How mr-aicoder/ai_trip_planner works
Project type and purpose
• This is a Python-based “AI Trip Planner” tool that exposes two interfaces:
a) a FastAPI-based HTTP service (main.py launched with uvicorn)
b) an interactive Streamlit UI (streamlit_app.py)
• It uses an agentic workflow to orchestrate calls to various helper “tools” (e.g., currency conversion, weather lookup, place search) and returns trip plans or documents.
File structure
Top-level directories and files:
• .devcontainer/ – VS Code dev-container config
• main.py – FastAPI app entry point
• streamlit_app.py – Streamlit front-end entry point
• agent/ – business-logic orchestrator (agentic_workflow.py)
• tools/ – domain-specific tool modules (weather, currency, expenses, place_search, arithmetic)
• utils/ – low-level helper functions (model loading, API wrappers, document saving)
• config/ – YAML configuration loader
• exception/ – custom exception handling
• logger/ – centralized logging
• prompt_library/ – templates for LLM prompts
• notebook/ – Jupyter notebook for experiments
• setup.py, pyproject.toml, requirements.txt – packaging and dependencies
README for additional insights
• Uses the “uv” tool to manage virtual environments and dependencies.
• FastAPI is served via uvicorn on port 8000.
• Streamlit is a second front-end.
• No database folder—likely ephemeral in-memory or file-based storage via utils/save_to_document.