Separate apps into tabs

README.md

@@ -1,6 +1,6 @@
 ---
 title: npb
-app_file: demo.py
+app_file: app.py
 sdk: gradio
 sdk_version: 4.44.1
 ---

app.py

@@ -0,0 +1,25 @@
+import gradio as gr
+
+from pitcher_dashboard import create_pitcher_dashboard
+from pitch_leaderboard import create_pitch_leaderboard
+from css import css
+
+
+with gr.Blocks(
+    css=css
+) as demo:
+  gr.Markdown('''
+  # NPB data visualization demo
+  [Data from SportsNavi](https://sports.yahoo.co.jp/)
+  ''')
+
+  with gr.Tab('Pitcher Dashboard'):
+    pitcher_dashboard = create_pitcher_dashboard()
+  
+  with gr.Tab('Pitch Leaderboard'):
+    pitch_leaderboard_app = create_pitch_leaderboard()
+
+demo.launch(
+    share=True,
+    debug=True
+)

css.py

@@ -0,0 +1,12 @@
+css = '''
+.pitch-usage {height: 256px}
+.pitch-usage .js-plotly-plot {height: 100%}
+
+.pitch-velo {height: 100px}
+.pitch-velo .js-plotly-plot {height: 100%}
+
+.pitch-loc {height: 320px}
+.pitch-loc .js-plotly-plot {height: 100%}
+
+.pitch-velo-summary div.plotly-notifier {visibility: hidden}
+'''

data.py

@@ -100,8 +100,8 @@ pitch_df = (
         pl.col('pitch_name').alias('jp_pitch_name')
     )
     .with_columns(
-        pl.col('jp_pitch_name').map_elements(lambda pitch_name: jp_pitch_to_en_pitch[pitch_name], return_dtype=str).alias('pitch_name'),
-        pl.col('jp_pitch_name').map_elements(lambda pitch_name: jp_pitch_to_pitch_code[pitch_name], return_dtype=str).alias('pitch_type'),
+        pl.col('jp_pitch_name').replace_strict(jp_pitch_to_en_pitch).alias('pitch_name'),
+        pl.col('jp_pitch_name').replace_strict(jp_pitch_to_pitch_code).alias('pitch_type'),
         pl.col('description').str.split(' ').list.first().map_elements(translate_pitch_outcome, return_dtype=str),
         pl.when(
             pl.col('release_speed') != '-'
@@ -168,7 +168,6 @@ pitch_stats, rhb_pitch_stats, lhb_pitch_stats = [
             pl.len().alias('Count')
         )
         .sort(['name', 'Count'], descending=[False, True])
-        .rename({'name': 'Player', 'pitch_name': 'Pitch'})
     )
     for _df
     in (

pitch_leaderboard.py

@@ -0,0 +1,40 @@
+import gradio as gr
+import polars as pl
+
+from math import ceil
+import os
+
+from data import pitch_stats
+from gradio_function import *
+from css import css
+
+pitch_stats = pitch_stats.rename({'name': 'Player', 'pitch_name': 'Pitch'})
+
+def filter_pitch_leaderboard(min_pitches):
+  return pitch_stats.filter(pl.col('Count') >= min_pitches).sort('CSW%', descending=True)
+
+def create_pitch_leaderboard():
+	with gr.Blocks(
+	    css=css
+	) as pitch_leaderboard_app:
+		init_min_pitches = 100
+		pitch_stats.write_csv('pitch_leaderboard.csv')
+		pitch_leaderboard_df = gr.State(filter_pitch_leaderboard(init_min_pitches))
+
+		min_pitches = gr.Number(init_min_pitches, precision=0, label='Min. Pitches')
+		pitch_leaderboard_download_file = gr.DownloadButton(value='pitch_leaderboard.csv', label='Download leaderboard')
+		pitch_leaderboard = gr.Dataframe(value=pitch_leaderboard_df.value)
+
+		min_pitches.change(filter_pitch_leaderboard, inputs=min_pitches, outputs=pitch_leaderboard_df)
+		(
+		    pitch_leaderboard_df
+		    .change(create_set_download_file_fn('files/pitch_leaderboard.csv'), inputs=pitch_leaderboard_df, outputs=pitch_leaderboard_download_file)
+		    .then(lambda df: df, inputs=pitch_leaderboard_df, outputs=pitch_leaderboard)
+		)
+	return pitch_leaderboard_app
+
+if __name__ == '__main__':
+	create_pitch_leaderboard().launch(
+	    share=True,
+	    debug=True
+	)

pitcher_dashboard.py

@@ -0,0 +1,118 @@
+
+import gradio as gr
+# import pandas as pd
+import polars as pl
+
+from math import ceil
+import os
+
+from data import df, pitch_stats, league_pitch_stats, player_df
+from gradio_function import *
+from translate import jp_pitch_to_en_pitch, max_pitch_types
+from css import css
+
+os.makedirs('files', exist_ok=True)
+
+def create_pitcher_dashboard():
+  with gr.Blocks(
+      css=css
+  ) as pitcher_dashboard_app:
+    gr.Markdown('''
+    # NPB data visualization demo
+    [Data from SportsNavi](https://sports.yahoo.co.jp/)
+    ''')
+
+    source_df = gr.State(df)
+    app_df = gr.State(df)
+    app_league_df = gr.State(df)
+    app_pitch_stats = gr.State(pitch_stats)
+    app_league_pitch_stats = gr.State(league_pitch_stats)
+
+    with gr.Row():
+      player = gr.Dropdown(value=None, choices=sorted(player_df.filter(pl.col('name').is_not_null())['name'].to_list()), label='Player')
+      handedness = gr.Radio(value='Both', choices=['Both', 'Left', 'Right'], type='value', interactive=False, label='Batter Handedness')
+
+    # preview = gr.DataFrame()
+    download_file = gr.DownloadButton(label='Download player data')
+
+    with gr.Group():
+      with gr.Row():
+        usage = gr.Plot(label='Pitch usage')
+        velo_summary = gr.Plot(label='Velocity summary', elem_classes='pitch-velo-summary')
+        loc_summary = gr.Plot(label='Overall location')
+
+    max_locs = len(jp_pitch_to_en_pitch)
+    locs_per_row = 4
+    max_rows = ceil(max_locs/locs_per_row)
+
+    gr.Markdown('''
+    ## Pitch Locations
+    Pitcher's persective
+    <br>
+    `NPB` refers to the top 10% of pitches thrown across the league with the current search constraints e.g. handedness
+    <br>
+    Note: To speed up the KDE, we restrict the league-wide pitches to 5,000 pitches
+    ''')
+    pitch_rows = []
+    pitch_groups = []
+    pitch_names = []
+    pitch_infos = []
+    pitch_velos = []
+    pitch_locs = []
+    for row in range(max_rows):
+      visible = row==0
+      pitch_row = gr.Row(visible=visible)
+      pitch_rows.append(pitch_row)
+      with pitch_row:
+        _locs_per_row = locs_per_row if row < max_rows-1 else max_locs - locs_per_row * (max_rows - 1)
+        for col in range(_locs_per_row):
+          with gr.Column(min_width=256):
+            pitch_group = gr.Group(visible=visible)
+            pitch_groups.append(pitch_group)
+            with pitch_group:
+              pitch_names.append(gr.Markdown(f'### Pitch {col+1}', visible=visible))
+              pitch_infos.append(gr.DataFrame(pl.DataFrame([{'Whiff%': None, 'CSW%': None}]), interactive=False, visible=visible))
+              pitch_velos.append(gr.Plot(show_label=False, elem_classes='pitch-velo', visible=visible))
+              pitch_locs.append(gr.Plot(label='Pitch Location', elem_classes='pitch-loc', visible=visible))
+
+    gr.Markdown('## Pitch Velocity')
+    velo_stats = gr.DataFrame(pl.DataFrame([{'Avg. Velo': None, 'League Avg. Velo': None}]), interactive=False, label='Pitch Velocity')
+
+    (
+        player
+        .input(update_dfs, inputs=[player, handedness, source_df], outputs=[app_df, app_league_df, app_pitch_stats, app_league_pitch_stats])
+        .then(lambda : gr.update(value='Both', interactive=True), outputs=handedness)
+    )
+    handedness.input(update_dfs, inputs=[player, handedness, source_df], outputs=[app_df, app_league_df, app_pitch_stats, app_league_pitch_stats])
+
+    # app_df.change(preview_df, inputs=app_df, outputs=preview)
+    # app_df.change(set_download_file, inputs=app_df, outputs=download_file)
+    # app_df.change(plot_usage, inputs=[app_df, player], outputs=usage)
+    # app_df.change(plot_velo_summary, inputs=[app_df, app_league_df, player], outputs=velo_summary)
+    # app_df.change(lambda df: plot_loc(df), inputs=app_df, outputs=loc_summary)
+    # app_df.change(plot_pitch_cards, inputs=[app_df, app_pitch_stats], outputs=pitch_rows+pitch_groups+pitch_names+pitch_infos+pitch_velos+pitch_locs)
+    app_pitch_stats.change(update_velo_stats, inputs=[app_pitch_stats, app_league_pitch_stats], outputs=velo_stats)
+
+    (
+        app_df
+        .change(create_set_download_file_fn('files/player.csv'), inputs=app_df, outputs=download_file)
+        .then(plot_usage, inputs=[app_df, player], outputs=usage)
+        .then(plot_velo_summary, inputs=[app_df, app_league_df, player], outputs=velo_summary)
+        .then(lambda df: plot_loc(df), inputs=app_df, outputs=loc_summary)
+        .then(plot_pitch_cards, inputs=[app_df, app_league_df, app_pitch_stats], outputs=pitch_rows+pitch_groups+pitch_names+pitch_infos+pitch_velos+pitch_locs)
+    )
+
+    gr.Markdown('## Bugs and other notes')
+    with gr.Accordion('Click to open', open=False):
+      gr.Markdown('''
+      - Y axis ticks messy when no velocity distribution is plotted
+      - DataFrame precision inconsistent
+      '''
+      )
+  return pitcher_dashboard_app
+
+if __name__ == '__main__':
+  create_pitcher_dashboard().launch(
+      share=True,
+      debug=True
+  )