如何有效地为数组调用函数，而不是遍历数组的值？

Question

我有一个定义的函数，可以从自变量x内插因变量y（x）。 （已知的值在下面的函数中显示。）我希望对整个值[512x512]（甚至[512x512x133]）数组执行该操作，而不是对单个值x0执行它。是否有比通过嵌套的for loop2 in range(0, len(target(loop1,loop2)))循环遍历每一行更好的方法？

我将添加嵌套的for循环来重写此函数，但是首先，我将阅读更多Python教程。当前在section 4.6, "Defining Functions"上。

def interpSPR(target):
"""Interpolate SPR given CT value according to Kanematsu et al. 2003

We use the lookup table provided by Dr. Matsumura to interpolate SPR values
from each pixel's CT Number."""
## Input CT Number to Stopping Power Ratio (SPR) Lookup Table from XiO-N
CTvalues = [-992.516,-399.723,-152.726,  94.27054,291.868,1971.447,11081,24414]  
SPR =      [       0,   0.602,   0.913822,1.078872, 1.151456,1.900912, 6,   12]
# Check first to see if the value is one of those already in the table.
if target in CTvalues:
    for num in range(0,len(CTvalues)):
        if target == CTvalues[num]:
            return SPR[num]
# If it isn't, interpolate the value.        
else:
    for num in range(0,len(CTValues)):
        if target < CTvalues[0]:
            y1,y2,x0,x1,x2 = SPR[0],SPR[1],target,CTvalues[0],CTvalues[1]
            answer = y1 - ((x1-x0)/(x2 - x1))*(y2 - y1)
            return answer
        elif target > CTvalues[num] and target < CTValues[num+1]:
            y1,y2,x0,x1,x2 = SPR[num],SPR[num+1],target,CTvalues[num],CTvalues[num+1]
            answer = y1 - ((x1-x0)/(x2 - x1))*(y2 - y1)
            return answer
        elif target > CTvalues[-1]:
            y1,y2,x0,x1,x2 = SPR[-2],SPR[-1],target,CTvalues[-2],CTvalues[-1]
            answer = y1 - ((x1-x0)/(x2 - x1))*(y2 - y1)
            return answer

Answer 1

numpy为此np.interp()提供了内置功能：

import numpy as np   

def interpSPR(target):
    CTvalues = [-992.516, -399.723,   -152.726, 94.27054,  291.868, 1971.447, 11081, 24414]
    SPR =      [       0,    0.602,   0.913822, 1.078872, 1.151456, 1.900912,     6,   12]
    return np.interp(target, CTvalues, SPR)

如果我们稍微改变数据，则在图中更容易看到：

CTvalues = [0, 5, 15, 20, 40]
SPR =      [2, 7,  20, 21, 50]

def interpSPR(target):
    CTvalues = [0, 5, 15, 20, 40]
    SPR =      [2, 7, 20, 21, 50]
    return np.interp(target, CTvalues, SPR)

fig, ax = plt.subplots()
ax.scatter(CTvalues, SPR)
target = np.linspace(-10, 50, 100)
ax.plot(target, interpSPR(target))