How tattle-made/feluda works
Main Components
a. Core System Components:
• API Server (src/server.py)
• Indexer Service (src/indexer.py)
• Reporter Service (src/reporter.py)
• Worker Services (src/worker/*)
• Audio Vector Worker
• Video Vector Worker
• Media Worker
• Hash Worker
• Clustering Media Worker
b. External Services:
• Elasticsearch (for vector storage)
• RabbitMQ (for job queuing)
• S3 Storage (indicated by s3_utils.py)
c. Core Framework Components:
• Operator System
• Media Factory
• Queue Management
• Store Adapters
Key Relationships to Highlight
a. Data Flow:
• Client → API Server → RabbitMQ → Workers → Storage
• Search Flow: Client → API Server → Elasticsearch → Client
b. Component Integration:
• Worker Services ↔ Operators
• Store Adapters ↔ Elasticsearch
• Media Factory ↔ Workers
Implementation Details
Use these symbols and conventions:
• Rectangles: Services and Components
• Hexagons: Operators
• Double Lines: API Boundaries
Additional Elements
Annotations for:
• Configuration Management
• Environment Variables
• Docker Container Boundaries
• Scaling Points
• Language Support
• Media Type Support
The separation of concerns between indexing, searching, and processing
This system design should reflect Feluda's core strength as a flexible, modular system for handling multi-modal content analysis with swappable components for different use cases.