Multiprocessing queues for numpy arrays using shared memory
This package provides a drop-in replacement for the Python multiprocessing Queue class which handles transport of large numpy arrays. It avoids pickling and uses the multiprocessing Array class in the background. The major difference between this implementation and the normal queue is that the maximal amount of memory that the queue can have must be specified beforehand.
Attempting to send an array of a different shape or datatype of the previously inserted one resets the queue. Only passing of numpy arrays is supported, optionally annotated with timestamps if using the TimestampedArrayQueue class, but other object types can be supported by extending the class.
The package has been tested on Python 3.6/3/7 on Windows and MacOS and Linux with Travis. Python 2.7 is not supported.
from arrayqueues.shared_arrays import ArrayQueue from multiprocessing import Process import numpy as np class ReadProcess(Process): def __init__(self, source_queue): super().__init__() self.source_queue = source_queue def run(self): print(self.source_queue.get()) if __name__ == "__main__": q = ArrayQueue(1) # intitialises an ArrayQueue which can hold 1MB of data n = np.full((5,5), 5) q.put(n) r = ReadProcess(q) r.start() r.join()
Further examples can be found in tests.
Release history Release notifications | RSS feed
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
|Filename, size||File type||Python version||Upload date||Hashes|
|Filename, size arrayqueues-1.3.1-py3-none-any.whl (6.2 kB)||File type Wheel||Python version py3||Upload date||Hashes View|
|Filename, size arrayqueues-1.3.1.tar.gz (8.6 kB)||File type Source||Python version None||Upload date||Hashes View|
Hashes for arrayqueues-1.3.1-py3-none-any.whl