def train_logreg(X, y, random_state=42):
    model = LogisticRegression(solver='liblinear', random_state=random_state)
    model.fit(X, y)
    return model


def predict_logreg(inX, model):
    return int(model.predict([inX])[0])


def load_data(filepath):
    data = np.genfromtxt(filepath, delimiter=' ', missing_values='?', filling_values=0)
    labels = data[:, -1].astype(int)
    data = data[:, :-1]
    labels = np.where((labels == 2) | (labels == 3), 0, labels)
    if data.shape[1] >= 3:
        data = np.delete(data, 2, axis=1)
    scaler = StandardScaler()
    data = scaler.fit_transform(data)
    return data, labels

    # 去掉不相关特征
    data = np.delete(data, [2] + list(range(23, 28)), axis=1)

    # 数据标准化
    scaler = StandardScaler()
    data = scaler.fit_transform(data)
    return data, labels