How 2023-konkuk-graduation-project-aiku/answerlink works
Main Components Identification
The project has three distinct main components:
• Frontend (React-based web application)
• Backend (Spring Boot Java application)
• Data-AI (Python-based ML services)
Frontend Layer (Top)
• Sub-components:
• Chat Interface (QnAContainer)
• Headline Display (HeadlineContainer)
• User Input System (QInput)
Backend Layer (Middle)
• Sub-components:
• Chat Controller
• Headline Controller
• Question Answer Service
Data-AI Layer (Bottom)
a) Crawler System:
• News Downloader
• Trend Topic Analyzer
• Naver News Crawler
b) NLP Server:
• Generator
• Retriever
• Document Converter
User Interaction Flow
• Frontend → Backend: REST API calls
• Backend → NLP Server: RAG processing requests
• Crawler → Backend: News data updates
Data Processing Flow
• Crawler → Document Storage
• Document Storage → Retriever
• Retriever → Generator
• Generator → Backend → Frontend
C. Special Components:
Storage Systems
• Document Index Storage
• News Headlines Storage
• Context Storage
External Services
• News Sources (Naver News)
• OpenAI Integration (noted from OpenAIRestTemplateConfig)
• Blue: Frontend components
• Green: Backend services
• Purple: Data-AI components
• Yellow: Storage systems
• Gray: External services
E. Important Architectural Patterns to Highlight:
RAG Architecture
• Clear separation between retrieval and generation components
• Document processing pipeline
Microservices Architecture
• Separate services for NLP, crawling, and main backend
• Independent scaling capabilities
State Management
• Recoil atoms for frontend state
• Chat state flow
F. Technical Details to Include:
• WebSocket connections for real-time chat
• REST API endpoints
• Document processing pipeline
• News crawling schedule
• Data transformation flow
• Component types (Frontend, Backend, AI Services)
• Storage types
• External service connections
Remember to indicate the asynchronous nature of the news crawling and document processing pipeline, as this appears to be a key feature for maintaining up-to-date information in the RAG system.