global_func/exposure_spread.py

import random
import numpy as np
import math

#### Goal is to choose a player and adjust the amount of lineups that have them
#### First thing you need to do is find comparable players in the projections, so any player in the projections that is within $500 of the player and within 10% of the projection
#### Take that list of players and create a list that can be accessed for random insertion into the portfolio
#### Find the player and the amount of rows that contain them and then find an exposure rate which is the percentage of total rows
#### Use the exposure target argument and try to replace the player from as many rows as necessary to be at or just under the target
#### makes sure to check if the player is eligible for the position before replacing them

def check_nba_position_eligibility(column_name, player_positions):
    if 'PG' in column_name:
        return 'PG' in player_positions
    elif 'SG' in column_name:
        return 'SG' in player_positions
    elif 'SF' in column_name:
        return 'SF' in player_positions
    elif 'PF' in column_name:
        return 'PF' in player_positions
    elif 'C' in column_name:
        return 'C' in player_positions
    elif 'G' in column_name:
        return any(pos in ['PG', 'SG'] for pos in player_positions)
    elif 'F' in column_name:
        return any(pos in ['SF', 'PF'] for pos in player_positions)
    elif 'UTIL' in column_name:
        return True  # UTIL can be any position
    return False

def check_lol_position_eligibility(column_name, player_positions):
    if 'TOP' in column_name:
        return 'TOP' in player_positions
    elif 'JNG' in column_name:
        return 'JNG' in player_positions
    elif 'MID' in column_name:
        return 'MID' in player_positions
    elif 'ADC' in column_name:
        return 'ADC' in player_positions
    elif 'SUP' in column_name:
        return 'SUP' in player_positions
    elif 'Team' in column_name:
        return 'Team' in player_positions
    elif 'CPT' in column_name:
        return any(pos in ['TOP', 'JNG', 'MID', 'ADC', 'SUP'] for pos in player_positions)
    return False

def check_mlb_position_eligibility(column_name, player_positions):
    if any(pos in column_name for pos in ['P', 'SP', 'RP']):
        return any(pos in ['P', 'SP', 'RP'] for pos in player_positions)
    elif 'C' in column_name:
        return 'C' in player_positions
    elif '1B' in column_name:
        return '1B' in player_positions
    elif '2B' in column_name:
        return '2B' in player_positions
    elif '3B' in column_name:
        return '3B' in player_positions
    elif 'SS' in column_name:
        return 'SS' in player_positions
    elif 'OF' in column_name:
        return 'OF' in player_positions
    return False

def check_nfl_position_eligibility(column_name, player_positions):
    if 'QB' in column_name:
        return 'QB' in player_positions
    elif 'RB' in column_name:
        return 'RB' in player_positions
    elif 'WR' in column_name:
        return 'WR' in player_positions
    elif 'TE' in column_name:
        return 'TE' in player_positions
    elif 'DST' in column_name:
        return 'DST' in player_positions
    elif 'FLEX' in column_name:
        return any(pos in ['RB', 'WR', 'TE'] for pos in player_positions)
    elif 'UTIL' in column_name:
        return any(pos in ['RB', 'WR', 'TE'] for pos in player_positions)
    return False

def check_golf_position_eligibility(column_name, player_positions):
    if 'FLEX' in column_name:
        return any(pos in ['G'] for pos in player_positions)
    return True

def check_tennis_position_eligibility(column_name, player_positions):
    if 'FLEX' in column_name:
        return any(pos in ['T'] for pos in player_positions)
    return True

def check_mma_position_eligibility(column_name, player_positions):
    if 'FLEX' in column_name:
        return any(pos in ['F'] for pos in player_positions)
    return True

def check_nascar_position_eligibility(column_name, player_positions):
    if 'FLEX' in column_name:
        return any(pos in ['D'] for pos in player_positions)
    return True

def check_ncaaf_position_eligibility(column_name, player_positions):
    if 'QB' in column_name:
        return 'QB' in player_positions
    elif 'RB' in column_name:
        return 'RB' in player_positions
    elif 'WR' in column_name:
        return 'WR' in player_positions
    elif 'FLEX' in column_name:
        return any(pos in ['RB', 'WR'] for pos in player_positions)
    elif 'SFLEX' in column_name:
        return any(pos in ['RB', 'WR', 'QB'] for pos in player_positions)
    return False

def check_nhl_position_eligibility(column_name, player_positions):
    if 'C' in column_name:
        return 'C' in player_positions
    elif 'W' in column_name:
        return 'W' in player_positions
    elif 'D' in column_name:
        return 'D' in player_positions
    elif 'G' in column_name:
        return 'G' in player_positions
    elif 'FLEX' in column_name:
        return any(pos in ['C', 'W', 'D'] for pos in player_positions)
    elif 'UTIL' in column_name:
        return any(pos in ['C', 'W', 'D'] for pos in player_positions)
    return False

def check_position_eligibility(sport, column_name, player_positions):
    if sport == 'NBA':
        return check_nba_position_eligibility(column_name, player_positions)
    elif sport == 'MLB':
        return check_mlb_position_eligibility(column_name, player_positions)
    elif sport == 'NFL':
        return check_nfl_position_eligibility(column_name, player_positions)
    elif sport == 'NHL':
        return check_nhl_position_eligibility(column_name, player_positions)
    elif sport == 'MMA':
        return check_mma_position_eligibility(column_name, player_positions)
    elif sport == 'GOLF':
        return check_golf_position_eligibility(column_name, player_positions)
    elif sport == 'TENNIS':
        return check_tennis_position_eligibility(column_name, player_positions)
    elif sport == 'LOL':
        return check_lol_position_eligibility(column_name, player_positions)
    else:
        # Default fallback - assume exact position match
        return column_name in player_positions

def exposure_spread(working_frame, exposure_player, exposure_target, ignore_stacks, remove_teams, specific_replacements, specific_columns, projections_df, sport_var, type_var, salary_max, stacking_sports):
    comparable_players = projections_df[projections_df['player_names'] == exposure_player]
    
    comparable_players = comparable_players.reset_index(drop=True)
    comp_salary_high = comparable_players['salary'][0]
    if type_var == 'Showdown':
        comp_salary_low = comparable_players['salary'][0] - 1000
    else:
        comp_salary_low = comparable_players['salary'][0] - 500
    comp_projection_high = comparable_players['median'][0]
    if type_var == 'Showdown':
        comp_projection_low = comparable_players['median'][0] - (comparable_players['median'][0] * .5)
    else:
        comp_projection_low = comparable_players['median'][0] - (comparable_players['median'][0] * .75)
    # players can be eligible at multiple positions, so we need to find all the positions the player is eligible at
    # the position column can have positions designated as 1B/OF which means they are eligible at 1B and OF
    comp_player_position = comparable_players['position'].tolist()
    comp_team = comparable_players['team'].tolist()
    try:
        comp_player_position = [pos.split('/') for pos in comp_player_position]
        comp_player_position = [item for sublist in comp_player_position for item in sublist]
        comp_player_position = list(set(comp_player_position))
    except:
        comp_player_position = comparable_players['position'].tolist()

    def has_position_overlap(player_positions, target_positions):
        player_pos_list = player_positions.split('/')
        return any(pos in target_positions for pos in player_pos_list)

    # find the exposure rate of the player in the working frame
    if specific_columns != []:
        player_mask = working_frame[specific_columns].apply(
            lambda row: exposure_player in list(row), axis=1
        )
    else:
        player_mask = working_frame[working_frame.columns].apply(
            lambda row: exposure_player in list(row), axis=1
        )
    
    if specific_columns != []:
        replace_mask = working_frame[specific_columns].apply(
                lambda row: exposure_player not in list(row), axis=1
            )
    else:
        replace_mask = working_frame[working_frame.columns].apply(
            lambda row: exposure_player not in list(row), axis=1
        )

    player_exposure = player_mask.sum() / len(working_frame)
    replace_exposure = replace_mask.sum() / len(working_frame)

    # find the number of lineups that need to be removed to reach the target exposure
    if exposure_target == 0:
        lineups_to_remove = (player_exposure * len(working_frame))
    else:
        lineups_to_remove = ((player_exposure - exposure_target) * len(working_frame)) * 1.01
        lineups_to_add = ((exposure_target - player_exposure) * (len(working_frame) - (player_exposure * len(working_frame)))) * 1.10
        
    # isolate the rows that contain the player
    player_rows = working_frame[player_mask]
    replace_rows = working_frame[replace_mask]
    if ignore_stacks != []:
        player_rows = player_rows[~player_rows['Stack'].isin(ignore_stacks)]
        replace_rows = replace_rows[~replace_rows['Stack'].isin(ignore_stacks)]
    
    change_counter = 0

    random_row_indices_insert = list(player_rows.index)
    random_row_indices_replace = list(replace_rows.index)
    random.shuffle(random_row_indices_insert)
    random.shuffle(random_row_indices_replace)

    # for each row to the the number of lineups to remove, replace with random choice from comparable player list if they can be inserted

    # we will need to use two separate functions here, one for an exposure player who has a lineups to remove above 0 and one for below 0
    # key concept here is if they have a lineups to remove above 0 it means that we are trying to replace them with comparable players
    # if the lineups to remove is below zero it means we want to find comparable players and replace them with the exposure player
    if lineups_to_remove > 0:
        for row in random_row_indices_insert:
            if change_counter < math.ceil(lineups_to_remove):
                if specific_replacements != []:
                    comparable_players = projections_df[(projections_df['player_names'].isin(specific_replacements)) &
                        (projections_df['salary'] <= comp_salary_high + (salary_max - working_frame['salary'][row]))
                    ]
                else:
                    comparable_players = projections_df[
                        (projections_df['salary'] >= comp_salary_low) &
                        (projections_df['salary'] <= comp_salary_high + (salary_max - working_frame['salary'][row])) &
                        (projections_df['median'] >= comp_projection_low) &
                        (projections_df['position'].apply(lambda x: has_position_overlap(x, comp_player_position)))
                    ]

                if exposure_target == 0:
                    comparable_players = comparable_players[comparable_players['player_names'] != exposure_player]
                
                if remove_teams is not None:
                    remove_mask = comparable_players.apply(
                            lambda row: not any(team in list(row) for team in remove_teams), axis=1
                        )
                    comparable_players = comparable_players[remove_mask]

                # Get the current row data to check for existing players
                if specific_columns != []:
                    current_row_data = working_frame.iloc[row][specific_columns]
                else:
                    current_row_data = working_frame.iloc[row]
                
                # Filter out players that are already present in this row
                existing_players = set(current_row_data.values)
                try:
                    comparable_players = comparable_players[~comparable_players['player_names'].isin(existing_players)]
                    comparable_player_list = comparable_players['player_names'].tolist()
                except:
                    comparable_player_list = []

                if comparable_player_list:
                    insert_player = random.choice(comparable_player_list)
                    # Find which column contains the exposure_player
                    if specific_columns != []:
                        row_data = working_frame.iloc[row][specific_columns]
                        working_columns = specific_columns
                    else:
                        row_data = working_frame.iloc[row]
                        working_columns = working_frame.columns
                    
                    for col in working_columns:
                        if row_data[col] == exposure_player:
                            # Get the replacement player's positions
                            replacement_player_positions = projections_df[projections_df['player_names'] == insert_player]['position'].iloc[0].split('/')
                            
                            # Check if the replacement player is eligible for this column
                            if type_var == 'Classic':
                                if check_position_eligibility(sport_var, col, replacement_player_positions):
                                    working_frame.at[row, col] = insert_player
                                    break
                            else:
                                working_frame.at[row, col] = insert_player
                                break
                    change_counter += 1
    else:
        for row in random_row_indices_replace:
            if change_counter < math.ceil(lineups_to_add):
                if specific_replacements != []:
                    comparable_players = projections_df[(projections_df['player_names'].isin(specific_replacements))
                    ]
                else:
                    if type_var == 'Showdown':
                        comparable_players = projections_df[
                            (projections_df['salary'] >= comp_salary_low) &
                            (projections_df['salary'] <= comp_salary_high + (salary_max - working_frame['salary'][row]))
                        ]
                    else:
                        comparable_players = projections_df[
                            (projections_df['salary'] >= comp_salary_low) &
                            (projections_df['salary'] <= comp_salary_high + (salary_max - working_frame['salary'][row])) &
                            (projections_df['position'].apply(lambda x: has_position_overlap(x, comp_player_position)))
                        ]
                    if sport_var in stacking_sports:
                        if working_frame.iloc[row]['Size'] == 5 and comp_team != working_frame.iloc[row]['Stack']:
                            remove_mask = comparable_players.apply(
                                lambda player_row: not any(team in list(player_row) for team in [working_frame.iloc[row]['Stack']]), axis=1
                            )
                            comparable_players = comparable_players[remove_mask]
                
                if remove_teams is not None:
                    remove_mask = comparable_players.apply(
                            lambda row: not any(team in list(row) for team in remove_teams), axis=1
                        )
                    comparable_players = comparable_players[remove_mask]

                comparable_players = comparable_players[comparable_players['player_names'] != exposure_player]

                # Create a list of comparable players
                comparable_player_list = comparable_players['player_names'].tolist()
                if comparable_player_list:
                    # Find which column contains the exposure_player
                    if specific_columns != []:
                        row_data = working_frame.iloc[row][specific_columns]
                        working_columns = specific_columns
                    else:
                        row_data = working_frame.iloc[row]
                        working_columns = working_frame.columns
    
                    for col in working_columns:
                        if row_data[col] in comparable_player_list:
                            if working_frame.iloc[row]['salary'] - projections_df[projections_df['player_names'] == row_data[col]]['salary'].iloc[0] + projections_df[projections_df['player_names'] == exposure_player]['salary'].iloc[0] <= salary_max:
                                if type_var == 'Classic':
                                    replacement_player_positions = projections_df[projections_df['player_names'] == row_data[col]]['position'].iloc[0].split('/')
                                    exposure_player_positions = projections_df[projections_df['player_names'] == exposure_player]['position'].iloc[0].split('/')
                                    
                                    # Check if the replacement player is eligible for this column
                                
                                    if check_position_eligibility(sport_var, col, exposure_player_positions):
                                        working_frame.at[row, col] = exposure_player
                                        change_counter += 1
                                        break
                                else:
                                    working_frame.at[row, col] = exposure_player
                                    change_counter += 1
                                    break
                            else:
                                continue
    return working_frame