import pandas as pd
from sklearn.linear_model import LinearRegression
from sklearn.model_selection import train_test_split

pd.set_option('display.max_columns', None)
all = pd.read_csv('winequalityN.csv', header=0)
all.isnull().any()
all = all.fillna(method='ffill')

print(all.shape[0])

mask = all['type'] == "white"
white, red = all[mask], all[~mask]
print(all)

print("\t\t\tОбщая выборка:")
df = all

df = df.drop(df.columns[0], axis='columns')
df = df.astype('float')
x_norm = (df - df.min()) / (df.max() - df.min())
#x_norm = df
print(x_norm)

for k in range(10):
    print("Разбивка №", k+1, "")
    x_norm = x_norm.sample(frac=1)
    x = x_norm[['fixed acidity', 'volatile acidity', 'citric acid', 'residual sugar', 'chlorides', 'free sulfur dioxide', 'total sulfur dioxide', 'density', 'pH', 'sulphates','alcohol']].values
    y = x_norm['quality'].values

    X_train, X_test, y_train, y_test = train_test_split(x, y, test_size=0.3, random_state=0)
    regressor = LinearRegression()
    regressor.fit(X_train, y_train)

    y_pred = regressor.predict(X_test)
    summ = 0

    for i in range(y_pred.shape[0]):
        print(y_test[i], " - ",  y_pred[i])
        summ = summ + abs(y_test[i] - y_pred[i])

    mistake = summ / y_pred.shape[0]
    good = 1 - mistake
    print("Процент верного предсказания = ",float(good), "%\n")

print("\t\t\tКрасное вино:")
df = red

df = df.drop(df.columns[0], axis='columns')
df = df.astype('float')
x_norm = (df - df.min()) / (df.max() - df.min())

for k in range(10):
    print("Разбивка №", k+1, "")
    x_norm = x_norm.sample(frac=1)
    x = x_norm[['fixed acidity', 'volatile acidity', 'citric acid', 'residual sugar', 'chlorides', 'free sulfur dioxide', 'total sulfur dioxide', 'density', 'pH', 'sulphates','alcohol']].values
    y = x_norm['quality'].values

    X_train, X_test, y_train, y_test = train_test_split(x, y, test_size=0.3, random_state=0)
    regressor = LinearRegression()
    regressor.fit(X_train, y_train)

    y_pred = regressor.predict(X_test)
    summ = 0
    for i in range(y_pred.shape[0]):
        summ = summ + abs(y_test[i] - y_pred[i])

    mistake = summ / y_pred.shape[0]
    good = 1 - mistake
    print("Процент верного предсказания = ",float(good), "%\n")

print("\t\t\tБелое вино:")
df = white

df = df.drop(df.columns[0], axis='columns')
df = df.astype('float')
x_norm = (df - df.min()) / (df.max() - df.min())

for k in range(10):
    print("Разбивка №", k+1, "")
    x_norm = x_norm.sample(frac=1)
    x = x_norm[['fixed acidity', 'volatile acidity', 'citric acid', 'residual sugar', 'chlorides', 'free sulfur dioxide', 'total sulfur dioxide', 'density', 'pH', 'sulphates','alcohol']].values
    y = x_norm['quality'].values

    X_train, X_test, y_train, y_test = train_test_split(x, y, test_size=0.3, random_state=0)
    regressor = LinearRegression()
    regressor.fit(X_train, y_train)

    y_pred = regressor.predict(X_test)
    summ = 0
    for i in range(y_pred.shape[0]):
        summ = summ + abs(y_test[i] - y_pred[i])

    mistake = summ / y_pred.shape[0]
    good = 1 - mistake
    print("Процент верного предсказания = ",float(good), "%\n")