Add colors to pitch leaderboard

convert.py

@@ -311,7 +311,7 @@ game_kind = {
 }
 
 ball_kind_code_to_color = {
-    '-': '',
+    '-': 'C0',
     '4S': 'crimson',
     'SL': 'gold',
     'VS': 'khaki',
@@ -326,6 +326,7 @@ ball_kind_code_to_color = {
     'SI': 'orange',
     'SB': 'lightgreen',
     'PB': 'yellowgreen',
+    'KN': 'mediumblue',
     'SH': 'tomato',
     '2S': 'orangered',
     '1S': 'lightsalmon',
@@ -336,9 +337,30 @@ ball_kind_code_to_color = {
 }
 
 
-ball_kind_code_to_color = {k: v if v else 'C0' for k, v in ball_kind_code_to_color.items()}
-
 def get_text_color_from_color(color):
-  if color in ['gold', 'paleturquoise']:
+  if color in ['gold', 'paleturquoise', 'turquoise']:
     return 'black'
   return 'white'
+
+ball_kind_to_color = {ball_kind: ball_kind_code_to_color[ball_kind_code[code]] for code, ball_kind in ball_kind.items()}
+
+team_to_color = {
+    'G': '#f69727',
+    'S': '#abcd05', 
+    'DB': '#004a8f', 
+    'D': '#08b1e8',
+    'T': '#ffe100', 
+    'C': '#f92b20',
+    'F': '#016299', 
+    'E': '#7d001a', 
+    'L': '#00214b', 
+    'M': '#efefef', 
+    'B': '#baa834', 
+    'H': '#fcc700'
+}
+
+def get_text_color_from_team(team):
+  if team in ['DB', 'F', 'L', 'E']:
+    return 'white'
+  else:
+    return 'black'

pitch_leaderboard.py

@@ -1,16 +1,19 @@
 import gradio as gr
 import polars as pl
+import numpy as np
 
 from datetime import datetime
 # from itertools import chain
 
 from data import data_df
 from stats import compute_pitch_stats, filter_data_by_date_and_game_kind
-from convert import ball_kind
+from convert import ball_kind, ball_kind_to_color, get_text_color_from_color, team_to_color, get_text_color_from_team
+from plotting import stat_cmap
 
 STATS = ['Count', 'Usage', '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%']
+STATS_WITH_PCTLS = ['Swing%', 'Z-Swing%', 'Chase%', 'Contact%', 'Z-Contact%', 'O-Contact%', 'SwStr%', 'Whiff%', 'CSW%', 'GB%', 'FB%', 'LD%']
+COLUMNS = ['Pitcher', 'Team', 'Pitch', 'Pitch (General)'] + STATS
 
 PITCH_TYPES = [pitch_type for pitch_type in ball_kind.values() if pitch_type != '-']
 TEAMS  = ['G', 'S', 'DB', 'D', 'T', 'C', 'F', 'E', 'L', 'M', 'B', 'H']
@@ -59,14 +62,50 @@ def gr_create_pitch_leaderboard(start_date, end_date, min_pitches, pitcher_lr='B
         compute_pitch_stats(data, player_type='pitcher', min_pitches=min_pitches, pitch_class_type='specific')
         .filter(pl.col('qualified') & (pl.col('ballKind').is_in(include_pitches)))
         .drop('pitId', 'ballKind_code', 'qualified')
-        .rename({'pitcher_name': 'Pitcher', 'pitcher_team': 'Team', 'count': 'Count', 'usage': 'Usage', 'ballKind': 'Pitch', 'general_ballKind': 'Pitch (General)'} | {f'{stat}_pctl': f'{stat} (Pctl)' for stat in STATS_WITH_PCTLS})
-        .with_columns(
-          pl.col(stat).mul(100).round(1)
-          for stat in PCT_STATS + [f'{stat} (Pctl)' for stat in STATS_WITH_PCTLS]
-        )
-        [['Pitcher', 'Team', 'Pitch', 'Pitch (General)'] + STATS]
+        .rename({'pitcher_name': 'Pitcher', 'pitcher_team': 'Team', 'count': 'Count', 'usage': 'Usage', 'ballKind': 'Pitch', 'general_ballKind': 'Pitch (General)'})
+        # .with_columns(
+          # pl.col(stat).mul(100).round(1)
+          # for stat in PCT_STATS + [f'{stat}_pctl' for stat in STATS_WITH_PCTLS]
+        # )
+        # [['Pitcher', 'Team', 'Pitch', 'Pitch (General)'] + STATS + [f'{stat}_pctl' for stat in STATS_WITH_PCTLS]]
   )
-  return pitch_stats
+
+  styling = []
+  for i, row in enumerate(pitch_stats[COLUMNS].iter_rows()):
+    styling_row = []
+    for col, item in zip(pitch_stats[COLUMNS].columns, row):
+      if f'{col}_pctl' in pitch_stats:
+        r, g, b = (stat_cmap([pitch_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_to_color[item]}')
+      elif col in ['Pitch', 'Pitch (General)']:
+        color = ball_kind_to_color[item]
+        styling_row.append(f'color: {get_text_color_from_color(color)}; background-color: {color}')
+      else:
+        styling_row.append('')
+    styling.append(styling_row)
+
+  display_value = []
+  for row in pitch_stats[COLUMNS].iter_rows():
+    display_value_row = []
+    for item in row:
+      if isinstance(item, float):
+        display_value_row.append(f'{item:.1%}')
+      else:
+        display_value_row.append(item)
+    display_value.append(display_value_row)
+
+  value = {
+      'data': pitch_stats[COLUMNS].rows(),
+      'headers': COLUMNS,
+      'metadata': {
+          'styling': styling,
+          'display_value': display_value,
+      }
+  }
+  
+  return value
 
 
 def create_pitch_leaderboard():

pitcher_overview.py

@@ -10,11 +10,12 @@ notes = '''**Limitations**
 - Foreign players names are in Hebpurn romanization. Contact me if you need a card for a foreign player.
 
 **To-do**
-- Fix names of foreign playeres
+- Fix names of foreign players
 - Add teams insignias
 - Measure percentiles per pitcher handedness
 - Allow for arbitrary date ranges
 - Improve readability of pitch velocities
+- Add post-season
 '''
 
 def dummy(*inputs):
@@ -22,9 +23,10 @@ def dummy(*inputs):
 
 def gr_create_pitcher_overview_card(name, season):
   # pit_id = name.split(' | ')[-1]
-  pit_id = data_df.filter(pl.col('pitcher_name') == name)['pitId'].unique()
+  # hard-coded to only do regular season
+  pit_id = data_df.filter((pl.col('pitcher_name') == name) & (pl.col('date').dt.year() == season) & (pl.col('coarse_game_kind') == 'Regular Season'))['pitId'].unique()
   if len(pit_id) == 0:
-    raise gr.Error(f"No data found for {name}. If the name looks strangely spelled or formatted there's a possibility that's what causing the error.")
+    raise gr.Error(f"No data found for {name} in {season}. If the name looks strangely spelled or formatted there's a possibility that's what causing the error.")
   elif len(pit_id) > 1:
     raise gr.Error(f'Multiple IDs for {name}')
   else:
@@ -49,7 +51,7 @@ def create_pitcher_overview(data_df):
         # names = [f'{pit_name} | {pit_id}' for pit_name, pit_id in data_df[['pitcher_name', 'pitId']].unique().sort('pitId').iter_rows()]
         names = data_df['pitcher_name'].unique().sort().to_list()
         name = gr.Dropdown(names, label='Name')
-        season = gr.Dropdown(SEASONS, label='Season')
+        season = gr.Dropdown(SEASONS, value=max(SEASONS), label='Season')
         # season_start = gr.Dropdown(SEASONS, label='Season start')
         # season_end = gr.Dropdown(SEASONS, label='Season end')
         # game_type = gr.Dropdown(['Spring Training', 'Regular Season', 'Postseason'], label='Game Type'])

stats.py

@@ -99,7 +99,7 @@ def compute_pitch_stats(data, player_type, pitch_class_type, min_pitches=1):
       .drop('G', 'F', 'B', 'P', 'L', 'null')
       .with_columns(
           (pl.when(pl.col('qualified')).then(pl.col(stat)).rank(descending=((stat in ['FB%', 'LD%'] or 'Contact%' in stat)))/pl.when(pl.col('qualified')).then(pl.col(stat)).count()).alias(f'{stat}_pctl')
-          for stat in ['Zone%', 'Swing%', 'Z-Swing%', 'Chase%', 'Contact%', 'Z-Contact%', 'O-Contact%', 'SwStr%', 'Whiff%', 'CSW%', 'GB%', 'FB%', 'LD%']
+          for stat in ['Swing%', 'Z-Swing%', 'Chase%', 'Contact%', 'Z-Contact%', 'O-Contact%', 'SwStr%', 'Whiff%', 'CSW%', 'GB%', 'FB%', 'LD%']
       )
       .rename({pitch_col: 'ballKind_code', pitch_name_col: 'ballKind'} if pitch_class_type == 'general' else {})
       .sort(id_col, 'count', descending=[False, True])