This uses SciPy, NumPy, and illustrates how NumPy arrays can be used to manipulate png images.
{{{c
"""
pngcat

Concatenates PNG files

Usage:
import pngcat
pngcat_v(["file1.png","file2.png",...,"filen.png"],"output.png") # above each other
pngcat_h(["file1.png","file2.png",...,"filen.png"],"output.png") # beside each other
"""

import numpy as np
from imageio import imread,imwrite

def pngcat_v(input_filenames, output_filename, bgcolor = [255,255,255,255]):
    """pngcat_v([filenames],output_filename)"""
    input_data = list(map(imread,input_filenames))
    input_heights = list(map(lambda t: t.shape[0],input_data))
    input_widths = list(map(lambda t: t.shape[1],input_data))
    input_dtype = input_data[0].dtype
    output_height = sum(input_heights)
    output_width = max(input_widths)
    output_array = np.zeros((output_height,output_width,4),dtype=input_dtype)
    output_array[:,:] = bgcolor
    start_row = 0
    for src in input_data:
        h,w,_ = src.shape
        output_array[start_row:start_row+h,:w,:] = src
        start_row += h
    imwrite(output_filename,output_array)
    
def pngcat_h(input_filenames, output_filename, bgcolor = [255,255,255,255]):
    """pngcat_h([filenames],output_filename)"""
    input_data = list(map(imread,input_filenames))
    input_heights = list(map(lambda t: t.shape[0],input_data))
    input_widths = list(map(lambda t: t.shape[1],input_data))
    input_dtype = input_data[0].dtype
    output_height = max(input_heights)
    output_width = sum(input_widths)
    output_array = np.zeros((output_height,output_width,4),dtype=input_dtype)
    output_array[:,:] = bgcolor
    start_col = 0
    for src in input_data:
        h,w,_ = src.shape
        output_array[0:h,start_col:start_col+w,:] = src
        start_col += w
    imwrite(output_filename,output_array)
}}}%TIME=1616260412