Merge branch 'Dev' of https://github.com/AbductiveLearning/ABL-Package into Dev

abl/abducer/abducer_base.py

@@ -10,9 +10,10 @@
 #
 # ================================================================#
 
+import os
 import abc
 import numpy as np
-import multiprocessing
+from multiprocessing import Pool
 from zoopt import Dimension, Objective, Parameter, Opt
 from ..utils.utils import confidence_dist, flatten, reform_idx, hamming_dist
 
@@ -110,10 +111,8 @@ class AbducerBase(abc.ABC):
         return self.abduce((z, prob, y), max_address, require_more_address)
 
     def batch_abduce(self, Z, Y, max_address=-1, require_more_address=0):
-        pool = multiprocessing.Pool(processes=len(Z))
-        results = pool.map(self._batch_abduce_helper, [(z, prob, y, max_address, require_more_address) for z, prob, y in zip(Z['cls'], Z['prob'], Y)])
-        pool.close()
-        pool.join()
+        with Pool(processes=os.cpu_count()) as pool:
+            results = pool.map(self._batch_abduce_helper, [(z, prob, y, max_address, require_more_address) for z, prob, y in zip(Z['cls'], Z['prob'], Y)])
         return results
     
     def __call__(self, Z, Y, max_address=-1, require_more_address=0):