Updated mround to return float

Counting stats nearly complete for batters

creation_helpers.py

@@ -570,7 +570,8 @@ def get_pitching_peripherals(season: int):
 
 
 def mround(x, prec=2, base=.05):
-    return round(base * round(float(x) / base), prec)
+    return float(round(Decimal(str(x)) / Decimal(str(base))) * Decimal(str(base)))
+    # return round(base * round(float(x) / base), prec)
 
 
 def chances_from_row(row_num):
@@ -913,7 +914,6 @@ def legacy_sanitize_chance_output(total_chances: float, min_chances: float = 1.0
             return x
 
 
-
 def mlbteam_and_franchise(mlbam_playerid):
     api_url = f'https://statsapi.mlb.com/api/v1/people/{mlbam_playerid}?hydrate=currentTeam'
     logging.info(f'Calling {api_url}')

retrosheet_data.py

@@ -8,6 +8,8 @@ from typing import Literal
 import pandas as pd
 import pybaseball as pb
 
+from creation_helpers import get_all_pybaseball_ids
+
 date = f'{datetime.datetime.now().year}-{datetime.datetime.now().month}-{datetime.datetime.now().day}'
 log_level = logging.INFO
 logging.basicConfig(
@@ -20,8 +22,8 @@ EVENTS_FILENAME = 'retrosheets_events_1998_short.csv'  # Removed last few column
 PERSONNEL_FILENAME = 'retrosheets_personnel.csv'
 
 
-def get_events_by_date(start_date: int, end_date: int) -> pd.DataFrame:
-    all_plays = pd.read_csv(f'{FILE_PATH}{EVENTS_FILENAME}', dtype={'game_id': 'str'})
+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[7:-1].astype(int)
     date_plays = all_plays[(all_plays.date >= start_date) & (all_plays.date <= end_date)]
     return date_plays
@@ -32,70 +34,111 @@ def get_result_series(plays: pd.DataFrame, event_type: str, pitcher_hand: Litera
     return this_series
 
 
-def get_batting_stats_by_date(start_date: int, end_date: int) -> pd.DataFrame:
-    all_plays = get_events_by_date(start_date, end_date)
+# def get_batting_handedness(plays: pd.DataFrame) -> pd.DataFrame:
     
-    bs = pd.DataFrame()
-    bs['batter_id'] = all_plays['batter_id'].unique()
-    bs = bs.set_index('batter_id')
 
-    pal_series = all_plays[(all_plays.batter_event == 't') & (all_plays.pitcher_hand == 'l')].groupby('batter_id').count()['event_type'].astype(int).rename('PAvL')
+
+def get_player_ids(plays: pd.DataFrame, which: Literal['batters', 'pitchers']) -> pd.DataFrame:
+    RETRO_PLAYERS = pd.read_csv(f'{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 time: {(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)
+
+    return players
+
+
+def get_base_batting_df(all_plays: pd.DataFrame) -> pd.DataFrame:
+    bs = get_player_ids(all_plays, 'batters')
+
+    # bs['key_mlbam'] = bs.apply()
+
+    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('PAvR')
+    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('ABvL')
+    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('ABvR')
+    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)
 
-    # Basic counting stats
-    for event_type, vs_hand, col_name in [
-        ('home run', 'r', 'HRvR'),
-        ('home run', 'l', 'HRvL'),
-        ('single', 'r', '1BvR'),
-        ('single', 'l', '1BvL'),
-        ('double', 'r', '2BvR'),
-        ('double', 'l', '2BvL'),
-        ('triple', 'r', '3BvR'),
-        ('triple', 'l', '3BvL'),
-        ('walk', 'r', 'BBvR'),
-        ('walk', 'l', 'BBvL'),
-        ('strikeout', 'r', 'SOvR'),
-        ('strikeout', 'l', 'SOvL'),
-        ('hit by pitch', 'r', 'HBPvR'),
-        ('hit by pitch', 'l', 'HBPvL')
-    ]:
-        this_series = get_result_series(all_plays, event_type, vs_hand, col_name)
-        bs = pd.concat([bs, this_series], axis=1)
-
-    # Bespoke queries
-    
-    # fill na to 0 following counting stats
-    bs = bs.fillna(0)
-
     return bs
 
 
-# def get_batting_stat_range(start_month: int, start_day: int, end_month: int, end_day: int):
-#     return get_batting_stats_by_date(
-#         start_date=start_month * 100 + start_day, 
-#         end_date=end_month * 100 + end_day
-#     )
+def get_batting_stats_by_date(file_path, start_date: int, end_date: int) -> pd.DataFrame:
+    all_plays = get_events_by_date(file_path, start_date, end_date)
+
+    batting_stats = get_base_batting_df(all_plays)    
+
+    # Basic counting stats
+    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 = pd.concat([batting_stats, this_series], axis=1)
+
+    # Bespoke queries
+    # Remaining: 
+    
+    # fill na to 0 following counting stats
+    batting_stats = batting_stats.fillna(0)
+
+    return batting_stats
 
 
 async def main(args):
     print(f'Running the calcs...')
     start = datetime.datetime.now()
-    data = get_batting_stats_by_date(start_date=101, end_date=430)
+    data = get_batting_stats_by_date(f'{FILE_PATH}{EVENTS_FILENAME}', start_date=101, end_date=430)
     end_calc = datetime.datetime.now()
 
-    print(f'Saving to csv...')
     data.to_csv(f'batting_stats.csv')
     end = datetime.datetime.now()
 
-    print(f'Done!\n\nCalc time: {(end_calc - start).total_seconds()}s\nSave time: {(end - end_calc).total_seconds()}s\nTotal: {(end - start).total_seconds()}s')
+    print(f'Stat calc time: {(end_calc - start).total_seconds():.2f}s\nSave time: {(end - end_calc).total_seconds():.2f}s\nTotal: {(end - start).total_seconds():.2f}s\n\nDone!')
 
 
 if __name__ == '__main__':

tests/test_batter_calcs.py

@@ -1,56 +1,9 @@
 from decimal import ROUND_HALF_EVEN, Decimal
 import math
 
-from batters.calcs_batter import bp_singles, wh_singles, sanitize_chance_output
-from creation_helpers import sanitize_chance_output, mround
+from batters.calcs_batter import bp_singles, wh_singles
 
 
-def test_sanitize():
-    # def my_round_decimal(i: float):
-    #     return Decimal(i).quantize(Decimal('0.05'), ROUND_HALF_EVEN)
-
-    # def my_round(num: float, to: float = 0.05):
-    #     num, to = Decimal(str(num)), Decimal(str(to))
-    #     return float(round(num / to) * to)
-    
-    # assert my_round(6) == 6
-    # assert my_round(5.96) == 5.95
-    # assert my_round(5.84) == 5.85
-    # assert my_round(3.123) == 3.1
-    # assert math.floor(my_round(6)) == 6
-    # assert math.floor(my_round(5.96)) == 5
-
-    assert sanitize_chance_output(6) == 6.0
-    assert sanitize_chance_output(1.21) == 1.2
-    assert sanitize_chance_output(4.77) == 4.75
-    assert sanitize_chance_output(4.78) == 4.8
-
-    # step_1 = Decimal(6) / Decimal(0.05)
-    # step_1_5 = round(step_1, )
-    # step_2 = round(step_1)
-    # step_3 = float(step_2 * Decimal(0.05))
-    # step_4 = Decimal(step_3)
-
-    # assert Decimal(6) == Decimal(6).quantize(Decimal('0.05'), ROUND_HALF_EVEN)
-
-    # assert round(step_1) == 120
-    # # assert step_1 == 120
-    # assert step_1_5 == 120
-    # assert step_2 == 120
-    # assert step_3 == 6.0
-    # assert step_4 == Decimal('6.0')
-    
-    # rounded_val = step_4.quantize(Decimal("0.05"), ROUND_HALF_EVEN)
-    # assert rounded_val == 6
-    # assert sanitize_chance_output(6) == 6
-
-
-def test_mround():
-    assert mround(6) == 6.0
-    assert mround(1.21) == 1.2
-    assert mround(4.77) == 4.75
-    assert mround(4.78) == 4.8
-
 
 def test_decimals():
     assert Decimal(8) == 8

tests/test_helpers.py

@@ -1,7 +1,52 @@
-from creation_helpers import pd_positions_df
+from creation_helpers import pd_positions_df, mround, sanitize_chance_output
 
 
 def test_positions_df():
     cardset_19_pos = pd_positions_df(19)
 
     assert True == True
+
+
+def test_mround():
+    assert mround(6.4) == 6.4
+    assert mround(6.66) == 6.65
+
+
+def test_sanitize():
+    # def my_round_decimal(i: float):
+    #     return Decimal(i).quantize(Decimal('0.05'), ROUND_HALF_EVEN)
+
+    # def my_round(num: float, to: float = 0.05):
+    #     num, to = Decimal(str(num)), Decimal(str(to))
+    #     return float(round(num / to) * to)
+    
+    # assert my_round(6) == 6
+    # assert my_round(5.96) == 5.95
+    # assert my_round(5.84) == 5.85
+    # assert my_round(3.123) == 3.1
+    # assert math.floor(my_round(6)) == 6
+    # assert math.floor(my_round(5.96)) == 5
+
+    assert sanitize_chance_output(6) == 6.0
+    assert sanitize_chance_output(1.21) == 1.2
+    assert sanitize_chance_output(4.77) == 4.75
+    assert sanitize_chance_output(4.78) == 4.8
+
+    # step_1 = Decimal(6) / Decimal(0.05)
+    # step_1_5 = round(step_1, )
+    # step_2 = round(step_1)
+    # step_3 = float(step_2 * Decimal(0.05))
+    # step_4 = Decimal(step_3)
+
+    # assert Decimal(6) == Decimal(6).quantize(Decimal('0.05'), ROUND_HALF_EVEN)
+
+    # assert round(step_1) == 120
+    # # assert step_1 == 120
+    # assert step_1_5 == 120
+    # assert step_2 == 120
+    # assert step_3 == 6.0
+    # assert step_4 == Decimal('6.0')
+    
+    # rounded_val = step_4.quantize(Decimal("0.05"), ROUND_HALF_EVEN)
+    # assert rounded_val == 6
+    # assert sanitize_chance_output(6) == 6