Create app.py

app.py

@@ -0,0 +1,324 @@
+import pulp
+import numpy as np
+import pandas as pd
+import random
+import sys
+import openpyxl
+import re
+import time
+import streamlit as st
+import matplotlib
+from  matplotlib.colors import LinearSegmentedColormap
+from st_aggrid import GridOptionsBuilder, AgGrid, GridUpdateMode, DataReturnMode
+import json
+import requests
+import gspread
+import plotly.figure_factory as ff
+
+scope = ['https://www.googleapis.com/auth/spreadsheets',
+          "https://www.googleapis.com/auth/drive"]
+
+credentials = {
+  "type": "service_account",
+  "project_id": "sheets-api-connect-378620",
+  "private_key_id": "1005124050c80d085e2c5b344345715978dd9cc9",
+  "private_key": "-----BEGIN PRIVATE KEY-----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\nJuElxA33rFEjbgbzdyrE768=\n-----END PRIVATE KEY-----\n",
+  "client_email": "gspread-connection@sheets-api-connect-378620.iam.gserviceaccount.com",
+  "client_id": "106625872877651920064",
+  "auth_uri": "https://accounts.google.com/o/oauth2/auth",
+  "token_uri": "https://oauth2.googleapis.com/token",
+  "auth_provider_x509_cert_url": "https://www.googleapis.com/oauth2/v1/certs",
+  "client_x509_cert_url": "https://www.googleapis.com/robot/v1/metadata/x509/gspread-connection%40sheets-api-connect-378620.iam.gserviceaccount.com"
+}
+
+gc = gspread.service_account_from_dict(credentials)
+
+st.set_page_config(layout="wide")
+
+roo_format = {'Top_finish': '{:.2%}','Top_5_finish': '{:.2%}', 'Top_10_finish': '{:.2%}',
+              '120+%': '{:.2%}','10x%': '{:.2%}','11x%': '{:.2%}','12x%': '{:.2%}','Own': '{:.2%}','LevX': '{:.2%}'}
+stat_format = {'Win%': '{:.2%}'}
+
+game_betting_model = 'https://docs.google.com/spreadsheets/d/1Yq0vGriWK-bS79e-bD6_u9pqrYE6Yrlbb_wEkmH-ot0/edit#gid=172632260'
+props_overall = 'DK_NBA_Props'
+player_overall = 'https://docs.google.com/spreadsheets/d/1Yq0vGriWK-bS79e-bD6_u9pqrYE6Yrlbb_wEkmH-ot0/edit#gid=172632260'
+points_overall = 'DK_Points_Props'
+assists_overall = 'DK_Assists_Props'
+rebounds_overall = 'DK_Rebounds_Props'
+pa_overall = 'DK_PA_Props'
+pr_overall = 'DK_PR_Props'
+pra_overall = 'DK_PRA_Props'
+
+@st.cache_data
+def create_player_props(URL):
+    sh = gc.open_by_url(URL)
+    worksheet = sh.get_worksheet(8)
+    load_display = pd.DataFrame(worksheet.get_all_records())
+    overall_data = load_display[['Name', 'Position', 'Team', '3P', 'Points', 'Rebounds', 'Assists', 'Steals', 'Blocks']]
+    overall_data.rename(columns={"Name": "player"}, inplace = True)
+    overall_data['Points + Rebounds'] = overall_data['Points'] + overall_data['Rebounds']
+    overall_data['Points + Assists'] = overall_data['Points'] + overall_data['Assists']
+    overall_data['Points + Rebounds + Assists'] = overall_data['Points'] + overall_data['Rebounds'] + overall_data['Assists']
+          
+    return overall_data
+
+@st.cache_data
+def load_game_betting(URL):
+    sh = gc.open_by_url(URL)
+    worksheet = sh.get_worksheet(1)
+    raw_display = pd.DataFrame(worksheet.get_all_records())
+
+    return raw_display
+
+@st.cache_data
+def load_props(URL):
+    sh = gc.open(URL)
+    worksheet = sh.get_worksheet(0)
+    raw_display = pd.DataFrame(worksheet.get_all_records())
+    raw_display.rename(columns={"player": "Player"}, inplace = True)
+
+    return raw_display
+
+@st.cache_data
+def load_player_baselines(URL):
+    sh = gc.open(URL)
+    worksheet = sh.get_worksheet(0)
+    raw_display = pd.DataFrame(worksheet.get_all_records())
+
+    return raw_display
+
+@st.cache_data
+def load_stat_specific(URL):
+    sh = gc.open(URL)
+    worksheet = sh.get_worksheet(0)
+    raw_display = pd.DataFrame(worksheet.get_all_records())
+    raw_display.rename(columns={"player": "Player"}, inplace = True)
+    raw_display = raw_display.drop(columns=['Model Probability', 'short%', 'mid%', 'long%', 's_weighted%', 'm_weighted%', 'l_weighted%', 'weighted prob%'])
+
+    return raw_display
+
+team_frame = load_game_betting(game_betting_model)
+props_frame = create_player_props(player_overall)
+
+tab1, tab2, tab3, tab4 = st.tabs(["Game Betting Model", "Player Prop Baselines", "Stat Specific Props Projections", "Player Prop Simulations"])
+
+def convert_df_to_csv(df):
+    return df.to_csv().encode('utf-8')
+
+with tab1:
+    if st.button("Reset Data/Load Data", key='reset1'):
+              # Clear values from *all* all in-memory and on-disk data caches:
+              # i.e. clear values from both square and cube
+              st.cache_data.clear()
+    st.dataframe(team_frame.style.background_gradient(axis=0).background_gradient(cmap='RdYlGn').format(precision=2), use_container_width = True)
+    st.download_button(
+            label="Export Projections",
+            data=convert_df_to_csv(team_frame),
+            file_name='NBA_DFS_team_frame.csv',
+            mime='text/csv',
+            key='team_frame',
+    )
+
+with tab2:
+    if st.button("Reset Data/Load Data", key='reset2'):
+              # Clear values from *all* all in-memory and on-disk data caches:
+              # i.e. clear values from both square and cube
+              st.cache_data.clear()
+    team_var1 = st.multiselect('View specific team?', options = props_frame['Team'].unique(), key = 'prop_teamvar')
+    if team_var1:
+          props_frame = props_frame[props_frame['Team'].isin(team_var1)]
+    props_frame = props_frame.set_index('player')
+    st.dataframe(props_frame.style.background_gradient(axis=0).background_gradient(cmap='RdYlGn').format(precision=2), use_container_width = True)
+    st.download_button(
+            label="Export Projections",
+            data=convert_df_to_csv(props_frame),
+            file_name='NBA_DFS_props_frame.csv',
+            mime='text/csv',
+            key='props_frame',
+    )
+
+with tab3:
+    st.write("The Stat specific models are currently not accurate due to an API issue. Apoligies!")
+    st.info('The Over and Under percentages are a compositve percentage based on simulations, historical performance, and implied probabilities, and may be different than you would expect based purely on the median projection. Likewise, the Edge of a bet is not the only indicator of if you should make the bet or not as the suggestion is using a base acceptable threshold to determine how much edge you should have for each stat category.')
+    if st.button("Reset Data/Load Data", key='reset3'):
+              # Clear values from *all* all in-memory and on-disk data caches:
+              # i.e. clear values from both square and cube
+              st.cache_data.clear()
+    col1, col2 = st.columns([1, 5])
+    
+    with col2:
+        df_hold_container = st.empty()
+        info_hold_container = st.empty()
+        plot_hold_container = st.empty()
+        export_container = st.empty()
+    
+    with col1:
+        prop_type_var = st.selectbox('Select prop category', options = ['Points', 'Assists', 'Rebounds', 'Points + Assists', 'Points + Rebounds', 'Points + Rebounds + Assists'])
+
+        if st.button('Simulate Prop Category'):
+            with col2:
+                
+                with st.spinner('Wait for it...'):
+                   
+                    with df_hold_container.container():
+
+                        if prop_type_var == "Points":
+                            player_df = load_stat_specific(points_overall)
+                            prop_df = load_props(props_overall)
+                            prop_df = prop_df[['Player', 'points', 'over_points_line', 'under_points_line']]
+                            prop_df = prop_df.loc[prop_df['points'] > 0]
+                            prop_df['Over'] = np.where(prop_df['over_points_line'] < 0, (-(prop_df['over_points_line'])/((-(prop_df['over_points_line']))+100)), 100/(prop_df['over_points_line']+100))
+                            prop_df['Under'] = np.where(prop_df['under_points_line'] < 0, (-(prop_df['under_points_line'])/((-(prop_df['under_points_line']))+100)), 100/(prop_df['under_points_line']+100))
+                            prop_df.rename(columns={"points": "Prop"}, inplace = True)
+                            df = pd.merge(player_df, prop_df, how='left', left_on=['Player'], right_on = ['Player'])
+                            df.rename(columns={"weighted%": "weighted"}, inplace = True)
+                        elif prop_type_var == "Assists":
+                            player_df = load_stat_specific(assists_overall)
+                            prop_df = load_props(props_overall)
+                            prop_df = prop_df[['Player', 'assists', 'over_assists_line', 'under_assists_line']]
+                            prop_df = prop_df.loc[prop_df['assists'] > 0]
+                            prop_df['Over'] = np.where(prop_df['over_assists_line'] < 0, (-(prop_df['over_assists_line'])/((-(prop_df['over_assists_line']))+100)), 100/(prop_df['over_assists_line']+100))
+                            prop_df['Under'] = np.where(prop_df['under_assists_line'] < 0, (-(prop_df['under_assists_line'])/((-(prop_df['under_assists_line']))+100)), 100/(prop_df['under_assists_line']+100))
+                            prop_df.rename(columns={"assists": "Prop"}, inplace = True)
+                            df = pd.merge(player_df, prop_df, how='left', left_on=['Player'], right_on = ['Player'])
+                            df.rename(columns={"weighted%": "weighted"}, inplace = True)
+                        elif prop_type_var == "Rebounds":
+                            player_df = load_stat_specific(rebounds_overall)
+                            prop_df = load_props(props_overall)
+                            prop_df = prop_df[['Player', 'rebounds', 'over_rebounds_line', 'under_rebounds_line']]
+                            prop_df = prop_df.loc[prop_df['rebounds'] > 0]
+                            prop_df['Over'] = np.where(prop_df['over_rebounds_line'] < 0, (-(prop_df['over_rebounds_line'])/((-(prop_df['over_rebounds_line']))+100)), 100/(prop_df['over_rebounds_line']+100))
+                            prop_df['Under'] = np.where(prop_df['under_rebounds_line'] < 0, (-(prop_df['under_rebounds_line'])/((-(prop_df['under_rebounds_line']))+100)), 100/(prop_df['under_rebounds_line']+100))
+                            prop_df.rename(columns={"rebounds": "Prop"}, inplace = True)
+                            df = pd.merge(player_df, prop_df, how='left', left_on=['Player'], right_on = ['Player'])
+                            df.rename(columns={"weighted%": "weighted"}, inplace = True)
+                        elif prop_type_var == "Points + Assists":
+                            player_df = load_stat_specific(pa_overall)
+                            prop_df = load_props(props_overall)
+                            prop_df = prop_df[['Player', 'points_assists', 'over_points_assists_line', 'under_points_assists_line']]
+                            prop_df = prop_df.loc[prop_df['points_assists'] > 0]
+                            prop_df['Over'] = np.where(prop_df['over_points_assists_line'] < 0, (-(prop_df['over_points_assists_line'])/((-(prop_df['over_points_assists_line']))+100)), 100/(prop_df['over_points_assists_line']+100))
+                            prop_df['Under'] = np.where(prop_df['under_points_assists_line'] < 0, (-(prop_df['under_points_assists_line'])/((-(prop_df['under_points_assists_line']))+100)), 100/(prop_df['under_points_assists_line']+100))
+                            prop_df.rename(columns={"points_assists": "Prop"}, inplace = True)
+                            df = pd.merge(player_df, prop_df, how='left', left_on=['Player'], right_on = ['Player'])
+                            df.rename(columns={"weighted%": "weighted"}, inplace = True)
+                        elif prop_type_var == "Points + Rebounds":
+                            player_df = load_stat_specific(pr_overall)
+                            prop_df = load_props(props_overall)
+                            prop_df = prop_df[['Player', 'points_rebounds', 'over_points_rebounds_line', 'under_points_rebounds_line']]
+                            prop_df = prop_df.loc[prop_df['points_rebounds'] > 0]
+                            prop_df['Over'] = np.where(prop_df['over_points_rebounds_line'] < 0, (-(prop_df['over_points_rebounds_line'])/((-(prop_df['over_points_rebounds_line']))+100)), 100/(prop_df['over_points_rebounds_line']+100))
+                            prop_df['Under'] = np.where(prop_df['under_points_rebounds_line'] < 0, (-(prop_df['under_points_rebounds_line'])/((-(prop_df['under_points_rebounds_line']))+100)), 100/(prop_df['under_points_rebounds_line']+100))
+                            prop_df.rename(columns={"points_rebounds": "Prop"}, inplace = True)
+                            prop_df = prop_df[['Player', 'Prop', 'Over', 'Under']]
+                            df = pd.merge(player_df, prop_df, how='left', left_on=['Player'], right_on = ['Player'])
+                            df.rename(columns={"weighted%": "weighted"}, inplace = True)
+                        elif prop_type_var == "Points + Rebounds + Assists":
+                            player_df = load_stat_specific(pra_overall)
+                            prop_df = load_props(props_overall)
+                            prop_df = prop_df[['Player', 'points_rebounds_assists', 'over_points_rebounds_assists_line', 'under_points_rebounds_assists_line']]
+                            prop_df = prop_df.loc[prop_df['points_rebounds_assists'] > 0]
+                            prop_df['Over'] = np.where(prop_df['over_points_rebounds_assists_line'] < 0, (-(prop_df['over_points_rebounds_assists_line'])/((-(prop_df['over_points_rebounds_assists_line']))+100)), 100/(prop_df['over_points_rebounds_assists_line']+100))
+                            prop_df['Under'] = np.where(prop_df['under_points_rebounds_assists_line'] < 0, (-(prop_df['under_points_rebounds_assists_line'])/((-(prop_df['under_points_rebounds_assists_line']))+100)), 100/(prop_df['under_points_rebounds_assists_line']+100))
+                            prop_df.rename(columns={"points_rebounds_assists": "Prop"}, inplace = True)
+                            prop_df = prop_df[['Player', 'Prop', 'Over', 'Under']]
+                            df = pd.merge(player_df, prop_df, how='left', left_on=['Player'], right_on = ['Player'])
+                            df.rename(columns={"weighted%": "weighted"}, inplace = True)
+                        
+                        prop_dict = dict(zip(df.Player, df.Prop))
+                        over_dict = dict(zip(df.Player, df.Over))
+                        under_dict = dict(zip(df.Player, df.Under))
+                        weighted_dict = dict(zip(df.Player, df.weighted))
+                        
+                        total_sims = 1000
+
+                        df.replace("", 0, inplace=True)
+
+                        if prop_type_var == "Points":
+                            df['Median'] = df['Points']
+                        elif prop_type_var == "Assists":
+                            df['Median'] = df['Assists']
+                        elif prop_type_var == "Rebounds":
+                            df['Median'] = df['Rebounds']
+                        elif prop_type_var == "Points + Assists":
+                            df['Median'] = df['Points + Assists']
+                        elif prop_type_var == "Points + Rebounds":
+                            df['Median'] = df['Points + Rebounds']
+                        elif prop_type_var == "Points + Rebounds + Assists":
+                            df['Median'] = df['Points + Rebounds + Assists']
+
+                        flex_file = df
+                        flex_file['Floor'] = flex_file['Median'] * .20
+                        flex_file['Ceiling'] = flex_file['Median'] + (flex_file['Median'] * .20)
+                        flex_file['STD'] = (flex_file['Median'] / 4)
+                        flex_file['Prop'] = flex_file['Player'].map(prop_dict)
+                        flex_file = flex_file[['Player', 'Prop', 'Floor', 'Median', 'Ceiling', 'STD']]
+
+                        hold_file = flex_file
+                        overall_file = flex_file
+                        prop_file = flex_file
+
+                        overall_players = overall_file[['Player']]
+
+                        for x in range(0,total_sims):    
+                            prop_file[x] = prop_file['Prop']
+
+                        prop_file = prop_file.drop(['Player', 'Floor', 'Median', 'Ceiling', 'STD'], axis=1)
+                        prop_file.astype('int').dtypes
+
+                        for x in range(0,total_sims):
+                            overall_file[x] = np.random.normal(overall_file['Median'],overall_file['STD'])
+
+                        overall_file=overall_file.drop(['Player', 'Prop', 'Floor', 'Median', 'Ceiling', 'STD'], axis=1)
+                        overall_file.astype('int').dtypes
+
+                        players_only = hold_file[['Player']]
+
+                        player_outcomes = pd.merge(players_only, overall_file, left_index=True, right_index=True)
+
+                        prop_check = (overall_file - prop_file)
+
+                        players_only['Mean_Outcome'] = overall_file.mean(axis=1)
+                        players_only['Weighted_over'] = players_only['Player'].map(weighted_dict)
+                        players_only['Weighted_under'] = 1 - players_only['Player'].map(weighted_dict)
+                        players_only['10%'] = overall_file.quantile(0.1, axis=1)
+                        players_only['90%'] = overall_file.quantile(0.9, axis=1)
+                        players_only['Over'] = prop_check[prop_check >= 1].count(axis=1)/float(total_sims)
+                        players_only['Imp Over'] = players_only['Player'].map(over_dict)
+                        players_only['Over%'] = players_only[["Over", "Weighted_over", "Imp Over"]].mean(axis=1)
+                        players_only['Under'] = prop_check[prop_check < 1].count(axis=1)/float(total_sims)
+                        players_only['Imp Under'] = players_only['Player'].map(under_dict)
+                        players_only['Under%'] = players_only[["Under", "Weighted_under", "Imp Under"]].mean(axis=1)
+                        players_only['Prop'] = players_only['Player'].map(prop_dict)
+                        players_only['Prop_avg'] = players_only['Prop'].mean() / 100
+                        players_only['prop_threshold'] = np.where(.25 - players_only['Prop_avg'] < .10, .10, .25 - players_only['Prop_avg'])
+                        players_only = players_only.loc[players_only['Mean_Outcome'] > 0]
+                        players_only['Over_diff'] = players_only['Over%'] - players_only['Imp Over']
+                        players_only['Under_diff'] = players_only['Under%'] - players_only['Imp Under']
+                        players_only['Bet_check'] = np.where(players_only['Over_diff'] > players_only['Under_diff'], players_only['Over_diff'] , players_only['Under_diff'])
+                        players_only['Bet_suggest'] = np.where(players_only['Over_diff'] > players_only['Under_diff'], "Over" , "Under")
+                        players_only['Bet?'] = np.where(players_only['Bet_check'] >= players_only['prop_threshold'], players_only['Bet_suggest'], "No Bet")
+                        players_only['Edge'] = players_only['Bet_check']
+
+                        players_only['Player'] = hold_file[['Player']]
+
+                        final_outcomes = players_only[['Player', 'Prop', 'Mean_Outcome', 'Imp Over', 'Over%', 'Imp Under', 'Under%', 'Bet?', 'Edge']]
+                        
+                        final_outcomes = final_outcomes.sort_values(by='Edge', ascending=False)
+                    
+                        final_outcomes = final_outcomes.set_index('Player')
+
+                        with df_hold_container:
+                            df_hold_container = st.empty()
+                            st.dataframe(final_outcomes.style.background_gradient(axis=0).background_gradient(cmap='RdYlGn').format(precision=2), use_container_width = True)
+                        with export_container:
+                            export_container = st.empty()
+                            st.download_button(
+                                label="Export Projections",
+                                data=convert_df_to_csv(final_outcomes),
+                                file_name='NBA_DFS_prop_proj.csv',
+                                mime='text/csv',
+                                key='prop_proj',
+                            )
+with tab4:
+    st.info('Coming soon!')