AFirooz · May 30, 2025 16:29
diff --git a/draw-nn-graph.ipynb b/draw-nn-graph.ipynb
diff --git a/pandas.ipynb b/pandas.ipynb
diff --git a/pool-processing.py b/pool-processing.py
 import multiprocessing as mp
 import time
 from tqdm import tqdm


 """ See Obsidian for more details.
 A more complicated example:
 https://github.com/mCodingLLC/VideosSampleCode/blob/master/videos/104_multiprocessing_pool/multiprocessing_pool.py
 """

 def process_task(func_args):
    data, other = func_args
    other = other if other else 0
    # Replace this with the actual computation for each loop iteration
    t = 25-data if data < 25 else data - 25
    time.sleep(0.25*t)
    # print(f"Processed {data=}")
    return data*other


 if __name__ == "__main__":
    print("Number of cpu : ", mp.cpu_count())
    
    # List of inputs for the loop
    inputs = list(range(50))
    other = list(range(0, 100, 2))
    func_args = [(x, y) for x, y in zip(inputs, other)]
    results = []

    start = time.time()
    
    # Create a pool of workers (defaults to all available cores)
    with mp.Pool(processes=30) as pool:
        # Map the function to the inputs, distributing tasks across CPUs
        # results = pool.map(process_task, func_args)
        # results = pool.map_async(process_task, func_args).get()
        # results = list(pool.imap(process_task, func_args))  # Convert iterator to list. Much slower than map
        # results = list(pool.imap_unordered(process_task, func_args))

        # tqdm need to be passed an iterable, not a generator
        # Source: https://stackoverflow.com/a/41921948/12315585
        with tqdm(total=len(inputs)) as pbar:
            for result in pool.imap_unordered(process_task, func_args):
                pbar.update()
                pbar.refresh()
                # Do something with the result
                results.append(result)

    print('\ntime taken: ', time.time() - start, '\n')
    
    if isinstance(results, list):
        print(results)
    else:
        for r in results:
            print(r)
diff --git a/sklearn-pipline.ipynb b/sklearn-pipline.ipynb
diff --git a/SMOTE.ipynb b/SMOTE.ipynb
diff --git a/SWE Best Practices.md b/SWE Best Practices.md
	import multiprocessing as mp
	import time
	from tqdm import tqdm


	""" See Obsidian for more details.
	A more complicated example:
	https://github.com/mCodingLLC/VideosSampleCode/blob/master/videos/104_multiprocessing_pool/multiprocessing_pool.py
	"""

	def process_task(func_args):
	data, other = func_args
	other = other if other else 0
	# Replace this with the actual computation for each loop iteration
	t = 25-data if data < 25 else data - 25
	time.sleep(0.25*t)
	# print(f"Processed {data=}")
	return data*other


	if __name__ == "__main__":
	print("Number of cpu : ", mp.cpu_count())

	# List of inputs for the loop
	inputs = list(range(50))
	other = list(range(0, 100, 2))
	func_args = [(x, y) for x, y in zip(inputs, other)]
	results = []

	start = time.time()

	# Create a pool of workers (defaults to all available cores)
	with mp.Pool(processes=30) as pool:
	# Map the function to the inputs, distributing tasks across CPUs
	# results = pool.map(process_task, func_args)
	# results = pool.map_async(process_task, func_args).get()
	# results = list(pool.imap(process_task, func_args)) # Convert iterator to list. Much slower than map
	# results = list(pool.imap_unordered(process_task, func_args))

	# tqdm need to be passed an iterable, not a generator
	# Source: https://stackoverflow.com/a/41921948/12315585
	with tqdm(total=len(inputs)) as pbar:
	for result in pool.imap_unordered(process_task, func_args):
	pbar.update()
	pbar.refresh()
	# Do something with the result
	results.append(result)

	print('\ntime taken: ', time.time() - start, '\n')

	if isinstance(results, list):
	print(results)
	else:
	for r in results:
	print(r)