How azure-samples/azure-search-openai-demo-java works
Main Components
a. Frontend Layer:
• React-based web application (TypeScript)
• Authentication components
• Chat and Q&A interfaces
• Settings and configuration UI
b. Backend Services:
• Java Spring Boot API service
• Document Indexer service
• Authentication/Authorization service
c. Azure Services:
• Azure OpenAI Service
• Azure AI Search
• Azure Document Intelligence
• Azure Blob Storage
• Azure Monitor/Application Insights
• Azure Active Directory
Key Architectural Patterns to Highlight
a. Microservices Architecture:
• Clear separation between frontend, backend, and indexer services
• Containerization (Docker) for each service
• Deployment options (App Service, Container Apps, or AKS)
b. RAG Pattern Implementation:
• Document the flow from user input to response generation
• Different RAG approaches (Plain Java vs Semantic Kernel)
• Vector search and hybrid search paths
[Web Frontend (React/TypeScript)]
[Authentication Layer (Azure AD)]
[Backend API Service]──────[Indexer Service]
[RAG Orchestration]────────[Document Processing]
[Azure OpenAI]──[Azure AI Search]──[Blob Storage]
[App Insights]──[Document Intelligence]──[Key Vault]
• User interaction flow
• Document ingestion pipeline
• Search and retrieval paths
• Authentication flows
• Monitoring and logging paths
Additional Elements
• Scaling components (load balancers, etc.)
• Security boundaries and network isolation
Implementation Details
• Plain Java OpenAI SDK implementation
• Semantic Kernel implementation
• Vector store integration
• Prompt flow and construction
Deployment Views
• App Service deployment
• Azure Container Apps deployment
• AKS deployment
Remember to include:
• Version information
• Date of last update
• Author/team information
• Reference to source documentation
• Any relevant configuration details
Tools Recommendation:
• Microsoft Visio for enterprise environments