Parallel

class Parallel(n_cores: int = 3, initializer=None, initargs=(), activate: bool = False)

Bases: object

Utility wrapper for paralellizing tasks across multiple CPU cores.

This class simplifies the management of a multiprocessing.Pool, providing a clean interface to activate, execute, and shut down parallel workers.

Parameters:

• n_cores (int, default multiprocessing.cpu_count() - 1) – Number of cores to use.

• initializer (callable, optional) – A function to be called by each worker process when it starts.

• initargs (tuple, optional) – Arguments to be passed to the initializer.

• activate (bool, default False) – Whether to start the pool immediately upon instantiation.

Methods

activate	Create and start the multiprocessing pool.
deactivate	Gracefully shut down the process pool.
perform	Execute a function in parallel.

activate()

Create and start the multiprocessing pool.

This method allocates system resources and spawns the worker processes. It must be called before perform can be used.

deactivate()

Gracefully shut down the process pool.

Closes the pool to new tasks and waits for all current worker processes to exit. Resources are freed and the internal pool is set back to None.

perform(func: Callable, *args)

Execute a function in parallel.

If a single iterable is provided in args, pool.map is used. If multiple iterables are provided, they are zipped together and executed via pool.starmap.

Parameters:

• func (callable) – The function to execute in parallel.

• *args (iterable) – The input data to be distributed among workers.

Returns:

list – The collected results from all worker processes.

Raises:

AssertionError – If func is not callable or no arguments are provided.

Attributes

max_cores = 4

Total number of logical CPU cores available on the system.

Calculated by multiprocessing.cpu_count()