How to return something from a method to a class in Python

Question

I am still novice in python and I am trying to write my first class that will eventually contain a bunch of methods for image processing.

My problem is that even though my method computes something I find it hard "return" this outcome to the class. So depending where I place the "return" command - either I get an error that method is unused either it does not get the reference to the returned variable (unresolved reference).

The code looks like this (version with unresolved reference) :

from sklearn.feature_extraction import image
import numpy as np


class ImagePreprocessing():


    def __init__(self, X=None, Y=None,
                 tile_size=None, label_computation_method = "majority"):
        self.X = X
        self.Y = Y
        self.tile_size = tile_size
        self.label_computation_method = label_computation_method


        def image_tilling_odd(X,Y,tile_size,label_computation_method):

          """
            This method takes a set of image data and their respectice tiles and constructs a
            tilebased vector image - with each tile per row - so the overall size is
                (size[0] x size[1] x number of bands)
          """

          # asser if tile dimension is the same is all directions
          assert tile_size[0] == tile_size[1]

          # test if tiles are even or odd
          assert tile_size[0] % 2 != 0
          assert tile_size[1] % 2 != 0


          # tile images and labels
          image_tiles=image.extract_patches_2d(X,tile_size)
          label_tiles=image.extract_patches_2d(Y,tile_size)

          # construct column vector of zeros to store label
          Y=np.zeros(label_tiles.size[0])

          if label_computation_method == "central_value":

              # index of middle element
              idx=tile_size[0]/2 +1

              # detect middle element and store labels --- central element labeling
              for i,rr in enumerate(label_tiles):

                 # construct vector with middle labels
                 Y[i]= rr[idx]

                 COM=np.append(image_tiles,Y,1)

        return COM

Answer 1

My guess is that your indentation is wrong or you don't need to return anything and just set a class attribute value.

class MyClass(object):
    def __init__(self, x=0, y=0):
        super().__init__()

        self.x = x
        self.y = y
        self.result = None
    # end Constructor

    def method_return(self):
        return self.x + self.y
    # end method_return

    def method_for_attribute(self):
        self.result = self.x + self.y
    # end method_for_attribute
# end class MyClass

if __name__ == "__main__":
    my_instance = MyClass(1, 1)

    result = my_instance.method_return() # Returns a value to result
    print(result)

    print(my_instance.result)
    my_instance.method_for_attribute() # Sets the value for my_instance.result
    print(my_instance.result)

Answer 2

Ok I think I found it - solution below !

from sklearn.feature_extraction import image
import numpy as np


class ImagePreprocessing():


        def __init__(self, X=None, Y=None,
                 tile_size=None, label_computation_method = "majority"):
           self.X = X
           self.Y = Y
           self.tile_size = tile_size
           self.label_computation_method = label_computation_method


        def image_tilling_odd(X,Y,tile_size,label_computation_method):

          """
            This method takes a set of image data and their respectice tiles and constructs a
            tilebased vector image - with each tile per row - so the overall size is
                (size[0] x size[1] x number of bands)
          """

          # asser if tile dimension is the same is all directions
          assert tile_size[0] == tile_size[1]

          # test if tiles are even or odd
          assert tile_size[0] % 2 != 0
          assert tile_size[1] % 2 != 0


          # tile images and labels
          image_tiles=image.extract_patches_2d(X,tile_size)
          label_tiles=image.extract_patches_2d(Y,tile_size)

          # construct column vector of zeros to store label
          Y=np.zeros(label_tiles.size[0])

          if label_computation_method == "central_value":

              # index of middle element
              idx=tile_size[0]/2 +1

              # detect middle element and store labels --- central element labeling
              for i,rr in enumerate(label_tiles):

                 # construct vector with middle labels
                 Y[i]= rr[idx]

                 COM=np.append(image_tiles,Y,1)

          return COM