streamlit-development

name: streamlit-development description: Best practices for building, structuring, and optimizing Streamlit applications in SimUCI.

Streamlit Development Skills

1. Architecture Pattern

• Logic Separation: Keep app.py for View/Controller logic only.
  • Move complex calculations to uci/.
  • Move helper UI functions (plots, formatting) to utils/visuals/.
• Monolith: app.py is the single entry point. Do not create multi-page apps unless architecture changes.

2. Performance & Caching

• Data: Use @st.cache_data for loading CSVs and heavy dataframes.

  @st.cache_data
  def load_data(path: Path) -> pd.DataFrame:
      return pd.read_csv(path)
  

• Resources: Use @st.cache_resource for loading ML models or DB connections.

3. Session State Management

• Persistence: Use st.session_state to store simulation results between re-runs.
• Initialization: Always check if key exists before access.

  if 'simulation_results' not in st.session_state:
      st.session_state.simulation_results = None
  

4. User Experience (UI/UX)

• Language: All visible text (titles, buttons, warnings) must be in Spanish.
• Variables: Load text from utils/constants/messages.py.
• Feedback:
  • Use st.spinner() for running simulations.
  • Use st.progress() for batch experiments.
  • Use st.error/warning/info for status updates.

5. Visualizations

• Library: plotly (interactive) or matplotlib/seaborn (static).
• Theme: Respect user theme preference (Light/Dark) where possible, or stick to project defaults defined in utils/constants/theme.py.