How paulrobello/par_gpt works
Main Components
a) Core Application Components:
• CLI Interface Layer
• Agent System
• AI Provider Interface
• Tool Management System
• Cache Management
• Voice/TTS Systems
b) External Services:
• AI Providers (OpenAI, Anthropic, Groq, Google, etc.)
• Search Services (Tavily, Google, Brave, etc.)
• Weather API
• GitHub API
• Other External APIs
c) Execution Environments:
• Local REPL Environment
• Docker Sandbox Environment
User Interface Layer
↓
Core Processing Layer
↓
Service Integration Layer
↓
External Services Layer
Specific Implementation Details
a) CLI Interface Layer:
• Command parser (Typer-based)
• Output formatter (Rich-based)
• Configuration manager
b) Core Processing Layer:
• Agent system (src/par_gpt/agents.py)
• Message handler (src/par_gpt/agent_messages.py)
• Cache manager (src/par_gpt/cache_manager.py)
• Voice/TTS managers (src/par_gpt/tts_manager.py, voice_input_manager.py)
c) Tool Integration Layer:
• AI Tools module (src/par_gpt/ai_tools/)
• Python REPL (src/par_gpt/ai_tools/par_python_repl.py)
• Repository manager (src/par_gpt/repo/)
d) Execution Environments:
• Local REPL environment
• Docker sandbox (sandbox/code_runner/)
Key Interactions to Highlight
• How user input flows from CLI through the system
• Demonstrate how the agent system interacts with various tools
• Security boundary between main application and sandbox
• Relationship between cache manager and external services
Notes for Implementation
• Keep the main flow clear and centered
• Related components together
• Use clear labels for all components
• Version information
• Notes about optional components
Remember to include a small overview section explaining the main data and control flows, and note any significant architectural decisions (like the separation between local REPL and sandboxed execution environments).