Refactor player selection logic in app.py

- Changed the method of selecting top players from using `nlargest` to `head`, improving performance and clarity in the code.
- This adjustment ensures that the top players are selected based on their ranking in the dataframe, enhancing data processing efficiency.

app.py

@@ -228,10 +228,10 @@ with tab2:
 
         for col in player_columns:
             contest_players = working_df.copy()
-            players_1per = working_df.nlargest(n=int(len(working_df) * 0.01), columns='actual_fpts')
-            players_5per = working_df.nlargest(n=int(len(working_df) * 0.05), columns='actual_fpts')
-            players_10per = working_df.nlargest(n=int(len(working_df) * 0.10), columns='actual_fpts')
-            players_20per = working_df.nlargest(n=int(len(working_df) * 0.20), columns='actual_fpts')
+            players_1per = working_df.head(int(len(working_df) * 0.01))
+            players_5per = working_df.head(int(len(working_df) * 0.05))
+            players_10per = working_df.head(int(len(working_df) * 0.10))
+            players_20per = working_df.head(int(len(working_df) * 0.20))
         player_counts = pd.Series(list(contest_players[player_columns].values.flatten())).value_counts()
         player_1per_counts = pd.Series(list(players_1per[player_columns].values.flatten())).value_counts()
         player_5per_counts = pd.Series(list(players_5per[player_columns].values.flatten())).value_counts()