Update app.py

app.py

@@ -5,15 +5,12 @@ for name in dir():
     if not name.startswith('_'):
         del globals()[name]
 
-import pulp
 import numpy as np
 import pandas as pd
 import streamlit as st
 import gspread
-import time
 import random
-import scipy.stats
-import os
+import gc
 
 @st.cache_resource
 def init_conn():
@@ -36,7 +33,7 @@ def init_conn():
           gc = gspread.service_account_from_dict(credentials)
           return gc
 
-gc = init_conn()
+gcservice_account  = init_conn()
 
 game_format = {'Win Percentage': '{:.2%}','First Inning Lead Percentage': '{:.2%}',
               'Fifth Inning Lead Percentage': '{:.2%}', '8+ runs': '{:.2%}', 'DK LevX': '{:.2%}', 'FD LevX': '{:.2%}'}
@@ -644,12 +641,6 @@ with tab1:
                             static_exposure['Exposure'] = static_exposure['count'] / len(split_portfolio)
                             static_exposure = static_exposure[['Player', 'Exposure']]
                             
-                            del player_salary_dict
-                            del player_proj_dict
-                            del player_own_dict
-                            del player_team_dict
-                            del static_col_raw
-                            del static_col
     with st.container():
           col1, col2 = st.columns([3, 3])
           
@@ -686,13 +677,14 @@ with tab1:
                               display_portfolio = split_portfolio[['Lineup', 'QB', 'RB1', 'RB2', 'WR1', 'WR2', 'WR3', 'FLEX', 'S_FLEX', 'Salary', 'Projection', 'Ownership']]
                               display_portfolio = display_portfolio.set_index('Lineup')
                               st.dataframe(display_portfolio.style.background_gradient(axis=0).background_gradient(cmap='RdYlGn').background_gradient(cmap='RdYlGn_r', subset=['Ownership']).format(precision=2))
-                              del split_portfolio
-                              del exposure_col_raw
-                              del exposure_col
+                              
 with tab2:
     col1, col2 = st.columns([1, 7])
     with col1:
-              
+        if st.button("Load/Reset Data", key='reset1'):
+            st.cache_data.clear()
+            for key in st.session_state.keys():
+                del st.session_state[key]      
         slate_var1 = 'User'
         site_var1 = 'Draftkings'
         if proj_file is not None:
@@ -727,18 +719,6 @@ with tab2:
             
     with col2:
         if st.button("Simulate Contest"):
-            try:
-                del dst_freq
-                del flex_freq
-                del te_freq
-                del wr_freq
-                del rb_freq
-                del qb_freq
-                del player_freq
-                del Sim_Winner_Export
-                del Sim_Winner_Frame
-            except:
-                pass
             with st.container():
                 st.write('Contest Simulation Starting')
                 seed_depth1 = 10
@@ -780,8 +760,6 @@ with tab2:
                         OwnFrame['Own%'] = np.where(OwnFrame['Own%'] > 75, 75, OwnFrame['Own%'])
                         OwnFrame['Own'] = OwnFrame['Own%'] * (800 / OwnFrame['Own%'].sum())
                     Overall_Proj = OwnFrame[['Player', 'Team', 'Position', 'Median', 'Own', 'Salary']]
-                    
-                    del OwnFrame
                 
                 if insert_port == 1:
                     UserPortfolio = portfolio_dataframe[['QB', 'RB1', 'RB2', 'WR1', 'WR2', 'WR3', 'FLEX', 'S_FLEX']]
@@ -806,8 +784,6 @@ with tab2:
                 Teams_used = Teams_used.drop(columns=['index'])
                 Teams_used_dictraw = Teams_used.drop(columns=['team_item'])
                 Teams_used_dict = Teams_used_dictraw.to_dict()
-                
-                del Teams_used_dictraw
         
                 team_list = Teams_used['Team'].to_list()
                 item_list = Teams_used['team_item'].to_list()
@@ -815,8 +791,6 @@ with tab2:
                 FieldStrength_raw = Strength_var + ((30 - len(Teams_used)) * .01)
                 FieldStrength = FieldStrength_raw - (FieldStrength_raw * (20000 / Contest_Size))
                 
-                del FieldStrength_raw
-                
                 if FieldStrength < 0:
                     FieldStrength = Strength_var
                 field_split = Strength_var
@@ -859,11 +833,6 @@ with tab2:
                 s_pos_players = pd.concat([qbs_raw, rbs_raw, wrs_raw])
                 pos_players.dropna(subset=['Median']).reset_index(drop=True)
                 pos_players = pos_players.reset_index(drop=True)
-                
-                del qbs_raw
-                del defs_raw
-                del rbs_raw
-                del wrs_raw
         
                 if insert_port == 1:
                     try:
@@ -956,10 +925,6 @@ with tab2:
                     'team_check_map':dict(zip(cleaport_players.Player,nerf_frame.Team))
                     }
                 
-                del cleaport_players
-                del Overall_Proj
-                del nerf_frame
-                
                 st.write('Seed frame creation')
                 FinalPortfolio, maps_dict = run_seed_frame(seed_depth1, Strength_var, strength_grow, Teams_used, Total_Runs)
                 
@@ -1011,174 +976,191 @@ with tab2:
                     best_lineup = final_array[final_array[:, -1].argsort(kind='stable')[::-1][:1]]
                     Sim_Winners.append(best_lineup)
                     SimVar += 1
-
-                        
-                # del smple_arrays
-                # del smple_arrays1
-                # del smple_arrays2
-                # del final_array
-                # del best_lineup
+                    
                 st.write('Contest simulation complete')
                 # Initial setup
                 Sim_Winner_Frame = pd.DataFrame(np.concatenate(Sim_Winners), columns=FinalPortfolio.columns.tolist() + ['Fantasy'])
                 Sim_Winner_Frame['GPP_Proj'] = (Sim_Winner_Frame['Projection'] + Sim_Winner_Frame['Fantasy']) / 2
+                Sim_Winner_Frame['unique_id'] = Sim_Winner_Frame['Projection'].astype(str) + Sim_Winner_Frame['Salary'].astype(str) + Sim_Winner_Frame['Own'].astype(str)
+                Sim_Winner_Frame = Sim_Winner_Frame.assign(win_count=Sim_Winner_Frame['unique_id'].map(Sim_Winner_Frame['unique_id'].value_counts()))
                 
                 # Type Casting
-                type_cast_dict = {'Salary': int, 'Projection': np.float16, 'Fantasy': np.float16, 'GPP_Proj': np.float16}
+                type_cast_dict = {'Salary': int, 'Projection': np.float16, 'Fantasy': np.float16, 'GPP_Proj': np.float32}
                 Sim_Winner_Frame = Sim_Winner_Frame.astype(type_cast_dict)
                 
+                del FinalPortfolio, insert_port, type_cast_dict
+                
                 # Sorting
-                Sim_Winner_Frame = Sim_Winner_Frame.sort_values(by='GPP_Proj', ascending=False)
+                st.session_state.Sim_Winner_Frame = Sim_Winner_Frame.sort_values(by=['win_count', 'GPP_Proj'], ascending= [False, False]).copy().drop_duplicates(subset='unique_id').head(100)
+                st.session_state.Sim_Winner_Frame.drop(columns='unique_id', inplace=True)
+                
+                # Data Copying
+                st.session_state.Sim_Winner_Export = Sim_Winner_Frame.copy()
                 
                 # Data Copying
-                Sim_Winner_Export = Sim_Winner_Frame.copy()
+                st.session_state.Sim_Winner_Display = Sim_Winner_Frame.copy()
                 
                 # Conditional Replacement
                 columns_to_replace = ['QB', 'RB1', 'RB2', 'WR1', 'WR2', 'WR3', 'FLEX', 'S_FLEX']
+                    
+                del Sim_Winner_Frame, Sim_Winners
                 
-                player_freq = pd.DataFrame(np.column_stack(np.unique(Sim_Winner_Frame.iloc[:,[0, 1, 2, 3, 4, 5, 6, 7]].values, return_counts=True)),
+                st.session_state.player_freq = pd.DataFrame(np.column_stack(np.unique(st.session_state.Sim_Winner_Display.iloc[:,[0, 1, 2, 3, 4, 5, 6, 7]].values, return_counts=True)),
                                             columns=['Player','Freq']).sort_values('Freq', ascending=False).reset_index(drop=True)
-                player_freq['Freq'] = player_freq['Freq'].astype(int)
-                player_freq['Position'] = player_freq['Player'].map(maps_dict['Pos_map'])
-                player_freq['Salary'] = player_freq['Player'].map(maps_dict['Salary_map'])
-                player_freq['Proj Own'] = player_freq['Player'].map(maps_dict['Own_map']) / 100
-                player_freq['Exposure'] = player_freq['Freq']/(Sim_size)
-                player_freq['Edge'] = player_freq['Exposure'] - player_freq['Proj Own']
-                player_freq['Team'] = player_freq['Player'].map(maps_dict['Team_map'])
+                st.session_state.player_freq['Freq'] = st.session_state.player_freq['Freq'].astype(int)
+                st.session_state.player_freq['Position'] = st.session_state.player_freq['Player'].map(maps_dict['Pos_map'])
+                st.session_state.player_freq['Salary'] = st.session_state.player_freq['Player'].map(maps_dict['Salary_map'])
+                st.session_state.player_freq['Proj Own'] = st.session_state.player_freq['Player'].map(maps_dict['Own_map']) / 100
+                st.session_state.player_freq['Exposure'] = st.session_state.player_freq['Freq']/(Sim_size)
+                st.session_state.player_freq['Edge'] = st.session_state.player_freq['Exposure'] - st.session_state.player_freq['Proj Own']
+                st.session_state.player_freq['Team'] = st.session_state.player_freq['Player'].map(maps_dict['Team_map'])
                 for checkVar in range(len(team_list)):
-                                    player_freq['Team'] = player_freq['Team'].replace(item_list, team_list)
-
-                player_freq = player_freq[['Player', 'Position', 'Team', 'Salary', 'Proj Own', 'Exposure', 'Edge']]
+                                    st.session_state.player_freq['Team'] = st.session_state.player_freq['Team'].replace(item_list, team_list)
                 
-                qb_freq = pd.DataFrame(np.column_stack(np.unique(Sim_Winner_Frame.iloc[:,[0]].values, return_counts=True)),
+                st.session_state.qb_freq = pd.DataFrame(np.column_stack(np.unique(st.session_state.Sim_Winner_Display.iloc[:,[0]].values, return_counts=True)),
                                             columns=['Player','Freq']).sort_values('Freq', ascending=False).reset_index(drop=True)
-                qb_freq['Freq'] = qb_freq['Freq'].astype(int)
-                qb_freq['Position'] = qb_freq['Player'].map(maps_dict['Pos_map'])
-                qb_freq['Salary'] = qb_freq['Player'].map(maps_dict['Salary_map'])
-                qb_freq['Proj Own'] = qb_freq['Player'].map(maps_dict['Own_map']) / 100
-                qb_freq['Exposure'] = qb_freq['Freq']/(Sim_size)
-                qb_freq['Edge'] = qb_freq['Exposure'] - qb_freq['Proj Own']
-                qb_freq['Team'] = qb_freq['Player'].map(maps_dict['Team_map'])
+                st.session_state.qb_freq['Freq'] = st.session_state.qb_freq['Freq'].astype(int)
+                st.session_state.qb_freq['Position'] = st.session_state.qb_freq['Player'].map(maps_dict['Pos_map'])
+                st.session_state.qb_freq['Salary'] = st.session_state.qb_freq['Player'].map(maps_dict['Salary_map'])
+                st.session_state.qb_freq['Proj Own'] = st.session_state.qb_freq['Player'].map(maps_dict['Own_map']) / 100
+                st.session_state.qb_freq['Exposure'] = st.session_state.qb_freq['Freq']/(Sim_size)
+                st.session_state.qb_freq['Edge'] = st.session_state.qb_freq['Exposure'] - st.session_state.qb_freq['Proj Own']
+                st.session_state.qb_freq['Team'] = st.session_state.qb_freq['Player'].map(maps_dict['Team_map'])
                 for checkVar in range(len(team_list)):
-                                    qb_freq['Team'] = qb_freq['Team'].replace(item_list, team_list)
+                                    st.session_state.qb_freq['Team'] = st.session_state.qb_freq['Team'].replace(item_list, team_list)
 
-                qb_freq = qb_freq[['Player', 'Team', 'Position', 'Salary', 'Proj Own', 'Exposure', 'Edge']]
-
-                rb_freq = pd.DataFrame(np.column_stack(np.unique(Sim_Winner_Frame.iloc[:,[1, 2]].values, return_counts=True)),
+                st.session_state.rb_freq = pd.DataFrame(np.column_stack(np.unique(st.session_state.Sim_Winner_Display.iloc[:,[1, 2]].values, return_counts=True)),
                                            columns=['Player','Freq']).sort_values('Freq', ascending=False).reset_index(drop=True)
-                rb_freq['Freq'] = rb_freq['Freq'].astype(int)
-                rb_freq['Position'] = rb_freq['Player'].map(maps_dict['Pos_map'])
-                rb_freq['Salary'] = rb_freq['Player'].map(maps_dict['Salary_map'])
-                rb_freq['Proj Own'] = rb_freq['Player'].map(maps_dict['Own_map']) / 100
-                rb_freq['Exposure'] = rb_freq['Freq']/Sim_size
-                rb_freq['Edge'] = rb_freq['Exposure'] - rb_freq['Proj Own']
-                rb_freq['Team'] = rb_freq['Player'].map(maps_dict['Team_map'])
+                st.session_state.rb_freq['Freq'] = st.session_state.rb_freq['Freq'].astype(int)
+                st.session_state.rb_freq['Position'] = st.session_state.rb_freq['Player'].map(maps_dict['Pos_map'])
+                st.session_state.rb_freq['Salary'] = st.session_state.rb_freq['Player'].map(maps_dict['Salary_map'])
+                st.session_state.rb_freq['Proj Own'] = st.session_state.rb_freq['Player'].map(maps_dict['Own_map']) / 100
+                st.session_state.rb_freq['Exposure'] = st.session_state.rb_freq['Freq']/Sim_size
+                st.session_state.rb_freq['Edge'] = st.session_state.rb_freq['Exposure'] - st.session_state.rb_freq['Proj Own']
+                st.session_state.rb_freq['Team'] = st.session_state.rb_freq['Player'].map(maps_dict['Team_map'])
                 for checkVar in range(len(team_list)):
-                                    rb_freq['Team'] = rb_freq['Team'].replace(item_list, team_list)
-
-                rb_freq = rb_freq[['Player', 'Team', 'Position', 'Salary', 'Proj Own', 'Exposure', 'Edge']]
+                                    st.session_state.rb_freq['Team'] = st.session_state.rb_freq['Team'].replace(item_list, team_list)
                 
-                wr_freq = pd.DataFrame(np.column_stack(np.unique(Sim_Winner_Frame.iloc[:,[3, 4, 5]].values, return_counts=True)),
+                st.session_state.wr_freq = pd.DataFrame(np.column_stack(np.unique(st.session_state.Sim_Winner_Display.iloc[:,[3, 4, 5]].values, return_counts=True)),
                                            columns=['Player','Freq']).sort_values('Freq', ascending=False).reset_index(drop=True)
-                wr_freq['Freq'] = wr_freq['Freq'].astype(int)
-                wr_freq['Position'] = wr_freq['Player'].map(maps_dict['Pos_map'])
-                wr_freq['Salary'] = wr_freq['Player'].map(maps_dict['Salary_map'])
-                wr_freq['Proj Own'] = wr_freq['Player'].map(maps_dict['Own_map']) / 100
-                wr_freq['Exposure'] = wr_freq['Freq']/Sim_size
-                wr_freq['Edge'] = wr_freq['Exposure'] - wr_freq['Proj Own']
-                wr_freq['Team'] = wr_freq['Player'].map(maps_dict['Team_map'])
+                st.session_state.wr_freq['Freq'] = st.session_state.wr_freq['Freq'].astype(int)
+                st.session_state.wr_freq['Position'] = st.session_state.wr_freq['Player'].map(maps_dict['Pos_map'])
+                st.session_state.wr_freq['Salary'] = st.session_state.wr_freq['Player'].map(maps_dict['Salary_map'])
+                st.session_state.wr_freq['Proj Own'] = st.session_state.wr_freq['Player'].map(maps_dict['Own_map']) / 100
+                st.session_state.wr_freq['Exposure'] = st.session_state.wr_freq['Freq']/Sim_size
+                st.session_state.wr_freq['Edge'] = st.session_state.wr_freq['Exposure'] - st.session_state.wr_freq['Proj Own']
+                st.session_state.wr_freq['Team'] = st.session_state.wr_freq['Player'].map(maps_dict['Team_map'])
                 for checkVar in range(len(team_list)):
-                                    wr_freq['Team'] = wr_freq['Team'].replace(item_list, team_list)
-
-                wr_freq = wr_freq[['Player', 'Team', 'Position', 'Salary', 'Proj Own', 'Exposure', 'Edge']]
+                                    st.session_state.wr_freq['Team'] = st.session_state.wr_freq['Team'].replace(item_list, team_list)
                 
-                flex_freq = pd.DataFrame(np.column_stack(np.unique(Sim_Winner_Frame.iloc[:,[6]].values, return_counts=True)),
+                st.session_state.flex_freq = pd.DataFrame(np.column_stack(np.unique(st.session_state.Sim_Winner_Display.iloc[:,[6]].values, return_counts=True)),
                                            columns=['Player','Freq']).sort_values('Freq', ascending=False).reset_index(drop=True)
-                flex_freq['Freq'] = flex_freq['Freq'].astype(int)
-                flex_freq['Position'] = flex_freq['Player'].map(maps_dict['Pos_map'])
-                flex_freq['Salary'] = flex_freq['Player'].map(maps_dict['Salary_map'])
-                flex_freq['Proj Own'] = flex_freq['Player'].map(maps_dict['Own_map']) / 100
-                flex_freq['Exposure'] = flex_freq['Freq']/Sim_size
-                flex_freq['Edge'] = flex_freq['Exposure'] - flex_freq['Proj Own']
-                flex_freq['Team'] = flex_freq['Player'].map(maps_dict['Team_map'])
+                st.session_state.flex_freq['Freq'] = st.session_state.flex_freq['Freq'].astype(int)
+                st.session_state.flex_freq['Position'] = st.session_state.flex_freq['Player'].map(maps_dict['Pos_map'])
+                st.session_state.flex_freq['Salary'] = st.session_state.flex_freq['Player'].map(maps_dict['Salary_map'])
+                st.session_state.flex_freq['Proj Own'] = st.session_state.flex_freq['Player'].map(maps_dict['Own_map']) / 100
+                st.session_state.flex_freq['Exposure'] = st.session_state.flex_freq['Freq']/Sim_size
+                st.session_state.flex_freq['Edge'] = st.session_state.flex_freq['Exposure'] - st.session_state.flex_freq['Proj Own']
+                st.session_state.flex_freq['Team'] = st.session_state.flex_freq['Player'].map(maps_dict['Team_map'])
                 for checkVar in range(len(team_list)):
-                                    flex_freq['Team'] = flex_freq['Team'].replace(item_list, team_list)
-
-                flex_freq = flex_freq[['Player', 'Team', 'Position', 'Salary', 'Proj Own', 'Exposure', 'Edge']]
+                                    st.session_state.flex_freq['Team'] = st.session_state.flex_freq['Team'].replace(item_list, team_list)
                 
-                dst_freq = pd.DataFrame(np.column_stack(np.unique(Sim_Winner_Frame.iloc[:,[7]].values, return_counts=True)),
+                st.session_state.dst_freq = pd.DataFrame(np.column_stack(np.unique(st.session_state.Sim_Winner_Display.iloc[:,[7]].values, return_counts=True)),
                                            columns=['Player','Freq']).sort_values('Freq', ascending=False).reset_index(drop=True)
-                dst_freq['Freq'] = dst_freq['Freq'].astype(int)
-                dst_freq['Position'] = dst_freq['Player'].map(maps_dict['Pos_map'])
-                dst_freq['Salary'] = dst_freq['Player'].map(maps_dict['Salary_map'])
-                dst_freq['Proj Own'] = dst_freq['Player'].map(maps_dict['Own_map']) / 100
-                dst_freq['Exposure'] = dst_freq['Freq']/Sim_size
-                dst_freq['Edge'] = dst_freq['Exposure'] - dst_freq['Proj Own']
-                dst_freq['Team'] = dst_freq['Player'].map(maps_dict['Team_map'])
+                st.session_state.dst_freq['Freq'] = st.session_state.dst_freq['Freq'].astype(int)
+                st.session_state.dst_freq['Position'] = st.session_state.dst_freq['Player'].map(maps_dict['Pos_map'])
+                st.session_state.dst_freq['Salary'] = st.session_state.dst_freq['Player'].map(maps_dict['Salary_map'])
+                st.session_state.dst_freq['Proj Own'] = st.session_state.dst_freq['Player'].map(maps_dict['Own_map']) / 100
+                st.session_state.dst_freq['Exposure'] = st.session_state.dst_freq['Freq']/Sim_size
+                st.session_state.dst_freq['Edge'] = st.session_state.dst_freq['Exposure'] - st.session_state.dst_freq['Proj Own']
+                st.session_state.dst_freq['Team'] = st.session_state.dst_freq['Player'].map(maps_dict['Team_map'])
                 for checkVar in range(len(team_list)):
-                                    dst_freq['Team'] = dst_freq['Team'].replace(item_list, team_list)
-
-                dst_freq = dst_freq[['Player', 'Team', 'Position', 'Salary', 'Proj Own', 'Exposure', 'Edge']]
+                                    st.session_state.dst_freq['Team'] = st.session_state.dst_freq['Team'].replace(item_list, team_list)
                 
             with st.container():
-                simulate_container = st.empty()
-                st.dataframe(Sim_Winner_Frame.head(100).style.background_gradient(axis=0).background_gradient(cmap='RdYlGn').background_gradient(cmap='RdYlGn_r', subset=['Own']).format(precision=2), use_container_width = True)
-            
-            st.download_button(
-                label="Export Tables",
-                data=convert_df_to_csv(Sim_Winner_Export),
-                file_name='NFL_consim_export.csv',
-                mime='text/csv',
-            )
+                if 'player_freq' in st.session_state: 
+                    player_split_var2 = st.radio("Are you wanting to isolate any lineups with specific players?", ('Full Players', 'Specific Players'), key='player_split_var2')
+                    if player_split_var2 == 'Specific Players':
+                              find_var2 = st.multiselect('Which players must be included in the lineups?', options = st.session_state.player_freq['Player'].unique())
+                    elif player_split_var2 == 'Full Players':
+                              find_var2 = st.session_state.player_freq.Player.values.tolist()
+        
+                    if player_split_var2 == 'Specific Players':
+                              st.session_state.Sim_Winner_Display = st.session_state.Sim_Winner_Frame[np.equal.outer(st.session_state.Sim_Winner_Frame.to_numpy(), find_var2).any(axis=1).all(axis=1)]
+                    if player_split_var2 == 'Full Players':
+                              st.session_state.Sim_Winner_Display = st.session_state.Sim_Winner_Frame
+                if 'Sim_Winner_Display' in st.session_state:
+                    st.dataframe(st.session_state.Sim_Winner_Display.style.background_gradient(axis=0).background_gradient(cmap='RdYlGn').background_gradient(cmap='RdYlGn_r', subset=['Own']).format(precision=2), use_container_width = True)
+                if 'Sim_Winner_Export' in st.session_state:
+                    st.download_button(
+                        label="Export Tables",
+                        data=convert_df_to_csv(st.session_state.Sim_Winner_Export),
+                        file_name='NFL_consim_export.csv',
+                        mime='text/csv',
+                    )
             
             with st.container():
                 freq_container = st.empty()
                 tab1, tab2, tab3, tab4, tab5, tab6 = st.tabs(['Overall Exposures', 'QB Exposures', 'RB Exposures', 'WR Exposures', 'FLEX Exposures', 'S_FLEX Exposures'])
                 with tab1:
-                    st.dataframe(player_freq.style.background_gradient(axis=0).background_gradient(cmap='RdYlGn').format(freq_format, precision=2), use_container_width = True)
-                    st.download_button(
-                        label="Export Exposures",
-                        data=convert_df_to_csv(player_freq),
-                        file_name='player_freq_export.csv',
-                        mime='text/csv',
-                    )
+                    if 'player_freq' in st.session_state:
+                        st.dataframe(st.session_state.player_freq.style.background_gradient(axis=0).background_gradient(cmap='RdYlGn').format(freq_format, precision=2), use_container_width = True)
+                        st.download_button(
+                            label="Export Exposures",
+                            data=st.session_state.player_freq.to_csv().encode('utf-8'),
+                            file_name='player_freq_export.csv',
+                            mime='text/csv',
+                        )
                 with tab2:
-                    st.dataframe(qb_freq.style.background_gradient(axis=0).background_gradient(cmap='RdYlGn').format(freq_format, precision=2), use_container_width = True)
-                    st.download_button(
-                        label="Export Exposures",
-                        data=convert_df_to_csv(qb_freq),
-                        file_name='qb_freq_export.csv',
-                        mime='text/csv',
-                    )
+                    if 'qb_freq' in st.session_state:
+                        st.dataframe(st.session_state.qb_freq.style.background_gradient(axis=0).background_gradient(cmap='RdYlGn').format(freq_format, precision=2), use_container_width = True)
+                        st.download_button(
+                            label="Export Exposures",
+                            data=st.session_state.qb_freq.to_csv().encode('utf-8'),
+                            file_name='qb_freq_export.csv',
+                            mime='text/csv',
+                        )
                 with tab3:
-                    st.dataframe(rb_freq.style.background_gradient(axis=0).background_gradient(cmap='RdYlGn').format(freq_format, precision=2), use_container_width = True)
-                    st.download_button(
-                        label="Export Exposures",
-                        data=convert_df_to_csv(rb_freq),
-                        file_name='rb_freq_export.csv',
-                        mime='text/csv',
-                    )
+                    if 'rb_freq' in st.session_state:
+                        st.dataframe(st.session_state.rb_freq.style.background_gradient(axis=0).background_gradient(cmap='RdYlGn').format(freq_format, precision=2), use_container_width = True)
+                        st.download_button(
+                            label="Export Exposures",
+                            data=st.session_state.rb_freq.to_csv().encode('utf-8'),
+                            file_name='rb_freq_export.csv',
+                            mime='text/csv',
+                        )
                 with tab4:
-                    st.dataframe(wr_freq.style.background_gradient(axis=0).background_gradient(cmap='RdYlGn').format(freq_format, precision=2), use_container_width = True)
-                    st.download_button(
-                        label="Export Exposures",
-                        data=convert_df_to_csv(wr_freq),
-                        file_name='wr_freq_export.csv',
-                        mime='text/csv',
-                    )
+                    if 'wr_freq' in st.session_state:
+                        st.dataframe(st.session_state.wr_freq.style.background_gradient(axis=0).background_gradient(cmap='RdYlGn').format(freq_format, precision=2), use_container_width = True)
+                        st.download_button(
+                            label="Export Exposures",
+                            data=st.session_state.wr_freq.to_csv().encode('utf-8'),
+                            file_name='wr_freq_export.csv',
+                            mime='text/csv',
+                        )
                 with tab5:
-                    st.dataframe(flex_freq.style.background_gradient(axis=0).background_gradient(cmap='RdYlGn').format(freq_format, precision=2), use_container_width = True)
-                    st.download_button(
-                        label="Export Exposures",
-                        data=convert_df_to_csv(flex_freq),
-                        file_name='flex_freq_export.csv',
-                        mime='text/csv',
-                    )
+                    if 'flex_freq' in st.session_state:
+                        st.dataframe(st.session_state.flex_freq.style.background_gradient(axis=0).background_gradient(cmap='RdYlGn').format(freq_format, precision=2), use_container_width = True)
+                        st.download_button(
+                            label="Export Exposures",
+                            data=st.session_state.flex_freq.to_csv().encode('utf-8'),
+                            file_name='flex_freq_export.csv',
+                            mime='text/csv',
+                        )
                 with tab6:
-                    st.dataframe(dst_freq.style.background_gradient(axis=0).background_gradient(cmap='RdYlGn').format(freq_format, precision=2), use_container_width = True)
-                    st.download_button(
-                        label="Export Exposures",
-                        data=convert_df_to_csv(dst_freq),
-                        file_name='s_flex_freq_export.csv',
-                        mime='text/csv',
-                    )
+                    if 'dst_freq' in st.session_state:
+                        st.dataframe(st.session_state.dst_freq.style.background_gradient(axis=0).background_gradient(cmap='RdYlGn').format(freq_format, precision=2), use_container_width = True)
+                        st.download_button(
+                            label="Export Exposures",
+                            data=st.session_state.dst_freq.to_csv().encode('utf-8'),
+                            file_name='dst_freq_export.csv',
+                            mime='text/csv',
+                        )
+
+del gcservice_account
+del static_exposure, overall_exposure
+del insert_port1, Contest_Size, sharp_split, Strength_var, scaling_var, Sort_function, Sim_function, strength_grow, field_growth
+del raw_baselines
+del freq_format
+
+gc.collect()