lambda function for slicing a numpy array to specific index and reshape it?-CodePudding

I want to slice a numpy array with its index and reshape that with one-liner lambda function. Let me explain with minimal producible example:

This is my numpy array.

In [1]: composite                                                                                                                                                                                                                           
Out[1]: 
array([ 1,  1, -1,  1,  1,  1,  1,  1, -1, -1,  1, -1, -1,  1,  1, -1,  3,
       -1,  3,  1,  3,  1,  1,  3,  1, -1,  1, -1, -1,  3,  1,  1,  1, -1,
        1,  1, -3, -3, -3, -1, -1, -3, -1, -1,  1, -3, -1, -1, -1, -1, -3,
       -1, -3,  1, -1, -3, -1, -3, -1, -1, -1, -3, -3, -3, -1, -3, -3,  1,
        1,  1, -1, -1,  1, -1, -1, -3,  1, -1, -1, -1, -1,  1, -1,  1, -3,
       -1,  1, -1,  1, -1, -1, -1,  1,  1,  1,  3,  1,  1,  1,  1,  1,  3,
        3,  1,  3,  3,  1,  1,  3, -1,  3, -1,  1, -1,  1,  1, -1,  1,  3,
        1,  3,  3, -1,  1, -1, -1,  1, -1, -1, -1, -1, -1,  1,  1, -1,  1,
        1, -1, -1,  1, -3,  1, -3, -1, -3, -1, -1, -3, -1,  1, -1,  1,  1,
       -3, -1, -3, -3, -1, -3, -3,  1,  1,  1, -1, -1,  1, -1, -1, -3,  1,
       -1, -1, -1, -1,  1, -1,  1, -3, -1,  1, -1,  1, -1, -1, -1,  1, -1,
       -1,  1, -1, -1, -1, -1, -1,  1,  1, -1,  1,  1, -1, -1,  1, -3,  1,
       -3, -1, -3, -1, -1, -3, -1,  1, -1,  1,  1, -3, -1,  1,  1, -1,  1,
        1,  1,  1,  1, -1, -1,  1, -1, -1,  1,  1, -1,  3, -1,  3,  1,  3,
        1,  1,  3,  1, -1,  1, -1, -1,  3,  1])

I want to slice every 32 element in the array and make them as one element in array, something like this:

# I know i used composite[:32] and get below
array([ 1,  1, -1,  1,  1,  1,  1,  1, -1, -1,  1, -1, -1,  1,  1, -1,  3,
       -1,  3,  1,  3,  1,  1,  3,  1, -1,  1, -1, -1,  3,  1,  1])

This is just an example but the actual length of the array is unknown in my case so i don't know how big my array is going to be. But i want something like this:

# I know i used composite[:32] and get below
array([ 1,  1, -1,  1,  1,  1,  1,  1, -1, -1,  1, -1, -1,  1,  1, -1,  3,
       -1,  3,  1,  3,  1,  1,  3,  1, -1,  1, -1, -1,  3,  1,  1],
      [-1,  1,  1, -3, -3, -3, -1, -1, -3, -1, -1,  1, -3, -1, -1, 
      -1, -1,-3, -1, -3,  1, -1, -3, -1, -3, -1, -1, -1, -3, -3, -3], ... and so
      on...till the end of the array))

Is there an intelligent way of doing this ? Perhaps a lambda function ?

presently i am manually doing it using below function.

def de_repeater(seq, size):
    chunks = []
    index = 0
    while index < len(seq):
        #print(list(seq[index:index size]))
        #print(len(list(seq[index:index size])))
        chunks.append(list(seq[index:index size]))
        index =size

    bits = []
    for List in chunks:
        #print(max(set(List), key = List.count)) 
        bits.append(max(set(List), key = List.count))

    return bits
a1 = de_repeater(composite, 8)
len(a1) #31

CodePudding user response：

You can use np.reshape() if you know that the length of the array is a multiple of 32 (as described in a comment):

np.reshape(composite, (-1, 32))

If you pad the original composite array with eight ones, you'll get:

[[ 1  1 -1  1  1  1  1  1 -1 -1  1 -1 -1  1  1 -1  3 -1  3  1  3  1  1  3
   1 -1  1 -1 -1  3  1  1]
 [ 1 -1  1  1 -3 -3 -3 -1 -1 -3 -1 -1  1 -3 -1 -1 -1 -1 -3 -1 -3  1 -1 -3
  -1 -3 -1 -1 -1 -3 -3 -3]
 [-1 -3 -3  1  1  1 -1 -1  1 -1 -1 -3  1 -1 -1 -1 -1  1 -1  1 -3 -1  1 -1
   1 -1 -1 -1  1  1  1  3]
 [ 1  1  1  1  1  3  3  1  3  3  1  1  3 -1  3 -1  1 -1  1  1 -1  1  3  1
   3  3 -1  1 -1 -1  1 -1]
 [-1 -1 -1 -1  1  1 -1  1  1 -1 -1  1 -3  1 -3 -1 -3 -1 -1 -3 -1  1 -1  1
   1 -3 -1 -3 -3 -1 -3 -3]
 [ 1  1  1 -1 -1  1 -1 -1 -3  1 -1 -1 -1 -1  1 -1  1 -3 -1  1 -1  1 -1 -1
  -1  1 -1 -1  1 -1 -1 -1]
 [-1 -1  1  1 -1  1  1 -1 -1  1 -3  1 -3 -1 -3 -1 -1 -3 -1  1 -1  1  1 -3
  -1  1  1 -1  1  1  1  1]
 [ 1 -1 -1  1 -1 -1  1  1 -1  3 -1  3  1  3  1  1  3  1 -1  1 -1 -1  3  1
   1  1  1  1  1  1  1  1]]

CodePudding user response：

Try using .reshape():

size = round(len(composite)/32) # 8
composite.reshape((size,-1))

Output:

array([[ 1,  1, -1,  1,  1,  1,  1,  1, -1, -1,  1, -1, -1,  1,  1, -1,
         3, -1,  3,  1,  3,  1,  1,  3,  1, -1,  1, -1, -1,  3,  1],
       [ 1,  1, -1,  1,  1, -3, -3, -3, -1, -1, -3, -1, -1,  1, -3, -1,
        -1, -1, -1, -3, -1, -3,  1, -1, -3, -1, -3, -1, -1, -1, -3],
       [-3, -3, -1, -3, -3,  1,  1,  1, -1, -1,  1, -1, -1, -3,  1, -1,
        -1, -1, -1,  1, -1,  1, -3, -1,  1, -1,  1, -1, -1, -1,  1],
       [ 1,  1,  3,  1,  1,  1,  1,  1,  3,  3,  1,  3,  3,  1,  1,  3,
        -1,  3, -1,  1, -1,  1,  1, -1,  1,  3,  1,  3,  3, -1,  1],
       [-1, -1,  1, -1, -1, -1, -1, -1,  1,  1, -1,  1,  1, -1, -1,  1,
        -3,  1, -3, -1, -3, -1, -1, -3, -1,  1, -1,  1,  1, -3, -1],
       [-3, -3, -1, -3, -3,  1,  1,  1, -1, -1,  1, -1, -1, -3,  1, -1,
        -1, -1, -1,  1, -1,  1, -3, -1,  1, -1,  1, -1, -1, -1,  1],
       [-1, -1,  1, -1, -1, -1, -1, -1,  1,  1, -1,  1,  1, -1, -1,  1,
        -3,  1, -3, -1, -3, -1, -1, -3, -1,  1, -1,  1,  1, -3, -1],
       [ 1,  1, -1,  1,  1,  1,  1,  1, -1, -1,  1, -1, -1,  1,  1, -1,
         3, -1,  3,  1,  3,  1,  1,  3,  1, -1,  1, -1, -1,  3,  1]])