production of a matrix and a tensor

now I have below two objects:

A = np.array(('a', 'b', 'c', 'd'), dtype=object)
x = np.array(range(1,9)).reshape(2,2,2).transpose(0,2,1)

With this setting, I want to have a 3D matrix as follows:

array([[['abb' 'aaabbbb'],
        ['cdd' 'cccdddd']],

       [['aaaaabbbbbb' 'aaaaaaabbbbbbbb'],
        ['cccccdddddd' 'cccccccdddddddd']]]

I have tried np.tensordot(A, x, 1) with some options, but I didn't have a desired result. Hope someone helps me.

Solution 1:

You can try:

A.reshape(2, 2)@x

It gives:

array([[['abb', 'aaabbbb'],
        ['cdd', 'cccdddd']],

       [['aaaaabbbbbb', 'aaaaaaabbbbbbbb'],
        ['cccccdddddd', 'cccccccdddddddd']]], dtype=object)