How abanteai/rawdog works
Main Components
a) User Interface Layer:
• CLI Interface (entry point through __main__.py)
• Configuration Handler (config.py)
b) Core Processing Layer:
• Script Execution Engine (execute_script.py)
• LLM Integration Service (llm_client.py)
• Prompt Management System (prompts.py)
• Parser Component (parsing.py)
c) External Services:
• LLM Providers (OpenAI, Anthropic, Ollama, etc.)
• File System Interface
• System Resources Interface
User Input Flow
• CLI Interface → Parser → LLM Integration → Script Generation → Execution Engine
Configuration Flow
• Config Files (.rawdog/config.yaml) → Configuration Handler → All Components
Execution Flow
• Execution Engine → File System/System Resources → Output → LLM Integration (for recursive operations)
Specific Design Elements
Use these visual elements:
• Hexagons for external services
• Rounded rectangles for configuration elements
Detailed Component Breakdown
User Interface Layer (Blue)
• CLI Interface
• Command Processing
• Configuration Interface
• YAML Config Parser
Core Processing Layer (Green)
• LLM Integration Service
• Provider Selection
• API Communication
• Script Execution Engine
• Python Script Generator
• Safety Checks
• Prompt Management
• Template Storage
• Context Management
• Parser Component
• Input/Output Processing
External Services Layer (Yellow)
• LLM Providers
• File System
• System Resources
Additional Notes
• Keep the main flow clear and centered
• Related components together
• Clear boundaries between system and external components
• Example flows for common operations