comparison.py

import numpy as np
from PIL import Image
from sklearn.metrics.pairwise import euclidean_distances

from io import BytesIO


def image_comparison(my_image, image_objects):
    """
    This function is to rank all images according to euclidean function

    :param image_objects: the photos from all the photographers
    :param my_image:  image to be compared
    :return :return a list of tuple structure
    (binary_image,euclidean_difference,photographer_id)
    """
    # open my image
    my_image = Image.open(BytesIO(my_image.file))
    # resize the image
    my_image = my_image.resize((400, 400))
    # convert the image to nd-array
    my_image = np.array(my_image)
    # flatten the image
    my_image = my_image.flatten()
    # create a list for compared images
    compared_images = []

    for image_object in image_objects:
        image = Image.open(BytesIO(image_object.file))
        # resize the image
        image = image.resize((400, 400))
        # turn image into numpy nd array
        image = np.array(image)
        # flatten the image
        image = image.flatten()
        # get the euclidean distance
        difference = euclidean_distances([image], [my_image])
        # add a tuple of the photo id and the euclidean difference
        compared_images.append(
            (image, difference[0][0], image_object.photographer_id)
        )
    # return sorted  according to the euclidean difference
    return sorted(compared_images, key=lambda x: x[1])