How tzolov/playground-flight-booking works
Main Components
a) Frontend Layer:
• React/TypeScript frontend components (Message.tsx, MessageList.tsx)
• Theming system (customer-support-agent theme)
b) Backend Layer:
• Spring Boot Application core
• AI Services integration
• Booking System components
• Data models
c) External Services:
• Multiple AI Provider options (OpenAI, VertexAI Gemini, Azure OpenAI, Groq, Anthropic Claude)
• PostgreSQL with pgvector (for vector embeddings)
[Client Layer]
• ↓↑
[API Layer]
• ↓↑
[Service Layer]
• ↓↑
[Data Layer]
• ↓↑
[External Services]
Component Details
Client Layer:
• Components: MessageList, Message components
• Theme: Customer Support Agent UI
API Layer:
• Entry point: Application.java
• Controllers (implicit from the structure)
Service Layer (highlight these key services):
• CustomerSupportAssistant
• FlightBookingService
• BookingTools
• AssistantService
• LoggingAdvisor
Data Layer:
• Models: Booking, BookingClass, Customer
• Enums: BookingStatus
• RAG: terms-of-service.txt
External Services Layer:
• AI Providers (with configurable options)
• Database (PostgreSQL with pgvector)
Important Relationships
• Frontend to Backend communication
• Service layer interactions
• AI service integration points
• Database interactions
• RAG system data flow
The configuration flexibility for different AI providers
This system design should clearly communicate how the application integrates AI capabilities with traditional booking system functionality while maintaining flexibility in AI provider selection.