0
|
1 |
#!/usr/bin/env python2.6
|
|
2 |
import sys
|
|
3 |
import logging
|
|
4 |
import threading
|
|
5 |
import Queue
|
|
6 |
|
|
7 |
##
|
|
8 |
# A basic worker pool... should be replaced with a resizable one.
|
|
9 |
class ThreadPool(object):
|
|
10 |
|
|
11 |
class _Worker(threading.Thread):
|
|
12 |
def __init__(self, uid, pool):
|
|
13 |
threading.Thread.__init__(self, name = "Worker%d" % uid)
|
|
14 |
self.pool = pool
|
|
15 |
def run(self):
|
|
16 |
logging.info("%s started", self.getName())
|
|
17 |
while True:
|
|
18 |
job = self.pool.jobs.get()
|
|
19 |
if job:
|
|
20 |
logging.debug("%s is running %r", self.getName(), job)
|
|
21 |
try:
|
|
22 |
job()
|
|
23 |
except:
|
|
24 |
logging.exception("Job %r return an exception", job)
|
|
25 |
logging.debug("Job %r completed.", job)
|
|
26 |
else:
|
|
27 |
self.pool.jobs.put(job)
|
|
28 |
break
|
|
29 |
logging.info("%s stopped", self.getName())
|
|
30 |
|
|
31 |
def __init__(self, size):
|
|
32 |
super(ThreadPool, self).__init__()
|
|
33 |
self.jobs = Queue.Queue()
|
|
34 |
self._workers = [self._Worker(x+1, self) for x in xrange(size)]
|
|
35 |
|
|
36 |
def Start(self):
|
|
37 |
logging.info("Starting Workers pool")
|
|
38 |
for worker in self._workers:
|
|
39 |
worker.start()
|
|
40 |
logging.info("Workers started")
|
|
41 |
|
|
42 |
def Stop(self, now = False):
|
|
43 |
logging.info("Stopping Workers")
|
|
44 |
if now:
|
|
45 |
self.jobs = Queue.Queue()
|
|
46 |
self.jobs.put(None)
|
|
47 |
for worker in self._workers:
|
|
48 |
worker.join()
|
|
49 |
logging.info("Workers stopped")
|
|
50 |
|
|
51 |
def AddJob(self, job):
|
|
52 |
self.jobs.put(job)
|
|
53 |
|
|
54 |
##
|
|
55 |
# A task, representing a series of linked pairs of callback and error
|
|
56 |
# handler Concept is similar to twisted, but you need to put all your
|
|
57 |
# callback on the task before giving it to the Scheduler. For this
|
|
58 |
# reason, you shouldnt call the callback/errorback method yourself (or
|
|
59 |
# at least, don't put it back in the scheduler after that).
|
|
60 |
|
|
61 |
class Task(object):
|
|
62 |
|
|
63 |
@staticmethod
|
|
64 |
def DefaultCallback(result):
|
|
65 |
return result
|
|
66 |
|
|
67 |
@staticmethod
|
|
68 |
def DefaultErrorback(error):
|
|
69 |
return error
|
|
70 |
|
|
71 |
def __init__(self, func = None, *args, **kwargs):
|
|
72 |
super(Task, self).__init__()
|
|
73 |
self.callbacks = []
|
|
74 |
if func:
|
|
75 |
def callback(result):
|
|
76 |
return func(*args, **kwargs)
|
|
77 |
self.AddCallback(callback)
|
|
78 |
|
|
79 |
def AddCallback(self, callback, errorback = None):
|
|
80 |
if errorback == None:
|
|
81 |
errorback = self.DefaultErrorback
|
|
82 |
self.callbacks.append((callback, errorback))
|
|
83 |
# permit chained calls
|
|
84 |
return self
|
|
85 |
|
|
86 |
def ChainTask(self, task):
|
|
87 |
return self.AddCallback(task.Callback, task.Errorback)
|
|
88 |
|
|
89 |
def Callback(self, result, error = None, traceback = None):
|
|
90 |
logging.debug("Handling task %r callbacks", self)
|
|
91 |
for cb, eb in self.callbacks:
|
|
92 |
try:
|
|
93 |
if error:
|
|
94 |
error = eb(error)
|
|
95 |
else:
|
|
96 |
result = cb(result)
|
|
97 |
except:
|
|
98 |
errtype, error, traceback = sys.exc_info()
|
|
99 |
if error:
|
|
100 |
raise error, None, traceback
|
|
101 |
return result
|
|
102 |
|
|
103 |
##
|
|
104 |
# Consume the callbacks with an error. Notes that error will
|
|
105 |
# raise if not handled and return value would be a result
|
|
106 |
def Errorback(self, error, traceback = None):
|
|
107 |
return self.Callback(None, error, traceback)
|
|
108 |
|
|
109 |
class Scheduler(threading.Thread):
|
|
110 |
|
|
111 |
class _SchedulerStop(Exception):
|
|
112 |
pass
|
|
113 |
|
|
114 |
def __init__(self, poolSize):
|
|
115 |
threading.Thread.__init__(self, name = "Scheduler", target = self.Run)
|
|
116 |
self.pool = ThreadPool(poolSize)
|
|
117 |
self.tasks = Queue.Queue()
|
|
118 |
|
|
119 |
def ExecuteOne(self, blocking = True):
|
|
120 |
logging.debug("Looking for next task...")
|
|
121 |
try:
|
|
122 |
task = self.tasks.get(blocking)
|
|
123 |
except Queue.Empty:
|
|
124 |
logging.debug("No task to run")
|
|
125 |
return
|
|
126 |
return task.Callback(None)
|
|
127 |
|
|
128 |
def Run(self):
|
|
129 |
logging.info("Scheduler start")
|
|
130 |
while True:
|
|
131 |
try:
|
|
132 |
self.ExecuteOne()
|
|
133 |
except self._SchedulerStop:
|
|
134 |
break
|
|
135 |
except:
|
|
136 |
logging.exception("Unhandled task exception")
|
|
137 |
logging.info("Scheduler stop")
|
|
138 |
|
|
139 |
def Start(self):
|
|
140 |
self.pool.Start()
|
|
141 |
return self.start()
|
|
142 |
|
|
143 |
def Stop(self, now = False):
|
|
144 |
self.pool.Stop(now)
|
|
145 |
if now:
|
|
146 |
self.tasks = Queue.Queue()
|
|
147 |
def RaiseSchedulerStop():
|
|
148 |
raise self._SchedulerStop
|
|
149 |
self.AddTask(Task(RaiseSchedulerStop))
|
|
150 |
self.join()
|
|
151 |
|
|
152 |
def AddTask(self, task, blocking = True):
|
|
153 |
self.tasks.put(task, blocking)
|
|
154 |
|
|
155 |
##
|
|
156 |
# A job is a task run in a seperated thread. After the job run, a
|
|
157 |
# new Task is add to the scheduler either with a result or an error,
|
|
158 |
# so that only the task, not the callbacks, is run in the worker
|
|
159 |
# thread. Note the passed task is consumed after this call.
|
|
160 |
# A better design would have to allow Task to suspend or
|
|
161 |
# resume themself and to run in the thread if it want to, but this will
|
|
162 |
# required the callback to know about its Task and the scheduler
|
|
163 |
# itself, which solve nothing.
|
|
164 |
def AddJob(self, task, func, *args, **kwargs):
|
|
165 |
def Job():
|
|
166 |
try:
|
|
167 |
result = func(*args, **kwargs)
|
|
168 |
def returnResult():
|
|
169 |
return result
|
|
170 |
jobTask = Task(returnResult)
|
|
171 |
except:
|
|
172 |
errtype, error, traceback = sys.exc_info()
|
|
173 |
def raiseError():
|
|
174 |
raise error, None, traceback
|
|
175 |
jobTask = Task(raiseError)
|
|
176 |
jobTask.ChainTask(task)
|
|
177 |
self.AddTask(jobTask)
|
|
178 |
self.pool.AddJob(Job)
|
|
179 |
|
|
180 |
##
|
|
181 |
# This basically allow one callback to run in a seperated thread
|
|
182 |
# and then get the result to another deferred callback. Helas,
|
|
183 |
# this create two Tasks, one for the caller, one for the
|
|
184 |
# callbacks, with only the first one having to be passed to the
|
|
185 |
# scheduler. This should be more transparent: a callback which
|
|
186 |
# required to be run as a job should simply be mark like it and
|
|
187 |
# the current task suspend until the job is finished, then resume.
|
|
188 |
def CreateCallbackAsJob(self, src, cb):
|
|
189 |
dst = Task()
|
|
190 |
def CreateJobCallback(result):
|
|
191 |
def Job():
|
|
192 |
cb(result)
|
|
193 |
self.AddJob(dst, Job)
|
|
194 |
src.AddCallback(CreateJobCallback)
|
|
195 |
return dst
|
|
196 |
|
|
197 |
# The global scheduler
|
|
198 |
scheduler = None
|
|
199 |
|
|
200 |
def StartScheduler(size):
|
|
201 |
global scheduler
|
|
202 |
if scheduler:
|
|
203 |
StopScheduler()
|
|
204 |
scheduler = Scheduler(size)
|
|
205 |
scheduler.Start()
|
|
206 |
|
|
207 |
def StopScheduler(now = False):
|
|
208 |
global scheduler
|
|
209 |
if scheduler:
|
|
210 |
scheduler.Stop(now)
|
|
211 |
scheduler = None
|
|
212 |
|
|
213 |
if __name__ == '__main__':
|
|
214 |
from time import sleep
|
|
215 |
logging.getLogger().setLevel(logging.DEBUG)
|
|
216 |
# This function is a sample and shouldn't know about the scheduler
|
|
217 |
count = 0
|
|
218 |
def AsyncCall():
|
|
219 |
global count
|
|
220 |
count += 1
|
|
221 |
def Initialize(name, sleep):
|
|
222 |
print "Here", name
|
|
223 |
return name, sleep
|
|
224 |
def Blocking(args):
|
|
225 |
name, time = args
|
|
226 |
print name, "goes to bed"
|
|
227 |
sleep(time)
|
|
228 |
print name, ": ZZZ..."
|
|
229 |
return name
|
|
230 |
def Finalize(name):
|
|
231 |
global count
|
|
232 |
print name, "wakes up!"
|
|
233 |
count -= 1
|
|
234 |
|
|
235 |
dinit = Task(Initialize, "Toto", 5)
|
|
236 |
# How can I remove the scheduler from the picture ?
|
|
237 |
# The only way I can see is to have a special kind of Task
|
|
238 |
# and a suspended queue... May be this will also be more clean
|
|
239 |
dfinal = scheduler.CreateCallbackAsJob(dinit, Blocking)
|
|
240 |
dfinal.AddCallback(Finalize)
|
|
241 |
# This is confusing but the reason is that the dfinal callback
|
|
242 |
# will be added to the scheduler by the job itself at the end of
|
|
243 |
# its execution.
|
|
244 |
return dinit
|
|
245 |
|
|
246 |
logging.info("Starting scheduler with 10 workers")
|
|
247 |
StartScheduler(10)
|
|
248 |
dasync = AsyncCall()
|
|
249 |
logging.info("Adding asynccall task")
|
|
250 |
scheduler.AddTask(dasync)
|
|
251 |
while count > 0:
|
|
252 |
logging.debug("Count = %d", count)
|
|
253 |
sleep(1)
|
|
254 |
logging.info("Stopping scheduler")
|
|
255 |
StopScheduler()
|
|
256 |
logging.info("The End.")
|