Check out Part I here
Check out Part II 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 compatible
- Request Handlers
- Routing: simple and parameterized
- Check for duplicate routes
- Class Based Handlers
- Unit tests
In this part, we will add a few awesome features to the list:
- Test Client
- Alternative way to add routes (like Django)
- Support for templates
Test Client
In the part 2, we wrote a couple of unit tests. However, we stopped when we needed to send HTTP requests to our handlers because we didn't have a test client that could do that. Let's add one then.
By far the most popular way of sending HTTP requests in Python is the Requests
library by Kenneth Reitz. However, for us to be able to use it in the unit tests, we should always have our app up and running (i.e. start gunicorn before running tests). The reason is that Requests only ships with a single Transport Adapter, the HTTPAdapter. That defeats the purpose of unit tests. Unit tests should be self sustained. Fortunately for us, Sean Brant wrote a WSGI Transport Adapter for Requests that we can use to create a test client. Let's write the code first and then discuss.
Add the following method to the main API
class in api.py
:
# api.py
...
from requests import Session as RequestsSession
from wsgiadapter import WSGIAdapter as RequestsWSGIAdapter
class API:
...
def test_session(self, base_url="http://testserver"):
session = RequestsSession()
session.mount(prefix=base_url, adapter=RequestsWSGIAdapter(self))
return session
...
As written here, to use the Requests WSGI Adapter, we need to mount it to a Session object. This way, any request made using this test_session
whose URL starts with the given prefix will use the given RequestsWSGIAdapter. Great, now we can use this test_session
to create a test client. Create a conftest.py
file and move the api
fixture to this file so that it looks like this:
# conftest.py
import pytest
from api import API
@pytest.fixture
def api():
return API()
In case you didn't know, this file is where pytest
looks for fixtures by default. Now, let's create the test client fixture here:
# conftest.py
...
@pytest.fixture
def client(api):
return api.test_session()
Our client
needs the api
fixture and returns the test_session
that we wrote earlier. Now we can use this client
fixture in our unit tests. Let's go right ahead to the test_bumbo.py
file and write a unit test that tests if the client
can send a request:
# test_bumbo.py
...
def test_bumbo_test_client_can_send_requests(api, client):
RESPONSE_TEXT = "THIS IS COOL"
@api.route("/hey")
def cool(req, resp):
resp.text = RESPONSE_TEXT
assert client.get("http://testserver/hey").text == RESPONSE_TEXT
Run the unit tests by pytest test_bumbo.py
and voila. We see that all the tests pass. Let's add a couple more unit tests for the most important parts:
# test_bumbo.py
...
def test_parameterized_route(api, client):
@api.route("/{name}")
def hello(req, resp, name):
resp.text = f"hey {name}"
assert client.get(url("/matthew")).text == "hey matthew"
assert client.get(url("/ashley")).text == "hey ashley"
This tests that the parameters that we send in the url are working.
# test_bumbo.py
...
def test_default_404_response(client):
response = client.get(url("/doesnotexist"))
assert response.status_code == 404
assert response.text == "Not found."
This one tests that if a request is sent to a non existent route, 404(Not Found) response is returned.
The rest I will leave to you. Try to write a couple more tasks and let me know in the comments if you need any help. Here are some ideas for unit tests:
- test that class based handlers are working with a GET request
- test that class based handlers are working with a POST request
- test that class based handlers are returning
Method Not Allowed.
response if an invalid request method is used - test that status code is being returned properly
Alternative way to add routes
Right now, here is how routes are added:
@api.route("/home")
def handler(req, resp):
resp.text = "YOLO"
That is, routes are added as decorators, like in Flask. Some people may like the Django way of registering urls. So, let's give them a choice to add routes like this:
def handler(req, resp):
resp.text = "YOLO"
def handler2(req, resp):
resp.text = "YOLO2"
api.add_route("/home", handler)
api.add_route("/about", handler2)
This add_route
method should do two things. Check if the route is already registered or not and if not, register it:
# api.py
class API:
...
def add_route(self, path, handler):
assert path not in self.routes, "Such route already exists."
self.routes[path] = handler
Pretty simple. Does this code look familiar to you? It is because we already wrote such code in the route
decorator. We can now follow the DRY principle and use this add_route
method inside the route
decorator:
# api.py
class API:
...
def add_route(self, path, handler):
assert path not in self.routes, "Such route already exists."
self.routes[path] = handler
def route(self, pattern):
def wrapper(handler):
self.add_route(pattern, handler)
return handler
return wrapper
And let's add a unit test to check if it is working:
# test_bumbo.py
def test_alternative_route(api, client):
response_text = "Alternative way to add a route"
def home(req, resp):
resp.text = response_text
api.add_route("/alternative", home)
assert client.get("http://testserver/alternative").text == response_text
Run your tests and you will see that all of them pass.
Templates support
When I am implementing something new, I like to do something called README driven development. It is a technique where you write down how you want your API to look like before implementing. Let's do just that for this feature. Say we have this template that we want to use in our handler:
<html>
<header>
<title></title>
</header>
<body>
The name of the framework is
</body>
</html>
and are variables that are sent from a handler and here is how a handler looks like:
api = API(templates_dir="templates")
@api.route("/home")
def handler(req, resp):
resp.body = api.template("home.html", context={"name": "Alcazar"})
I want it to be as simple as possible so I just need one method that takes template name and context as params and renders that template with the given params. Also, we want templates directory to be configurable just like above.
With the API designed, we can now implement it.
For templates support, I think that Jinja2 is the best choice. It is a modern and designer-friendly templating language for Python, modelled after Django’s templates. So, if you know Django it should feel right at home.
Jinja2
uses a central object called the template Environment
. We will configure this environment upon application initialization and load templates with the help of this environment. Here is how to create and configure one:
from jinja2 import Environment, FileSystemLoader
templates_env = Environment(loader=FileSystemLoader(os.path.abspath("templates")))
FileSystemLoader
loads templates from the file system. This loader can find templates in folders on the file system and is the preferred way to load them. It takes the path to the templates directory as a parameter. Now we can use this templates_env
like so:
templates_env.get_template("index.html").render({"title": "Awesome Framework", "name": "Alcazar"})
Now that we understand how everything works in Jinja2
, let's add it to our own framework. First, let's install Jinja2
:
pip install Jinja2
Then, create the Environment
object in the __init__
method of our API
class:
# api.py
from jinja2 import Environment, FileSystemLoader
class API:
def __init__(self, templates_dir="templates"):
self.routes = {}
self.templates_env = Environment(loader=FileSystemLoader(os.path.abspath(templates_dir)))
...
We did almost the same thing as above except that we gave templates_dir
a default value of templates
so that users don't have to write it if they don't want to. Now we have everything to implement the template
method we designed earlier:
# api.py
class API:
def template(self, template_name, context=None):
if context is None:
context = {}
return self.templates_env.get_template(template_name).render(**context)
I don't think there is a need to explain anything here. The only thing you may wonder about is why I gave context
a default value of None
, checked if it is None
and then set the value to an empty dictionary {}
. You may say I could have given it the default value of {}
in the declaration. But dict
is a mutable object and it is a bad practice to set a mutable object as a default value in Python. Read more about this here.
With everything ready, we can create templates and handlers. First, create the templates
folder:
mkdir templates
Create the index.html
file by doing touch templates/index.html
and put the following inside:
<html>
<header>
<title></title>
</header>
<body>
<h1>The name of the framework is </h1>
</body>
</html>
Now we can create a handler in our app.py
:
# app.py
@api.route("/template")
def template_handler(req, resp):
resp.body = app.template("index.html", context={"name": "Alcazar", "title": "Best Framework"})
That's it (well, almost). Start gunicorn
and go to http://localhost:8000/template
. You will see a big bold Internal Server Error
. That's because resp.body
expects bytes and our template
method returns a unicode string. Thus, we will need to encode it:
# app.py
@api.route("/template")
def template_handler(req, resp):
resp.body = app.template("index.html", context={"name": "Alcazar", "title": "Best Framework"}).encode()
Restart gunicorn and you will see our template in all its glory. In the future posts, we will remove the need to encode
and make our API prettier.
Conclusion
We have implemented three new features in this post:
- Test Client
- Alternative way to add routes (like Django)
- Support for templates
Make sure to let me know in the comments what other features we should implement in this series. For the next part, we will definitely add support for static files but I am not sure what other features we should add.
Check out Part I here
Check out Part II 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 for today!
Fight on!
from Planet Python
via read more
No comments:
Post a Comment