helpers.py

import numpy as np

def grayscale(pixels):
    avg = np.mean(pixels, axis=2, keepdims=True)
    return np.repeat(avg, 3, axis=2).astype(np.int16)

def reflect(pixels):
    return pixels[:, ::-1].copy()

def blur(pixels):
    padded = np.pad(pixels, ((1,1), (1,1), (0,0)), mode='edge')
    total = np.zeros_like(pixels, dtype=np.float32)
    for r in range(3):
        for c in range(3):
            total += padded[r:r+pixels.shape[0], c:c+pixels.shape[1]]
            
    return np.round(total / 9.0).astype(np.int16)

def edges(pixels):
    img_float = pixels.astype(np.float32)
    
    gx_kernel = np.array([[-1, 0, 1], [-2, 0, 2], [-1, 0, 1]], dtype=np.float32)
    gy_kernel = np.array([[-1, -2, -1], [0, 0, 0], [1, 2, 1]], dtype=np.float32)
    
    padded = np.pad(img_float, ((1,1), (1,1), (0,0)), mode='edge')
    
    gx_total = np.zeros_like(img_float)
    gy_total = np.zeros_like(img_float)
    
    for r in range(3):
        for c in range(3):
            slice_zone = padded[r:r+pixels.shape[0], c:c+pixels.shape[1]]
            gx_total += slice_zone * gx_kernel[r, c]
            gy_total += slice_zone * gy_kernel[r, c]
            
    magnitude = np.sqrt(gx_total**2 + gy_total**2)
    return np.round(magnitude).astype(np.int16)