Upload 2 files

pitch_summary_functions.py

@@ -228,43 +228,59 @@ def percentile(n):
     return percentile_
 
 ### TJ STUFF+ DF CLEAN ###
+
 def df_clean(df):
     df_copy = df.copy()
-
-    df_copy = df_copy[(df_copy['spin_rate']>0)&(df_copy['extension']>0)]
-    
     df_copy.loc[df_copy['pitcher_hand'] == 'L','hb'] *= -1
     df_copy.loc[df_copy['pitcher_hand'] == 'L','x0'] *= -1
     df_copy.loc[df_copy['pitcher_hand'] == 'L','spin_direction'] = 360 - df_copy.loc[df_copy['pitcher_hand'] == 'L','spin_direction']
 
     df_copy['pitch_l'] = [1 if x == 'L' else 0 for x in df_copy['pitcher_hand']]
     df_copy['bat_l'] = [1 if x == 'L' else 0 for x in df_copy['batter_hand']]
-    #df_copy = df_copy[~df_copy.pitch_type.isin(["EP", "PO", "KN", "CS", "SC", "FA"])].reset_index(drop=True)
-    #df_copy = df_copy[~df_copy.pitch_type.isin(["EP", "PO", "CS", "SC", "FA"])].reset_index(drop=True)
-
-
-    df_copy['pitch_type'] = df_copy['pitch_type'].replace({'FT':'SI',
-                                                           'FA':'FF',
-                                                           'SV':'SL',
-                                                           'FO':'FS'})
-
-    df_copy_fb_sum = df_copy[df_copy.pitch_type.isin(["FF", "FC", "SI"])].groupby(['pitcher_id']).agg(
-        fb_velo = ('start_speed','mean'),
-        fb_max_ivb  = ('ivb',percentile(0.9)),
-        fb_max_x  = ('hb',percentile(0.9)),
-        fb_min_x  = ('hb',percentile(0.1)),
-        fb_max_velo  = ('start_speed',percentile(0.9)),
-        fb_axis  = ('spin_direction','mean'),
-    )
-
-    df_copy = df_copy.merge(df_copy_fb_sum,left_on='pitcher_id',right_index=True,how='left')
-
-    df_copy['fb_velo_diff'] = df_copy['start_speed']- df_copy['fb_velo']
-    df_copy['fb_max_ivb_diff'] = df_copy['ivb']- df_copy['fb_max_ivb']
-    df_copy['fb_max_hb_diff'] = -abs(df_copy['hb']- df_copy['fb_max_x'])
-    df_copy['fb_min_hb_diff'] = df_copy['hb']- df_copy['fb_min_x']
-    df_copy['fb_max_velo_diff'] = df_copy['start_speed']- df_copy['fb_max_velo']
-    df_copy['fb_axis_diff'] = df_copy['spin_direction']- df_copy['fb_axis']
+    df_copy = df_copy[~df_copy.pitch_type.isin(["EP", "PO", "KN", "CS", "SC", "FA"])].reset_index(drop=True)
+    df_copy['pitch_type'] = df_copy['pitch_type'].replace({'FT':'SI','KC':'CU','SV':'SL','FO':'FS'})
+
+    # df_copy['des_new'] = df_copy['play_description'].map(des_dict)
+    # df_copy['ev_new'] = df_copy.loc[df_copy['des_new'] == 'hit_into_play','event_type'].map(ev_dict)
+    # df_copy.loc[df_copy['des_new']=='hit_into_play','des_new'] = df_copy.loc[df_copy['des_new']=='hit_into_play','ev_new']
+    # df_copy = df_copy.dropna(subset=['des_new'])
+    # des_values = df_copy.groupby(['des_new'])['delta_run_exp'].mean()
+    # df_copy = df_copy.merge(des_values,left_on='des_new',right_on='des_new',suffixes=['','_mean'])
+
+
+
+
+    # df_copy_fb_sum = df_copy[df_copy.pitch_type.isin(["FF", "FC", "SI"])].groupby(['pitcher_id']).agg(
+    #     fb_velo = ('start_speed','mean'),
+    #     fb_max_ivb  = ('ivb',percentile(0.9)),
+    #     fb_max_x  = ('hb',percentile(0.9)),
+    #     fb_min_x  = ('hb',percentile(0.1)),
+    #     fb_max_velo  = ('start_speed',percentile(0.9)),
+    #     fb_axis  = ('spin_direction','mean'),
+    # )
+    df_grouping = df_copy.groupby(['pitcher_id'])[['pitch_type']].value_counts().sort_values(ascending=False).reset_index()
+    df_grouping_dupl = df_grouping.drop_duplicates(subset=['pitcher_id'])#.set_index(['pitcher_id','pitch_type'])
+
+    df_copy_fb_sum = df_copy.groupby(['pitcher_id','pitch_type']).agg(
+        primary_mean_velo = ('start_speed','mean'),
+        primary_mean_ivb  = ('ivb','mean'),
+        primary_mean_x  = ('hb','mean'),
+        # fb_max_velo  = ('start_speed','mean'),
+        # fb_axis  = ('spin_direction','mean'),
+    ).reset_index()
+
+    df_copy_fb_sum = df_grouping_dupl.merge(df_copy_fb_sum,left_on=['pitcher_id','pitch_type'],
+                                        right_on=['pitcher_id','pitch_type']).reset_index()
+    
+    df_copy = df_copy.merge(df_copy_fb_sum,left_on='pitcher_id',right_on='pitcher_id',how='right',suffixes=['','_primary'])
+    # return df_copy_fb_sum
+    
+    df_copy['primary_mean_velo_diff'] = df_copy['start_speed']- df_copy['primary_mean_velo']
+    df_copy['primary_mean_ivb_diff'] = df_copy['ivb']- df_copy['primary_mean_ivb']
+    df_copy['primary_mean_hb_diff'] = abs(df_copy['hb']- df_copy['primary_mean_x'])
+    #df_copy['primary_min_hb_diff'] = df_copy['hb']- df_copy['fb_min_x']
+    #df_copy['primary_max_velo_diff'] = df_copy['start_speed']- df_copy['fb_max_velo']
+    #df_copy['primary_axis_diff'] = df_copy['spin_direction']- df_copy['fb_axis']
 
     # df_copy.loc[df_copy.pitch_type.isin(["FF", "FC", "SI"]),'fb_velo_diff'] = 0
     # df_copy.loc[df_copy.pitch_type.isin(["FF", "FC", "SI"]),'fb_max_ivb_diff'] = 0
@@ -295,10 +311,11 @@ def df_clean(df):
 
     # df_copy['vaa'] = np.arctan(df_copy['z_diff'] / df_copy['release_pos_y']) * 360 / np.pi
     # df_copy['haa'] = np.arctan(-df_copy['x_diff'] / df_copy['release_pos_y']) * 360 / np.pi
-
-    df_copy = df_copy.dropna(subset=['pitch_type'])#.fillna(0)
+    df_copy['log_extension'] = np.log(df_copy['extension'])
+    df_copy = df_copy.dropna(subset=['pitch_type']).fillna(0)
     return df_copy
 
+
 ### PITCH COLOURS ###
 pitch_colours = {
         'Four-Seam Fastball':'#FF007D',#BC136F
@@ -641,6 +658,36 @@ def tj_stuff_roling(df,
         # ax.set_xlim(left=1)
         ax.xaxis.set_major_locator(MaxNLocator(integer=True))
 
+### TJ STUFF+ ROLLING ###
+def tj_velocity_roling(df,
+                    window,
+                    ax):
+        ## Velocity Plot
+        sorted_value_counts = df['pitch_type'].value_counts().sort_values(ascending=False)
+
+        # Get the list of items ordered from most to least frequent
+        items_in_order = sorted_value_counts.index.tolist()
+        
+
+        for i in items_in_order:
+            if max(df[df['pitch_type']==i]['pitch_type_count_each']) >= window:
+                sns.lineplot(x=range(1,max(df[df['pitch_type']==i]['pitch_type_count_each'])+1),
+                        y=df[df['pitch_type']==i]['start_speed'].rolling(window).sum()/window,
+                    color=pitch_colours[df[df['pitch_type']==i]['pitch_description'].values[0]],
+                    ax=ax,linewidth=3)
+
+        # Adjust x-axis limits to start from 1
+        ax.set_xlim(window,max(df['pitch_type_count_each']))
+        ax.set_ylim(65,105)
+        #ax.get_legend().remove()
+        ax.set_xlabel('Pitches', fontdict=font_properties_axes)
+        ax.set_ylabel('Velocity (mph)', fontdict=font_properties_axes)
+        ax.set_title(f"{window} Pitch Rolling Velocity",fontdict=font_properties_titles)
+        # ax.axis('square')
+        # ax.set_xlim(left=1)
+        ax.xaxis.set_major_locator(MaxNLocator(integer=True))
+
+                        
 ### BREAK PLOT ###
 def break_plot(df,
                ax): 
@@ -1044,4 +1091,30 @@ def location_plot(df,ax,hand):
         if len(pitch_location_group['px'])>0:
             ax.get_legend().remove()
         ax.grid(False)
-        ax.set_title(f"Pitch Locations vs {hand}HB\n{pitch_location_group['pitches'].sum()} Pitches",fontdict=font_properties_titles)
+        ax.set_title(f"Pitch Locations vs {hand}HH\n{pitch_location_group['pitches'].sum()} Pitches",fontdict=font_properties_titles)
+
+
+
+
+####LHH
+def location_pitch_plot(df,ax,hand):
+        #label_labels = df.sort_values(by=['prop','pitch_type'],ascending=[False,True]).pitch_description.unique()
+        ## Location Plot
+        df_pitch= df[df['batter_hand']==hand]
+        sns.scatterplot(ax=ax,x=df_pitch['px'],
+                                    y=df_pitch['pz'],
+                                    hue=df_pitch['pitch_description'],
+                                    palette=pitch_colours,ec='black',
+                                    s=175,
+                                    linewidth=2,
+                                    zorder=2)
+
+        ax.axis('square')
+        draw_line(ax,alpha_spot=0.75,catcher_p=False)
+        ax.axis('off')
+        ax.set_xlim((-2.75,2.75))
+        ax.set_ylim((-0.5,5))
+        if len(df_pitch['px'])>0:
+            ax.get_legend().remove()
+        ax.grid(False)
+        ax.set_title(f"Pitch Locations vs {hand}HH\n{df_pitch['pitches'].sum()} Pitches",fontdict=font_properties_titles)

stuff_values_20240428.py

@@ -0,0 +1,4 @@
+# dictionary.py
+
+my_dict = {"y_pred_mean":-0.0022258872631937265,
+	   "y_pred_std": 0.007943911477923393}