Python multiprocessing.Pool.map()删除子类ndarray的属性
在Python multiprocessing.Pool.map()删除子类ndarray的属性,python,numpy,subclass,python-multiprocessing,Python,Numpy,Subclass,Python Multiprocessing,在numpy.ndarray-子类的实例列表上使用multiprocessing.Pool()中的map()时,会删除自己类的新属性 以下基于的最小示例再现了该问题: from multiprocessing import Pool import numpy as np class MyArray(np.ndarray): def __new__(cls, input_array, info=None): obj = np.asarray(input_array).
numpy.ndarray
-子类的实例列表上使用multiprocessing.Pool()
中的map()
时,会删除自己类的新属性
以下基于的最小示例再现了该问题:
from multiprocessing import Pool
import numpy as np
class MyArray(np.ndarray):
def __new__(cls, input_array, info=None):
obj = np.asarray(input_array).view(cls)
obj.info = info
return obj
def __array_finalize__(self, obj):
if obj is None: return
self.info = getattr(obj, 'info', None)
def sum_worker(x):
return sum(x) , x.info
if __name__ == '__main__':
arr_list = [MyArray(np.random.rand(3), info=f'foo_{i}') for i in range(10)]
with Pool() as p:
p.map(sum_worker, arr_list)
属性info
被删除
AttributeError: 'MyArray' object has no attribute 'info'
使用内置的map()
arr_list = [MyArray(np.random.rand(3), info=f'foo_{i}') for i in range(10)]
list(map(sum_worker, arr_list2))
方法\u\u数组\u finalize\u()
的目的是对象在切片后保留属性
arr = MyArray([1,2,3], info='foo')
subarr = arr[:2]
print(subarr.info)
但是对于Pool.map()
这个方法不知何故是不起作用的…因为多处理使用pickle
将数据序列化到单独的进程中或从单独的进程中序列化,这本质上是一个复制
根据该问题调整公认的解决方案,您的示例变成:
from multiprocessing import Pool
import numpy as np
class MyArray(np.ndarray):
def __new__(cls, input_array, info=None):
obj = np.asarray(input_array).view(cls)
obj.info = info
return obj
def __array_finalize__(self, obj):
if obj is None: return
self.info = getattr(obj, 'info', None)
def __reduce__(self):
pickled_state = super(MyArray, self).__reduce__()
new_state = pickled_state[2] + (self.info,)
return (pickled_state[0], pickled_state[1], new_state)
def __setstate__(self, state):
self.info = state[-1]
super(MyArray, self).__setstate__(state[0:-1])
def sum_worker(x):
return sum(x) , x.info
if __name__ == '__main__':
arr_list = [MyArray(np.random.rand(3), info=f'foo_{i}') for i in range(10)]
with Pool() as p:
p.map(sum_worker, arr_list)
注意,第二个答案建议您可以使用pathos.multi-processing
和您未经调整的原始代码,因为pathos使用dill
而不是pickle
。但是,当我测试它时,它不起作用