nova的wsgi介绍【WIP】
有关openstack的所有的帖子。
https://www.ustack.com/blog/openstack_hacker/#Nova_Workflow
今天liyong同学问起pecan的原理。
有必要写篇博客介绍， nova的wsgi。
改博客边走读代码，边写，是一个野路子开发者的流程。
不习惯使用 tox.
$ cat /usr/bin/pyunittest
python -m unittest `tr / . <<< ${@%.py*}`
1. 首先对wsgi有个形象的了解：
nova/tests/unit/api/test_wsgi.py
""" Test WSGI basics and provide some helper functions for other WSGI tests. """ from nova import test import routes import webob from nova import wsgi class Test(test.NoDBTestCase): def test_debug(self): class Application(wsgi.Application): """Dummy application to test debug.""" def __call__(self, environ, start_response): start_response("200", [("X-Test", "checking")]) return ['Test result']
application = wsgi.Debug(Application()) result = webob.Request.blank('/').get_response(application) self.assertEqual(result.body, "Test result") def test_router(self): class Application(wsgi.Application): """Test application to call from router.""" def __call__(self, environ, start_response): start_response("200", []) return ['Router result'] class Router(wsgi.Router): """Test router.""" def __init__(self): mapper = routes.Mapper() mapper.connect("/test", controller=Application()) super(Router, self).__init__(mapper) result = webob.Request.blank('/test').get_response(Router()) self.assertEqual(result.body, "Router result") result = webob.Request.blank('/bad').get_response(Router()) self.assertNotEqual(result.body, "Router result")
自己可以print打桩，了解有疑惑的代码。 然后运行。
$ pyunittest nova/tests/unit/api/test_wsgi.py
查看wsgi的实现。
$ git grep -n3 "class Router"
--
nova/wsgi.py-440- print()
nova/wsgi.py-441-
nova/wsgi.py-442-
nova/wsgi.py:443:class Router(object):
nova/wsgi.py-444- """WSGI middleware that maps incoming requests to WSGI apps."""
nova/wsgi.py-445-
nova/wsgi.py-446- def __init__(self, mapper):
class Router(object): """WSGI middleware that maps incoming requests to WSGI apps.""" def __init__(self, mapper): """Create a router for the given routes.Mapper. Each route in `mapper` must specify a 'controller', which is a WSGI app to call. You'll probably want to specify an 'action' as well and have your controller be an object that can route the request to the action-specific method. Examples: mapper = routes.Mapper() sc = ServerController() # Explicit mapping of one route to a controller+action mapper.connect(None, '/svrlist', controller=sc, action='list') # Actions are all implicitly defined mapper.resource('server', 'servers', controller=sc) # Pointing to an arbitrary WSGI app. You can specify the # {path_info:.*} parameter so the target app can be handed just that # section of the URL. mapper.connect(None, '/v1.0/{path_info:.*}', controller=BlogApp()) """ self.map = mapper self._router = routes.middleware.RoutesMiddleware(self._dispatch, self.map) @webob.dec.wsgify(RequestClass=Request) def __call__(self, req): """Route the incoming request to a controller based on self.map. If no match, return a 404. """ return self._router @staticmethod @webob.dec.wsgify(RequestClass=Request) def _dispatch(req): """Dispatch the request to the appropriate controller. Called by self._router after matching the incoming request to a route and putting the information into req.environ. Either returns 404 or the routed WSGI app's response. """ match = req.environ['wsgiorg.routing_args'][1] if not match: return webob.exc.HTTPNotFound() app = match['controller'] return app
routes.middleware.RoutesMiddleware的帮助文档。Constructor information: Definition:routes.middleware.RoutesMiddleware(self, wsgi_app, mapper, use_method_override=True, path_info=True, singleton=True) Docstring: Create a Route middleware object Using the use_method_override keyword will require Paste to be installed, and your application should use Paste's WSGIRequest object as it will properly handle POST issues with wsgi.input should Routes check it. If path_info is True, then should a route var contain path_info, the SCRIPT_NAME and PATH_INFO will be altered accordingly. This should be used with routes like: .. code-block:: python map.connect('blog/*path_info', controller='blog', path_info='')
文档中明确说明： use Paste's WSGIRequest
nova/wsgi.py 实现了load app功能， 从
class Loader(object): """Used to load WSGI applications from paste configurations.""" def __init__(self, config_path=None): """Initialize the loader, and attempt to find the config. :param config_path: Full or relative path to the paste config. :returns: None """ self.config_path = None config_path = config_path or CONF.api_paste_config if not os.path.isabs(config_path): self.config_path = CONF.find_file(config_path) elif os.path.exists(config_path): self.config_path = config_path if not self.config_path: raise exception.ConfigNotFound(path=config_path) def load_app(self, name): """Return the paste URLMap wrapped WSGI application. :param name: Name of the application to load. :returns: Paste URLMap object wrapping the requested application. :raises: `nova.exception.PasteAppNotFound` """ try: LOG.debug("Loading app %(name)s from %(path)s", {'name': name, 'path': self.config_path}) return deploy.loadapp("config:%s" % self.config_path, name=name) except LookupError: LOG.exception(_LE("Couldn't lookup app: %s"), name) raise exception.PasteAppNotFound(name=name, path=self.config_path)
至于 server 怎么加载wsgi的，请参考：
nova/tests/unit/test_wsgi.py
$ pyunittest nova/tests/unit/test_wsgi.TestWSGIServerWithSSL.test_ssl_server
最终urlmap的事情交给 paste
请参考：
http://blog.csdn.net/sonicatnoc/article/details/6539716
一以下纯拷贝：
首先python paste是一个WSGI工具包，在WSGI的基础上包装了几层，让应用管理和实现变得方便。说实话，Python Paste的文档做的真差劲！加之python代码可读性本来就不怎么滴，真费劲。
paste.deploy关键部分留个抓印：
1）python paste.deploy不能只装个paste.deploy包就可以工作了，还需要paste.script包
2）python paste.deploy中loadapp给的路径可用os.path.abspath(配置文件相对路径)得到配置文件的绝对路径，否则报找不到relative_to path...没搞明白怎么回事，目前不重要，放过。
3）python paste.deploy中filter,filter_factory,app,app_factory的规范在文档中都没怎么写清楚，我来给你补上吧：
- app是一个callable object，接受的参数(environ,start_response)，这是paste系统交给application的，符合
WSGI规范的参数. app需要完成的任务是响应envrion中的请求，准备好响应头和消息体，然后交给start_response处理，并返回响应消息体。
- filter是一个callable object，其唯一参数是(app)，这是WSGI的application对象，见（1），filter需要完成的工作是将application包 装成另一个application（“过滤”），并返回这个包装后的application。
- app_factory是一个callable object，其接受的参数是一些关于application的配置信息：(global_conf,**kwargs)，global_conf是在 ini文件中default section中定义的一系列key-value对，而**kwargs，即一些本地配置，是在ini文件中，app:xxx section中定义的一 系列key-value对。app_factory返回值是一个application对象
- filter_factory是一个callable object，其接受的参数是一系列关于filter的配置信息：(global_conf,**kwargs)，global_conf是在ini文件 中default section中定义的一系列key-value对，而**kwargs，即一些本地配置，是在ini文件中，filter:xxx section中定 义的一系列key-value对。filter_factory返回一个filter对象
给个例子：
pastedeploylab.ini:
[DEFAULT] key1=value1 key2=value2 key3=values [composite:pdl] use=egg:Paste#urlmap /:root /calc:calc [pipeline:root] pipeline = logrequest showversion [pipeline:calc] pipeline = logrequest calculator [filter:logrequest] username = root password = root123 paste.filter_factory = pastedeploylab:LogFilter.factory [app:showversion] version = 1.0.0 paste.app_factory = pastedeploylab:ShowVersion.factory [app:calculator] description = This is an "+-*/" Calculator paste.app_factory = pastedeploylab:Calculator.factory
pastedeploylab.py
''''' Created on 2011-6-12 @author: Sonic ''' import os import webob from webob import Request from webob import Response from paste.deploy import loadapp from wsgiref.simple_server import make_server #Filter class LogFilter(): def __init__(self,app): self.app = app pass def __call__(self,environ,start_response): print "filter:LogFilter is called." return self.app(environ,start_response) @classmethod def factory(cls, global_conf, **kwargs): print "in LogFilter.factory", global_conf, kwargs return LogFilter class ShowVersion(): def __init__(self): pass def __call__(self,environ,start_response): start_response("200 OK",[("Content-type", "text/plain")]) return ["Paste Deploy LAB: Version = 1.0.0",] @classmethod def factory(cls,global_conf,**kwargs): print "in ShowVersion.factory", global_conf, kwargs return ShowVersion() class Calculator(): def __init__(self): pass def __call__(self,environ,start_response): req = Request(environ) res = Response() res.status = "200 OK" res.content_type = "text/plain" # get operands operator = req.GET.get("operator", None) operand1 = req.GET.get("operand1", None) operand2 = req.GET.get("operand2", None) print req.GET opnd1 = int(operand1) opnd2 = int(operand2) if operator == u'plus': opnd1 = opnd1 + opnd2 elif operator == u'minus': opnd1 = opnd1 - opnd2 elif operator == u'star': opnd1 = opnd1 * opnd2 elif operator == u'slash': opnd1 = opnd1 / opnd2 res.body = "%s /nRESULT= %d" % (str(req.GET) , opnd1) return res(environ,start_response) @classmethod def factory(cls,global_conf,**kwargs): print "in Calculator.factory", global_conf, kwargs return Calculator() if __name__ == '__main__': configfile="pastedeploylab.ini" appname="pdl" wsgi_app = loadapp("config:%s" % os.path.abspath(configfile), appname) server = make_server('localhost',8088,wsgi_app) server.serve_forever() pass
使用：
http://127.0.0.1:8080/
输出：
Paste Deploy LAB: Version = 1.0.0
http://127.0.0.1:8080/calc?operator=plus&operand1=12&operand2=23
输出：
UnicodeMultiDict([('operator', u'plus'), ('operand1', u'12'), ('operand2', u'23')])
RESULT= 35
====================================================
进一步猜测filter的使用过程：在paste deploy库中应该有类似这样的一段代码对application进行重组包装：
#
# 假设在ini文件中, 某条pipeline的顺序是filter1, filter2, filter3
# app, 那么，最终运行的app_real是这样组织的：
#
app_real = filter1(filter2(filter3(app)))
# 在app真正被调用的过程中，filter1.__call__(environ,start_response)被首先调用，若某种检查未通过，filter1做出反应；否则交给filter2__call__(environ,start_response)进一步处理，若某种检查未通过，filter2做出反应，中断链条，否则交给filter3.__call__(environ,start_response)处理，若filter3的某种检查都通过了，最后交给app.__call__(environ,start_response)进行处理。
看到令贤写了一篇更简单的。