From 9334ab0cdbd54b8f52b244f90a896196da9693ee Mon Sep 17 00:00:00 2001
From: Raymond Yee <raymond.yee@gmail.com>
Date: Mon, 6 Feb 2012 20:16:37 -0800
Subject: [PATCH] I've learned how to use a multiprocessing.Manager and dict to
 share results.  Now can I extend it?

---
 test/parallel.py              | 38 +++------------
 test/parallel_with_manager.py | 90 +++++++++++++++++++++++++++++++++++
 2 files changed, 97 insertions(+), 31 deletions(-)
 create mode 100644 test/parallel_with_manager.py

diff --git a/test/parallel.py b/test/parallel.py
index 885de929..4203d690 100644
--- a/test/parallel.py
+++ b/test/parallel.py
@@ -33,8 +33,6 @@ def tripler(n):
     return 3*n
 
 
-
-
 class Consumer(multiprocessing.Process):
 
     def __init__(self, task_queue, result_queue):
@@ -63,32 +61,17 @@ class NetTask(object):
         self.n = n
     def __call__(self):
         print "NetTask %s" % (self.n)
-        global doubler_pool, negator_pool, tripler_pool
-        print doubler_pool, negator_pool, tripler_pool
         
-        async_results = [doubler_pool.apply_async(doubler, (self.n,)),
-                         negator_pool.apply_async(negator, (self.n,)),
-                         tripler_pool.apply_async(tripler, (self.n,))]
-        print async_results
-        print "NetTask about to return async_results for %s" % (self.n)
-        return (self.n, async_results)
+        results = [apply(doubler, (self.n,)),
+                   apply(negator, (self.n,)),
+                   apply(tripler, (self.n,))]
+        print results
+        print "NetTask about to return results for %s" % (self.n)
+        return (self.n, results)
         #return (self.n, sum(r.get() for r in async_results))
     def __str__(self):
         return 'Totaler (%d)' % (self.n)
 
-def pooler():
-    
-    global doubler_pool, negator_pool, tripler_pool
-    
-    TO_CALC = 10
-    results = []
-    
-    for n in range(TO_CALC):
-        async_results = [doubler_pool.apply_async(doubler, (n,)), negator_pool.apply_async(negator, (n,)), tripler_pool.apply_async(tripler, (n,)),]
-        results.append(async_results)
-        
-    for result in results:
-        print(sum(r.get() for r in result))
         
 def main():
     
@@ -96,11 +79,7 @@ def main():
     tasks = multiprocessing.JoinableQueue()
     results_queue = multiprocessing.Queue()
     
-    global doubler_pool, negator_pool, tripler_pool
-    
-    doubler_pool = multiprocessing.Pool(1) #use one core for now
-    negator_pool = multiprocessing.Pool(1)
-    tripler_pool = multiprocessing.Pool(1)
+    doubler_queue = multiprocessing.JoinableQueue()
     
     random.seed()
 
@@ -111,9 +90,6 @@ def main():
                   for i in xrange(num_consumers) ]
     for w in consumers:
         w.start()    
-
-    # demonstrate that we can do stuff with the pools
-    pooler() 
         
     n_tasks = 2
     
diff --git a/test/parallel_with_manager.py b/test/parallel_with_manager.py
new file mode 100644
index 00000000..4d7017a1
--- /dev/null
+++ b/test/parallel_with_manager.py
@@ -0,0 +1,90 @@
+import multiprocessing
+from functools import partial
+
+def wrap_with_args(f):
+    def f1(*args, **kwargs):
+        r = f(*args, **kwargs)
+        return ((args, tuple(sorted(kwargs.items()))), r) # return hashable components
+    return f1
+
+@wrap_with_args
+def doubler(n):
+    #time.sleep(random.uniform(0,0.1))
+    print "in doubler %s " % (n)
+    return 2*n
+
+class DoubleTask(object):
+    def __init__(self, n):
+        self.n = n
+    def __call__(self):
+        print "DoubleTask %s" % (self.n)
+        return (self.n, 2*self.n)
+    def __str__(self):
+        return 'DoubleTask (%d)' % (self.n)
+
+class ConsumerWithResultDict(multiprocessing.Process):
+
+    def __init__(self, task_queue, result_dict):
+        multiprocessing.Process.__init__(self)
+        self.task_queue = task_queue
+        self.result_dict = result_dict
+
+    def run(self):
+        proc_name = self.name
+        while True:
+            next_task = self.task_queue.get()
+            if next_task is None:
+                # Poison pill means shutdown
+                print '%s: Exiting' % proc_name
+                self.task_queue.task_done()
+                break
+            print '%s: %s' % (proc_name, next_task)
+            answer = next_task()
+            print "answer: ", answer
+            self.task_queue.task_done()
+            self.result_dict[answer[0]] = answer[1]
+        return
+    
+
+class Consumer(multiprocessing.Process):
+
+    def __init__(self, task_queue, result_queue):
+        multiprocessing.Process.__init__(self)
+        self.task_queue = task_queue
+        self.result_queue = result_queue
+
+    def run(self):
+        proc_name = self.name
+        while True:
+            next_task = self.task_queue.get()
+            if next_task is None:
+                # Poison pill means shutdown
+                print '%s: Exiting' % proc_name
+                self.task_queue.task_done()
+                break
+            print '%s: %s' % (proc_name, next_task)
+            answer = next_task()
+            self.task_queue.task_done()
+            self.result_queue.put(answer)
+        return
+
+if __name__ == '__main__':
+    
+    manager = multiprocessing.Manager()
+    
+    doubler_queue = multiprocessing.JoinableQueue()
+    results = manager.dict()
+    
+    doubler_processor = ConsumerWithResultDict(doubler_queue, results)
+    doubler_processor.start()
+    
+    n_tasks = 10
+    for k in xrange(n_tasks):
+        doubler_queue.put(DoubleTask(k))
+    doubler_queue.put(None) # mark the end
+    
+    doubler_queue.join()
+    
+    print results
+    
+    
\ No newline at end of file