Check out Part I here
Check out Part II here
Check out Part III here
A little reminder that this series is based on the Alcazar framework that I am writing for learning purposes. If you liked this series, show some love by starring the repo.
In the previous blog posts in the series, we started writing our own Python framework and implemented the following features:
- WSGI compatibility
- Request Handlers
- Routing: simple and parameterized
- Check for duplicate routes
- Class Based Handlers
- Unit tests
- Test Client
- Alternative way to add routes (like Django)
- Support for templates
In this part, we will add a few more awesome features to the list:
- Custom exception handler
- Support for static files
- Middleware
Custom exception handler
Exceptions inevitably happen. Users may do something that we didn't expect. We may write some code that doesn't work on some occasions. Users may go to a non existent page. With what we have right now, if some exception happens, we show a big ugly Internal Server Error
message. Instead, we could show some nice one. Something along the lines of Oops! Something went wrong.
or Please, contact our customer support
. For that, we need to be able to catch those exceptions and handle them however we want.
It will look like this:
# app.py
from api import API
app = API()
def custom_exception_handler(request, response, exception_cls):
response.text = "Oops! Something went wrong. Please, contact our customer support at +1-202-555-0127."
app.add_exception_handler(custom_exception_handler)
Here we create a custom exception handler. It looks almost like our simple request handlers, except that it has exception_cls
as its third argument. Now, if we have a request handler that throws an exception, this above-mentioned custom exception handler should be called.
# app.py
@app.route("/home")
def exception_throwing_handler(request, response):
raise AssertionError("This handler should not be user")
If we go to http://localhost:8000/home
, instead of our previous big ugly Internal Server Error
, we should be able to see our custom message of Oops! Something went wrong. Please, contact our customer support at +1-202-555-0127.
. Does it look good enough? Let's go ahead and implement it.
The first thing we need is a variable inside our main API class where we will store our exception handler:
# api.py
class API:
def __init__(self, templates_dir="templates"):
...
self.exception_handler = None
Now we need to add the add_exception_handler
method:
# api.py
class API:
...
def add_exception_handler(self, exception_handler):
self.exception_handler = exception_handler
Having registered our custom exception handler, we need to call when an exception happens. Where do exceptions happen? That's right: when handlers are called. We call the handlers inside our handle_request
method. So, we need to wrap it with a try/except clause and call our custom exception handler in the except
part:
# api.py
class API:
...
def handle_request(self, request):
response = Response()
handler, kwargs = self.find_handler(request_path=request.path)
try:
if handler is not None:
if inspect.isclass(handler):
handler = getattr(handler(), request.method.lower(), None)
if handler is None:
raise AttributeError("Method now allowed", request.method)
handler(request, response, **kwargs)
else:
self.default_response(response)
except Exception as e:
if self.exception_handler is None:
raise e
else:
self.exception_handler(request, response, e)
return response
We also need to make sure that if no exception handler has been registered, the exception is propagated.
We have everything in place. Go ahead and restart your gunicorn and go to http://localhost:8000/home
. You should see our little cute message instead of the big ugly default one. Of course, make sure that you have the above mentioned exception handler and the errorful request handler in the app.py
.
If you want to go one step further, create a nice template and use our api.template()
method inside the exception handler. However, our framework doesn't support static files and thus you will have hard time designing your template with CSS and JavaScript. Don't get sad because this is exactly what we are doing next.
Support for static files
Templates are not truly templates without good CSS and JavaScript, are they? Shall we add a support for such files then?
Just like we used Jinja2 for template support, we will use WhiteNoise for static file serving. Install it:
pip install whitenoise
WhiteNoise is pretty simple. The only thing that we need to do is wrap our WSGI app and give it the static folder path as a parameter. Before we do that, let's remember how our __call__
method looks like:
# api.py
class API:
...
def __call__(self, environ, start_response):
request = Request(environ)
response = self.handle_request(request)
return response(environ, start_response)
...
This is basically an entrypoint to our WSGI app and this is exactly what we need to wrap with WhiteNoise. Thus, let's refactor its content to a separate method so that it will be easier to wrap it with WhiteNoise:
# api.py
class API:
...
def wsgi_app(self, environ, start_response):
request = Request(environ)
response = self.handle_request(request)
return response(environ, start_response)
def __call__(self, environ, start_response):
return self.wsgi_app(environ, start_response)
Now, in our constructor, we can initialize a WhiteNoise instance:
# api.py
...
from whitenoise import WhiteNoise
class API:
...
def __init__(self, templates_dir="templates", static_dir="static"):
self.routes = {}
self.templates_env = Environment(loader=FileSystemLoader(templates_dir))
self.exception_handler = None
self.whitenoise = WhiteNoise(self.wsgi_app, root=static_dir)
As you can see, we wrapped our wsgi_app
with WhiteNoise and gave it a path to the static folder as the second param. The only thing left to do is make this self.whitenoise
an entrypoint to our framework:
# api.py
class API:
...
def __call__(self, environ, start_response):
return self.whitenoise(environ, start_response)
With everything in place, create static
folder in the project root, create the main.css
file inside and put the following into it:
body {
background-color: chocolate;
}
In the third blog post, we created the templates/index.html
. Now we can put our newly created css file inside this template:
<html>
<header>
<title>{{ title }}</title>
<link href="/main.css" type="text/css" rel="stylesheet">
</header>
<body>
<h1>The name of the framework is {{ name }}</h1>
</body>
</html>
Restart your gunicorn and go to http://localhost/template
. You should see that the color of the whole background is chocolate, not white, meaning that our static file is being served. Awesome!
Middleware
If you need a little recap of what middlewares are and how they work, go read this post first. Otherwise, this part may seem a little confusing. I will wait. Back? Great. Let's go.
You know what they are and how they work but you may be wondering what they are used for. Basically, middleware is a component that can modify an HTTP request and/or response and is designed to be chained together to form a pipeline of behavioral changes during request processing. Examples of middleware tasks can be request logging and HTTP authentication. The main point is that none of these is fully responsible for responding to a client. Instead, each middleware changes the behavior in some way as part of the pipeline, leaving the actual response to come from something later in the pipeline. In our case, that something that actually responds to a client is our request handlers. Middlewares are wrappers around our WSGI app that have the ability to modify requests and responses.
From the bird's eye view, the code will look like this:
FirstMiddleware(SecondMiddleware(our_wsgi_app))
So, when a request comes in, it first goes to FirstMiddleware
. It modifies the request and sends it over to SecondMiddleware
. Now, SecondMiddleware
modifies the request and sends it over to our_wsgi_app
. Our app handles the request, prepares the response and sends it back to SecondMiddleware
. It can modify the response if it wants and send it back to FirstMiddleware
. It modifies the response and sends it back to the web server (e.g. gunicorn).
Let's go ahead and create a Middleware
class that other middlewares will inherit from and that wraps our wsgi app.
Create a middleware.py
file first:
touch middleware.py
Now, we can begin our Middleware class
:
# middleware.py
class Middleware:
def __init__(self, app):
self.app = app
As we mentioned above, it should wrap a wsgi app and in case of multiple middlewares that app
can also be another middleware.
As a base middleware class, it should also have the ability to add another middleware to the stack:
# middleware.py
class Middleware:
...
def add(self, middleware_cls):
self.app = middleware_cls(self.app)
It is simply wrapping the given middleware class around our current app.
It should also have its main methods which are request processing and response processing. For now, they will do nothing. The child classes that will inherit from this class will implement these methods:
# middleware.py
class Middleware:
...
def process_request(self, req):
pass
def process_response(self, req, resp):
pass
Now, the most important part, the method that handles incoming requests:
# middleware.py
class Middleware:
...
def handle_request(self, request):
self.process_request(request)
response = self.app.handle_request(request)
self.process_response(request, response)
return response
It first calls the self.process_request
to do something with the request. Then delegates the response creation to the app that it is wrapping. Finally, it calls the process_response
to do something with the response object. Then simply returns the response upward.
As middlewares are the first entrypoint to our app now, they are the ones called by our web server (e.g. gunicorn). Thus, middlewares should implement the WSGI entrypoint interface:
# middleware.py
from webob import Request
class Middleware:
def __call__(self, environ, start_response):
request = Request(environ)
response = self.app.handle_request(request)
return response(environ, start_response)
It is just a copy of the wsgi_app
function we created above.
With our Middleware class implemented, let's add it to our main API
class:
# api.py
...
from middleware import Middleware
class API:
def __init__(self, templates_dir="templates", static_dir="static"):
...
self.middleware = Middleware(self)
It wraps around self
which is our wsgi app. Now, let's give it the ability to add middlewares:
# api.py
class API:
...
def add_middleware(self, middleware_cls):
self.middleware.add(middleware_cls)
The only thing left to do is call this middleware in the entrypoint instead of our own wsgi app:
# api.py
class API:
...
def __call__(self, environ, start_response):
return self.middleware(environ, start_response)
Why do you ask? Because we are delegating the job of being an entrypoint to the middlewares now. Remember that we implemented WSGI entrypoint interface inside our Middleware
class. Let's go ahead now and create a simple middleware that simply prints to the console:
# app.py
from api import API
from middleware import Middleware
app = API()
...
class SimpleCustomMiddleware(Middleware):
def process_request(self, req):
print("Processing request", req.url)
def process_response(self, req, res):
print("Processing response", req.url)
app.add_middleware(SimpleCustomMiddleware)
...
Restart your gunicorn and go to any url (e.g. http://localhost:8000/home
). Everything should work just like before. The only exception is that those texts should appear in the console. Open your console and you should see the following:
Processing request http://localhost:8000/home
Processing response http://localhost:8000/home
There is a catch. Have you found it? Static files don't work now. The reason is that we stopped using WhiteNoise
. We removed it. Instead of calling WhiteNoise
, we are calling the middleware. Here is what we should do. We need to distinguish between requests for static files and the others. When a request is coming in for a static file, we should call WhiteNoise
. For others, we should call the middleware. The question is how do we distinguish between them. Right now, requests for static files look like this: http://localhost:8000/main.css
. Other requests look like this http://localhost:8000/home
. They look the same for our API
class. Thus we will add a root to the URLs of static files so that they look like this http://localhost:8000/static/main.css
. We will check if the request path starts with /static
. If so, we will call WhiteNoise
, otherwise we will call the middleware. We should also make sure to cut the /static
part. Otherwise WhiteNoise
won't find the files:
# api.py
class API:
...
def __call__(self, environ, start_response):
path_info = environ["PATH_INFO"]
if path_info.startswith("/static"):
environ["PATH_INFO"] = path_info[len("/static"):]
return self.whitenoise(environ, start_response)
return self.middleware(environ, start_response)
Now, in the templates, we should call static files like so:
<link href="/static/main.css" type="text/css" rel="stylesheet">
Go ahead and change your index.html
.
Restart your gunicorn and check that everything is working properly.
We will use this middleware feature in the future posts to add authentication to our apps.
I think that this middleware part is more difficult to understand compared to others. I also think that I didn't do a great job explaining it. Thus, please write the code, let it sink in and ask me questions in the comments if something is not clear.
Check out Part I here
Check out Part II here
Check out Part III here
A little reminder that this series is based on the Alcazar framework that I am writing for learning purposes. If you liked this series, show some love by starring the repo.
That's it from me today.
Fight on!