Update elo.py

elo.py

@@ -1,23 +1,11 @@
-import pandas as pd
-
 import math
 
 def update_elo_ratings(ratings_dict, winner, loser):
-    # Check if the ratings_dict is empty
-    if not ratings_dict:
-        # Create a blank DataFrame with the required columns
-        ratings_df = pd.DataFrame(columns=['bot_name', 'elo_rating', 'games_played'])
-    else:
-        # Convert the dictionary to a pandas DataFrame
-        ratings_df = pd.DataFrame.from_dict(ratings_dict, orient='index')
-        ratings_df.reset_index(inplace=True)
-        ratings_df.columns = ['bot_name', 'elo_rating', 'games_played']
-
-    # Check and add new players if they don't exist in the dataset
-    for player in [winner, loser]:
-        if player not in ratings_df['bot_name'].values:
-            new_player = {'bot_name': player, 'elo_rating': 1200, 'games_played': 0}
-            ratings_df = pd.concat([ratings_df, pd.DataFrame([new_player])], ignore_index=True)
+    # Extract old ratings and games played
+    winner_old_rating = ratings_dict.get(winner, {}).get('elo_rating', 1200)
+    loser_old_rating = ratings_dict.get(loser, {}).get('elo_rating', 1200)
+    winner_games_played = ratings_dict.get(winner, {}).get('games_played', 0)
+    loser_games_played = ratings_dict.get(loser, {}).get('games_played', 0)
 
     # Function to determine the K-factor based on games played
     def determine_k_factor(games_played):
@@ -28,6 +16,11 @@ def update_elo_ratings(ratings_dict, winner, loser):
             return 20
         else:
             return 10
+
+    # Determine K-factors
+    winner_k_factor = determine_k_factor(winner_games_played)
+    loser_k_factor = determine_k_factor(loser_games_played)
+
     def elo(winner_rating, loser_rating, k_factor_winner=32, k_factor_loser=32):
         # Calculate the expected scores for each player
         winner_expected = 1 / (1 + 10 ** ((loser_rating - winner_rating) / 400))
@@ -39,34 +32,11 @@ def update_elo_ratings(ratings_dict, winner, loser):
 
         return winner_new_rating, loser_new_rating
 
-    # ...
-
     # Calculate new ratings
     winner_new_rating, loser_new_rating = elo(winner_old_rating, loser_old_rating, k_factor_winner=winner_k_factor, k_factor_loser=loser_k_factor)
 
+    # Update ratings and games played in the dictionary
+    ratings_dict[winner] = {'elo_rating': winner_new_rating, 'games_played': winner_games_played + 1}
+    ratings_dict[loser] = {'elo_rating': loser_new_rating, 'games_played': loser_games_played + 1}
 
-    # Update games played
-    ratings_df.loc[ratings_df['bot_name'] == winner, 'games_played'] += 1
-    ratings_df.loc[ratings_df['bot_name'] == loser, 'games_played'] += 1
-
-    # Extract old ratings and games played
-    winner_old_rating = ratings_df.loc[ratings_df['bot_name'] == winner, 'elo_rating'].iloc[0]
-    loser_old_rating = ratings_df.loc[ratings_df['bot_name'] == loser, 'elo_rating'].iloc[0]
-    winner_games_played = ratings_df.loc[ratings_df['bot_name'] == winner, 'games_played'].iloc[0]
-    loser_games_played = ratings_df.loc[ratings_df['bot_name'] == loser, 'games_played'].iloc[0]
-
-    # Determine K-factors
-    winner_k_factor = determine_k_factor(winner_games_played)
-    loser_k_factor = determine_k_factor(loser_games_played)
-
-    # Calculate new ratings
-    winner_new_rating, loser_new_rating = elo(winner_old_rating, loser_old_rating, k_factor_winner=winner_k_factor, k_factor_loser=loser_k_factor)
-
-    # Update the DataFrame
-    ratings_df.loc[ratings_df['bot_name'] == winner, 'elo_rating'] = winner_new_rating
-    ratings_df.loc[ratings_df['bot_name'] == loser, 'elo_rating'] = loser_new_rating
-
-    # Convert the DataFrame to a dictionary
-    updated_ratings_dict = ratings_df.set_index('bot_name').to_dict(orient='index')
-
-    return updated_ratings_dict
+    return ratings_dict