Enhance trimming variable options in app.py and update portfolio trimming logic in trim_portfolio.py to accommodate 'Similarity Score'. This change allows for more flexible data handling and improves the trimming process based on the selected variable.

app.py

@@ -1079,7 +1079,7 @@ with tab2:
                     with perf_var:
                         performance_type = st.selectbox("Sorting variable", ['median', 'Finish_percentile'], key='sort_var')
                     with own_var:
-                        own_type = st.selectbox("Trimming variable", ['Own', 'Geomean', 'Weighted Own'], key='trim_var')
+                        own_type = st.selectbox("Trimming variable", ['Own', 'Geomean', 'Weighted Own', 'Similarity Score'], key='trim_var')
 
                     trim_slack_var = st.number_input("Trim slack (percentile addition to trimming variable ceiling)", value=0.0, min_value=0.0, max_value=1.0, step=0.1, key='trim_slack')
 

global_func/trim_portfolio.py

@@ -9,14 +9,22 @@ def trim_portfolio(portfolio: pd.DataFrame, trim_slack: float, performance_type:
     curr_own_type_max = working_portfolio.loc[0, own_type] + (trim_slack * working_portfolio.loc[0, own_type])
 
     for i in range(1, len(working_portfolio)):
-        if working_portfolio.loc[i, own_type] > curr_own_type_max and \
-        working_portfolio.loc[i, performance_type] > performance_threshold_low and \
-        working_portfolio.loc[i, performance_type] <= performance_threshold_high and \
-        working_portfolio.loc[i, own_type] > own_threshold_low and \
-        working_portfolio.loc[i, own_type] <= own_threshold_high:
-            rows_to_drop.append(i)
+        if own_type == 'Similarity Score':
+            if working_portfolio.loc[i, own_type] > curr_own_type_max and \
+            working_portfolio.loc[i, performance_type] > performance_threshold_low and \
+            working_portfolio.loc[i, performance_type] <= performance_threshold_high and \
+            working_portfolio.loc[i, own_type] > own_threshold_low and \
+            working_portfolio.loc[i, own_type] <= own_threshold_high:
+                rows_to_drop.append(i)
+            else:
+                curr_own_type_max = working_portfolio.loc[i, own_type] + (trim_slack * working_portfolio.loc[i, own_type])
         else:
-            curr_own_type_max = working_portfolio.loc[i, own_type] + (trim_slack * working_portfolio.loc[i, own_type])
-
+            if working_portfolio.loc[i, own_type] > curr_own_type_max and \
+            working_portfolio.loc[i, performance_type] > performance_threshold_low and \
+            working_portfolio.loc[i, performance_type] <= performance_threshold_high and \
+            working_portfolio.loc[i, own_type] > own_threshold_low and \
+            working_portfolio.loc[i, own_type] <= own_threshold_high:
+                rows_to_drop.append(i)
+            
     working_portfolio = working_portfolio.drop(rows_to_drop).reset_index(drop=True)
     return working_portfolio