Add pitcher handedness filtering

README.md

@@ -4,7 +4,7 @@ emoji: ⚾️
 colorFrom: blue
 colorTo: red
 sdk: gradio
-sdk_version: 5.37.0
+sdk_version: 5.44.1
 app_file: app.py
 pinned: false
 python_version: 3.13.5

pitch_leaderboard.py

@@ -13,7 +13,7 @@ from plotting import stat_cmap
 STATS = ['Count', 'Usage', 'Avg Velo', 'Max Velo', 'Swing%', 'Z-Swing%', 'Chase%', 'Contact%', 'Z-Contact%', 'O-Contact%', 'SwStr%', 'Whiff%', 'CSW%', 'GB%', 'FB%', 'LD%', 'Zone%', 'Arm%', 'Glove%', 'High%', 'Low%', 'MM%']
 PCT_STATS = ['Usage', 'Swing%', 'Z-Swing%', 'Chase%', 'Contact%', 'Z-Contact%', 'O-Contact%', 'SwStr%', 'Whiff%', 'CSW%', 'GB%', 'FB%', 'LD%', 'Zone%', 'Arm%', 'Glove%', 'High%', 'Low%', 'MM%']
 STATS_WITH_PCTLS = ['Swing%', 'Z-Swing%', 'Chase%', 'Contact%', 'Z-Contact%', 'O-Contact%', 'SwStr%', 'Whiff%', 'CSW%', 'GB%', 'FB%', 'LD%', 'Zone%']
-COLUMNS = ['Pitcher', 'Team', 'Pitch', 'Pitch (General)'] + STATS
+COLUMNS = ['Pitcher', 'Team', 'Throws', 'Pitch', 'Pitch (General)'] + STATS
 
 PITCH_TYPES = [pitch_type for pitch_type in ball_kind.values() if pitch_type != '-']
 TEAMS  = [
@@ -38,17 +38,23 @@ notes = '''**Limitations**
 '''
 
 
-def gr_create_pitch_leaderboard(start_date, end_date, min_pitches, pitcher_lr='Both', include_pitches=PITCH_TYPES, include_teams=None):
+def gr_create_pitch_leaderboard(start_date, end_date, min_pitches, pitcher_lr='Both', batter_lr='Both', include_pitches=PITCH_TYPES, include_teams=None):
   assert pitcher_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')
-  if pitcher_lr != 'Both':
-    data = data.filter(pl.col('batLR') == pitcher_lr[0].lower())
     
   pitch_stats = (
-        compute_pitch_stats(data, player_type='pitcher', min_pitches=min_pitches, pitch_class_type='specific', group_by_team=True)
+        compute_pitch_stats(
+          data, 
+          player_type='pitcher', 
+          min_pitches=min_pitches, 
+          pitch_class_type='specific', 
+          pitcher_lr='both' if pitcher_lr=='Both' else pitcher_lr[0].lower(), 
+          batter_lr='both' if batter_lr == 'Both' else batter_lr[0].lower(), 
+          group_by_team=True
+        )
         .filter(pl.col('qualified') & (pl.col('ballKind').is_in(include_pitches)))
         .drop('pitId', 'ballKind_code', 'qualified')
         .rename({
@@ -123,10 +129,12 @@ def create_pitch_leaderboard():
       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():
-      include_pitches = gr.CheckboxGroup(PITCH_TYPES, value=PITCH_TYPES, label='Pitches', scale=3)
-      with gr.Column(scale=1):
+      with gr.Column(scale=3):
+        include_pitches = gr.CheckboxGroup(PITCH_TYPES, value=PITCH_TYPES, label='Pitches')
         all_pitches = gr.Button('Select/Deselect all pitches')
+      with gr.Column(scale=1):
         min_pitches = gr.Number(100, label='Min. Pitches', precision=0, minimum=0)
+        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)
@@ -136,7 +144,7 @@ def create_pitch_leaderboard():
     pin_columns = gr.Button('Pin columns')
     leaderboard = gr.DataFrame(
       pl.DataFrame({'Pitcher': [], 'Pitch': []}),
-      column_widths=[125, 75, 125, 125] + [max(50, 10*len(stat)) for stat in STATS],
+      column_widths=[125, 75, 60, 125, 125] + [max(50, 10*len(stat)) for stat in STATS],
       show_copy_button=True,
       show_search='filter',
       pinned_columns=3,
@@ -145,7 +153,7 @@ def create_pitch_leaderboard():
 
     gr.Markdown(notes)
 
-    search.click(gr_create_pitch_leaderboard, inputs=[start_date, end_date, min_pitches, batter_lr, include_pitches, include_teams], outputs=leaderboard)
+    search.click(gr_create_pitch_leaderboard, inputs=[start_date, end_date, min_pitches, pitcher_lr, batter_lr, include_pitches, include_teams], outputs=leaderboard)
     all_pitches.click(lambda _pitch_types : [] if _pitch_types == PITCH_TYPES else PITCH_TYPES, inputs=include_pitches, outputs=include_pitches)
     all_teams.click(lambda _teams : [] if _teams == TEAMS else TEAMS, inputs=include_teams, outputs=include_teams)
     # pin_columns.input(

pitcher_leaderboard.py

@@ -13,7 +13,7 @@ from plotting import stat_cmap
 STATS = ['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%']
 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%']
-COLUMNS = ['Pitcher', 'Team', 'IP', 'TBF'] + STATS
+COLUMNS = ['Pitcher', 'Team', 'Throws', 'IP', 'TBF'] + STATS
 
 TEAMS  = [
     'Yomiuri',
@@ -38,17 +38,21 @@ notes = '''**Limitations**
 '''
 
 
-def gr_create_pitcher_leaderboard(start_date, end_date, min_ip, pitcher_lr='Both', include_teams=None):
+def gr_create_pitcher_leaderboard(start_date, end_date, min_ip, pitcher_lr='Both', batter_lr='Both', include_teams=None):
   assert pitcher_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')
-  if pitcher_lr != 'Both':
-    data = data.filter(pl.col('batLR') == pitcher_lr[0].lower())
     
   pitcher_stats = (
-        compute_player_stats(data, player_type='pitcher', min_ip=min_ip, group_by_team=True)
+        compute_player_stats(
+          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=min_ip, 
+          group_by_team=True
+        )
         .filter(pl.col('qualified'))
         .drop('pitId', 'qualified')
         .rename({
@@ -117,7 +121,9 @@ def create_pitcher_leaderboard():
         min_ip_state = gr.State('qualified')
         min_ip = gr.Number(100, label='Min. IP', precision=0, minimum=0, interactive=False)
         qualified = gr.Checkbox(True, label='Qualified')
-      batter_lr = gr.Radio(['Both', 'Left', 'Right'], value='Both', label='Batter handedness')
+      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')
@@ -126,7 +132,7 @@ def create_pitcher_leaderboard():
     pin_columns = gr.Button('Pin columns')
     leaderboard = gr.DataFrame(
       pl.DataFrame({'Pitcher': [], 'Pitch': []}),
-      column_widths=[125, 75, 50, 50] + [max(50, 10*len(stat)) for stat in STATS],
+      column_widths=[125, 75, 60, 50, 50] + [max(50, 10*len(stat)) for stat in STATS],
       show_copy_button=True,
       show_search='filter',
       pinned_columns=2,
@@ -135,9 +141,9 @@ def create_pitcher_leaderboard():
 
     gr.Markdown(notes)
 
-    search.click(gr_create_pitcher_leaderboard, inputs=[start_date, end_date, min_ip_state, batter_lr, include_teams], outputs=leaderboard)
+    search.click(gr_create_pitcher_leaderboard, inputs=[start_date, end_date, min_ip_state, pitcher_lr, batter_lr, include_teams], outputs=leaderboard)
     all_teams.click(lambda _teams : [] if _teams == TEAMS else TEAMS, inputs=include_teams, outputs=include_teams)
-    min_ip_state_kwargs = dict(fn=lambda min_ip, qualified: (qualified if qualified else min_ip, gr.Number(interactive=not qualified)), inputs=[min_ip, qualified], outputs=[min_ip_state, min_ip])
+    min_ip_state_kwargs = dict(fn=lambda min_ip, qualified: ('qualified' if qualified else min_ip, gr.Number(interactive=not qualified)), inputs=[min_ip, qualified], outputs=[min_ip_state, min_ip])
     min_ip.change(**min_ip_state_kwargs)
     qualified.change(**min_ip_state_kwargs)
     # pin_columns.input(

stats.py

@@ -69,9 +69,18 @@ def compute_team_games(data):
   )
 
 
-def compute_pitch_stats(data, player_type, pitch_class_type, min_pitches=1, group_by_team=False):
+def compute_pitch_stats(data, player_type, pitch_class_type, min_pitches=1, 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')
   assert pitch_class_type in ('general', 'specific')
+
+  if pitcher_lr != 'both':
+    data = data.filter(pl.col('pitLR') == pitcher_lr)
+
+  if batter_lr != 'both':
+    data = data.filter(pl.col('batLR') == batter_lr)
+    
   id_cols = ['pitId' if player_type == 'pitcher' else 'batId']
   if group_by_team:
     id_cols.append('pitcher_team_name_short')
@@ -84,6 +93,7 @@ def compute_pitch_stats(data, player_type, pitch_class_type, min_pitches=1, grou
       .group_by(*id_cols, pitch_col)
       .agg(
           pl.first(name_col),
+          pl.col('pitLR').first().str.to_uppercase().alias('Throws'),
           *([pl.first('general_ballKind')] if pitch_class_type == 'specific' else []),
           pl.first(pitch_name_col),
           pl.len().alias('count'),
@@ -131,7 +141,10 @@ def compute_pitch_stats(data, player_type, pitch_class_type, min_pitches=1, grou
   )
   return pitch_stats
 
-def compute_player_stats(data, player_type, min_ip='qualified', group_by_team=False):
+def compute_player_stats(data, player_type, min_ip='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')
+  
   data = (
       compute_team_games(data)
       .with_columns(
@@ -152,6 +165,12 @@ def compute_player_stats(data, player_type, min_ip='qualified', group_by_team=Fa
   else:
     data = data.with_columns((pl.col('IP') >= min_ip).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)
+
   assert player_type in ('pitcher', 'batter') 
   id_cols = ['pitId' if player_type == 'pitcher' else 'batId']
   if group_by_team:
@@ -164,6 +183,7 @@ def compute_player_stats(data, player_type, min_ip='qualified', group_by_team=Fa
     .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('IP').first(),
         pl.col('pa_code').unique().len().alias('PA'),
         pl.col('FB Velo').max(),
@@ -206,26 +226,18 @@ def compute_player_stats(data, player_type, min_ip='qualified', group_by_team=Fa
   return player_stats
 
 
-def get_pitcher_stats(id, lr=None, game_kind=None, start_date=None, end_date=None, min_ip=1, min_pitches=1, pitch_class_type='specific'):
-  # source_data = data_df.filter(pl.col('ballKind_code') != '-')
-
-  # if start_date is not None:
-    # source_data = source_data.filter(pl.col('date') >= start_date)
-  # if end_date is not None:
-    # source_data = source_data.filter(pl.col('date') <= end_date)
-# 
-  # if game_kind is not None:
-    # source_data = source_data.filter(pl.col('coarse_game_kind') == game_kind)
+def get_pitcher_stats(id, lr='both', game_kind=None, start_date=None, end_date=None, min_ip=1, min_pitches=1, pitch_class_type='specific'):
+  
   source_data = data_df
   source_data = filter_data_by_date_and_game_kind(source_data, start_date=start_date, end_date=end_date, game_kind=game_kind)
 
-  if lr is not None:
-    source_data = source_data.filter(pl.col('batLR') == lr)
+  # if lr is not None:
+    # source_data =
 
-  pitch_stats = compute_pitch_stats(source_data, player_type='pitcher', pitch_class_type=pitch_class_type, min_pitches=min_pitches, group_by_team=False).filter(pl.col('pitId') == id)
+  pitch_stats = compute_pitch_stats(source_data, player_type='pitcher', pitch_class_type=pitch_class_type, min_pitches=min_pitches, batter_lr=lr, group_by_team=False).filter(pl.col('pitId') == id)
 
   pitch_shapes = (
-      source_data
+      (source_data.filter(pl.col('batLR') == lr) if lr != 'both' else source_data)
       .filter(
           (pl.col('pitId') == id) &
           pl.col('x').is_not_null() &
@@ -236,6 +248,5 @@ def get_pitcher_stats(id, lr=None, game_kind=None, start_date=None, end_date=Non
       .with_columns((pl.col('ballSpeed')/1.609).alias('ballSpeed_mph'))
   )
 
-  pitcher_stats = compute_player_stats(source_data, player_type='pitcher', min_ip=min_ip, group_by_team=False).filter(pl.col('pitId') == id)
-
+  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)
   return SimpleNamespace(pitcher_stats=pitcher_stats, pitch_stats=pitch_stats, pitch_shapes=pitch_shapes)