AsyncProcessScheduler 0.3.0b1

GUI-friendly library for scheduling processes using coroutines.

Async Process Scheduler

Current status: beta.

Introduction

Async Process Scheduler is a small Python library which provides a simple, GUI-friendly way to efficiently run many processes while avoiding a callback-based data flow.

Async Process Scheduler is compatible with multiprocessing from the standard library, and equivalent implementations such as multiprocess.

Note: To use Async Process Scheduler in a GUI program, you'll need a library which implements a compatible event loop. For example, asyncqt can be used with PyQt5.

Installation

To install Async Process Scheduler, use pip:

pip install AsyncProcessScheduler --user

Usage

To use Scheduler in Python, import from scheduler:

from scheduler.Scheduler import Scheduler

After adding processes to a Scheduler instance, running the processes and getting their results is simple:

results: List[Tuple] = await Scheduler.run()

Detailed examples are in the examples folder, including a GUI example using PyQt5. Here is a code snippet which shows the general workflow:

def long_calculation(queue: Queue):
    time.sleep(5) # Simulate a long calculation.
    queue.put((
        "x", "y", "z" # These values will be returned by the scheduler for each process.
    ))


async def run():
    scheduler = Scheduler()
    num_processes = 16

    for i in range(num_processes):
        queue = Queue()
        process = Process(target=long_calculation, args=(queue,))

        scheduler.add(process, queue)

    # Run all processes and get an ordered list containing the results from each.
    results: List[Tuple] = await scheduler.run()

    # Do something with the results.
    print(results)

# Start the coroutine.
asyncio.ensure_future(run()) 

Design

This diagram demonstrates the implementation of Scheduler.

Developer notes

Packaging the project

From the documentation:

rm -r dist/
python setup.py sdist
python setup.py bdist_wheel
twine upload dist/*