import random
import numpy as np
import data as gl_data


def linear_function(coefficient, x):  # 构造一次函数y = a * X + b
    return coefficient[0] * x + coefficient[1]


def quadratic_function(coefficient, x):  # 构造二次函数y = a * X^2 + b * X + c
    return coefficient[0] * x ** 2 + coefficient[1] * x + coefficient[2]


def qubic_function(coefficient, x):  # 构造三次函数y = a* X^3 + b * X^2 + c * X + d
    return coefficient[0] * x ** 3 + coefficient[1] * x ** 2 + coefficient[2] * x + coefficient[3]


def quartic_function(coefficient, x):  # 构造四次函数
    return coefficient[0] * x ** 4 + coefficient[1] * x ** 3 + coefficient[2] * x ** 2 + coefficient[3] * x + \
           coefficient[4]


# 编程4.1 -----------------------------------------
def random_points(sampleNum, Low, High):
    sx, sy=[], []
    for i in range(sampleNum):
        sx.append(random.uniform(Low, High))    # 生成随机浮点数
        sy.append(random.uniform(Low, High))
    sx, sy = np.array(sx), np.array(sy)	        # 列表转数组
    gl_data.SampleData = np.array(list(zip(sx, sy)))    # 拼接成二维数组并存入全局变量
    return gl_data.SampleData
# if __name__ == '__main__':
#     sampleData = random_points(20, 0, 1000)
#     print(sampleData )
# 编程4.1 END-----------------------------------------


# 编程4.2 -----------------------------------------
def compute_curveData(Low, High, step, coefficient, m):  # coefficient代表二次函数系数
    x_values = np.arange(Low, High, step)
    if m == 1:
        y_values = linear_function(coefficient, x_values)
    if m == 2:
        y_values = quadratic_function(coefficient, x_values)		  # 调用quadratic_function(x)函数得到y值
    if m == 3:
        y_values = qubic_function(coefficient, x_values)
    if m == 4:
        y_values = quartic_function(coefficient, x_values)
    gl_data.CurveData = np.column_stack((x_values, y_values))  # 将x,y堆叠成二维数组
    return gl_data.CurveData
# if __name__ == '__main__':
#     coefficient = [1, 2, 3]
#     curveData = compute_curveData(gl_data.LOW, gl_data.HIGH, 1, coefficient,2)
#     print(curveData)


# 由于出现了数据稳定性的问题，因此在这里多了一个需要进行标准化处理的步骤
def compute_curveData2(low, high, step, theta, m, x_mean, x_std):
    x_fit = np.arange(low, high, step)
    # 将x_fit标准化
    x_fit_normalized = (x_fit - x_mean) / x_std
    A_fit = np.vander(x_fit_normalized, m + 1, increasing=True)
    y_fit = A_fit @ theta
    gl_data.CurveData = np.column_stack((x_fit, y_fit))
    return gl_data.CurveData


# 在X5中由于拓展了函数因此修改
def compute_curveData_X5(func, popt):
    x_values = np.arange(gl_data.LOW, gl_data.HIGH, 1)
    y_values = func(x_values, *popt)
    gl_data.CurveData = np.column_stack((x_values, y_values))
    return gl_data.CurveData

# 进行标准化的X5函数
def compute_curveData_X5_2(func, popt, sampleData):
    sx, sy = sampleData[:, 0], sampleData[:, 1]
    x_values = np.arange(gl_data.LOW, gl_data.HIGH, 1)
    x_values_normalized = (x_values - np.mean(sx)) / np.std(sx)
    y_values = func(x_values_normalized, *popt)
    gl_data.CurveData = np.column_stack((x_values, y_values))
    return gl_data.CurveData
# 编程4.2 END-----------------------------------------