src/streamlit_app.py

import streamlit as st
import numpy as np
import pandas as pd
import pymongo
import re
import os
from itertools import combinations

st.set_page_config(layout="wide")

@st.cache_resource
def init_conn():
         # Try to get from environment variable first, fall back to secrets
        uri = os.getenv('MONGO_URI')
        if not uri:
            uri = st.secrets['mongo_uri']
        client = pymongo.MongoClient(uri, retryWrites=True, serverSelectionTimeoutMS=500000)
        db = client["NFL_Database"]

        return db
    
db = 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%}'}

team_roo_format = {'Top Score%': '{:.2%}','0 Runs': '{:.2%}', '1 Run': '{:.2%}', '2 Runs': '{:.2%}', '3 Runs': '{:.2%}', '4 Runs': '{:.2%}',
                   '5 Runs': '{:.2%}','6 Runs': '{:.2%}', '7 Runs': '{:.2%}', '8 Runs': '{:.2%}', '9 Runs': '{:.2%}', '10 Runs': '{:.2%}'}

wrong_acro = ['WSH', 'AZ', 'CHW']
right_acro = ['WAS', 'ARI', 'CWS']

st.markdown("""
<style>
    /* Tab styling */
    .stElementContainer [data-baseweb="button-group"] {
        gap: 8px;
        padding: 4px;
    }
    .stElementContainer [kind="segmented_control"] {
        height: 45px;
        white-space: pre-wrap;
        background-color: #DAA520;
        color: white;
        border-radius: 10px;
        gap: 1px;
        padding: 10px 20px;
        font-weight: bold;
        transition: all 0.3s ease;
    }
    .stElementContainer [kind="segmented_controlActive"] {
        height: 50px;
        background-color: #DAA520;
        border: 3px solid #FFD700;
        color: white;
    }
    .stElementContainer [kind="segmented_control"]:hover {
        background-color: #FFD700;
        cursor: pointer;
    }

    div[data-baseweb="select"] > div {
        background-color: #DAA520;
        color: white;
    }

</style>""", unsafe_allow_html=True)

@st.cache_resource(ttl=60)
def init_baselines():

    collection = db["Player_Baselines"] 
    cursor = collection.find()

    raw_display = pd.DataFrame(list(cursor))
    raw_display = raw_display[['name', 'Team', 'Opp', 'Position', 'Salary', 'team_plays', 'team_pass', 'team_rush', 'team_tds', 'team_pass_tds', 'team_rush_tds', 'dropbacks', 'pass_yards', 'pass_tds',
                               'rush_att', 'rush_yards', 'rush_tds', 'targets', 'rec', 'rec_yards', 'rec_tds', 'PPR', 'Half_PPR', 'Own']]
    player_stats = raw_display[raw_display['Position'] != 'K']

    collection = db["DK_NFL_ROO"] 
    cursor = collection.find()

    raw_display = pd.DataFrame(list(cursor))
    raw_display = raw_display.rename(columns={'player_ID': 'player_id'})
    raw_display = raw_display[['Player', 'Position', 'Team', 'Opp', 'Salary', 'Floor', 'Median', 'Ceiling', 'Top_finish', 'Top_5_finish', 'Top_10_finish', '20+%', '2x%', '3x%', '4x%',
                               'Own', 'Small_Field_Own', 'Large_Field_Own', 'Cash_Field_Own', 'CPT_Own', 'LevX', 'version', 'slate', 'timestamp', 'player_id', 'site']]
    load_display = raw_display[raw_display['Position'] != 'K']
    dk_roo_raw = load_display.dropna(subset=['Median'])

    collection = db["FD_NFL_ROO"] 
    cursor = collection.find()

    raw_display = pd.DataFrame(list(cursor))
    raw_display = raw_display.rename(columns={'player_ID': 'player_id'})
    raw_display = raw_display[['Player', 'Position', 'Team', 'Opp', 'Salary', 'Floor', 'Median', 'Ceiling', 'Top_finish', 'Top_5_finish', 'Top_10_finish', '20+%', '2x%', '3x%', '4x%',
                               'Own', 'Small_Field_Own', 'Large_Field_Own', 'Cash_Field_Own', 'CPT_Own', 'LevX', 'version', 'slate', 'timestamp', 'player_id', 'site']]
    load_display = raw_display[raw_display['Position'] != 'K']
    fd_roo_raw = load_display.dropna(subset=['Median'])

    collection = db["DK_DFS_Stacks"] 
    cursor = collection.find()

    raw_display = pd.DataFrame(list(cursor))
    raw_display = raw_display[['Team', 'QB', 'WR1_TE', 'WR2_TE', 'Total', 'Salary', 'Floor', 'Median', 'Ceiling', 'Top_finish', 'Top_5_finish', 'Top_10_finish', '60+%', '2x%', '3x%', '4x%', 'Own', 'LevX', 'slate', 'version']]
    dk_stacks_raw = raw_display.copy()

    collection = db["FD_DFS_Stacks"] 
    cursor = collection.find()

    raw_display = pd.DataFrame(list(cursor))
    raw_display = raw_display[['Team', 'QB', 'WR1_TE', 'WR2_TE', 'Total', 'Salary', 'Floor', 'Median', 'Ceiling', 'Top_finish', 'Top_5_finish', 'Top_10_finish', '60+%', '2x%', '3x%', '4x%', 'Own', 'LevX', 'slate', 'version']]
    fd_stacks_raw = raw_display.copy()

    return player_stats, dk_stacks_raw, fd_stacks_raw, dk_roo_raw, fd_roo_raw

@st.cache_data
def convert_df_to_csv(df):
    return df.to_csv().encode('utf-8')

player_stats, dk_stacks_raw, fd_stacks_raw, dk_roo_raw, fd_roo_raw = init_baselines()

app_load_reset_column, app_view_site_column = st.columns([1, 9])
with app_load_reset_column:
    if st.button("Load/Reset Data", key='reset_data_button'):
        st.cache_data.clear()
        player_stats, dk_stacks_raw, fd_stacks_raw, dk_roo_raw, fd_roo_raw = init_baselines()
        for key in st.session_state.keys():
            del st.session_state[key]
with app_view_site_column:
    with st.container():
        app_view_column, app_site_column = st.columns([3, 3])
        with app_view_column:
            view_var = st.selectbox("Select view", ["Simple", "Advanced"], key='view_selectbox')
        with app_site_column:
            site_var = st.selectbox("What site do you want to view?", ('Draftkings', 'Fanduel'), key='site_selectbox')

selected_tab = st.segmented_control(
    "Select Tab",
    options=["Stack Finder", "User Upload"],
    selection_mode='single',
    default='Stack Finder',
    width='stretch',
    label_visibility='collapsed',
    key='tab_selector'
)

if selected_tab == 'Stack Finder':
    with st.expander("Info and Filters"):
        app_info_column, slate_choice_column, filtering_column, stack_info_column = st.columns(4)
        with app_info_column:
            if st.button("Load/Reset Data", key='reset1'):
                st.cache_data.clear()
                player_stats, dk_stacks_raw, fd_stacks_raw, dk_roo_raw, fd_roo_raw = init_baselines()
                for key in st.session_state.keys():
                    del st.session_state[key]
            st.info(f"Last Update: " + str(dk_roo_raw['timestamp'][0]) + f" CST")
        with slate_choice_column:
            slate_var1 = st.radio("What slate are you working with?", ('Main Slate', 'Secondary Slate', 'Late Slate', 'Thurs-Mon Slate', 'User Upload'), key='slate_var1')
            if slate_var1 == 'User Upload':
                slate_var1 = st.session_state['proj_dataframe']
            else:
                if site_var == 'Draftkings':
                    raw_baselines = dk_roo_raw
                    if slate_var1 == 'Main Slate':
                        raw_baselines = raw_baselines[raw_baselines['slate'] == 'Main Slate']
                    elif slate_var1 == 'Secondary Slate':
                        raw_baselines = raw_baselines[raw_baselines['slate'] == 'Secondary Slate']
                    elif slate_var1 == 'Late Slate':
                        raw_baselines = raw_baselines[raw_baselines['slate'] == 'Late Slate']
                    elif slate_var1 == 'Thurs-Mon Slate':
                        raw_baselines = raw_baselines[raw_baselines['slate'] == 'Thurs-Mon Slate']
                    raw_baselines = raw_baselines.sort_values(by='Own', ascending=False)
                    qb_lookup = raw_baselines[raw_baselines['Position'] == 'QB']
                elif site_var == 'Fanduel':
                    raw_baselines = fd_roo_raw
                    if slate_var1 == 'Main Slate':
                        raw_baselines = raw_baselines[raw_baselines['slate'] == 'Main Slate']
                    elif slate_var1 == 'Secondary Slate':
                        raw_baselines = raw_baselines[raw_baselines['slate'] == 'Secondary Slate']
                    elif slate_var1 == 'Late Slate':
                        raw_baselines = raw_baselines[raw_baselines['slate'] == 'Late Slate']
                    elif slate_var1 == 'Thurs-Mon Slate':
                        raw_baselines = raw_baselines[raw_baselines['slate'] == 'Thurs-Mon Slate']
                    raw_baselines = raw_baselines.sort_values(by='Own', ascending=False)
                    qb_lookup = raw_baselines[raw_baselines['Position'] == 'QB']
        with filtering_column:
            split_var2 = st.radio("Would you like to run stack analysis for the full slate or individual teams?", ('Full Slate Run', 'Specific Teams'), key='split_var2')
            if split_var2 == 'Specific Teams':
                team_var2 = st.multiselect('Which teams would you like to include in the analysis?', options = raw_baselines['Team'].unique(), key='team_var2')
            elif split_var2 == 'Full Slate Run':
                team_var2 = raw_baselines.Team.unique().tolist()
            pos_var2 = st.multiselect('What Positions would you like to view?', options = ['WR', 'TE', 'RB'], default = ['WR', 'TE', 'RB'], key='pos_var2')
        with stack_info_column:
            if site_var == 'Draftkings':
                max_sal2 = st.number_input('Max Salary', min_value = 5000, max_value = 50000, value = 35000, step = 100, key='max_sal2')
            elif site_var == 'Fanduel':
                max_sal2 = st.number_input('Max Salary', min_value = 5000, max_value = 35000, value = 25000, step = 100, key='max_sal2')
            size_var2 = st.selectbox('What size of stacks are you analyzing?', options = ['3-man', '4-man', '5-man'])
            if size_var2 == '3-man':
                stack_size = 3
            if size_var2 == '4-man':
                stack_size = 4
            if size_var2 == '5-man':
                stack_size = 5

            team_dict = dict(zip(raw_baselines.Player, raw_baselines.Team))
            proj_dict = dict(zip(raw_baselines.Player, raw_baselines.Median))
            own_dict = dict(zip(raw_baselines.Player, raw_baselines.Own))
            cost_dict = dict(zip(raw_baselines.Player, raw_baselines.Salary))
            qb_dict = dict(zip(qb_lookup.Team, qb_lookup.Player))
    if st.button("Run Stack Analysis", key='run_stack_analysis'):
        if site_var == 'Draftkings':
            position_limits = {
                'QB': 1,
                'RB': 2,
                'WR': 3,
                'TE': 1,
                'UTIL': 1,
                'DST': 1,
                # Add more as needed
            }
            max_salary = max_sal2
            max_players = 9
        else:
            position_limits = {
                'QB': 1,
                'RB': 2,
                'WR': 3,
                'TE': 1,
                'UTIL': 1,
                'DST': 1,
                # Add more as needed
            }
            max_salary = max_sal2
            max_players = 9

        stack_hold_container = st.empty()
        comb_list = []
        raw_baselines = raw_baselines[raw_baselines['Position'].str.contains('|'.join(pos_var2 + ['QB']))]

        # Create a position dictionary mapping players to their eligible positions
        pos_dict = dict(zip(raw_baselines.Player, raw_baselines.Position))

        def is_valid_combination(combo):
            # Count positions in this combination
            position_counts = {pos: 0 for pos in position_limits.keys()}
            
            # For each player in the combination
            for player in combo:
                # Get their eligible positions
                player_positions = pos_dict[player].split('/')
                
                # For each position they can play
                for pos in player_positions:
                    if pos == 'UTIL':
                        # UTIL can be filled by any position
                        for p in position_counts:
                            position_counts[p] += 1
            
            # Check if any position exceeds its limit
            for pos, limit in position_limits.items():
                if position_counts[pos] > limit:
                    return False
            
            return True

        # Modify the combination generation code
        comb_list = []
        for cur_team in team_var2:
            working_baselines = raw_baselines
            working_baselines = working_baselines[working_baselines['Team'] == cur_team]
            working_baselines = working_baselines[working_baselines['Position'] != 'DST']
            working_baselines = working_baselines[working_baselines['Position'] != 'K']
            qb_var = qb_dict[cur_team]
            order_list = working_baselines['Player']

            comb = combinations(order_list, stack_size)

            for i in list(comb):
                if qb_var in i:
                    comb_list.append(i)

            # Only add combinations that satisfy position limits
            for i in list(comb):
                if is_valid_combination(i):
                    comb_list.append(i)

        comb_DF = pd.DataFrame(comb_list)

        if stack_size == 3:
            comb_DF['Team'] = comb_DF[0].map(team_dict)

            comb_DF['Proj'] = sum([comb_DF[0].map(proj_dict),
                    comb_DF[1].map(proj_dict),
                    comb_DF[2].map(proj_dict)])

            comb_DF['Salary'] = sum([comb_DF[0].map(cost_dict),
                    comb_DF[1].map(cost_dict),
                    comb_DF[2].map(cost_dict)])

            comb_DF['Own%'] = sum([comb_DF[0].map(own_dict),
                    comb_DF[1].map(own_dict),
                    comb_DF[2].map(own_dict)])
        elif stack_size == 4:
            comb_DF['Team'] = comb_DF[0].map(team_dict)

            comb_DF['Proj'] = sum([comb_DF[0].map(proj_dict),
                    comb_DF[1].map(proj_dict),
                    comb_DF[2].map(proj_dict),
                    comb_DF[3].map(proj_dict)])

            comb_DF['Salary'] = sum([comb_DF[0].map(cost_dict),
                    comb_DF[1].map(cost_dict),
                    comb_DF[2].map(cost_dict),
                    comb_DF[3].map(cost_dict)])

            comb_DF['Own%'] = sum([comb_DF[0].map(own_dict),
                    comb_DF[1].map(own_dict),
                    comb_DF[2].map(own_dict),
                    comb_DF[3].map(own_dict)])
        elif stack_size == 5:
            comb_DF['Team'] = comb_DF[0].map(team_dict)

            comb_DF['Proj'] = sum([comb_DF[0].map(proj_dict),
                    comb_DF[1].map(proj_dict),
                    comb_DF[2].map(proj_dict),
                    comb_DF[3].map(proj_dict),
                    comb_DF[4].map(proj_dict)])

            comb_DF['Salary'] = sum([comb_DF[0].map(cost_dict),
                    comb_DF[1].map(cost_dict),
                    comb_DF[2].map(cost_dict),
                    comb_DF[3].map(cost_dict),
                    comb_DF[4].map(cost_dict)])

            comb_DF['Own%'] = sum([comb_DF[0].map(own_dict),
                    comb_DF[1].map(own_dict),
                    comb_DF[2].map(own_dict),
                    comb_DF[3].map(own_dict),
                    comb_DF[4].map(own_dict)])

        comb_DF = comb_DF.sort_values(by='Proj', ascending=False)
        comb_DF = comb_DF.loc[comb_DF['Salary'] <= max_sal2]

        cut_var = 0

        if stack_size == 2:
                    while cut_var <= int(len(comb_DF)):
                        try:
                            if int(cut_var) == 0:
                                cur_proj = float(comb_DF.iat[cut_var, 3])
                                cur_own = float(comb_DF.iat[cut_var, 5])
                            elif int(cut_var) >= 1:
                                check_own = float(comb_DF.iat[cut_var, 5])
                                if check_own > cur_own:
                                    comb_DF = comb_DF.drop([cut_var])
                                    cur_own = cur_own
                                    cut_var = cut_var - 1
                                    comb_DF = comb_DF.reset_index()
                                    comb_DF = comb_DF.drop(['index'], axis=1)
                                elif check_own <= cur_own:
                                    cur_own = float(comb_DF.iat[cut_var, 5])
                                    cut_var = cut_var
                            cut_var += 1
                        except:
                            cut_var += 1
        elif stack_size == 3:
            while cut_var <= int(len(comb_DF)):
                try:
                    if int(cut_var) == 0:
                        cur_proj = float(comb_DF.iat[cut_var,4])
                        cur_own = float(comb_DF.iat[cut_var,6])
                    elif int(cut_var) >= 1:
                        check_own = float(comb_DF.iat[cut_var,6])
                        if check_own > cur_own:
                            comb_DF = comb_DF.drop([cut_var])
                            cur_own = cur_own
                            cut_var = cut_var - 1
                            comb_DF = comb_DF.reset_index()
                            comb_DF = comb_DF.drop(['index'], axis=1)
                        elif check_own <= cur_own:
                            cur_own = float(comb_DF.iat[cut_var,6])
                            cut_var = cut_var
                    cut_var += 1
                except:
                    cut_var += 1
        elif stack_size == 4:
            while cut_var <= int(len(comb_DF)):
                try:
                    if int(cut_var) == 0:
                        cur_proj = float(comb_DF.iat[cut_var,5])
                        cur_own = float(comb_DF.iat[cut_var,7])
                    elif int(cut_var) >= 1:
                        check_own = float(comb_DF.iat[cut_var,7])
                        if check_own > cur_own:
                            comb_DF = comb_DF.drop([cut_var])
                            cur_own = cur_own
                            cut_var = cut_var - 1
                            comb_DF = comb_DF.reset_index()
                            comb_DF = comb_DF.drop(['index'], axis=1)
                        elif check_own <= cur_own:
                            cur_own = float(comb_DF.iat[cut_var,7])
                            cut_var = cut_var
                    cut_var += 1
                except:
                    cut_var += 1
        elif stack_size == 5:
            while cut_var <= int(len(comb_DF)):
                try:
                    if int(cut_var) == 0:
                        cur_proj = float(comb_DF.iat[cut_var,6])
                        cur_own = float(comb_DF.iat[cut_var,8])
                    elif int(cut_var) >= 1:
                        check_own = float(comb_DF.iat[cut_var,8])
                        if check_own > cur_own:
                            comb_DF = comb_DF.drop([cut_var])
                            cur_own = cur_own
                            cut_var = cut_var - 1
                            comb_DF = comb_DF.reset_index()
                            comb_DF = comb_DF.drop(['index'], axis=1)
                        elif check_own <= cur_own:
                            cur_own = float(comb_DF.iat[cut_var,8])
                            cut_var = cut_var
                    cut_var += 1
                except:
                    cut_var += 1
        st.session_state['display_frame'] = comb_DF
    if 'display_frame' in st.session_state:
        st.dataframe(st.session_state['display_frame'].style.background_gradient(axis=0).background_gradient(cmap='RdYlGn').format(precision=2), hide_index=True, use_container_width = True)
        st.download_button(
                label="Export Tables",
                data=convert_df_to_csv(st.session_state['display_frame']),
                file_name='NFL_Stack_Options_export.csv',
                mime='text/csv',
        )
    else:
        st.info("When you run the stack analysis, the results will be displayed here. Open up the 'Info and Filters' tab to check the settings.")

if selected_tab == 'User Upload':
    st.info("The Projections file can have any columns in any order, but must contain columns explicitly named: 'Player', 'Salary', 'Position', 'Team', 'Opp', 'Median', and 'Own'.")
    col1, col2 = st.columns([1, 5])

    with col1:
        proj_file = st.file_uploader("Upload Projections File", key = 'proj_uploader')
    
        if proj_file is not None:
                  try:
                            st.session_state['proj_dataframe'] = pd.read_csv(proj_file)
                  except:
                            st.session_state['proj_dataframe'] = pd.read_excel(proj_file)
    with col2:
        if proj_file is not None:  
                  st.dataframe(st.session_state['proj_dataframe'].style.background_gradient(axis=0).background_gradient(cmap='RdYlGn').format(precision=2), use_container_width = True)