867 lines
38 KiB
Python
867 lines
38 KiB
Python
import asyncio
|
|
import datetime
|
|
import logging
|
|
import sys
|
|
|
|
from typing import Literal
|
|
|
|
import pandas as pd
|
|
import pybaseball as pb
|
|
from pybaseball import cache
|
|
import urllib
|
|
|
|
from creation_helpers import get_args, CLUB_LIST, FRANCHISE_LIST
|
|
from batters.stat_prep import DataMismatchError
|
|
from db_calls import DB_URL, db_get, db_patch, db_post, db_put
|
|
import batters.calcs_batter as cba
|
|
import defenders.calcs_defense as cde
|
|
|
|
cache.enable()
|
|
date = f'{datetime.datetime.now().year}-{datetime.datetime.now().month}-{datetime.datetime.now().day}'
|
|
log_level = logging.INFO
|
|
logging.basicConfig(
|
|
filename=f'logs/{date}.log',
|
|
format='%(asctime)s - retrosheet_data - %(levelname)s - %(message)s',
|
|
level=log_level
|
|
)
|
|
|
|
RETRO_FILE_PATH = 'data-input/retrosheet/'
|
|
EVENTS_FILENAME = 'retrosheets_events_1998_short.csv' # Removed last few columns which were throwing dtype errors
|
|
PERSONNEL_FILENAME = 'retrosheets_personnel.csv'
|
|
DATA_INPUT_FILE_PATH = 'data-input/1998 Season Cardset/'
|
|
CARD_BASE_URL = f'{DB_URL}/v2/players/'
|
|
|
|
start_time = datetime.datetime.now()
|
|
RELEASE_DIRECTORY = f'{start_time.year}-{start_time.month}-{start_time.day}'
|
|
|
|
MIN_PA_VL = 20
|
|
MIN_PA_VR = 40
|
|
MIN_TBF_VL = MIN_PA_VL
|
|
MIN_TBF_VR = MIN_PA_VR
|
|
CARDSET_ID = 20
|
|
PLAYER_DESCRIPTION = 'Live'
|
|
|
|
|
|
async def store_defense_to_csv(season: int):
|
|
for position in ['c', '1b', '2b', '3b', 'ss', 'lf', 'cf', 'rf', 'of', 'p']:
|
|
pos_df = cde.get_bbref_fielding_df(position, season)
|
|
pos_df.to_csv(f'{DATA_INPUT_FILE_PATH}defense_{position}.csv')
|
|
await asyncio.sleep(8)
|
|
|
|
|
|
def get_events_by_date(file_path: str, start_date: int, end_date: int) -> pd.DataFrame:
|
|
all_plays = pd.read_csv(f'{file_path}', dtype={'game_id': 'str'})
|
|
all_plays['date'] = all_plays['game_id'].str[3:-1].astype(int)
|
|
date_plays = all_plays[(all_plays.date >= start_date) & (all_plays.date <= end_date)]
|
|
return date_plays
|
|
|
|
|
|
def get_result_series(plays: pd.DataFrame, event_type: str, pitcher_hand: Literal['r', 'l'], col_name: str) -> pd.Series:
|
|
this_series = plays[(plays.event_type == event_type) & (plays.pitcher_hand == pitcher_hand)].groupby('batter_id').count()['event_type'].astype(int).rename(col_name)
|
|
return this_series
|
|
|
|
|
|
def get_run_stat_df(input_path: str):
|
|
run_data = pd.read_csv(f'{input_path}running.csv') #.set_index('Name-additional'))
|
|
# if 'Player' in run_data:
|
|
# run_data = run_data.rename(columns={'Player': 'Full Name'})
|
|
# if 'Name' in run_data:
|
|
# run_data = run_data.rename(columns={'Name': 'Full Name'})
|
|
if 'Player-additional' in run_data:
|
|
run_data = run_data.rename(columns={'Player-additional': 'key_bbref'})
|
|
if 'Name-additional' in run_data:
|
|
run_data = run_data.rename(columns={'Name-additional': 'key_bbref'})
|
|
|
|
run_data = run_data[['key_bbref', 'Tm', 'ROE', 'XI', 'RS%', 'SBO', 'SB', 'CS', 'SB%', 'SB2', 'CS2', 'SB3', 'CS3', 'SBH', 'CSH', 'PO', 'PCS', 'OOB', 'OOB1', 'OOB2', 'OOB3', 'OOBHm', 'BT', 'XBT%', '1stS', '1stS2', '1stS3', '1stD', '1stD3', '1stDH', '2ndS', '2ndS3', '2ndSH']]
|
|
|
|
run_data = run_data.fillna(0)
|
|
return run_data.set_index('key_bbref')
|
|
|
|
|
|
def get_player_ids(plays: pd.DataFrame, which: Literal['batters', 'pitchers']) -> pd.DataFrame:
|
|
RETRO_PLAYERS = pd.read_csv(f'{RETRO_FILE_PATH}{PERSONNEL_FILENAME}')
|
|
id_key = 'batter_id' if which == 'batters' else 'pitcher_id'
|
|
|
|
players = pd.DataFrame()
|
|
unique_players = pd.Series(plays[id_key].unique()).to_frame('id')
|
|
players = pd.merge(
|
|
left=RETRO_PLAYERS,
|
|
right=unique_players,
|
|
how='right',
|
|
left_on='id',
|
|
right_on='id'
|
|
).rename(columns={'id': id_key})
|
|
|
|
def get_pids(row):
|
|
# return get_all_pybaseball_ids([row[id_key]], 'retro', full_name=f'{row["use_name"]} {row["last_name"]}')
|
|
pull = pb.playerid_reverse_lookup([row[id_key]], key_type='retro')
|
|
if len(pull.values) == 0:
|
|
print(f'Could not find id {row[id_key]} in pybaseball lookup')
|
|
return pull.loc[0][['key_mlbam', 'key_retro', 'key_bbref', 'key_fangraphs']]
|
|
|
|
players = players[[id_key, 'last_name', 'use_name']]
|
|
start_time = datetime.datetime.now()
|
|
other_ids = players.apply(get_pids, axis=1)
|
|
end_time = datetime.datetime.now()
|
|
print(f'ID lookup: {(end_time - start_time).total_seconds():.2f}s')
|
|
|
|
players = pd.merge(
|
|
left=players,
|
|
right=other_ids,
|
|
left_on=id_key,
|
|
right_on='key_retro'
|
|
)
|
|
players = players.set_index(id_key)
|
|
|
|
def get_bat_hand(row):
|
|
pa_vl = plays[(plays.batter_id == row['key_retro']) & (plays.pitcher_hand == 'l')].groupby('result_batter_hand').count()['game_id'].astype(int)
|
|
pa_vr = plays[(plays.batter_id == row['key_retro']) & (plays.pitcher_hand == 'r')].groupby('result_batter_hand').count()['game_id'].astype(int)
|
|
|
|
l_vs_l = 0 if 'l' not in pa_vl else pa_vl['l']
|
|
l_vs_r = 0 if 'l' not in pa_vr else pa_vr['l']
|
|
r_vs_l = 0 if 'r' not in pa_vl else pa_vl['r']
|
|
r_vs_r = 0 if 'r' not in pa_vr else pa_vr['r']
|
|
|
|
if sum([l_vs_l, l_vs_r]) == 0 and sum([r_vs_l, r_vs_r]) > 0:
|
|
return 'R'
|
|
elif sum([l_vs_l, l_vs_r]) > 0 and sum([r_vs_l, r_vs_r]) == 0:
|
|
return 'L'
|
|
|
|
if sum([l_vs_l, l_vs_r, r_vs_l, r_vs_r]) < 10:
|
|
if sum([l_vs_l, l_vs_r]) > sum([r_vs_l, r_vs_r]):
|
|
return 'L'
|
|
else:
|
|
return 'R'
|
|
else:
|
|
return 'S'
|
|
|
|
if which == 'batters':
|
|
players['bat_hand'] = players.apply(get_bat_hand, axis=1)
|
|
|
|
return players
|
|
|
|
|
|
def get_base_batting_df(all_plays: pd.DataFrame) -> pd.DataFrame:
|
|
bs = get_player_ids(all_plays, 'batters')
|
|
|
|
pal_series = all_plays[(all_plays.batter_event == 't') & (all_plays.pitcher_hand == 'l')].groupby('batter_id').count()['event_type'].astype(int).rename('PA_vL')
|
|
bs = pd.concat([bs, pal_series], axis=1)
|
|
par_series = all_plays[(all_plays.batter_event == 't') & (all_plays.pitcher_hand == 'r')].groupby('batter_id').count()['event_type'].astype(int).rename('PA_vR')
|
|
bs = pd.concat([bs, par_series], axis=1)
|
|
|
|
abl_series = all_plays[(all_plays.ab == 't') & (all_plays.pitcher_hand == 'l')].groupby('batter_id').count()['event_type'].astype(int).rename('AB_vL')
|
|
bs = pd.concat([bs, abl_series], axis=1)
|
|
abr_series = all_plays[(all_plays.ab == 't') & (all_plays.pitcher_hand == 'r')].groupby('batter_id').count()['event_type'].astype(int).rename('AB_vR')
|
|
bs = pd.concat([bs, abr_series], axis=1)
|
|
|
|
return bs.dropna().query(f'PA_vL >= {MIN_PA_VL} & PA_vR >= {MIN_PA_VR}')
|
|
|
|
|
|
def get_batting_stats_by_date(retro_file_path, start_date: int, end_date: int) -> pd.DataFrame:
|
|
start = datetime.datetime.now()
|
|
all_plays = get_events_by_date(retro_file_path, start_date, end_date)
|
|
print(f'Pull events: {(datetime.datetime.now() - start).total_seconds():.2f}s')
|
|
|
|
start = datetime.datetime.now()
|
|
batting_stats = get_base_batting_df(all_plays)
|
|
print(f'Get base dataframe: {(datetime.datetime.now() - start).total_seconds():.2f}s')
|
|
|
|
start = datetime.datetime.now()
|
|
all_player_ids = batting_stats['key_retro']
|
|
all_plays = all_plays[all_plays['batter_id'].isin(all_player_ids)]
|
|
print(f'Shrink all_plays: {(datetime.datetime.now() - start).total_seconds():.2f}s')
|
|
|
|
# Basic counting stats
|
|
start = datetime.datetime.now()
|
|
for event_type, vs_hand, col_name in [
|
|
('home run', 'r', 'HR_vR'),
|
|
('home run', 'l', 'HR_vL'),
|
|
('single', 'r', '1B_vR'),
|
|
('single', 'l', '1B_vL'),
|
|
('double', 'r', '2B_vR'),
|
|
('double', 'l', '2B_vL'),
|
|
('triple', 'r', '3B_vR'),
|
|
('triple', 'l', '3B_vL'),
|
|
('walk', 'r', 'BB_vR'),
|
|
('walk', 'l', 'BB_vL'),
|
|
('strikeout', 'r', 'SO_vR'),
|
|
('strikeout', 'l', 'SO_vL'),
|
|
('hit by pitch', 'r', 'HBP_vR'),
|
|
('hit by pitch', 'l', 'HBP_vL')
|
|
]:
|
|
this_series = get_result_series(all_plays, event_type, vs_hand, col_name)
|
|
batting_stats[col_name] = this_series
|
|
print(f'Count basic stats: {(datetime.datetime.now() - start).total_seconds():.2f}s')
|
|
|
|
# Bespoke counting stats
|
|
start = datetime.datetime.now()
|
|
def get_fb_vl(row):
|
|
return all_plays[(all_plays.batter_id == row['key_retro']) & (all_plays.batted_ball_type == 'f') & (all_plays.pitcher_hand == 'l')].count()['event_type'].astype(int)
|
|
def get_fb_vr(row):
|
|
return all_plays[(all_plays.batter_id == row['key_retro']) & (all_plays.batted_ball_type == 'f') & (all_plays.pitcher_hand == 'r')].count()['event_type'].astype(int)
|
|
|
|
def get_gb_vl(row):
|
|
return all_plays[(all_plays.batter_id == row['key_retro']) & (all_plays.batted_ball_type == 'G') & (all_plays.pitcher_hand == 'l')].count()['event_type'].astype(int)
|
|
def get_gb_vr(row):
|
|
return all_plays[(all_plays.batter_id == row['key_retro']) & (all_plays.batted_ball_type == 'G') & (all_plays.pitcher_hand == 'r')].count()['event_type'].astype(int)
|
|
|
|
def get_ld_vl(row):
|
|
return all_plays[(all_plays.batter_id == row['key_retro']) & (all_plays.batted_ball_type == 'l') & (all_plays.pitcher_hand == 'l')].count()['event_type'].astype(int)
|
|
def get_ld_vr(row):
|
|
return all_plays[(all_plays.batter_id == row['key_retro']) & (all_plays.batted_ball_type == 'l') & (all_plays.pitcher_hand == 'r')].count()['event_type'].astype(int)
|
|
|
|
def get_gdp_vl(row):
|
|
dp = all_plays[(all_plays.batter_id == row['key_retro']) & (all_plays.batter_event == 't') & (all_plays.pitcher_hand == 'l') & (all_plays.dp == 't')].count()['event_type'].astype(int)
|
|
tp = all_plays[(all_plays.batter_id == row['key_retro']) & (all_plays.batter_event == 't') & (all_plays.pitcher_hand == 'l') & (all_plays.tp == 't')].count()['event_type'].astype(int)
|
|
return dp + tp
|
|
def get_gdp_vr(row):
|
|
dp = all_plays[(all_plays.batter_id == row['key_retro']) & (all_plays.batter_event == 't') & (all_plays.pitcher_hand == 'r') & (all_plays.dp == 't')].count()['event_type'].astype(int)
|
|
tp = all_plays[(all_plays.batter_id == row['key_retro']) & (all_plays.batter_event == 't') & (all_plays.pitcher_hand == 'r') & (all_plays.tp == 't')].count()['event_type'].astype(int)
|
|
return dp + tp
|
|
|
|
def get_bunt(row):
|
|
return all_plays[(all_plays.batter_id == row['key_retro']) & (all_plays.bunt == 't')].count()['event_type'].astype(int)
|
|
|
|
batting_stats['FB_vL'] = batting_stats.apply(get_fb_vl, axis=1)
|
|
batting_stats['FB_vR'] = batting_stats.apply(get_fb_vr, axis=1)
|
|
|
|
batting_stats['GB_vL'] = batting_stats.apply(get_gb_vl, axis=1)
|
|
batting_stats['GB_vR'] = batting_stats.apply(get_gb_vr, axis=1)
|
|
|
|
batting_stats['LD_vL'] = batting_stats.apply(get_ld_vl, axis=1)
|
|
batting_stats['LD_vR'] = batting_stats.apply(get_ld_vr, axis=1)
|
|
|
|
batting_stats['GDP_vL'] = batting_stats.apply(get_gdp_vl, axis=1)
|
|
batting_stats['GDP_vR'] = batting_stats.apply(get_gdp_vr, axis=1)
|
|
|
|
batting_stats['Bunts'] = batting_stats.apply(get_bunt, axis=1)
|
|
print(f'Custom counting stats: {(datetime.datetime.now() - start).total_seconds():.2f}s')
|
|
|
|
# Infield Hit %
|
|
ifh_vl = all_plays[(all_plays.hit_val.str.contains('1|2|3')) & (all_plays.pitcher_hand == 'l') & (all_plays.hit_location.str.contains('1|2|3|4|5|6')) & (~all_plays.hit_location.str.contains('D', na=False))].groupby('batter_id').count()['event_type'].astype(int).rename('ifh_vL')
|
|
ifh_vr = all_plays[(all_plays.hit_val.str.contains('1|2|3')) & (all_plays.pitcher_hand == 'r') & (all_plays.hit_location.str.contains('1|2|3|4|5|6')) & (~all_plays.hit_location.str.contains('D', na=False))].groupby('batter_id').count()['event_type'].astype(int).rename('ifh_vR')
|
|
|
|
batting_stats['ifh_vL'] = ifh_vl
|
|
batting_stats['ifh_vR'] = ifh_vr
|
|
|
|
def get_pull_vl(row):
|
|
pull_loc = '5|7' if row['bat_hand'] != 'L' else '3|9'
|
|
x = all_plays[(all_plays.batter_id == row['key_retro']) & (all_plays.pitcher_hand == 'l') & (all_plays.hit_location.str.contains(pull_loc))].count()['event_type'].astype(int)
|
|
return x
|
|
def get_pull_vr(row):
|
|
pull_loc = '5|7' if row['bat_hand'] == 'R' else '3|9'
|
|
x = all_plays[(all_plays.batter_id == row['key_retro']) & (all_plays.pitcher_hand == 'r') & (all_plays.hit_location.str.contains(pull_loc))].count()['event_type'].astype(int)
|
|
return x
|
|
|
|
# Bespoke Queries
|
|
batting_stats['pull_vL'] = batting_stats.apply(get_pull_vl, axis=1)
|
|
batting_stats['pull_vR'] = batting_stats.apply(get_pull_vr, axis=1)
|
|
|
|
center_vl = all_plays[(all_plays.pitcher_hand == 'l') & (all_plays.hit_location.str.contains('1|4|6|8'))].groupby('batter_id').count()['event_type'].astype(int).rename('center_vl')
|
|
center_vr = all_plays[(all_plays.pitcher_hand == 'r') & (all_plays.hit_location.str.contains('1|4|6|8'))].groupby('batter_id').count()['event_type'].astype(int).rename('center_vr')
|
|
|
|
batting_stats['center_vL'] = center_vl
|
|
batting_stats['center_vR'] = center_vr
|
|
|
|
oppo_vl = all_plays[(all_plays.pitcher_hand == 'l') & (all_plays.hit_location.str.contains('5|7'))].groupby('batter_id').count()['event_type'].astype(int).rename('oppo_vL')
|
|
oppo_vr = all_plays[(all_plays.pitcher_hand == 'r') & (all_plays.hit_location.str.contains('5|7'))].groupby('batter_id').count()['event_type'].astype(int).rename('oppo_vR')
|
|
|
|
batting_stats['oppo_vL'] = oppo_vl
|
|
batting_stats['oppo_vR'] = oppo_vr
|
|
|
|
# fill na to 0 following counting stats
|
|
batting_stats = batting_stats.fillna(0)
|
|
|
|
# Calculated Fields
|
|
start = datetime.datetime.now()
|
|
batting_stats['H_vL'] = batting_stats['1B_vL'] + batting_stats['2B_vL'] + batting_stats['3B_vL'] + batting_stats['HR_vL']
|
|
batting_stats['H_vR'] = batting_stats['1B_vR'] + batting_stats['2B_vR'] + batting_stats['3B_vR'] + batting_stats['HR_vR']
|
|
|
|
batting_stats['AVG_vL'] = round(batting_stats['H_vL'] / batting_stats['AB_vL'], 5)
|
|
batting_stats['AVG_vR'] = round(batting_stats['H_vR'] / batting_stats['AB_vR'], 5)
|
|
|
|
batting_stats['OBP_vL'] = round((batting_stats['H_vL'] + batting_stats['BB_vL'] + batting_stats['HBP_vL']) / batting_stats['PA_vL'], 5)
|
|
batting_stats['OBP_vR'] = round((batting_stats['H_vR'] + batting_stats['BB_vR'] + batting_stats['HBP_vR']) / batting_stats['PA_vR'], 5)
|
|
|
|
batting_stats['SLG_vL'] = round((batting_stats['1B_vL'] + batting_stats['2B_vL'] * 2 + batting_stats['3B_vL'] * 3 + batting_stats['HR_vL'] * 4) / batting_stats['AB_vL'], 5)
|
|
batting_stats['SLG_vR'] = round((batting_stats['1B_vR'] + batting_stats['2B_vR'] * 2 + batting_stats['3B_vR'] * 3 + batting_stats['HR_vR'] * 4) / batting_stats['AB_vR'], 5)
|
|
|
|
batting_stats['HR/FB_vL'] = round(batting_stats['HR_vL'] / batting_stats['FB_vL'], 5)
|
|
batting_stats['HR/FB_vR'] = round(batting_stats['HR_vR'] / batting_stats['FB_vR'], 5)
|
|
|
|
batting_stats['FB%_vL'] = round(batting_stats['FB_vL'] / (batting_stats['FB_vL'] + batting_stats['GB_vL'] + batting_stats['LD_vL']), 5)
|
|
batting_stats['FB%_vR'] = round(batting_stats['FB_vR'] / (batting_stats['FB_vR'] + batting_stats['GB_vR'] + batting_stats['LD_vR']), 5)
|
|
|
|
batting_stats['GB%_vL'] = round(batting_stats['GB_vL'] / (batting_stats['FB_vL'] + batting_stats['GB_vL'] + batting_stats['LD_vL']), 5)
|
|
batting_stats['GB%_vR'] = round(batting_stats['GB_vR'] / (batting_stats['FB_vR'] + batting_stats['GB_vR'] + batting_stats['LD_vR']), 5)
|
|
|
|
batting_stats['LD%_vL'] = round(batting_stats['LD_vL'] / (batting_stats['FB_vL'] + batting_stats['GB_vL'] + batting_stats['LD_vL']), 5)
|
|
batting_stats['LD%_vR'] = round(batting_stats['LD_vR'] / (batting_stats['FB_vR'] + batting_stats['GB_vR'] + batting_stats['LD_vR']), 5)
|
|
|
|
batting_stats['Hard%_vL'] = round(0.2 + batting_stats['SLG_vL'] - batting_stats['AVG_vL'], 5)
|
|
batting_stats['Hard%_vR'] = round(0.2 + batting_stats['SLG_vR'] - batting_stats['AVG_vR'], 5)
|
|
|
|
def get_med_vL(row):
|
|
high = 0.9 - row['Hard%_vL']
|
|
low = (row['SLG_vL'] - row['AVG_vL']) * 1.5
|
|
return round(max(min(high, low),0.1), 5)
|
|
def get_med_vR(row):
|
|
high = 0.9 - row['Hard%_vR']
|
|
low = (row['SLG_vR'] - row['AVG_vR']) * 1.5
|
|
return round(max(min(high, low),0.1), 5)
|
|
|
|
batting_stats['Med%_vL'] = batting_stats.apply(get_med_vL, axis=1)
|
|
batting_stats['Med%_vR'] = batting_stats.apply(get_med_vR, axis=1)
|
|
|
|
batting_stats['Soft%_vL'] = round(1 - batting_stats['Hard%_vL'] - batting_stats['Med%_vL'], 5)
|
|
batting_stats['Soft%_vR'] = round(1 - batting_stats['Hard%_vR'] - batting_stats['Med%_vR'], 5)
|
|
|
|
batting_stats['IFH%_vL'] = round(batting_stats['ifh_vL'] / batting_stats['H_vL'], 5)
|
|
batting_stats['IFH%_vR'] = round(batting_stats['ifh_vR'] / batting_stats['H_vR'], 5)
|
|
|
|
batting_stats['Pull%_vL'] = round(batting_stats['pull_vL'] / (batting_stats['pull_vL'] + batting_stats['center_vL'] + batting_stats['oppo_vL']), 5)
|
|
batting_stats['Pull%_vR'] = round(batting_stats['pull_vR'] / (batting_stats['pull_vR'] + batting_stats['center_vR'] + batting_stats['oppo_vR']), 5)
|
|
|
|
batting_stats['Cent%_vL'] = round(batting_stats['center_vL'] / (batting_stats['pull_vL'] + batting_stats['center_vL'] + batting_stats['oppo_vL']), 5)
|
|
batting_stats['Cent%_vR'] = round(batting_stats['center_vL'] / (batting_stats['pull_vR'] + batting_stats['center_vR'] + batting_stats['oppo_vR']), 5)
|
|
|
|
batting_stats['Oppo%_vL'] = round(1 - batting_stats['Pull%_vL'] - batting_stats['Cent%_vL'], 5)
|
|
batting_stats['Oppo%_vR'] = round(1 - batting_stats['Pull%_vR'] - batting_stats['Cent%_vR'], 5)
|
|
|
|
batting_stats = batting_stats.fillna(0)
|
|
|
|
print(f'Calculated fields: {(datetime.datetime.now() - start).total_seconds():.2f}s')
|
|
|
|
return batting_stats
|
|
|
|
|
|
def calc_batting_cards(bs: pd.DataFrame, season_pct: float) -> pd.DataFrame:
|
|
def create_batting_card(row):
|
|
steal_data = cba.stealing(
|
|
chances=int(row['SBO']),
|
|
sb2s=int(row['SB2']),
|
|
cs2s=int(row['CS2']),
|
|
sb3s=int(row['SB3']),
|
|
cs3s=int(row['CS3']),
|
|
season_pct=1.0
|
|
)
|
|
y = pd.DataFrame({
|
|
'key_bbref': [row['key_bbref']],
|
|
'steal_low': [steal_data[0]],
|
|
'steal_high': [steal_data[1]],
|
|
'steal_auto': [steal_data[2]],
|
|
'steal_jump': [steal_data[3]],
|
|
'hit_and_run': [cba.hit_and_run(
|
|
row['AB_vL'], row['AB_vR'], row['H_vL'], row['H_vR'],
|
|
row['HR_vL'], row['HR_vR'], row['SO_vL'], row['SO_vR']
|
|
)],
|
|
'bunt': [cba.bunting(row['Bunts'], season_pct)],
|
|
'running': [cba.running(row['XBT%'])],
|
|
'hand': [row['bat_hand']],
|
|
})
|
|
return y.loc[0]
|
|
|
|
all_cards = bs.apply(create_batting_card, axis=1)
|
|
all_cards = all_cards.set_index('key_bbref')
|
|
|
|
return all_cards
|
|
|
|
|
|
def calc_batter_ratings(bs: pd.DataFrame) -> pd.DataFrame:
|
|
def create_batting_rating(row):
|
|
ratings = cba.get_batter_ratings(row)
|
|
ops_vl = ratings[0]['obp'] + ratings[0]['slg']
|
|
ops_vr = ratings[1]['obp'] + ratings[1]['slg']
|
|
total_ops = (ops_vl + ops_vr + min(ops_vr, ops_vl)) / 3
|
|
|
|
def calc_cost(total_ops, base_cost, base_ops, max_delta) -> int:
|
|
delta = ((total_ops - base_ops) / 0.1) * 2
|
|
if delta < 1:
|
|
delta = (max_delta * (1 - (total_ops / base_ops))) * -0.1
|
|
|
|
final_cost = base_cost + (max_delta * delta)
|
|
|
|
return round(final_cost)
|
|
|
|
if total_ops >= 1.2:
|
|
rarity_id = 99
|
|
cost = calc_cost(total_ops, base_cost=2400, base_ops=1.215, max_delta=810)
|
|
elif total_ops >= 1:
|
|
rarity_id = 1
|
|
cost = calc_cost(total_ops, base_cost=810, base_ops=1.05, max_delta=270)
|
|
elif total_ops >= 0.9:
|
|
rarity_id = 2
|
|
cost = calc_cost(total_ops, base_cost=270, base_ops=0.95, max_delta=90)
|
|
elif total_ops >= 0.8:
|
|
rarity_id = 3
|
|
cost = calc_cost(total_ops, base_cost=90, base_ops=0.85, max_delta=30)
|
|
elif total_ops >= 0.7:
|
|
rarity_id = 4
|
|
cost = calc_cost(total_ops, base_cost=30, base_ops=0.75, max_delta=10)
|
|
else:
|
|
rarity_id = 5
|
|
cost = calc_cost(total_ops, base_cost=10, base_ops=0.61, max_delta=8)
|
|
|
|
x = pd.DataFrame({
|
|
'key_bbref': [row['key_bbref']],
|
|
'ratings_vL': [ratings[0]],
|
|
'ratings_vR': [ratings[1]],
|
|
'ops_vL': ops_vl,
|
|
'ops_vR': ops_vr,
|
|
'total_ops': total_ops,
|
|
'rarity_id': rarity_id,
|
|
'cost': cost
|
|
})
|
|
return x.loc[0]
|
|
|
|
all_ratings = bs.apply(create_batting_rating, axis=1)
|
|
all_ratings = all_ratings.set_index('key_bbref')
|
|
|
|
return all_ratings
|
|
|
|
|
|
def calc_positions(bs: pd.DataFrame) -> pd.DataFrame:
|
|
df_c = pd.read_csv(f'{DATA_INPUT_FILE_PATH}defense_c.csv').set_index('key_bbref')
|
|
df_1b = pd.read_csv(f'{DATA_INPUT_FILE_PATH}defense_1b.csv').set_index('key_bbref')
|
|
df_2b = pd.read_csv(f'{DATA_INPUT_FILE_PATH}defense_2b.csv').set_index('key_bbref')
|
|
df_3b = pd.read_csv(f'{DATA_INPUT_FILE_PATH}defense_3b.csv').set_index('key_bbref')
|
|
df_ss = pd.read_csv(f'{DATA_INPUT_FILE_PATH}defense_ss.csv').set_index('key_bbref')
|
|
df_lf = pd.read_csv(f'{DATA_INPUT_FILE_PATH}defense_lf.csv').set_index('key_bbref')
|
|
df_cf = pd.read_csv(f'{DATA_INPUT_FILE_PATH}defense_cf.csv').set_index('key_bbref')
|
|
df_rf = pd.read_csv(f'{DATA_INPUT_FILE_PATH}defense_rf.csv').set_index('key_bbref')
|
|
df_of = pd.read_csv(f'{DATA_INPUT_FILE_PATH}defense_of.csv').set_index('key_bbref')
|
|
season_pct = 1.0
|
|
|
|
all_pos = []
|
|
|
|
def process_pos(row):
|
|
no_data = True
|
|
# TODO: Add pos_1 through pos_8 to def df to be pulled in at post time
|
|
for pos_df, position in [(df_1b, '1b'), (df_2b, '2b'), (df_3b, '3b'), (df_ss, 'ss')]:
|
|
if row['key_bbref'] in pos_df.index:
|
|
logging.info(f'Running {position} stats for {row["use_name"]} {row["last_name"]}')
|
|
try:
|
|
if 'tz_runs_total' in row:
|
|
average_range = (int(pos_df.at[row["key_bbref"], 'tz_runs_total']) +
|
|
int(pos_df.at[row["key_bbref"], 'bis_runs_total']) +
|
|
min(
|
|
int(pos_df.at[row["key_bbref"], 'tz_runs_total']),
|
|
int(pos_df.at[row["key_bbref"], 'bis_runs_total'])
|
|
)) / 3
|
|
else:
|
|
average_range = pos_df.at[row["key_bbref"], 'tz_runs_total']
|
|
|
|
if float(pos_df.at[row["key_bbref"], 'Inn_def']) >= 10.0:
|
|
all_pos.append({
|
|
"key_bbref": row['key_bbref'],
|
|
"position": position.upper(),
|
|
"innings": float(pos_df.at[row["key_bbref"], 'Inn_def']),
|
|
"range": cde.get_if_range(
|
|
pos_code=position,
|
|
tz_runs=round(average_range),
|
|
r_dp=0,
|
|
season_pct=season_pct
|
|
),
|
|
"error": cde.get_any_error(
|
|
pos_code=position,
|
|
errors=int(pos_df.at[row["key_bbref"], 'E_def']),
|
|
chances=int(pos_df.at[row["key_bbref"], 'chances']),
|
|
season_pct=season_pct
|
|
)
|
|
})
|
|
no_data = False
|
|
except Exception as e:
|
|
logging.info(f'Infield position failed: {e}')
|
|
|
|
of_arms = []
|
|
of_payloads = []
|
|
for pos_df, position in [(df_lf, 'lf'), (df_cf, 'cf'), (df_rf, 'rf')]:
|
|
if row["key_bbref"] in pos_df.index:
|
|
try:
|
|
if 'tz_runs_total' in row:
|
|
average_range = (int(pos_df.at[row["key_bbref"], 'tz_runs_total']) +
|
|
int(pos_df.at[row["key_bbref"], 'bis_runs_total']) +
|
|
min(
|
|
int(pos_df.at[row["key_bbref"], 'tz_runs_total']),
|
|
int(pos_df.at[row["key_bbref"], 'bis_runs_total'])
|
|
)) / 3
|
|
else:
|
|
average_range = pos_df.at[row["key_bbref"], 'tz_runs_total']
|
|
|
|
if float(pos_df.at[row["key_bbref"], 'Inn_def']) >= 10.0:
|
|
of_payloads.append({
|
|
"key_bbref": row['key_bbref'],
|
|
"position": position.upper(),
|
|
"innings": float(pos_df.at[row["key_bbref"], 'Inn_def']),
|
|
"range": cde.get_of_range(
|
|
pos_code=position,
|
|
tz_runs=round(average_range),
|
|
season_pct=season_pct
|
|
)
|
|
})
|
|
of_run_rating = 'bis_runs_outfield' if 'bis_runs_outfield' in pos_df else 'tz_runs_outfield'
|
|
of_arms.append(int(pos_df.at[row["key_bbref"], of_run_rating]))
|
|
no_data = False
|
|
except Exception as e:
|
|
logging.info(f'Outfield position failed: {e}')
|
|
|
|
if row["key_bbref"] in df_of.index and len(of_arms) > 0 and len(of_payloads) > 0:
|
|
try:
|
|
error_rating = cde.get_any_error(
|
|
pos_code=position,
|
|
errors=int(df_of.at[row["key_bbref"], 'E_def']),
|
|
chances=int(df_of.at[row["key_bbref"], 'chances']),
|
|
season_pct=season_pct
|
|
)
|
|
arm_rating = cde.arm_outfield(of_arms)
|
|
for f in of_payloads:
|
|
f['error'] = error_rating
|
|
f['arm'] = arm_rating
|
|
all_pos.append(f)
|
|
no_data = False
|
|
except Exception as e:
|
|
logging.info(f'Outfield position failed: {e}')
|
|
|
|
if row["key_bbref"] in df_c.index:
|
|
try:
|
|
run_rating = 'bis_runs_catcher_sb' if 'bis_runs_catcher_sb' in df_c else 'tz_runs_catcher'
|
|
|
|
if df_c.at[row["key_bbref"], 'SB'] + df_c.at[row["key_bbref"], 'CS'] == 0:
|
|
arm_rating = 3
|
|
else:
|
|
arm_rating = cde.arm_catcher(
|
|
cs_pct=df_c.at[row["key_bbref"], 'caught_stealing_perc'],
|
|
raa=int(df_c.at[row["key_bbref"], run_rating]),
|
|
season_pct=season_pct
|
|
)
|
|
|
|
if float(df_c.at[row["key_bbref"], 'Inn_def']) >= 10.0:
|
|
all_pos.append({
|
|
"key_bbref": row['key_bbref'],
|
|
"position": 'C',
|
|
"innings": float(df_c.at[row["key_bbref"], 'Inn_def']),
|
|
"range": cde.range_catcher(
|
|
rs_value=int(df_c.at[row["key_bbref"], 'tz_runs_catcher']),
|
|
season_pct=season_pct
|
|
),
|
|
"error": cde.get_any_error(
|
|
pos_code='c',
|
|
errors=int(df_c.at[row["key_bbref"], 'E_def']),
|
|
chances=int(df_c.at[row["key_bbref"], 'chances']),
|
|
season_pct=season_pct
|
|
),
|
|
"arm": arm_rating,
|
|
"pb": cde.pb_catcher(
|
|
pb=int(df_c.at[row["key_bbref"], 'PB']),
|
|
innings=int(float(df_c.at[row["key_bbref"], 'Inn_def'])),
|
|
season_pct=season_pct
|
|
),
|
|
"overthrow": cde.ot_catcher(
|
|
errors=int(df_c.at[row["key_bbref"], 'E_def']),
|
|
chances=int(df_c.at[row["key_bbref"], 'chances']),
|
|
season_pct=season_pct
|
|
)
|
|
})
|
|
no_data = False
|
|
except Exception as e:
|
|
logging.info(f'Catcher position failed: {e}')
|
|
|
|
if no_data:
|
|
all_pos.append({
|
|
"key_bbref": row['key_bbref'],
|
|
"position": 'DH',
|
|
"innings": row['PA_vL'] + row['PA_vR']
|
|
})
|
|
|
|
bs.apply(process_pos, axis=1)
|
|
pos_df = pd.DataFrame(all_pos)
|
|
pos_df = pos_df.set_index('key_bbref')
|
|
|
|
return pos_df
|
|
|
|
|
|
async def get_or_post_players(stat_df: pd.DataFrame, bat_card_df: pd.DataFrame, bat_rat_df: pd.DataFrame, def_rat_df: pd.DataFrame) -> pd.DataFrame:
|
|
# Refactor this to support batters or pitchers, make stat dfs optional and send to getorpostbatters/getorpostpitchers
|
|
all_players = []
|
|
|
|
dev_count = 0
|
|
for index, row in stat_df.iterrows():
|
|
if dev_count < 0:
|
|
break
|
|
|
|
p_query = await db_get('players', params=[('bbref_id', row["key_bbref"]), ('cardset_id', CARDSET_ID)])
|
|
if p_query['count'] > 0:
|
|
this_record = p_query['players'][0]
|
|
# if 'id' in this_record:
|
|
# player_id = this_record['id']
|
|
# else:
|
|
# player_id = this_record['player_id']
|
|
|
|
# all_bbref_ids.append(row['key_bbref'])
|
|
# all_player_ids.append(player_id)
|
|
all_players.append(this_record)
|
|
else:
|
|
mlb_query = await db_get('mlbplayers', params=[('key_retro', row['key_retro'])])
|
|
if mlb_query['count'] > 0:
|
|
mlb_player = mlb_query['players'][0]
|
|
else:
|
|
mlb_player = await db_post(
|
|
'mlbplayers/one',
|
|
payload={
|
|
'first_name': row['use_name'],
|
|
'last_name': row['last_name'],
|
|
'key_mlbam': row['key_mlbam'],
|
|
'key_fangraphs': row['key_fangraphs'],
|
|
'key_bbref': row['key_bbref'],
|
|
'key_retro': row['key_retro']
|
|
}
|
|
)
|
|
|
|
player_payload = {
|
|
'p_name': f'{row["use_name"]} {row["last_name"]}',
|
|
'cost': f'{bat_rat_df.loc[row['key_bbref']]["cost"]}',
|
|
'image': f'change-me',
|
|
'mlbclub': CLUB_LIST[row['Tm']],
|
|
'franchise': FRANCHISE_LIST[row['Tm']],
|
|
'cardset_id': CARDSET_ID,
|
|
'set_num': int(float(row['key_fangraphs'])),
|
|
'rarity_id': int(bat_rat_df.loc[row['key_bbref']]['rarity_id']),
|
|
'description': PLAYER_DESCRIPTION,
|
|
'bbref_id': row['key_bbref'],
|
|
'fangr_id': row['key_fangraphs'],
|
|
'mlbplayer_id': mlb_player['id']
|
|
}
|
|
|
|
try:
|
|
count = 1
|
|
all_pos = def_rat_df.loc[row['key_bbref']].sort_values(by='innings', ascending=False)
|
|
for index, pos_row in all_pos.iterrows():
|
|
player_payload[f'pos_{count}'] = pos_row.position
|
|
count += 1
|
|
except KeyError:
|
|
logging.info(f'No positions found for {row['use_name']} {row['last_name']}')
|
|
player_payload['pos_1'] = 'DH'
|
|
except TypeError:
|
|
logging.info(f'Only one position found for {row['use_name']} {row['last_name']}')
|
|
player_payload['pos_1'] = def_rat_df.loc[row['key_bbref']].position
|
|
|
|
new_player = await db_post('players', payload=player_payload)
|
|
|
|
if 'id' in new_player:
|
|
player_id = new_player['id']
|
|
else:
|
|
player_id = new_player['player_id']
|
|
|
|
await db_patch('players', object_id=player_id, params=[('image', f'{CARD_BASE_URL}{player_id}/battingcard{urllib.parse.quote("?d=")}{RELEASE_DIRECTORY}')])
|
|
|
|
# all_bbref_ids.append(row['key_bbref'])
|
|
# all_player_ids.append(player_id)
|
|
all_players.append(new_player)
|
|
|
|
dev_count += 1
|
|
|
|
players_df = pd.DataFrame(all_players).set_index('bbref_id')
|
|
return players_df
|
|
|
|
|
|
async def post_batting_cards(cards_df: pd.DataFrame):
|
|
all_cards = []
|
|
|
|
cards_df.apply(lambda x: all_cards.append({
|
|
'player_id': int(x["player_id"]),
|
|
'steal_low': x['steal_low'],
|
|
'steal_high': x['steal_high'],
|
|
'steal_auto': x['steal_auto'],
|
|
'steal_jump': x['steal_jump'],
|
|
'bunting': x['bunt'],
|
|
'hit_and_run': x['hit_and_run'],
|
|
'running': x['running'],
|
|
'hand': x['hand']
|
|
}), axis=1)
|
|
resp = await db_put('battingcards', payload={'cards': all_cards}, timeout=6)
|
|
if resp is not None:
|
|
pass
|
|
else:
|
|
e_msg = 'Unable to post batting cards'
|
|
logging.debug(e_msg, stack_info=True)
|
|
raise ValueError(e_msg)
|
|
|
|
bc_query = await db_get('battingcards', params=[('cardset_id', CARDSET_ID)])
|
|
if bc_query['count'] > 0:
|
|
bc_data = bc_query['cards']
|
|
|
|
for line in bc_data:
|
|
line['player_id'] = line['player']['player_id']
|
|
line['key_bbref'] = line['player']['bbref_id']
|
|
line['battingcard_id'] = line['id']
|
|
|
|
return pd.DataFrame(bc_data).set_index('key_bbref')
|
|
|
|
|
|
async def post_batting_ratings(ratings_df: pd.DataFrame):
|
|
all_ratings = []
|
|
|
|
def append_ratings(row):
|
|
vl = row['ratings_vL']
|
|
vl['player_id'] = row['player_id']
|
|
vl['battingcard_id'] = row['battingcard_id']
|
|
|
|
vr = row['ratings_vR']
|
|
vr['player_id'] = row['player_id']
|
|
vr['battingcard_id'] = row['battingcard_id']
|
|
|
|
all_ratings.append(vl)
|
|
all_ratings.append(vr)
|
|
|
|
ratings_df.apply(append_ratings, axis=1)
|
|
resp = await db_put('battingcardratings', payload={'ratings': all_ratings}, timeout=6)
|
|
if resp is not None:
|
|
return True
|
|
else:
|
|
e_msg = 'Unable to post batting ratings'
|
|
logging.debug(e_msg, stack_info=True)
|
|
raise ValueError(e_msg)
|
|
|
|
|
|
async def post_positions(pos_df: pd.DataFrame):
|
|
all_pos = []
|
|
|
|
def append_positions(row):
|
|
clean_row = row.dropna()
|
|
new_val = clean_row.to_dict()
|
|
new_val['player_id'] = int(row['player_id'])
|
|
all_pos.append(new_val)
|
|
pos_df.apply(append_positions, axis=1)
|
|
|
|
resp = await db_put('cardpositions', payload={'positions': all_pos}, timeout=6)
|
|
if resp is not None:
|
|
return True
|
|
else:
|
|
e_msg = 'Unable to post positions'
|
|
logging.debug(e_msg, stack_info=True)
|
|
raise ValueError(e_msg)
|
|
|
|
|
|
async def post_batter_data(bs: pd.DataFrame, bc: pd.DataFrame, br: pd.DataFrame, dr: pd.date_range) -> int:
|
|
all_players = await get_or_post_players(bs, bc, br, dr)
|
|
|
|
# Post Batting Cards
|
|
bc = pd.merge(
|
|
left=bc,
|
|
right=all_players,
|
|
how='right', # 'left', TODO: switch back to left when all players are pulled
|
|
left_on='key_bbref',
|
|
right_on='bbref_id'
|
|
)
|
|
bc = await post_batting_cards(bc)
|
|
|
|
# Post Batting Ratings
|
|
br = pd.merge(
|
|
left=br,
|
|
right=bc,
|
|
how='right', #'left', TODO: switch back to left when all players are pulled
|
|
left_on='key_bbref',
|
|
right_on='key_bbref'
|
|
)
|
|
br = await post_batting_ratings(br)
|
|
|
|
# Post Positions
|
|
dr = pd.merge(
|
|
left=dr,
|
|
right=all_players,
|
|
how='right', # 'left',
|
|
left_on='key_bbref',
|
|
right_on='bbref_id'
|
|
)
|
|
await post_positions(dr)
|
|
|
|
return len(all_players)
|
|
|
|
|
|
async def run_batters(data_input_path: str, start_date: int, end_date: int, post_data: bool = False, season_pct: float = 1.0):
|
|
print(f'Running the batter calcs...')
|
|
batter_start = datetime.datetime.now()
|
|
|
|
# Get batting stats
|
|
batting_stats = get_batting_stats_by_date(f'{RETRO_FILE_PATH}{EVENTS_FILENAME}', start_date=start_date, end_date=end_date)
|
|
bs_len = len(batting_stats)
|
|
|
|
end_calc = datetime.datetime.now()
|
|
print(f'Combined batting stats: {(end_calc - batter_start).total_seconds():.2f}s\n')
|
|
running_start = datetime.datetime.now()
|
|
|
|
# Get running stats
|
|
running_stats = get_run_stat_df(data_input_path)
|
|
|
|
batting_stats = pd.merge(
|
|
left=batting_stats,
|
|
right=running_stats,
|
|
how='left',
|
|
left_on='key_bbref',
|
|
right_on='key_bbref'
|
|
)
|
|
end_calc = datetime.datetime.now()
|
|
print(f'Running stats: {(end_calc - running_start).total_seconds():.2f}s')
|
|
|
|
if len(batting_stats) != bs_len:
|
|
raise DataMismatchError(f'retrosheet_data - run_batters - We started with {bs_len} batting lines and have {len(batting_stats)} after merging with running_stats')
|
|
|
|
# Calculate batting cards
|
|
card_start = datetime.datetime.now()
|
|
all_batting_cards = calc_batting_cards(batting_stats, season_pct)
|
|
card_end = datetime.datetime.now()
|
|
|
|
print(f'Create batting cards: {(card_end - card_start).total_seconds():.2f}s')
|
|
|
|
# Calculate batting ratings
|
|
rating_start = datetime.datetime.now()
|
|
batting_stats['battingcard_id'] = batting_stats['key_fangraphs']
|
|
all_batting_ratings = calc_batter_ratings(batting_stats)
|
|
rating_end = datetime.datetime.now()
|
|
|
|
print(f'Create batting ratings: {(rating_end - rating_start).total_seconds():.2f}s')
|
|
|
|
# Calculate defense ratings
|
|
defense_start = datetime.datetime.now()
|
|
all_defense_ratings = calc_positions(batting_stats)
|
|
defense_end = datetime.datetime.now()
|
|
|
|
print(f'Create defense ratings: {(defense_end - defense_start).total_seconds():.2f}s')
|
|
|
|
# Post all data
|
|
if post_data:
|
|
print(f'Posting player data...')
|
|
post_start = datetime.datetime.now()
|
|
num_players = await post_batter_data(batting_stats, all_batting_cards, all_batting_ratings, all_defense_ratings)
|
|
post_end = datetime.datetime.now()
|
|
|
|
print(f'Post player data: {(post_end - post_start).total_seconds()}s')
|
|
|
|
post_msg = f'Posted {num_players} players to the database'
|
|
logging.info(post_msg)
|
|
print(post_msg)
|
|
else:
|
|
post_msg = f'Players are NOT being posted to the database'
|
|
logging.warning(post_msg)
|
|
print(post_msg)
|
|
|
|
return batting_stats
|
|
|
|
|
|
async def main(args):
|
|
batter_start = datetime.datetime.now()
|
|
batting_stats = await run_batters(f'{DATA_INPUT_FILE_PATH}', start_date=19980101, end_date=19980430, post_data=True, season_pct=0.16666667)
|
|
batting_stats.to_csv(f'batting_stats.csv')
|
|
batter_end = datetime.datetime.now()
|
|
|
|
pitcher_start = datetime.datetime.now()
|
|
pitcher_end = datetime.datetime.now()
|
|
|
|
print(f'\n\nBatter time: {(batter_end - batter_start).total_seconds():.2f}s \nPitcher time: {(pitcher_end - pitcher_start).total_seconds():.2f}s\nTotal: {(pitcher_end - batter_start).total_seconds():.2f}s\n\nDone!')
|
|
|
|
# await store_defense_to_csv(1998)
|
|
|
|
|
|
if __name__ == '__main__':
|
|
asyncio.run(main(sys.argv[1:]))
|