Numpy Array creation problems in class method

Question

The following class method has been giving me trouble:

def align(self,array,a):
   aligned=np.zeros(array.shape)
   b=[0,0,1]
   v=np.cross(a,b)
   c=np.dot(a,b)
   I=np.eye(3,3)
   vx=np.zeros((3,3))
   vx=[[0,-v[2],v[1]],[v[2],0,-v[0]],[-v[1],v[0],0]]
   rotmatrix=I + vx + (vx @ vx)/(1+c)
   aligned = rotmatrix @ array.T
   return aligned

The error message I get is:

TypeError: unsupported operand type(s) for +: 'float' and 'list'

I am certain that the operation that is breaking is my matrix multiplication (vx @ vx) after investigating each component of the rotation matrix.

To generate the vx matrix, I have tried a variety of different things in numpy such as np.asarray, np.vstack, np.array with and without initialized an array of zeros. When printing out vx, depending on what I have tried it is either an array of lists, or a regular python list of lists. It never has the correct numpy array formatting. The only way I have found to get the array right is by assigning each index of the array to the correct value, which works but I have a hard time believing that is the only way to do this.

You assign vx twice, once as a (3,3) array, and again as a list. You try the same with aligned. In Python we don't 'declare' a variable type. We make an object and assign it to a variable. If we assign another object to a variable, the initial assignment is lost. — hpaulj
– hpaulj, Commented Feb 10, 2020 at 18:00
Yeah I know that initializing a variable is unneccessary in python, however I have had issues in the past that were resolved by initializing arrays/variables so that's why I tried that — msincla01
– msincla01, Commented Feb 10, 2020 at 18:40
You only need to initialize arrays if you are going to modify them. Initializing an array is quite different from initializing a variable. In any case, the issue here is that the 2nd vx assignment is a list, not an array. — hpaulj
– hpaulj, Commented Feb 10, 2020 at 18:44
Ah I see, that did it! Thank you, I was unaware it could create this problem — msincla01
– msincla01, Commented Feb 10, 2020 at 18:53

mathfux · Accepted Answer · 2020-02-10 17:58:05Z

1

How about a classical way vx = np.array([[0,-v[2],v[1]],[v[2],0,-v[0]],[-v[1],v[0],0]])?

answered Feb 10, 2020 at 17:58

mathfux

5,9792 gold badges20 silver badges38 bronze badges

Sign up to request clarification or add additional context in comments.

1 Comment

msincla01 Over a year ago

I forgot to mention that was the first thing I tried. I am baffled as to why that returns an array of lists

Collectives™ on Stack Overflow

Numpy Array creation problems in class method

1 Answer 1

1 Comment

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

1 Comment

Your Answer

Sign up or log in

Post as a guest

Related