Add team pitching leaderboard

acknowledgements.md

@@ -6,4 +6,4 @@ The approach to visualizing pitch locations was inspired by [Stephen-Sutton Brow
 
 The attempt to be more aesthetically-minded (ex. more conscious focus on font choice and legibility) was inspired by the hockey player cards of [JFresh](https://x.com/JFreshHockey).
 
-Thanks to [ぼーのさん](https://bo-no05.hatenadiary.org/) for answering some of my questions while making this.
+Thanks to [ぼーのさん](https://bo-no05.hatenadiary.org/) for answering some of our questions while making this.

app.py

@@ -3,7 +3,8 @@ import matplotlib as mpl
 
 from data import data_df
 from pitcher_overview import create_pitcher_overview
-from pitcher_leaderboard import create_pitcher_leaderboard
+# from pitcher_leaderboard import create_pitcher_leaderboard
+from player_team_leaderboard import create_pitcher_leaderboard, create_team_pitching_leaderboard
 from pitch_leaderboard import create_pitch_leaderboard
 from daily_weekly_leaderboard import create_daily_weekly_leaderboard_app
 from css import css
@@ -16,7 +17,7 @@ with open('updated.txt') as f:
 
 limitations = '''**General Limitations**
 - As new players make their debut, some names may be missing or not translated/transliterated correctly.
-- IP is overestimated
+- IP is underestimated (does not count outs via caught stealing)
 '''
 
 with open('acknowledgements.md', 'r') as f:
@@ -30,11 +31,14 @@ if __name__ == '__main__':
       create_pitcher_leaderboard()
     with gr.Tab('Pitch Leaderboard'):
       create_pitch_leaderboard()
+    with gr.Tab('Team Pitching Leaderboard'):
+      create_team_pitching_leaderboard()
     with gr.Tab('Daily/Weekly Leaderboard'):
       create_daily_weekly_leaderboard_app(data_df)
     with gr.Tab('Acknowledgements'):
       gr.Markdown(acknowledgements)
 
+
     gr.Markdown('---')
     gr.Markdown(f'**Data up to:** {latest_data_date}')
     gr.Markdown(f'**Last updated:** {updated}')

data.py

@@ -148,7 +148,7 @@ data_df = (
     )
     .join(
         (
-            aux_df.filter(~pl.col('ibb'))[['universal_code', 'battingResult', 'inning_pas', 'pa_pitches']]
+            aux_df.filter(~pl.col('ibb'))[['universal_code', 'battingResult', 'inning_pas', 'pa_pitches', 'beforeBso', 'bso']]
             .rename({'battingResult': 'aux_bresult', 'inning_pas': 'aux_inning_pas', 'pa_pitches': 'aux_pa_pitches'})
         ),
         on='universal_code',
@@ -214,7 +214,6 @@ data_df = (
     .filter(pl.col('ballKind_code') != '-')
     .unique()
 )
-bar = data_df
 
 
 def select_name(names):

pitcher_leaderboard.py

@@ -52,7 +52,7 @@ def gr_create_pitcher_leaderboard(start_date, end_date, min_ip, qualified, pitch
           data, player_type='pitcher', 
           pitcher_lr='both' if pitcher_lr=='Both' else pitcher_lr[0].lower(), 
           batter_lr='both' if batter_lr == 'Both' else batter_lr[0].lower(), 
-          min_ip='qualified' if qualified else min_ip,
+          qual='qualified' if qualified else min_ip,
           group_by_team=True
         )
         .filter(pl.col('qualified'))
@@ -156,7 +156,7 @@ def create_pitcher_leaderboard():
     # )
     # pin_columns_state = gr.State(True)
     pin_columns.click(
-      lambda : gr.update(pinned_columns=None),
+      lambda : gr.Dataframe(pinned_columns=None),
       outputs=leaderboard
     )
     

player_team_leaderboard.py

@@ -0,0 +1,210 @@
+import gradio as gr
+import polars as pl
+import numpy as np
+
+from datetime import datetime
+from functools import partial
+
+from data import data_df
+from stats import compute_player_stats, filter_data_by_date_and_game_kind
+from convert import team_names_short_to_color, get_text_color_from_team
+from plotting import stat_cmap
+
+
+TEAMS  = [
+    'Yomiuri',
+    'Yakult',
+    'DeNA',
+    'Chunichi',
+    'Hanshin',
+    'Hiroshima',
+    'Nipponham',
+    'Rakuten',
+    'Seibu',
+    'Lotte',
+    'ORIX',
+    'SoftBank'
+]
+
+
+notes = '''**Limitations**
+- [Column widths get messed up when filtering](https://github.com/gradio-app/gradio/issues/11564)
+
+**To-do**
+- Fix IP calculation
+- Add post-season
+'''
+
+
+def get_col_width(col, player_team_type):
+  match col:
+    case 'Pitcher' | 'Batter':
+      return 125
+    case 'Team':
+      return 75
+    case 'Throws' | 'Bats':
+      return 60
+    case 'IP':
+      return 60 if player_team_type == 'team pitching' else 50
+    case _:
+      return max(50, 10*len(col))
+
+
+def create_player_team_leaderboard_app(player_team_type):
+  pitching = player_team_type in ('pitcher', 'team pitching')
+  team = 'team' in player_team_type
+
+  # stats
+  if pitching:
+    pct_stats = ['K%', 'BB%', 'Swing%', 'Z-Swing%', 'Chase%', 'Contact%', 'Z-Contact%', 'O-Contact%', 'SwStr%', 'Whiff%', 'CSW%', 'GB%', 'FB%', 'LD%', 'Zone%', 'Arm%', 'Glove%', 'High%', 'Low%', 'MM%']
+    stats_with_pctls = ['FB Velo', 'K%', 'BB%', 'Swing%', 'Z-Swing%', 'Chase%', 'Contact%', 'Z-Contact%', 'O-Contact%', 'SwStr%', 'Whiff%', 'CSW%', 'GB%', 'FB%', 'LD%', 'Zone%']
+    cols = ['Pitcher', 'Team', 'Throws', 'IP', 'TBF', 'FB Velo', 'K%', 'BB%', 'Swing%', 'Z-Swing%', 'Chase%', 'Contact%', 'Z-Contact%', 'O-Contact%', 'SwStr%', 'Whiff%', 'CSW%', 'GB%', 'FB%', 'LD%', 'Zone%', 'Arm%', 'Glove%', 'High%', 'Low%', 'MM%']
+    if team:
+      cols = [col for col in cols if col not in ('Pitcher', 'Throws')]
+  else:
+    pct_stats = ['K%', 'BB%', 'Swing%', 'Z-Swing%', 'Chase%', 'Contact%', 'Z-Contact%', 'O-Contact%', 'SwStr%', 'Whiff%', 'CSW%', 'GB%', 'FB%', 'LD%', 'Zone%', 'Arm%', 'Glove%', 'High%', 'Low%', 'MM%']
+    stats_with_pctls = ['K%', 'BB%', 'Swing%', 'Z-Swing%', 'Chase%', 'Contact%', 'Z-Contact%', 'O-Contact%', 'SwStr%', 'Whiff%', 'CSW%', 'GB%', 'FB%', 'LD%', 'Zone%']
+    cols = ['Batter', 'Team', 'Bats', 'PA', 'K%', 'BB%', 'Swing%', 'Z-Swing%', 'Chase%', 'Contact%', 'Z-Contact%', 'O-Contact%', 'SwStr%', 'Whiff%', 'CSW%', 'GB%', 'FB%', 'LD%']
+    if team:
+      cols = [col for col in cols if col not in ('Batter', 'Bats')]
+
+  # col names
+  player_type = 'pitcher' if pitching else 'batter'
+  id_col = f'{player_type[:3].lower()}Id' if not team else f'{player_type}_team_name_short'
+  qual_name = 'IP' if pitching else 'PA'
+  
+  def gr_create_player_team_leaderboard(start_date, end_date, min_qual, qualified, pitcher_lr='Both', batter_lr='Both', include_teams=None):
+    assert pitcher_lr in ['Both', 'Left', 'Right']
+    assert batter_lr in ['Both', 'Left', 'Right']
+
+    data = data_df.filter(pl.col('ballKind_code') != '-')
+    
+    data = filter_data_by_date_and_game_kind(data, start_date=start_date, end_date=end_date, game_kind='Regular Season')
+    
+    rename = {f'{player_type}_team_name_short': 'Team'}
+    if not team:
+      rename[f'{player_type}_name'] = player_type.title()
+    if pitching:
+      rename['PA'] = 'TBF'
+
+    # typically "qualified" should be a valid input for min_ip for the current function, 
+    # but we separate it from a numerical min_ip argument for API compabtibility
+    pitcher_stats = (
+          compute_player_stats(
+            data, player_type=player_team_type, 
+            pitcher_lr='both' if pitcher_lr=='Both' else pitcher_lr[0].lower(), 
+            batter_lr='both' if batter_lr == 'Both' else batter_lr[0].lower(), 
+            qual='qualified' if qualified else min_qual,
+            group_by_team=not team
+          )
+          .filter(pl.col('qualified'))
+          .drop(['qualified'] + ([id_col] if not team else []))
+          .rename(rename)
+          .with_columns(
+            pl.col(stat).mul(100)
+            for stat in pct_stats
+          )
+    )
+
+    # if not team:
+    if include_teams is not None:
+      pitcher_stats = pitcher_stats.filter(pl.col('Team').is_in(include_teams))
+
+    styling = []
+    for i, row in enumerate(pitcher_stats[cols].iter_rows()):
+      styling_row = []
+      for col, item in zip(pitcher_stats[cols].columns, row):
+        _styling = 'font-size: 0.75em; '
+        if col in stats_with_pctls:
+          r, g, b = (stat_cmap([pitcher_stats[f'{col}_pctl'][i]])[0, :3]*255).astype(np.uint8)
+          styling_row.append(f'background-color: rgba({r}, {g}, {b})')
+        elif col == 'Team':
+          styling_row.append(f'color: {get_text_color_from_team(item)}; background-color: {team_names_short_to_color[item]}')
+        else:
+          styling_row.append('')
+      styling.append(styling_row)
+
+    display_value = []
+    for row in pitcher_stats[cols].iter_rows():
+      display_value_row = []
+      for col, item in zip(cols, row):
+        if col in pct_stats:
+          display_value_row.append(f'{item:.1f}%')
+        elif isinstance(item, float):
+          display_value_row.append(f'{item:.1f}')
+        else:
+          display_value_row.append(item)
+      display_value.append(display_value_row)
+
+    value = {
+        'data': pitcher_stats[cols].rows(),
+        'headers': cols,
+        'metadata': {
+            'styling': styling,
+            'display_value': display_value,
+        }
+    }
+    
+    return value
+
+
+  now = datetime.now()
+  start_datetime_init = datetime(now.year, 1, 1)
+  end_datetime_init = now
+  with gr.Blocks() as app:
+    gr.Markdown(f'# {player_team_type.title()} Leaderboard')
+    with gr.Row():
+      start_date = gr.DateTime(start_datetime_init, include_time=False, type='datetime', label='Start')
+      end_date = gr.DateTime(end_datetime_init, include_time=False, type='datetime', label='End')
+    with gr.Row():
+      if not team:
+        with gr.Group():
+          min_ip = gr.Number(100, label=f'Min. {qual_name}', precision=0, minimum=0, interactive=False)
+          qualified = gr.Checkbox(True, label='Qualified')
+      else:
+        min_ip = gr.State(0)
+        qualified = gr.State(False)
+      with gr.Group():
+        pitcher_lr = gr.Radio(['Both', 'Left', 'Right'], value='Both', label='Pitcher handedness') 
+        batter_lr = gr.Radio(['Both', 'Left', 'Right'], value='Both', label='Batter handedness')
+    with gr.Row():
+      include_teams = gr.CheckboxGroup(TEAMS, value=TEAMS, label='Teams', scale=3)
+      all_teams = gr.Button('Select/Deselect all teams')
+        
+    search = gr.Button('Search')
+    pin_columns = gr.Checkbox(True, label='Pin columns')
+    leaderboard = gr.DataFrame(
+      pl.DataFrame({'Pitcher': [], 'Pitch': []}),
+      column_widths=[get_col_width(col, player_team_type) for col in cols],
+      show_copy_button=True,
+      show_search='filter',
+      pinned_columns=2,
+      elem_id='leaderboard'
+    )
+
+    gr.Markdown(notes)
+
+    search.click(gr_create_player_team_leaderboard, inputs=[start_date, end_date, min_ip, qualified, pitcher_lr, batter_lr, include_teams], outputs=leaderboard)
+    all_teams.click(lambda _teams : [] if _teams == TEAMS else TEAMS, inputs=include_teams, outputs=include_teams)
+    qualified.change(lambda qualified: gr.Number(interactive=not qualified), inputs=qualified, outputs=min_ip)
+    pin_columns.input(
+      lambda pin: gr.DataFrame(pinned_columns=2 if pin else None),
+      inputs=pin_columns,
+      outputs=leaderboard
+    )
+    
+  return app
+
+create_pitcher_leaderboard = partial(
+  create_player_team_leaderboard_app,
+  player_team_type='pitcher'
+)
+
+create_team_pitching_leaderboard = partial(
+  create_player_team_leaderboard_app,
+  player_team_type='team pitching'
+)
+
+if __name__ == '__main__':
+  app = foo()
+  app.launch()

stats.py

@@ -75,6 +75,8 @@ def compute_pitch_stats(data, player_type, pitch_class_type, min_pitches=1, pitc
   assert player_type in ('pitcher', 'batter')
   assert pitch_class_type in ('general', 'specific')
 
+  pitching = player_type in ('pitcher', )
+
   if pitcher_lr != 'both':
     data = data.filter(pl.col('pitLR') == pitcher_lr)
 
@@ -82,8 +84,9 @@ def compute_pitch_stats(data, player_type, pitch_class_type, min_pitches=1, pitc
     data = data.filter(pl.col('batLR') == batter_lr)
     
   id_cols = ['pitId' if player_type == 'pitcher' else 'batId']
+  team_col = 'pitcher_team_name_short' if pitching else 'batter_team_name_short'
   if group_by_team:
-    id_cols.append('pitcher_team_name_short')
+    id_cols.append(team_col)
   name_col = 'pitcher_name' if player_type == 'pitcher' else 'batter_name'
   pitch_col = 'ballKind_code' if pitch_class_type == 'specific' else 'general_ballKind_code'
   pitch_name_col = 'ballKind' if pitch_class_type == 'specific' else 'general_ballKind'
@@ -141,15 +144,32 @@ def compute_pitch_stats(data, player_type, pitch_class_type, min_pitches=1, pitc
   )
   return pitch_stats
 
-def compute_player_stats(data, player_type, min_ip='qualified', pitcher_lr='both', batter_lr='both', group_by_team=False):
+def compute_player_stats(data, player_type, qual='qualified', pitcher_lr='both', batter_lr='both', group_by_team=False):
   assert pitcher_lr in ('both', 'l', 'r')
   assert batter_lr in ('both', 'l', 'r')
-  
+  assert player_type in ('pitcher', 'batter', 'team pitching', 'team batting') 
+
+  # pitching or batting, player or team
+  pitching = player_type in ('pitcher', 'team pitching')
+  team = player_type in ('team pitching', 'team batting')
+
+  # handedness filters
+  if pitcher_lr != 'both':
+    data = data.filter(pl.col('pitLR') == pitcher_lr)
+  if batter_lr != 'both':
+    data = data.filter(pl.col('batLR') == batter_lr)
+
+  if pitching:
+    over_col = 'pitId' if not team else 'pitcher_team_name_short'
+  else:
+    over_col = 'batId' if not team else 'batter_team_name_short'
   data = (
       compute_team_games(data)
       .with_columns(
           pl.when(pl.col('half_inning').str.ends_with('1')).then('home_games').otherwise('visitor_games').first().over('pitId').alias('games'),
-          pl.col('inning_code').unique().len().over('pitId').alias('IP')
+          # pl.col('inning_code').unique().len().over(over_col).alias('IP'),
+          (pl.col('bso').struct.field('o').cast(pl.Int32) - pl.col('beforeBso').struct.field('o').cast(pl.Int32)).sum().mul(1/3).over(over_col).alias('IP'),
+          pl.col('pa_code').unique().len().over(over_col).alias('PA'),
           # pl.col('presult').is_in(verify_and_return_presult([
             # 'Groundout', 'Flyout', 'Lineout', 'Groundout (Double play)',
             # 'Foul fly', 'Foul line (?)',
@@ -160,32 +180,52 @@ def compute_player_stats(data, player_type, min_ip='qualified', pitcher_lr='both
       )
   )
 
-  if min_ip == 'qualified':
-    data = data.with_columns((pl.col('IP') >= pl.col('games')).alias('qualified'))
+  # qualifiers
+  qualified_factor = 1 if pitching else 3.1
+  qual_col = 'IP' if pitching else 'PA'
+  if qual == 'qualified':
+    data = data.with_columns((pl.col(qual_col) >= qualified_factor * pl.col('games')).alias('qualified'))
   else:
-    data = data.with_columns((pl.col('IP') >= min_ip).alias('qualified'))
+    data = data.with_columns((pl.col(qual_col) >= qual).alias('qualified'))
 
-  if pitcher_lr != 'both':
-    data = data.filter(pl.col('pitLR') == pitcher_lr)
-
-  if batter_lr != 'both':
-    data = data.filter(pl.col('batLR') == batter_lr)
+  # percentile ascending/descending
+  if pitching:
+    stat_descending_pctl = lambda stat: stat in ['BB%', 'FB%', 'LD%', 'Z-Swing%'] or 'Contact%' in stat
+  else:
+    stat_descending_pctl = lambda stat: not (stat in ['BB%', 'FB%', 'LD%', 'Swing%', 'Z-Swing%'] or 'Contact%' in stat)
 
-  assert player_type in ('pitcher', 'batter') 
-  id_cols = ['pitId' if player_type == 'pitcher' else 'batId']
-  if group_by_team:
-    id_cols.append('pitcher_team_name_short')
-  name_col = 'pitcher_name' if player_type == 'pitcher' else 'batter_name'
+  # col names
+  match player_type:
+    case 'pitcher':
+      id_cols = ['pitId']
+      name_col = 'pitcher_name'
+    case 'batter':
+      id_cols = ['batId']
+      name_col = 'batter_name'
+    case _:
+      id_cols = []
+      name_col = None
+      
+  team_col = 'pitcher_team_name_short' if pitching else 'batter_team_name_short'
+  if group_by_team or team:
+    id_cols.append(team_col)
+  
+  handedness_col = 'pitLR' if pitching else 'batLR'
+  new_handedness_col = 'Throws' if pitching else 'Bats'
   player_stats = (
     data
     .with_columns(pl.when(pl.col('general_ballKind_code').is_in(['4S', 'FC', 'SI'])).then(pl.when(valid_pitch).then('ballSpeed').mean().over('pitId', 'general_ballKind_code')).mul(1/1.609).round(1).alias('FB Velo'))
     .group_by(id_cols)
     .agg(
-        pl.col(name_col).first(),
-        *([] if group_by_team else [pl.col('pitcher_team_name_short').last()]),
-        pl.col('pitLR').first().str.to_uppercase().alias('Throws'),
+        *([pl.col(name_col).first()] if not team else []),
+        *([] if group_by_team or team else [pl.col(team_col).last()]),
+        *(
+          [pl.col(handedness_col).first().str.to_uppercase().alias(new_handedness_col) ]
+          if not (team and ((pitcher_lr == 'both') if pitching else (batter_lr == 'both')))
+          else []
+        ),
         pl.col('IP').first(),
-        pl.col('pa_code').unique().len().alias('PA'),
+        pl.col('PA').first(),
         pl.col('FB Velo').max(),
         (pl.when(pl.col('presult').str.contains('strikeout')).then(1).otherwise(0).sum() / pl.col('pa_code').unique().len()).alias('K%'),
         (pl.when(pl.col('presult') == 'Walk').then(1).otherwise(0).sum() / pl.col('pa_code').unique().len()).alias('BB%'),
@@ -218,10 +258,10 @@ def compute_player_stats(data, player_type, min_ip='qualified', pitcher_lr='both
     )
     .drop('G', 'F', 'B', 'P', 'L')
     .with_columns(
-        (pl.when(pl.col('qualified')).then(pl.col(stat)).rank(descending=stat in ['BB%', 'FB%', 'LD%'] or 'Contact%' in stat)/pl.when(pl.col('qualified')).then(pl.col(stat)).count()).alias(f'{stat}_pctl')
+        (pl.when(pl.col('qualified')).then(pl.col(stat)).rank(descending=stat_descending_pctl(stat))/pl.when(pl.col('qualified')).then(pl.col(stat)).count()).alias(f'{stat}_pctl')
         for stat in ['FB Velo', 'K%', 'BB%', 'Swing%', 'Z-Swing%', 'Chase%', 'Contact%', 'Z-Contact%', 'O-Contact%', 'SwStr%', 'Whiff%', 'CSW%', 'GB%', 'FB%', 'LD%', 'Zone%']
     )
-    .sort('IP', descending=True)
+    .sort(qual_col, descending=True)
   )
   return player_stats
 
@@ -248,5 +288,5 @@ def get_pitcher_stats(id, lr='both', game_kind=None, start_date=None, end_date=N
       .with_columns((pl.col('ballSpeed')/1.609).alias('ballSpeed_mph'))
   )
 
-  pitcher_stats = compute_player_stats(source_data, player_type='pitcher', min_ip=min_ip, batter_lr=lr, group_by_team=False).filter(pl.col('pitId') == id)
+  pitcher_stats = compute_player_stats(source_data, player_type='pitcher', qual=min_ip, batter_lr=lr, group_by_team=False).filter(pl.col('pitId') == id)
   return SimpleNamespace(pitcher_stats=pitcher_stats, pitch_stats=pitch_stats, pitch_shapes=pitch_shapes)