将法线矩阵转换为numpy

Question

The purpose of this code is to arrange the lines, using the third column as a parameter. If I use normal matrix, the program works just fine, but I need to use numpy, because it's part of a bigger program.

The desired output is : [[2,-2,7],[-1,1,4],[10,7,1]]

import numpy as np

y = np.matrix([[-1,1,4],[2,-2,7],[10,7,1]])

c = True

def OrdenaMatriz(y):
    matriz = []
    matriz.append(y[0])
    for a in range(2):
        if y[a,2] < y[a+1,2]:
            matriz.insert(a,y[a+1])
        else:
            matriz.append(y[a+1])
    return matriz

while c == True:
    a = OrdenaMatriz(y)
    if a == y:
        c = False
        print(a)
    y = a

The following error is showing:

DeprecationWarning: elementwise == comparison failed; this will raise an 
error in the future.
  if a == y:
Traceback (most recent call last):
  File "teste.py", line 26, in <module>
    a = OrdenaMatriz(y)
  File "teste.py", line 19, in OrdenaMatriz
    if y[a,2] < y[a+1,2]:
TypeError: list indices must be integers or slices, not tuple

Answer 1

I'll try to explain your errors and warning.

y = np.matrix([[-1,1,4],[2,-2,7],[10,7,1]])

could just as well be

y = np.array([[-1,1,4],[2,-2,7],[10,7,1]])

You don't, especially as a beginner, need to use np.matrix . np.array() produces the regular numpy array object. Using np.matrix is discouraged, since it doesn't add anything special now.

a = OrdenaMatriz(y) produces a Python list. You start with [] , and either insert or append values, so the result is still a list.

It's the

a == y

that produces the DeprecationWarning . It's the result of comparing a list with a numpy array.

Then you y=a . Now y is a list, not the original array (or matrix). So on the next loop, OrdenaMatriz is called with a list. That's when

y[a,2] < y[a+1,2]

raise the TypeError. That indexing is ok for array/matrix, but not for list.

So if you stick with this code, or something similar, start with an np.array() call, and make sure that OrdenaMatriz returns an array, not a list.

Answer 2

"The purpose of this code is to arrange the lines, using the third column as a parameter.":

>>> y = y[np.argsort(y[:,-1].T),:]
>>> y 
matrix([[[10,  7,  1],
     [-1,  1,  4],
     [ 2, -2,  7]]])

Like this?