How fynnfluegge/rocketnotes works
Main System Components
a. Frontend Layer:
• Angular Web Application (webapp directory)
• Electron Desktop Application
• Landing Page (Angular-based)
b. Backend Layer (Serverless):
• Lambda Functions (lambda-handler directory)
• Authentication (AWS Cognito)
• Storage Layer (DynamoDB)
• Vector Store (FAISS)
c. AI/ML Services:
• LLM Providers (OpenAI, Anthropic, Voyage)
• Local AI Options (Ollama)
• Vector Embedding Services
Component Relationships
a. Client-Side Flow:
Web Browser/Electron App -> API Gateway -> Lambda Functions
b. Authentication Flow:
Client -> AWS Cognito -> Protected Resources
c. Document Processing Flow:
Editor Component -> Save Document Handler -> DynamoDB
• > Vector Embedding Handler -> FAISS Index
d. Search/RAG Pipeline:
Search Request -> Semantic Search Handler -> FAISS Index
• > LLM Service -> Response
Specific Implementation Details
a. Frontend Components:
• Editor component
• Navigation/Sidenav
• Zettelkasten interface
• Dialog components (Config, LLM)
• Public document viewer
b. Lambda Functions (group by functionality):
• Document management handlers
• Zettel handlers
• AI/ML handlers (chat, semantic search, vector embeddings)
• User configuration handlers
c. Storage Systems:
• Document store (DynamoDB)
• Vector store (FAISS)
• User configurations
Infrastructure Layer
AWS infrastructure components:
• API Gateway
• Lambda
• DynamoDB
• Cognito
• CDK deployment
Notes for Implementation
• Use clear hierarchical layout
• Related components
• Primary and secondary data flows
• Version/environment indicators
• Component scaling indicators
Remember to include both the cloud-hosted and local Docker-based deployment scenarios, as they represent different architectural patterns within the same system.