
ATTRIBUTION.md		ATTRIBUTION.md
README.MD		README.MD
SETUP.md		SETUP.md
build.bat		build.bat
start.bat		start.bat
start.sh		start.sh

Folders and files

Latest commit

History

Repository files navigation

AI Search Solution

🏗️ Project Structure

📐 System Architecture

Architecture Overview:

✨ Features

Frontend (.NET 9 Razor Pages/MVC)

Backend (.NET 9 Web API)

🚀 Getting Started

Prerequisites

Frontend Setup (AISearch.Web)

Backend Setup (AISearch.Api)

📊 Data Flow Architecture

Processing Pipeline:

📚 API Endpoints

Documents

Search

Chat

Index Management

🛠️ Technologies Used

Frontend (AISearch.Web)

Backend (AISearch.Api)

Core (AISearch.Core)

🔧 Configuration

Development Environment

Authentication Configuration

🚀 Running the Application

Development Mode

Production Deployment

📖 Usage Examples

Access the Web Interface

Upload Documents via API

Search Documents via API

Chat with Documents via API

🖥️ User Interface Features

Main Pages

UI Components

🔒 Security Features

🏗️ Build Status

🔗 Related Projects

Original Python Implementation

Key Differences in .NET Version

Development Methodology

🤝 Contributing

📝 License

🆘 Support

📋 Prerequisites Checklist

🔄 .NET 9 Migration Notes

What's New in .NET 9

Migration Steps Completed

Post-Migration Steps

About

Resources

Uh oh!

Stars

Watchers

Forks

Releases

Packages 0

Uh oh!

Contributors

Uh oh!

Languages

Packages