Refactor distribute_preset function to improve filtering logic for 'Finish_percentile' by handling edge cases for the first and last finishing ranges, ensuring accurate portfolio selection and maintaining sorting by 'median' in descending order.

global_func/distribute_preset.py

@@ -8,7 +8,13 @@ def distribute_preset(portfolio: pd.DataFrame, lineup_target: int, exclude_cols:
         for finishing_range in range(1, 20):
             rows_to_drop = []
             working_portfolio = portfolio.copy()
-            working_portfolio = working_portfolio[(working_portfolio['Finish_percentile'] <= (finishing_range / 100)) & (working_portfolio['Finish_percentile'] >= ((finishing_range - 1) / 100))].sort_values(by='median', ascending = False)
+            if finishing_range == 1:
+                working_portfolio = working_portfolio[working_portfolio['Finish_percentile'] <= (finishing_range / 100)]
+            elif finishing_range == 20:
+                working_portfolio = working_portfolio[working_portfolio['Finish_percentile'] >= ((finishing_range - 1) / 100)]
+            else:
+                working_portfolio = working_portfolio[(working_portfolio['Finish_percentile'] <= (finishing_range / 100)) & (working_portfolio['Finish_percentile'] >= ((finishing_range - 1) / 100))]
+            working_portfolio = working_portfolio.sort_values(by='median', ascending = False)
             working_portfolio = working_portfolio.reset_index(drop=True)
             curr_own_type_max = working_portfolio.loc[0, 'Weighted Own'] + (slack_var / 20 * working_portfolio.loc[0, 'Weighted Own'])