multiprocessing without main

pip install multiprocnomain

Advantages:

• Allows easy parallelization of functions without the need for execution in the '__main__' block or at the top level.
• Facilitates the parallel execution of functions with different input parameters, providing flexibility.
• Improves performance by leveraging multiprocessing, especially for computationally intensive tasks.
• Automatically aggregates results into a dictionary, simplifying result mapping to input indices.
• Customizable parameters (processes and chunks) to adapt to specific requirements and hardware capabilities.
• Suitable for various use cases, including data science, machine learning, image processing, and scientific computing.

Parameters:
	- fu (callable): The function to be executed in parallel.
	- it (iterable): An iterable of dictionaries, each containing the input parameters for the function.
	- processes (int): The number of processes to use (default is 3).
	- chunks (int): The chunk size for multiprocessing.Pool.starmap (default is 1).

Returns:
	dict: A dictionary containing the results of the parallel executions, where keys correspond to the indices
		of the input iterable and values contain the corresponding function outputs.

Examples:
	import random
	from multiprocnomain import start_multiprocessing
	import subprocess
	from a_cv_imwrite_imread_plus import open_image_in_cv
	import numpy as np


	def somefu(q=100):
		exec(f"import random", globals())  # necessary
		y = random.randint(10, 20)
		for x in range(q):
			y = y + x

		return y


	# somefu=lambda r:1111
	it = [{"q": 100}, {"q": 100}, {"q": 100}, {"q": 10}]
	b2 = start_multiprocessing(fu=somefu, it=it, processes=3, chunks=1)
	print(b2)


	def somefu2(path):
		exec(f"import subprocess", globals()) # necessary
		y = subprocess.run([f"ls" ,f"{path}"],capture_output=True)
		return y

	allpath=[{'path':'c:\\windows'}, {'path':'c:\\cygwin'}]
	b1 = start_multiprocessing(fu=somefu2, it=allpath, processes=3, chunks=1)
	print(b1)
	def somefu3(q):
		exec(f"from a_cv_imwrite_imread_plus import open_image_in_cv", globals()) # necessary
		exec(f"import numpy as np", globals()) # necessary

		im = open_image_in_cv(q)
		r = im[..., 2]
		g = im[..., 1]
		b = im[..., 0]
		return np.where((r == 255) & (g == 255) & (b == 255))


	allimages = [
		{"q": r"C:\Users\hansc\Pictures\collage_2023_04_23_06_04_51_956747.png"},
		{"q": r"C:\Users\hansc\Pictures\bw_clickabutton.png"},
		{"q": r"C:\Users\hansc\Pictures\cgea.png"},
		{"q": r"C:\Users\hansc\Pictures\checkboxes.png"},
		{"q": r"C:\Users\hansc\Pictures\clickabutton.png"},
		{"q": r"C:\Users\hansc\Pictures\collage_2023_04_23_05_24_31_797203.png"},
		{"q": r"C:\Users\hansc\Pictures\collage_2023_04_23_05_25_48_657510.png"},
		{"q": r"C:\Users\hansc\Pictures\collage_2023_04_23_05_26_16_431863.png"},
		{"q": r"C:\Users\hansc\Pictures\collage_2023_04_23_05_27_07_483808.png"},
		{"q": r"C:\Users\hansc\Pictures\collage_2023_04_23_05_27_41_985343.png"},
		{"q": r"C:\Users\hansc\Pictures\collage_2023_04_23_05_28_16_529438.png"},
		{"q": r"C:\Users\hansc\Pictures\collage_2023_04_23_05_28_55_105250.png"},
		{"q": r"C:\Users\hansc\Pictures\collage_2023_04_23_05_29_11_492492.png"},
		{"q": r"C:\Users\hansc\Pictures\collage_2023_04_23_05_38_13_226848.png"},
		{"q": r"C:\Users\hansc\Pictures\collage_2023_04_23_06_04_14_676085.png"},
		{'q':r"C:\Users\hansc\Downloads\IMG-20230618-WA0000.jpeg"},
		{'q':r"C:\Users\hansc\Downloads\maxresdefault.jpg"},
		{'q':r"C:\Users\hansc\Downloads\panda-with-broom-600x500 (1).jpg"},
		{'q':r"C:\Users\hansc\Downloads\panda-with-broom-600x500.jpg"},
		{'q':r"C:\Users\hansc\Downloads\panda-with-broom-600x500222222222.jpg"},
		{'q':r"C:\Users\hansc\Downloads\pexels-alex-andrews-2295744.jpg"},
		{'q':r"C:\Users\hansc\Downloads\pexels-niki-nagy-1128416.jpg"},

	]
	b = start_multiprocessing(fu=somefu3, it=allimages, processes=3, chunks=5)
	print(b)