app.py

import streamlit as st
st.set_page_config(layout="wide")
import numpy as np
import pandas as pd
from rapidfuzz import process, fuzz
from collections import Counter
from pymongo.mongo_client import MongoClient
from pymongo.server_api import ServerApi
from datetime import datetime

def init_conn():
        
        uri = st.secrets['mongo_uri']
        client = MongoClient(uri, retryWrites=True, serverSelectionTimeoutMS=500000)
        db = client['Contest_Information']

        return db
    
def grab_contest_names(db, sport, type):
    if type == 'Classic':
        db_type = 'reg'
    elif type == 'Showdown':
        db_type = 'sd'
    collection = db[f'{sport}_{db_type}_contest_info']
    cursor = collection.find()

    curr_info = pd.DataFrame(list(cursor)).drop('_id', axis=1)
    curr_info['Date'] = pd.to_datetime(curr_info['Contest Date'].sort_values(ascending = False))
    curr_info['Date'] = curr_info['Date'].dt.strftime('%Y-%m-%d')
    contest_names = curr_info['Contest Name'] + ' - ' + curr_info['Date']
    
    return contest_names, curr_info

def grab_contest_player_info(db, sport, type, contest_date, contest_name, contest_id_map):
    if type == 'Classic':
        db_type = 'reg'
    elif type == 'Showdown':
        db_type = 'showdown'
    collection = db[f'{sport}_{db_type}_player_info']
    cursor = collection.find()

    player_info = pd.DataFrame(list(cursor)).drop('_id', axis=1)
    player_info = player_info[player_info['Contest Date'] == contest_date]
    player_info = player_info.rename(columns={'Display Name': 'Player'})
    player_info = player_info.sort_values(by='Salary', ascending=True).drop_duplicates(subset='Player', keep='first')

    info_maps = {
        'position_dict': dict(zip(player_info['Player'], player_info['Position'])),
        'salary_dict': dict(zip(player_info['Player'], player_info['Salary'])),
        'team_dict': dict(zip(player_info['Player'], player_info['Team'])),
        'opp_dict': dict(zip(player_info['Player'], player_info['Opp'])),
        'fpts_avg_dict': dict(zip(player_info['Player'], player_info['Avg FPTS']))
    }
    
    return player_info, info_maps

db = init_conn()

## import global functions
from global_func.load_contest_file import load_contest_file
from global_func.create_player_exposures import create_player_exposures
from global_func.create_stack_exposures import create_stack_exposures
from global_func.create_stack_size_exposures import create_stack_size_exposures
from global_func.create_general_exposures import create_general_exposures
from global_func.grab_contest_data import grab_contest_data

def is_valid_input(file):
    if isinstance(file, pd.DataFrame):
        return not file.empty
    else:
        return file is not None  # For Streamlit uploader objects

player_exposure_format = {'Exposure Overall': '{:.2%}', 'Exposure Top 1%': '{:.2%}', 'Exposure Top 5%': '{:.2%}', 'Exposure Top 10%': '{:.2%}', 'Exposure Top 20%': '{:.2%}'}
dupe_format = {'uniques%': '{:.2%}', 'under_5%': '{:.2%}', 'under_10%': '{:.2%}'}

tab1, tab2 = st.tabs(["Data Load", "Contest Analysis"])
with tab1:
    col1, col2 = st.columns(2)

    with col1:
        if st.button('Clear data', key='reset1'):
            st.session_state.clear()
        search_options, sport_options, date_options = st.columns(3)
        with search_options:
            parse_type = st.selectbox("Manual upload or DB search?", ['Manual', 'DB Search'], key='parse_type')
        with sport_options:
            sport_select = st.selectbox("Select Sport", ['MLB', 'MMA', 'GOLF', 'NBA', 'NHL'], key='sport_select')
            type_var = st.selectbox("Select Game Type", ['Classic', 'Showdown'], key='type_var')
            try:
                contest_names, curr_info = grab_contest_names(db, sport_select, type_var)
            except:
                st.error("No contests found for this sport and/or game type")
                st.stop()
            
        with date_options:
            date_list = curr_info['Date'].sort_values(ascending=False).unique()
            date_list = date_list[date_list != pd.Timestamp.today().strftime('%Y-%m-%d')]
            date_select = st.selectbox("Select Date", date_list, key='date_select')
            date_select2 = (pd.to_datetime(date_select) + pd.Timedelta(days=1)).strftime('%Y-%m-%d')
            
            name_parse = curr_info[curr_info['Date'] == date_select]['Contest Name'].reset_index(drop=True)
            contest_id_map = dict(zip(name_parse, curr_info[curr_info['Date'] == date_select]['Contest ID']))
            date_select = date_select.replace('-', '')
            date_select2 = date_select2.replace('-', '')

        contest_name_var = st.selectbox("Select Contest to load", name_parse)
        if parse_type == 'DB Search':
            if 'Contest_file_helper' in st.session_state:
                del st.session_state['Contest_file_helper']
            if 'Contest_file' in st.session_state:
                del st.session_state['Contest_file']
            if 'Contest_file' not in st.session_state:
                if st.button('Load Contest Data', key='load_contest_data'):
                    st.session_state['player_info'], st.session_state['info_maps'] = grab_contest_player_info(db, sport_select, type_var, date_select, contest_name_var, contest_id_map)
                    st.session_state['Contest_file'] = grab_contest_data(sport_select, contest_name_var, contest_id_map, date_select, date_select2)
            else:
                pass
    with col2:
        st.info(f"If you are manually loading and do not have the results CSV for the contest you selected, you can find it here: https://www.draftkings.com/contest/gamecenter/{contest_id_map[contest_name_var]}#/")
        if parse_type == 'Manual':
            if 'Contest_file_helper' in st.session_state:
                del st.session_state['Contest_file_helper']
            if 'Contest_file' in st.session_state:
                del st.session_state['Contest_file']
            if 'Contest_file' not in st.session_state:
                st.session_state['Contest_upload'] = st.file_uploader("Upload Contest File (CSV or Excel)", type=['csv', 'xlsx', 'xls'])
                st.session_state['player_info'], st.session_state['info_maps'] = grab_contest_player_info(db, sport_select, type_var, date_select, contest_name_var, contest_id_map)
                try:
                    st.session_state['Contest_file'] = pd.read_csv(st.session_state['Contest_upload'])
                except:
                    st.warning('Please upload a Contest CSV')
            else:
                pass

    if 'Contest_file' in st.session_state:
        st.session_state['Contest'], st.session_state['ownership_df'], st.session_state['actual_df'], st.session_state['entry_list'], check_lineups = load_contest_file(st.session_state['Contest_file'], type_var, st.session_state['player_info'], sport_select)
        st.session_state['Contest'] = st.session_state['Contest'].dropna(how='all')
        st.session_state['Contest'] = st.session_state['Contest'].reset_index(drop=True)
        if st.session_state['Contest'] is not None:
            st.success('Contest file loaded successfully!')
            st.dataframe(st.session_state['Contest'].head(100))
        
    if 'Contest_file' in st.session_state:
        st.session_state['ownership_dict'] = dict(zip(st.session_state['ownership_df']['Player'], st.session_state['ownership_df']['Own']))
        st.session_state['actual_dict'] = dict(zip(st.session_state['actual_df']['Player'], st.session_state['actual_df']['FPTS']))
        st.session_state['salary_dict'] = st.session_state['info_maps']['salary_dict']
        st.session_state['team_dict'] = st.session_state['info_maps']['team_dict']
        st.session_state['pos_dict'] = st.session_state['info_maps']['position_dict']

with tab2:
    excluded_cols = ['BaseName', 'EntryCount']
    if 'Contest' in st.session_state:
        player_columns = [col for col in st.session_state['Contest'].columns if col not in excluded_cols]
        for col in player_columns:
            st.session_state['Contest'][col] = st.session_state['Contest'][col].astype(str)
    
        # Create mapping dictionaries
        map_dict = {
            'pos_map': st.session_state['pos_dict'],
            'team_map': st.session_state['team_dict'],
            'salary_map': st.session_state['salary_dict'],
            'own_map': st.session_state['ownership_dict'],
            'own_percent_rank': dict(zip(st.session_state['ownership_df']['Player'], st.session_state['ownership_df']['Own'].rank(pct=True)))
        }
        # Create a copy of the dataframe for calculations
        working_df = st.session_state['Contest'].copy()

        if type_var == 'Classic':
            working_df['stack'] = working_df.apply(
                lambda row: Counter(
                    map_dict['team_map'].get(player, '') for player in row[4:]
                    if map_dict['team_map'].get(player, '') != ''
                ).most_common(1)[0][0] if any(map_dict['team_map'].get(player, '') for player in row[4:]) else '',
                axis=1
            )
            working_df['stack_size'] = working_df.apply(
                lambda row: Counter(
                    map_dict['team_map'].get(player, '') for player in row[4:]
                    if map_dict['team_map'].get(player, '') != ''
                ).most_common(1)[0][1] if any(map_dict['team_map'].get(player, '') for player in row[4:]) else '',
                axis=1
            )
            working_df['salary'] = working_df.apply(lambda row: sum(map_dict['salary_map'].get(player, 0) for player in row), axis=1)
            working_df['actual_fpts'] = working_df.apply(lambda row: sum(st.session_state['actual_dict'].get(player, 0) for player in row), axis=1)
            working_df['actual_own'] = working_df.apply(lambda row: sum(st.session_state['ownership_dict'].get(player, 0) for player in row), axis=1)
            working_df['sorted'] = working_df[player_columns].apply(
                lambda row: ','.join(sorted(row.values)),
                axis=1
            )
            working_df['dupes'] = working_df.groupby('sorted').transform('size')

            working_df['uniques'] = working_df.groupby('BaseName').apply(
                lambda x: (x['dupes'] == 1).sum()
            ).reindex(working_df['BaseName']).values

            working_df['under_5'] = working_df.groupby('BaseName').apply(
                lambda x: (x['dupes'] <= 5).sum()
            ).reindex(working_df['BaseName']).values

            working_df['under_10'] = working_df.groupby('BaseName').apply(
                lambda x: (x['dupes'] <= 10).sum()
            ).reindex(working_df['BaseName']).values

            working_df = working_df.reset_index()
            working_df['percentile_finish'] = working_df['index'].rank(pct=True)
            working_df['finish'] = working_df['index']
            working_df = working_df.drop(['sorted', 'index'], axis=1)
        
        elif type_var == 'Showdown':
            working_df['stack'] = working_df.apply(
                lambda row: Counter(
                    map_dict['team_map'].get(player, '') for player in row[2:]
                    if map_dict['team_map'].get(player, '') != ''
                ).most_common(1)[0][0] if any(map_dict['team_map'].get(player, '') for player in row[2:]) else '',
                axis=1
            )
            working_df['stack_size'] = working_df.apply(
                lambda row: Counter(
                    map_dict['team_map'].get(player, '') for player in row[2:]
                    if map_dict['team_map'].get(player, '') != ''
                ).most_common(1)[0][1] if any(map_dict['team_map'].get(player, '') for player in row[2:]) else '',
                axis=1
            )
            # Modified salary calculation with 1.5x multiplier for first player
            working_df['salary'] = working_df.apply(
                lambda row: (map_dict['salary_map'].get(row[2], 0) * 1.5) + 
                           sum(map_dict['salary_map'].get(player, 0) for player in row[3:]),
                axis=1
            )
            # Modified actual_fpts calculation with 1.5x multiplier for first player
            working_df['actual_fpts'] = working_df.apply(
                lambda row: (st.session_state['actual_dict'].get(row[2], 0) * 1.5) + 
                           sum(st.session_state['actual_dict'].get(player, 0) for player in row[3:]),
                axis=1
            )
            working_df['actual_own'] = working_df.apply(lambda row: sum(st.session_state['ownership_dict'].get(player, 0) for player in row), axis=1)
            working_df['sorted'] = working_df[player_columns].apply(
                lambda row: ','.join(sorted(row.values)),
                axis=1
            )
            working_df['dupes'] = working_df.groupby('sorted').transform('size')

            working_df['uniques'] = working_df.groupby('BaseName').apply(
                lambda x: (x['dupes'] == 1).sum()
            ).reindex(working_df['BaseName']).values

            working_df['under_5'] = working_df.groupby('BaseName').apply(
                lambda x: (x['dupes'] <= 5).sum()
            ).reindex(working_df['BaseName']).values

            working_df['under_10'] = working_df.groupby('BaseName').apply(
                lambda x: (x['dupes'] <= 10).sum()
            ).reindex(working_df['BaseName']).values

            working_df = working_df.reset_index()
            working_df['percentile_finish'] = working_df['index'].rank(pct=True)
            working_df['finish'] = working_df['index']
            working_df = working_df.drop(['sorted', 'index'], axis=1)
        # working_df['stack_size'] = working_df['stack_size'].fillna(1).astype(int)
        st.session_state['field_player_frame'] = create_player_exposures(working_df, player_columns)
        st.session_state['field_stack_frame'] = create_stack_exposures(working_df)

        with st.expander("Info and filters"):
            st.info("Note that any filtering here needs to be reset manually, i.e. if you parse down the specific users and want to reset the table, just backtrack your filtering by setting it back to 'All'")
            if st.button('Clear data', key='reset3'):
                st.session_state.clear()

            with st.form(key='filter_form'):
                users_var, entries_var, stack_var, stack_size_var, player_var = st.columns(5)
                with users_var:
                    entry_parse_var = st.selectbox("Do you want to view a specific user(s)?", ['All', 'Specific'], key = 'entry_parse_var')
                    entry_names = st.multiselect("Select players", options=st.session_state['entry_list'], default=[], key = 'entry_names')
                with entries_var:
                    low_entries_var = st.number_input("Low end of entries range", min_value=0, max_value=150, value=1, key = 'low_entries_var')
                    high_entries_var = st.number_input("High end of entries range", min_value=0, max_value=150, value=150, key = 'high_entries_var')
                with stack_var:
                    stack_parse_var = st.selectbox("Do you want to view lineups with specific team(s)?", ['All', 'Specific'], key = 'stack_parse_var')
                    stack_names = st.multiselect("Select teams", options=working_df['stack'].unique(), default=[], key = 'stack_names')
                with stack_size_var:
                    stack_size_parse_var = st.selectbox("Do you want to view a specific stack size(s)?", ['All', 'Specific'], key = 'stack_size_parse_var')
                    stack_size_names = st.multiselect("Select stack sizes", options=working_df['stack_size'].unique(), default=[], key = 'stack_size_names')
                with player_var:
                    unique_players = pd.unique(working_df[player_columns].values.ravel('K'))
                    unique_players = [p for p in unique_players if p != 'nan']  # Remove any NaN values
                    player_parse_var = st.selectbox("Do you want to view lineups with specific player(s)?", ['All', 'Specific'], key = 'player_parse_var')
                    player_names = st.multiselect("Select players", options=unique_players, default=[], key = 'player_names')
                submitted = st.form_submit_button("Submit")
                if submitted:
                    if 'player_frame' in st.session_state:
                        del st.session_state['player_frame']
                    if 'stack_frame' in st.session_state:
                        del st.session_state['stack_frame']

                    if entry_parse_var == 'Specific' and entry_names:
                        working_df = working_df[working_df['BaseName'].isin(entry_names)]
                    if stack_parse_var == 'Specific' and stack_names:
                        working_df = working_df[working_df['stack'].isin(stack_names)]
                    if stack_size_parse_var == 'Specific' and stack_size_names:
                        working_df = working_df[working_df['stack_size'].isin(stack_size_names)]
                    if player_parse_var == 'Specific' and player_names:
                        mask = working_df[player_columns].apply(lambda row: all(player in row.values for player in player_names), axis=1)
                        working_df = working_df[mask]
                    if low_entries_var and high_entries_var:
                        working_df = working_df[working_df['EntryCount'].between(low_entries_var, high_entries_var)]
            
        # Initialize pagination in session state if not exists
        if 'current_page' not in st.session_state:
            st.session_state.current_page = 1

        # Calculate total pages
        rows_per_page = 500
        total_rows = len(working_df)
        total_pages = (total_rows + rows_per_page - 1) // rows_per_page

        # Create pagination controls in a single row
        pagination_cols = st.columns([4, 1, 1, 1, 4])
        with pagination_cols[1]:
            if st.button(f"Previous Page"):
                if st.session_state['current_page'] > 1:
                    st.session_state.current_page -= 1
                else:
                    st.session_state.current_page = 1
                    if 'player_frame' in st.session_state:
                        del st.session_state['player_frame']
                    if 'stack_frame' in st.session_state:
                        del st.session_state['stack_frame']

        with pagination_cols[3]:
            if st.button(f"Next Page"):
                st.session_state.current_page += 1
                if 'player_frame' in st.session_state:
                    del st.session_state['player_frame']
                if 'stack_frame' in st.session_state:
                    del st.session_state['stack_frame']

        # Calculate start and end indices for current page
        start_idx = (st.session_state.current_page - 1) * rows_per_page
        end_idx = min((st.session_state.current_page) * rows_per_page, total_rows)
        st.dataframe(
            working_df.iloc[start_idx:end_idx].style
            .background_gradient(axis=0)
            .background_gradient(cmap='RdYlGn')
            .format(precision=2), 
            height=500,
            use_container_width=True,
            hide_index=True
        )

        with st.container():
            tab1, tab2, tab3, tab4, tab5 = st.tabs(['Player Used Info', 'Stack Used Info', 'Stack Size Info', 'General Info', 'Duplication Info'])
            with tab1:
                with st.form(key='player_info_pos_form'):
                    col1, col2 = st.columns(2)
                    with col1:
                        pos_var = st.selectbox("Which position(s) would you like to view?",  ['All', 'Specific'], key='pos_var')
                    with col2:
                        pos_select = st.multiselect("Select your position(s)", ['P', 'C', '1B', '2B', '3B', 'SS', 'OF'], key='pos_select')
                    submitted = st.form_submit_button("Submit")
                    if submitted:
                        if pos_var == 'Specific':
                            pos_select = pos_select
                        else:
                            pos_select = None

                if entry_parse_var == 'All':

                    st.session_state['player_frame'] = create_player_exposures(working_df, player_columns)
                    hold_frame = st.session_state['player_frame'].copy()
                    if sport_select == 'GOLF':
                        hold_frame['Pos'] = 'G'
                    else:
                        hold_frame['Pos'] = hold_frame['Player'].map(map_dict['pos_map'])
                    st.session_state['player_frame'].insert(1, 'Pos', hold_frame['Pos'])
                    st.session_state['player_frame'] = st.session_state['player_frame'].dropna(subset=['Pos'])
                    if pos_select:
                        position_mask = st.session_state['player_frame']['Pos'].apply(lambda x: any(pos in x for pos in pos_select))
                        st.session_state['player_frame'] = st.session_state['player_frame'][position_mask]
                    st.dataframe(st.session_state['player_frame'].
                        sort_values(by='Exposure Overall', ascending=False).
                        style.background_gradient(cmap='RdYlGn').
                        format(formatter='{:.2%}', subset=st.session_state['player_frame'].iloc[:, 2:].select_dtypes(include=['number']).columns),
                        hide_index=True)
                else:

                    st.session_state['player_frame'] = create_player_exposures(working_df, player_columns, entry_names)
                    hold_frame = st.session_state['player_frame'].copy()
                    if sport_select == 'GOLF':
                        hold_frame['Pos'] = 'G'
                    else:
                        hold_frame['Pos'] = hold_frame['Player'].map(map_dict['pos_map'])
                    st.session_state['player_frame'].insert(1, 'Pos', hold_frame['Pos'])
                    st.session_state['player_frame'] = st.session_state['player_frame'].dropna(subset=['Pos'])
                    if pos_select:
                        position_mask = st.session_state['player_frame']['Pos'].apply(lambda x: any(pos in x for pos in pos_select))
                        st.session_state['player_frame'] = st.session_state['player_frame'][position_mask]
                    st.dataframe(st.session_state['player_frame'].
                        sort_values(by='Exposure Overall', ascending=False).
                        style.background_gradient(cmap='RdYlGn').
                        format(formatter='{:.2%}', subset=st.session_state['player_frame'].iloc[:, 2:].select_dtypes(include=['number']).columns),
                        hide_index=True)
            with tab2:

                if entry_parse_var == 'All':
                    st.session_state['stack_frame'] = create_stack_exposures(working_df)
                    st.dataframe(st.session_state['stack_frame'].
                        sort_values(by='Exposure Overall', ascending=False).
                        style.background_gradient(cmap='RdYlGn').
                        format(formatter='{:.2%}', subset=st.session_state['stack_frame'].iloc[:, 1:].select_dtypes(include=['number']).columns),
                        hide_index=True)
                else:
                    st.session_state['stack_frame'] = create_stack_exposures(working_df, entry_names)
                    st.dataframe(st.session_state['stack_frame'].
                        sort_values(by='Exposure Overall', ascending=False).
                        style.background_gradient(cmap='RdYlGn').
                        format(formatter='{:.2%}', subset=st.session_state['stack_frame'].iloc[:, 1:].select_dtypes(include=['number']).columns),
                        hide_index=True)
            with tab3:
                
                if entry_parse_var == 'All':
                    st.session_state['stack_size_frame'] = create_stack_size_exposures(working_df)
                    st.dataframe(st.session_state['stack_size_frame'].
                        sort_values(by='Exposure Overall', ascending=False).
                        style.background_gradient(cmap='RdYlGn').
                        format(formatter='{:.2%}', subset=st.session_state['stack_size_frame'].iloc[:, 1:].select_dtypes(include=['number']).columns),
                        hide_index=True)
                else:
                    st.session_state['stack_size_frame'] = create_stack_size_exposures(working_df, entry_names)
                    st.dataframe(st.session_state['stack_size_frame'].
                        sort_values(by='Exposure Overall', ascending=False).
                        style.background_gradient(cmap='RdYlGn').
                        format(formatter='{:.2%}', subset=st.session_state['stack_size_frame'].iloc[:, 1:].select_dtypes(include=['number']).columns),
                        hide_index=True)
            
            with tab4:
                
                if entry_parse_var == 'All':
                    st.session_state['general_frame'] = create_general_exposures(working_df)
                    st.dataframe(st.session_state['general_frame'].style.background_gradient(cmap='RdYlGn', axis=1).format(precision=2), hide_index=True)
                    
                else:
                    st.session_state['general_frame'] = create_general_exposures(working_df, entry_names)
                    st.dataframe(st.session_state['general_frame'].style.background_gradient(cmap='RdYlGn', axis=1).format(precision=2), hide_index=True)

            with tab5:
                dupe_frame = working_df[['BaseName', 'EntryCount', 'dupes', 'uniques', 'under_5', 'under_10']]
                dupe_frame['average_dupes'] = dupe_frame['dupes'].mean()
                dupe_frame['uniques%'] = dupe_frame['uniques'] / dupe_frame['EntryCount']
                dupe_frame['under_5%'] = dupe_frame['under_5'] / dupe_frame['EntryCount']
                dupe_frame['under_10%'] = dupe_frame['under_10'] / dupe_frame['EntryCount']
                st.session_state['duplication_frame'] = dupe_frame[['BaseName', 'EntryCount', 'average_dupes', 'dupes', 'uniques', 'uniques%', 'under_5', 'under_5%', 'under_10', 'under_10%']].drop_duplicates(subset='BaseName', keep='first')
                st.dataframe(st.session_state['duplication_frame'].style.
                             background_gradient(cmap='RdYlGn', subset=['uniques%', 'under_5%', 'under_10%'], axis=0).
                             background_gradient(cmap='RdYlGn_r', subset=['uniques', 'under_5', 'under_10'], axis=0).
                             format(dupe_format, precision=2), hide_index=True)