import pandas as pd from datasets import Dataset def calculate_elo(old_rating, opponent_rating, score, k_factor=32): """ Calculate the new ELO rating for a player. :param old_rating: The current ELO rating of the player. :param opponent_rating: The ELO rating of the opponent. :param score: The score of the game (1 for win, 0.5 for draw, 0 for loss). :param k_factor: The K-factor used in ELO rating (default is 32). :return: The new ELO rating. """ expected_score = 1 / (1 + 10 ** ((opponent_rating - old_rating) / 400)) new_rating = old_rating + k_factor * (score - expected_score) return new_rating def update_elo_ratings(ratings_dataset, winner, loser, k_factor=32): """ Update ELO ratings for two players in a Hugging Face dataset. :param ratings_dataset: A Hugging Face dataset of current ELO ratings. :param winner: The name of the winning player. :param loser: The name of the losing player. :param k_factor: The K-factor used in ELO rating (default is 32). :return: Updated ELO ratings as a Hugging Face dataset. """ # Convert the Hugging Face dataset to a pandas DataFrame for easier manipulation ratings_df = pd.DataFrame(ratings_dataset) # Extract old ratings winner_old_rating = ratings_df.loc[ratings_df == winner, 'elo_rating'].iloc[0] loser_old_rating = ratings_df.loc[ratings_df == loser, 'elo_rating'].iloc[0] # Calculate new ratings winner_new_rating = calculate_elo(winner_old_rating, loser_old_rating, 1, k_factor) loser_new_rating = calculate_elo(loser_old_rating, winner_old_rating, 0, k_factor) # Update the DataFrame ratings_df.loc[ratings_df == winner, 'elo_rating'] = winner_new_rating ratings_df.loc[ratings_df == loser, 'elo_rating'] = loser_new_rating # Convert the DataFrame back to a Hugging Face dataset updated_ratings_dataset = Dataset.from_pandas(ratings_df) return updated_ratings_dataset