day-3 路由底层源码

1. 定义路由本质

比如在url.py定义以下路由，浏览器中输入http://192.168.0.1:8000/user/2003-04-21可以访问

意味着此url http://192.168.0.1:8000/user/2003-04-21 和url.py里的路由们做了路由匹配

如果匹配成功找到相应的试图函数

源码解析

ctrl+鼠标左键 点进re_path,会发现

如果re_path = partial(....)

那么上面的url re_path() = partial(), 这里的_path 和Pattern代表什么接着往后看

partial偏函数：

作用：_path()里面规定了要传递一下参数，而path = partial(...)里面传递了Pattern=RegxPatten

那么我们在使用时即不需要再次传递Pattern

点击_path 恍然大悟，_path接收route, view, kwargs, name,Pattern

route view 在url.py已经规定好了, Pattren在partial函数中规定好了

那么这里url.py里的re_path函数就可以写成

partial(_path(route, view, kwargs=None, name=None, Pattern=None), Pattern=RegexPattern)
替换掉route 和 view

partial(_path(r'user/(\d{4})-(\d{2})-(\d{2})', view.user, kwargs=None, name=None, Pattern=None), Pattern=RegexPattern)
因为partial是偏函数,最后路由变成

_path(r'user/(\d{4})-(\d{2})-(\d{2})', view.user, kwargs=None, name=None, Pattern=RegexPattern)

点进_path看看这个函数干啥呢

现在这是我们的路由

re_path(r'user/(\d{4})-(\d{2})-(\d{2})', view.user)就变成了下面这样

URLPattern(Pattern(r'user/(\d{4})-(\d{2})-(\d{2})', name=None, is_endpoint=True), view.user, None, None),

因为规定Pattern = RegxPattern

URLPattern(RegxPattern(r'user/(\d{4})-(\d{2})-(\d{2})', name=None, is_endpoint=True), view.user, None, None),

两个类的关系 URLPattern RegexPattern

URLPattern默认接受四个参数并封装

封装后的

self.pattern = RegxPattern(r'user/(\d{4})-(\d{2})-(\d{2})', name=None, is_endpoint=True）

self.callback = view.user

其他为None

最后到了RegxPattern类，这里会去封装self._regex

上面传递了RegxPattern(r'user/(\d{4})-(\d{2})-(\d{2})', name=None, is_endpoint=True）

那么self._regex= r'user/(\d{4})-(\d{2})-(\d{2})'

两层封装结束,现在

URLPattern.pattern._regex = r'user/(\d{4})-(\d{2})-(\d{2})'

self.callback = view.user

2.路由匹配流程

-启动项目

-请求到来

http://127.0.0.1:8000/user/

与定义好的路由逐一匹配

那么路由是怎么找到的呢

wsgi请求入口

往下找到wsgiHandler函数里的__call__方法，

此函数为类方法入口

environ为每次请求时浏览器携带的参数

request = self.request_class(environ)
因为规定

request_class = WSGIRequest
所以

request = WSGIRequest(environ)

封装好后可以通过request.POST request.GET request.method request.FILES 拿到想要的数据

request.path_info 可以拿到地址

比如 https://192.168.0.1:8000/user/, path_info 可以拿到/user/

下一步通过request 进行函数处理拿到response， 点进get_response

点进get_response发现又进行了_middleware_chain处理

再点 _middleware_chain = handler

往上翻找handler到底是什么

传入的mw_instance

mw_instance = middleware(adapted_handler)

又绕回到上面的adapt_method_mode方法

所以adapted_handler = handler

现在确定handler就是_get_response

下图描述有误，匹配后

callback为完整路径， callback_args为浏览器url传递进的参数

比如url.py规定 re_path('/user/(\d+)')

如访问url为 http://192.168.0.1:8000:/user/123

callback为路经所在内存地址， args为123

现在具体观察这个方法

这里分成两部分查看

1. get_resolver()

进一步查看_get_cached_resolver(urlconf)

URlResolver分析完毕，下面看URLResolver里的resolve方法， 图片注释有误

import_module相当于 from day006 import urls

urls.py被导入之后,patterns = urls.urlpatterns

然后就可进行循环

从最开始的讲解

urls.py里的每一条数据

比如 path('user/(\d{4})-(\d{2})-(\d{2})', view.user)也就等于

URLPattern(RegxPattern(r'user/(\d{4})-(\d{2})-(\d{2})', name=None, is_endpoint=True), view.user, None, None),

每次for pattern做匹配的时候也就是相当于

URLPattern(RegxPattern(r'user/(\d{4})-(\d{2})-(\d{2})', name=None, is_endpoint=True), view.user, None, None),

找到它下面的resolve方法 传入'user/'做验证

这里的resolve就在URLPattern.resolve(new_path)

因为一个re_path() = URLPattern(RegxPattern(r'user/(\d{4})-(\d{2})-(\d{2})', name=None, is_endpoint=True), view.user, None, None),

URLPattern()这里传递的第一个参数为RegxPattern(r'user/(\d{4})-(\d{2})-(\d{2})'

所以self.pattern = RegxPattern(r'user/(\d{4})-(\d{2})-(\d{2})'

这里return的返回值