Enhance portfolio trimming logic in app.py: add comprehensive filtering options based on salary, finish percentile, and player selections, and update predict_dupes function to adjust 'Product Own' calculation for improved accuracy.

Remove unused import of gmean from app.py and predict_dupes.py for cleaner code and improved performance.

Update trimming options in app.py: replace 'Geomean' with 'Product Own' for portfolio calculations, and ensure trimming logic is executed only upon form submission for improved user experience.

Refactor trimming options in app.py: encapsulate trimming selections within a form for better user interaction, and update the submission button for clarity in the trimming process.

Enhance portfolio filtering and trimming options in app.py: introduce expandable sections for filter and trimming options, update portfolio trimming logic to support geomean calculations, and improve export functionality with stack count adjustments.

Refactor portfolio filtering logic in app.py: streamline conditions for portfolio adjustments based on user selections, and ensure calculations are only performed when necessary. Add stack count adjustments for export options.

Update dataframe display logic in app.py: use current_page_data for improved clarity in pagination rendering.

Refactor pagination logic in app.py: move dataframe display to the correct position for improved user experience and clarity.

Add 'Trim Portfolio' button in app.py and implement trim_portfolio function in new global_func/trim_portfolio.py

Update site selection logic in app.py: move 'Select Site' dropdown above 'Clear data' button and adjust minimum lineup edge input precision.

Refactor portfolio loading logic in app.py: ensure portfolio is only loaded if not already in session state, and add minimum lineup edge input for filtering.

Refactor Streamlit UI in app.py: reorganize layout for clearer user interaction and restore 'Clear data' button functionality.

Add initial Streamlit application with data loading, portfolio management, and optimization features