encapsulated the preprocessing into usefull functions

preprocess.py

@@ -0,0 +1,90 @@
+import pandas as pd
+import numpy as np
+import matplotlib.pyplot as plt
+import seaborn as sns
+import warnings
+warnings.filterwarnings('ignore')
+
+
+def extract_games(df):
+    all_games = []
+    for games in df:
+        all_games.append(games)
+    return all_games
+
+def extract_moves(games):
+    all_white_moves = []
+    all_black_moves = []
+    for game in games:
+        game = game.split(",")
+        for move in game:
+            move = move.split(" ")
+            all_black_moves.append(move[1::2])
+            all_white_moves.append(move[0::2])
+    return all_white_moves, all_black_moves
+
+def number_of_takes(player_games):
+    all_takes = []
+    for game in player_games:
+        takes = 0  
+        for moves in game:
+            takes = takes + moves.count("x")
+        all_takes.append(takes)
+    return all_takes
+
+def mate_games(chess_data):
+    chess_data = chess_data[chess_data.rated != False]
+    chess_data = chess_data[chess_data.victory_status == "mate"]
+    chess_data = chess_data[chess_data.turns > 4]
+    chess_data.drop_duplicates(subset=['id'], inplace=True)
+    chess_data[['whites_opening','blacks_opening']] = chess_data.opening_name.str.split(": ", 1, expand=True)
+    chess_data[['time_limit','increment']] = chess_data.increment_code.str.split("+", 1, expand=True).astype('int')
+    chess_data.drop(['id', 'rated', 'white_id', 'black_id','opening_name', 'increment_code', 'victory_status'],axis=1,inplace=True)
+    from sklearn.preprocessing import LabelEncoder
+    le = LabelEncoder()
+    for column_name in ['winner','whites_opening','blacks_opening', 'opening_eco']:
+        chess_data[column_name] = le.fit_transform(chess_data[column_name])
+    games_df = extract_games(chess_data.moves)
+    white_moves, black_moves = extract_moves(games_df)
+    white_took = number_of_takes(white_moves)
+    wtdf = pd.DataFrame(data = white_took, columns=["white_took"])
+    black_took = number_of_takes(black_moves)
+    btdf = pd.DataFrame(data = black_took, columns=["black_took"])
+    chess_data = pd.concat([chess_data,wtdf],axis=1)
+    chess_data = pd.concat([chess_data,btdf],axis=1)
+    chess_data.drop(['moves', 'last_move_at', 'created_at'], axis=1, inplace=True)
+    chess_data = chess_data[chess_data.turns < 200]
+    chess_data.reset_index(inplace = True)
+    chess_data.drop(['index'],axis=1,inplace=True)
+    x = chess_data.iloc[:, chess_data.columns != 'winner']
+    y = chess_data.iloc[:, 2]
+    return x, y
+
+def draw_and_mate_games(chess_data):
+    chess_data = chess_data[chess_data.rated != False]
+    chess_data = chess_data[chess_data.victory_status != "outoftime"]
+    chess_data = chess_data[chess_data.victory_status != "resign"]
+    chess_data = chess_data[chess_data.turns > 4]
+    chess_data.drop_duplicates(subset=['id'], inplace=True)
+    chess_data[['whites_opening','blacks_opening']] = chess_data.opening_name.str.split(": ", 1, expand=True)
+    chess_data[['time_limit','increment']] = chess_data.increment_code.str.split("+", 1, expand=True).astype('int')
+    chess_data.drop(['id', 'rated', 'white_id', 'black_id','opening_name', 'increment_code'],axis=1,inplace=True)
+    from sklearn.preprocessing import LabelEncoder
+    le = LabelEncoder()
+    for column_name in ['winner','whites_opening','blacks_opening', 'opening_eco', 'victory_status']:
+        chess_data[column_name] = le.fit_transform(chess_data[column_name])
+    games_df = extract_games(chess_data.moves)
+    white_moves, black_moves = extract_moves(games_df)
+    white_took = number_of_takes(white_moves)
+    wtdf = pd.DataFrame(data = white_took, columns=["white_took"])
+    black_took = number_of_takes(black_moves)
+    btdf = pd.DataFrame(data = black_took, columns=["black_took"])
+    chess_data = pd.concat([chess_data,wtdf],axis=1)
+    chess_data = pd.concat([chess_data,btdf],axis=1)
+    chess_data.drop(['moves', 'last_move_at', 'created_at'], axis=1, inplace=True)
+    chess_data = chess_data[chess_data.turns < 200]
+    chess_data.reset_index(inplace = True)
+    chess_data.drop(['index'],axis=1,inplace=True)
+    x = chess_data.iloc[:, chess_data.columns != 'winner']
+    y = chess_data.iloc[:, 2]
+    return x, y