什么是stevedore？

stevedore是建立在setuptools的entry point的功能上的，用于python程序动态加载代码，在openstack中被多个组件使用：比如ceilometer，neutron的plugin。当然，你可以直接使用

python的某些黑魔法实现插件的加载，但太原始了。stevedore基于entry point提供了更高层次的封装。

stevedore的官方文档在此：http://docs.openstack.org/developer/stevedore/

学习和入门setuptools：http://www.360doc.com/content/14/0306/11/13084517_358166737.shtml

官方文档的部分翻译：http://www.360doc.com/content/14/0429/19/9482_373285413.shtml

来自华为孔令贤（源地址非该人博客）的setup.py详解：http://blog.sina.com.cn/s/blog_4951301d0101etvj.html

偶计划在sora项目中引入stevedore与oslo.config简化某些开发的组件，先是测试了stevedore，写了个简单的scheduler插件

环境准备：

安装stevedore库，及组织相关目录

pip install stevedore
mkdir sora
cd sora
mkdir scheduler  #scheduler在sora目录中

构建这样一个目录树：

步骤：

创建一个抽象类scheduler，新的plugin要继承scheduler并重写相关方法scheduler

#sora/scheduler/base.py
import abcclass scheduler(object):__metaclass__ = abc.ABCMeta@abc.abstractmethoddef scheduler(self,data):pass

继承base类创建插件simple与memory

#sora/scheduler/memory.py
import base
class memoryscheduler(base.scheduler):def scheduler(self,data):id = data[max(data['memory'])]return id

#sora/scheduler/simple.py
import base
import random
class simplescheduler(base.scheduler):def scheduler(self,data):id = data[random.choice(data['memory'])]return id

simple插件是随机选择一个节点id，而memory则选择内存剩余最多的节点id

编写setup.py

#sora/setup.py
from setuptools import setup, find_packages
setup(name='sora-scheduler',version='1.0',description='sora.scheduler',author='hochikong',author_email='hochikong',platforms=['Any'],scripts=[],# provides=['sora.scheduler',#           ],packages=find_packages(),include_package_data=True,entry_points={'sora.scheduler': ['memorybase = scheduler.memory:memoryscheduler','randombase = scheduler.simple:simplescheduler',],},zip_safe=False,
)

安装自己编写的包：

root@workgroup1:~/sora# python setup.py install
running install
running bdist_egg
running egg_info
creating sora_scheduler.egg-info
writing sora_scheduler.egg-info/PKG-INFO
writing top-level names to sora_scheduler.egg-info/top_level.txt
writing dependency_links to sora_scheduler.egg-info/dependency_links.txt
writing entry points to sora_scheduler.egg-info/entry_points.txt
writing manifest file 'sora_scheduler.egg-info/SOURCES.txt'
reading manifest file 'sora_scheduler.egg-info/SOURCES.txt'
writing manifest file 'sora_scheduler.egg-info/SOURCES.txt'
installing library code to build/bdist.linux-x86_64/egg
running install_lib
running build_py
creating build
creating build/lib.linux-x86_64-2.7
creating build/lib.linux-x86_64-2.7/scheduler
copying scheduler/__init__.py -> build/lib.linux-x86_64-2.7/scheduler
copying scheduler/base.py -> build/lib.linux-x86_64-2.7/scheduler
copying scheduler/memory.py -> build/lib.linux-x86_64-2.7/scheduler
copying scheduler/simple.py -> build/lib.linux-x86_64-2.7/scheduler
creating build/bdist.linux-x86_64
creating build/bdist.linux-x86_64/egg
creating build/bdist.linux-x86_64/egg/scheduler
copying build/lib.linux-x86_64-2.7/scheduler/__init__.py -> build/bdist.linux-x86_64/egg/scheduler
copying build/lib.linux-x86_64-2.7/scheduler/base.py -> build/bdist.linux-x86_64/egg/scheduler
copying build/lib.linux-x86_64-2.7/scheduler/memory.py -> build/bdist.linux-x86_64/egg/scheduler
copying build/lib.linux-x86_64-2.7/scheduler/simple.py -> build/bdist.linux-x86_64/egg/scheduler
byte-compiling build/bdist.linux-x86_64/egg/scheduler/__init__.py to __init__.pyc
byte-compiling build/bdist.linux-x86_64/egg/scheduler/base.py to base.pyc
byte-compiling build/bdist.linux-x86_64/egg/scheduler/memory.py to memory.pyc
byte-compiling build/bdist.linux-x86_64/egg/scheduler/simple.py to simple.pyc
creating build/bdist.linux-x86_64/egg/EGG-INFO
copying sora_scheduler.egg-info/PKG-INFO -> build/bdist.linux-x86_64/egg/EGG-INFO
copying sora_scheduler.egg-info/SOURCES.txt -> build/bdist.linux-x86_64/egg/EGG-INFO
copying sora_scheduler.egg-info/dependency_links.txt -> build/bdist.linux-x86_64/egg/EGG-INFO
copying sora_scheduler.egg-info/entry_points.txt -> build/bdist.linux-x86_64/egg/EGG-INFO
copying sora_scheduler.egg-info/not-zip-safe -> build/bdist.linux-x86_64/egg/EGG-INFO
copying sora_scheduler.egg-info/top_level.txt -> build/bdist.linux-x86_64/egg/EGG-INFO
creating dist
creating 'dist/sora_scheduler-1.0-py2.7.egg' and adding 'build/bdist.linux-x86_64/egg' to it
removing 'build/bdist.linux-x86_64/egg' (and everything under it)
Processing sora_scheduler-1.0-py2.7.egg
creating /usr/local/lib/python2.7/dist-packages/sora_scheduler-1.0-py2.7.egg
Extracting sora_scheduler-1.0-py2.7.egg to /usr/local/lib/python2.7/dist-packages
Adding sora-scheduler 1.0 to easy-install.pth fileInstalled /usr/local/lib/python2.7/dist-packages/sora_scheduler-1.0-py2.7.egg
Processing dependencies for sora-scheduler==1.0
Finished processing dependencies for sora-scheduler==1.0

尝试手动加载插件：

root@workgroup1:~/sora# python
Python 2.7.6 (default, Mar 22 2014, 22:59:56)
[GCC 4.8.2] on linux2
Type "help", "copyright", "credits" or "license" for more information.
>>> from scheduler.memory import memoryscheduler
>>> dt = {13:'id1',324:'id2',434:'id3','memory':[13,324,434]}
>>> driver = memoryscheduler()
>>> driver.scheduler(dt)
'id3'
>>>

不过手动加载并没有太大意义

使用stevedore的drivermanager：

>>> from stevedore import driver
>>> dt = {13:'id1',324:'id2',434:'id3','memory':[13,324,434]}
>>> mgr = driver.DriverManager(
...     namespace='sora.scheduler',
...     name='randombase',
...     invoke_on_load=True,         #设置为true，即载入后自动实例化插件类，如果是函数，则调用
... )
>>> mgr.driver.scheduler(dt)
'id3'
>>> mgr.driver.scheduler(dt)
'id2'
>>> mgr.driver.scheduler(dt)
'id3'
>>> mgr.driver.scheduler(dt)
'id2'
>>> mgr.driver.scheduler(dt)
'id1'
>>>

这里我导入了randombase，怎样调用plugin里的方法很明显了

顺带检查下python中我的包的安装状况：

root@workgroup1:~# cd /usr/local/lib/python2.7/dist-packages/
root@workgroup1:/usr/local/lib/python2.7/dist-packages# ls
amqp                                           OpenSSL
amqp-1.4.6.dist-info                           pbr
anyjson                                        pbr-1.3.0.dist-info
anyjson-0.3.3.egg-info                         pika
backports                                      pika-0.9.14.egg-info
backports.ssl_match_hostname-3.4.0.2.egg-info  psutil
billiard                                       psutil-2.2.1.egg-info
billiard-3.3.0.19.egg-info                     _psutil_linux.so
_billiard.so                                   _psutil_posix.so
bottle-0.12.8.egg-info                         pymongo
bottle.py                                      pymongo-2.8.egg-info
bottle.pyc                                     pyOpenSSL-0.15.1.dist-info
bson                                           python_etcd-0.3.3.egg-info
celery                                         pytz
celery-3.1.17.dist-info                        pytz-2015.2.dist-info
docker                                         six-1.9.0.dist-info
docker_py-1.0.0.egg-info                       six.py
easy-install.pth                               six.pyc
etcd                                           sora_scheduler-1.0-py2.7.egg
eventlet                                       SQLAlchemy-0.9.9-py2.7-linux-x86_64.egg
eventlet-0.17.1.dist-info                      stevedore
funtests                                       stevedore-1.6.0.dist-info
glances                                        tests
Glances-2.3.egg-info                           virtualenv-12.0.7.dist-info
greenlet-0.4.5.egg-info                        virtualenv.py
greenlet.so                                    virtualenv.pyc
gridfs                                         virtualenv_support
kombu                                          websocket
kombu-3.0.24.dist-info                         websocket_client-0.25.0.egg-info

可以看到里面有个sora_scheduler-1.0-py2.7.egg目录，查看一下：

我不太清楚为什么python在导入scheduler模块时，并不需要指明sora_scheduler-1.0-py2.7.egg，可能与egg有关，执行import scheduler时，导入的是sora_scheduler-1.0-py2.7.egg目录下

的scheduler包，同样的还有SQLAlchemy，有一个SQLAlchemy-0.9.9-py2.7-linux-x86_64.egg,导入时只需执行import sqlalchemy

疑问：

可能有人会问，如果我想添加新的插件要怎么办，我想，最好的方法就是修改setup.py，更新安装一次该包即可

参考：

http://docs.openstack.org/developer/stevedore/

http://www.360doc.com/content/14/0306/11/13084517_358166737.shtml

http://www.360doc.com/content/14/0429/19/9482_373285413.shtml