Refactoring to make testing more stable

This commit is contained in:
Ilan Steemers
2015-06-23 21:12:08 +02:00
parent 97a4b27f68
commit 0ed1da8f05
14 changed files with 327 additions and 71 deletions

View File

@@ -1,10 +1,7 @@
"""
A multiprocessing task queue application for Django
Author: Ilan Steemers (koed00@gmail.com
Github: https://github.com/Koed00/django-q
"""
from django_q.models import Task
from django_q.core import async, Cluster
from .main import Cluster, async, SignedPackage, Stat
default_app_config = 'django_q.apps.DjangoQConfig'
default_app_config = 'django_q.apps.SessionAdminConfig'
def result(name):
return Task.get_result(name)

View File

@@ -1,6 +1,6 @@
from django.contrib import admin
from django_q import async
from django_q.core import async
from .models import Success, Failure

View File

@@ -1,6 +1,6 @@
from django.apps import AppConfig
class SessionAdminConfig(AppConfig):
class DjangoQConfig(AppConfig):
name = 'django_q'
verbose_name = "Django Q"

View File

@@ -3,6 +3,14 @@ from multiprocessing import cpu_count
VERSION = '0.1.0'
"""
Prefixes the Redis keys. Defaults to django_q
"""
try:
PREFIX = settings.Q_PREFIX
except AttributeError:
PREFIX = 'django_q'
"""
Sets the logging level for the app
"""
@@ -41,3 +49,8 @@ try:
COMPRESSED = settings.Q_COMPRESSED
except AttributeError:
COMPRESSED = False
try:
USE_TZ = settings.USE_TZ
except AttributeError:
USE_TZ = False

View File

@@ -1,3 +1,18 @@
# Future
from __future__ import unicode_literals
from __future__ import print_function
from __future__ import division
from __future__ import absolute_import
from builtins import dict
from builtins import range
from datetime import datetime
from django.utils.timezone import make_aware
from future import standard_library
standard_library.install_aliases()
# Standard
import importlib
import logging
import os
@@ -6,6 +21,7 @@ from multiprocessing import Queue, Event, Process, current_process
import socket
import sys
from time import sleep
import gc
# External
import jsonpickle
@@ -15,22 +31,29 @@ import redis
# Django
from django.core import signing
from django.utils import timezone
# Local
from .conf import LOG_LEVEL, SECRET_KEY, SAVE_LIMIT, WORKERS, COMPRESSED, VERSION
from .conf import LOG_LEVEL, SECRET_KEY, SAVE_LIMIT, WORKERS, COMPRESSED, PREFIX, USE_TZ
from .humanhash import uuid
from .models import Task, Success
SIGNAL_NAMES = dict((getattr(signal, n), n) for n in dir(signal) if n.startswith('SIG') and '_' not in n)
prefix = 'django_q'
q_list = '{}:q'.format(prefix)
logger = logging.getLogger('django-q')
coloredlogs.install(level=getattr(logging, LOG_LEVEL))
Q_LIST = '{}:q'.format(PREFIX)
STARTING = 'Starting'
RUNNING = 'Running'
STOPPED = 'Stopped'
STOPPING = 'Stopping'
r = redis.StrictRedis()
# JsonPickle seems to have a problem with django's timezone.now
def time_zone(value):
if USE_TZ:
return make_aware(value)
return value
class Cluster(object):
def __init__(self):
@@ -41,6 +64,8 @@ class Cluster(object):
return
self.sentinel = None
self.stop_event = None
self.start_event = None
self.stopped_event = None
self.pid = current_process().pid
signal.signal(signal.SIGTERM, self.sig_handler)
signal.signal(signal.SIGINT, self.sig_handler)
@@ -54,22 +79,56 @@ class Cluster(object):
sys.stdin.close = dummy_close
# Start Sentinel
self.stop_event = Event()
self.sentinel = Process(target=Sentinel, args=(self.stop_event,))
self.start_event = Event()
self.sentinel = Process(target=Sentinel, args=(self.stop_event, self.start_event))
self.sentinel.start()
logger.info('Starting Q cluster version {} at {}'.format(VERSION, self.pid))
logger.info('Q Cluster-{} starting.'.format(self.pid))
return self.pid
def stop(self):
if not self.sentinel.is_alive():
return False
logger.info('Q Cluster-{} stopping.'.format(self.pid))
self.stop_event.set()
self.sentinel.join()
logger.info('Q cluster has stopped.')
logger.info('Q Cluster-{} has stopped.'.format(self.pid))
self.start_event = None
self.stop_event = None
return True
def sig_handler(self, signum, frame):
logger.debug('{} got signal {}'.format(current_process().name, SIGNAL_NAMES.get(signum, 'UNKNOWN')))
self.stop()
@property
def stat(self):
if self.sentinel:
return Stat.get(self.pid)
return Status(self.pid)
@property
def is_starting(self):
return self.stop_event and self.start_event and not self.start_event.is_set()
@property
def is_running(self):
return self.stop_event and self.start_event and self.start_event.is_set()
@property
def is_stopping(self):
return self.stop_event and self.start_event and self.start_event.is_set() and self.stop_event.is_set()
@property
def has_stopped(self):
return self.start_event is None and self.stop_event is None and self.sentinel
@property
def is_idle(self):
return self.sentinel is None
class Sentinel(object):
def __init__(self, event):
def __init__(self, stop_event, start_event):
signal.signal(signal.SIGINT, signal.SIG_IGN)
signal.signal(signal.SIGTERM, signal.SIG_DFL)
self.pid = current_process().pid
@@ -77,8 +136,9 @@ class Sentinel(object):
self.name = current_process().name
self.status = None
self.reincarnations = 0
self.tob = timezone.now()
self.stop_event = event
self.tob = datetime.utcnow()
self.stop_event = stop_event
self.start_event = start_event
self.pool_size = WORKERS
self.pool = []
self.task_queue = Queue()
@@ -124,7 +184,7 @@ class Sentinel(object):
self.reincarnations += 1
def spawn_cluster(self):
self.set_status('Starting')
self.set_status(STARTING)
for i in range(self.pool_size):
self.spawn_worker()
self.monitor_pid = self.spawn_monitor()
@@ -132,21 +192,23 @@ class Sentinel(object):
def guard(self):
logger.info('{} guarding cluster at {}'.format(current_process().name, self.pid))
self.set_status('Running')
while not self.stop_event.is_set():
self.start_event.set()
self.set_status(RUNNING)
logger.info('Q Cluster-{} running.'.format(self.parent_pid))
while True:
for p in list(self.pool):
if not p.is_alive():
# Be humane
p.terminate()
self.pool.remove(p)
# Replace it with a fresh one
self.reincarnate(p.pid)
Stat(self).publish()
Stat(self).save()
if self.stop_event.is_set():
break
sleep(2)
self.stop()
def stop(self):
self.set_status('Stopping')
self.set_status(STOPPING)
name = current_process().name
logger.info('{} stopping pool processes'.format(name))
# Stopping pusher
@@ -163,20 +225,22 @@ class Sentinel(object):
# Finally stop the monitor
self.done_queue.put('STOP')
self.pool = []
self.set_status('Stopped')
self.set_status(STOPPED)
def set_status(self, message=None):
Stat(self, message).publish()
Stat(self, message).save()
def pusher(task_queue, e):
logger.info('{} pushing tasks at {}'.format(current_process().name, current_process().pid))
while not e.is_set():
task = r.blpop(q_list, 1)
while True:
task = r.blpop(Q_LIST, 1)
if task:
task = task[1]
task_queue.put(task)
logger.debug('queueing {}'.format(task))
if e.is_set():
break
logger.info("{} stopped pushing tasks".format(current_process().name))
@@ -185,7 +249,7 @@ def monitor(done_queue):
logger.info("{} monitoring results at {}".format(name, current_process().pid))
for task in iter(done_queue.get, 'STOP'):
if task['success']:
logger.info("Finished [{}]".format(task['name']))
logger.info("Processed [{}]".format(task['name']))
else:
logger.error("Failed [{}] - {}".format(task['name'], task['result']))
save_task(task)
@@ -204,7 +268,7 @@ def worker(task_queue, done_queue):
continue
except signing.BadSignature as e:
task['name'] = task['name'].rsplit(":", 1)[0]
task['stopped'] = timezone.now()
task['stopped'] = datetime.utcnow()
task['result'] = e
task['success'] = False
done_queue.put(task)
@@ -215,12 +279,13 @@ def worker(task_queue, done_queue):
m = importlib.import_module(module)
f = getattr(m, func)
task['result'] = f(*task['args'], **task['kwargs'])
task['stopped'] = timezone.now()
task['stopped'] = datetime.utcnow()
task['success'] = True
done_queue.put(task)
gc.collect()
except Exception as e:
task['result'] = e
task['stopped'] = timezone.now()
task['stopped'] = datetime.utcnow()
task['success'] = False
done_queue.put(task)
logger.info('{} stopped doing work'.format(name))
@@ -234,19 +299,24 @@ def save_task(task):
hook=task['hook'],
args=task['args'],
kwargs=task['kwargs'],
started=task['started'],
stopped=task['stopped'],
started=time_zone(task['started']),
stopped=time_zone(task['stopped']),
result=task['result'],
success=task['success'])
def async(func, *args, hook=None, **kwargs):
def async(func, *args, **kwargs):
"""
Schedules a task
Schedules a task with optional hook
"""
task = {'name': uuid()[0], 'func': func, 'hook': hook, 'args': args, 'kwargs': kwargs, 'started': timezone.now()}
if 'hook' in kwargs:
hook = kwargs['hook']
del kwargs['hook']
else:
hook = None
task = {'name': uuid()[0], 'func': func, 'hook': hook, 'args': args, 'kwargs': kwargs, 'started': datetime.utcnow()}
pack = SignedPackage.dumps(task)
r.rpush(q_list, pack)
r.rpush(Q_LIST, pack)
logger.debug('Pushed {}'.format(pack))
return task['name']
@@ -287,32 +357,75 @@ class JSONPickleSerializer(object):
return jsonpickle.loads(data.decode('latin-1'))
class Stat(object):
def __init__(self, sentinel, message=None):
self.cluster_id = sentinel.parent_pid
class Status(object):
def __init__(self, pid):
self.workers = []
self.tob = None
self.reincarnations = 0
self.cluster_id = pid
self.sentinel = 0
self.status = 'Idle'
self.done_q_size = 0
self.host = socket.gethostname()
self.monitor = 0
self.task_q_size = 0
self.pusher = 0
self.timestamp = datetime.utcnow()
class Stat(Status):
def __init__(self, sentinel, message=None):
super().__init__(sentinel.parent_pid)
if message:
sentinel.status = message
self.status = sentinel.status
self.tob = sentinel.tob
self.reincarnations = sentinel.reincarnations
self.sentinel = sentinel.pid
self.status = sentinel.status
self.done_q_size = sentinel.done_queue.qsize()
self.monitor = sentinel.monitor_pid
self.task_q_size = sentinel.task_queue.qsize()
self.pusher = sentinel.pusher_pid
self.workers = []
for w in sentinel.pool:
self.workers.append(w.pid)
self.reincarnations = sentinel.reincarnations
self.task_q_size = sentinel.task_queue.qsize()
self.done_q_size = sentinel.done_queue.qsize()
self.tob = sentinel.tob
self.timestamp = timezone.now()
def uptime(self):
return (datetime.utcnow() - self.tob).total_seconds()
@property
def key(self):
return self.get_key()
return self.get_key(self.cluster_id)
@staticmethod
def get_key():
return '{}:cluster'.format(prefix)
def get_key(cluster_id):
return '{}:cluster:{}'.format(PREFIX, cluster_id)
def publish(self):
r.publish(self.key, SignedPackage.dumps(self, True))
def save(self):
r.set(self.key, SignedPackage.dumps(self, True), 3)
def empty_queues(self):
return self.done_q_size + self.task_q_size == 0
@staticmethod
def get(cluster_id):
key = Stat.get_key(cluster_id)
if r.exists(key):
pack = r.get(key)
try:
return SignedPackage.loads(pack)
except signing.BadSignature:
return None
return Status(cluster_id)
@staticmethod
def get_all():
stats = []
keys = r.keys(pattern='{}:cluster:*'.format(PREFIX))
if keys:
packs = r.mget(keys)
for pack in packs:
try:
stats.append(SignedPackage.loads(pack))
except signing.BadSignature:
continue
return stats

View File

@@ -1,9 +1,9 @@
from django.core.management.base import BaseCommand
from django_q import Cluster
from django_q.core import Cluster
class Command(BaseCommand):
help = "My shiny new management command."
help = "Starts a Django Q Cluster."
def handle(self, *args, **options):
q = Cluster()

View File

@@ -14,7 +14,7 @@ class Migration(migrations.Migration):
migrations.CreateModel(
name='Task',
fields=[
('id', models.AutoField(verbose_name='ID', serialize=False, auto_created=True, primary_key=True)),
('id', models.AutoField(auto_created=True, primary_key=True, verbose_name='ID', serialize=False)),
('name', models.CharField(max_length=100)),
('func', models.CharField(max_length=256)),
('hook', models.CharField(max_length=256, null=True)),
@@ -31,8 +31,8 @@ class Migration(migrations.Migration):
fields=[
],
options={
'verbose_name': 'Failed task',
'proxy': True,
'verbose_name': 'Failed task',
},
bases=('django_q.task',),
),
@@ -41,8 +41,8 @@ class Migration(migrations.Migration):
fields=[
],
options={
'verbose_name': 'Successful task',
'proxy': True,
'verbose_name': 'Successful task',
},
bases=('django_q.task',),
),

View File

@@ -20,7 +20,9 @@ class Task(models.Model):
@staticmethod
def get_result(name):
return Task.objects.get(name=name).result
if Task.objects.filter(name=name).exists():
return Task.objects.get(name=name).result
return None
def time_taken(self):
return (self.stopped - self.started).total_seconds()

View File

@@ -1 +0,0 @@
__author__ = 'ilan'

View File

@@ -103,3 +103,5 @@ USE_TZ = True
STATIC_URL = '/static/'
# Django Q specific
#Q_PREFIX = 'test_django_q'

130
django_q/tests/test_q.py Normal file
View File

@@ -0,0 +1,130 @@
import sys
import os
from time import sleep
from multiprocessing import Queue, Event
import pytest
myPath = os.path.dirname(os.path.abspath(__file__))
sys.path.insert(0, myPath + '/../')
from django_q.core import Cluster, r, async, Q_LIST, pusher, worker, monitor
from django_q.humanhash import DEFAULT_WORDLIST
from django_q import result
class WordClass(object):
def __init__(self):
self.word_list = DEFAULT_WORDLIST
def get_words(self):
return self.word_list
def test_admin_view(admin_client):
response = admin_client.get('/admin/django_q/')
assert response.status_code == 200
response = admin_client.get('/admin/django_q/failure/')
assert response.status_code == 200
response = admin_client.get('/admin/django_q/success/')
assert response.status_code == 200
def test_redis_connection():
assert r.ping() is True
def test_cluster_initial():
c = Cluster()
assert c.sentinel is None
assert c.is_idle
c.start()
while c.is_starting:
sleep(0.2)
assert c.sentinel.is_alive() is True
assert c.is_running
c.stop()
while c.is_stopping:
sleep(0.2)
assert c.sentinel.is_alive() is False
assert c.has_stopped
@pytest.mark.django_db
def test_cluster():
task = async('django_q.tests.tasks.count_letters', DEFAULT_WORDLIST)
task_count = r.llen(Q_LIST)
assert task_count >= 1
task_queue = Queue()
assert task_queue.qsize() == 0
result_queue = Queue()
assert result_queue.qsize() == 0
event = Event()
event.set()
# Test push
pusher(task_queue, event)
assert task_queue.qsize() == 1
assert r.llen(Q_LIST) == task_count - 1
# Test work
task_queue.put('STOP')
worker(task_queue, result_queue)
assert task_queue.qsize() == 0
assert result_queue.qsize() == 1
# Test monitor
result_queue.put('STOP')
monitor(result_queue)
assert result_queue.qsize() == 0
# check result
assert result(task) == 1506
@pytest.mark.django_db
def test_async():
a = async('django_q.tests.tasks.count_letters', DEFAULT_WORDLIST, hook='django_q.tests.test_q.assert_result')
b = async('django_q.tests.tasks.count_letters2', WordClass(), hook='django_q.tests.test_q.assert_result')
c = Cluster()
assert c.start() > 0
while not c.is_running:
sleep(0.5)
assert isinstance(a, str)
assert isinstance(b, str)
while c.stat.task_q_size > 0 and c.stat.done_q_size > 0:
sleep(0.5)
assert c.stop() is True
result_a = result(a)
assert result_a is not None
assert result_a.success is True
assert result_a.result == 1506
result_b = result(b)
assert result_b is not None
assert result_b.success is True
assert result_b.result == 1506
@pytest.mark.django_db
def assert_result(task):
assert task is not None
assert task.success is True
assert task.result == 1506
@pytest.mark.django_db
def broken_package():
a = async('django_q.tests.tasks.count_letters', DEFAULT_WORDLIST, 'oneargumentoomany',
hook='django_q.tests.test_q.assert_bad_result')
b = async('django_q.tests.tasks.does_not_exist', WordClass(), hook='django_q.tests.test_q.assert_bad_result')
assert isinstance(a, str)
assert isinstance(b, str)
sleep(5)
result_a = result(a)
assert result_a.success is False
result_b = result(b)
assert result_b.success is False
@pytest.mark.django_db
def assert_bad_result(task):
assert task is not None
assert task.success is False

View File

@@ -1,3 +1,2 @@
[pytest]
DJANGO_SETTINGS_MODULE=django_q.tests.settings
#django_find_project = false

View File

@@ -1,3 +1,5 @@
blessed
future
-e .
coloredlogs==1.0.1
django-picklefield==0.3.1

View File

@@ -1,16 +1,15 @@
import os
from setuptools import setup
with open(os.path.join(os.path.dirname(__file__), 'README.rst')) as readme:
README = readme.read()
os.chdir(os.path.normpath(os.path.join(os.path.abspath(__file__), os.pardir)))
from distutils.core import setup, Command
from setuptools import setup, Command
# you can also import from setuptools
class PyTest(Command):
user_options = []
def initialize_options(self):
pass
@@ -20,9 +19,11 @@ class PyTest(Command):
def run(self):
import subprocess
import sys
errno = subprocess.call([sys.executable, 'runtests.py'])
raise SystemExit(errno)
setup(
name='django-q',
version='0.1.0',
@@ -33,10 +34,8 @@ setup(
license='MIT',
description='A multiprocessing task queue for Django',
long_description=README,
include_package_data=True,
install_requires=['django>=1.7', 'redis', 'coloredlogs', 'django-picklefield', 'jsonpickle'],
test_suite='django_q.tests',
cmdclass = {'test': PyTest},
cmdclass={'test': PyTest},
classifiers=[
'Development Status :: 2 - PreAlpha',
'Environment :: Web Environment',