Fetch me that JSON from Django
The new Fetch API is similar to the renowned XMLHttpRequest that shook up the internet in 2005 with Ajax. Simply put, Fetch is better, providing more power and flexibility to the developer. It uses promises instead of callbacks, and defines Headers, Request, and Response objects to manipulate requests. To use the Fetch API to request data from a Django application, we also need to be aware of a few differences between XMLHttpRequest and Fetch.
The first noticible difference is the prominent use of Promises to make the asyncronous handling of requesting new data a breeze. Fetch uses this standard ES6 feature instead of callbacks, returning a promise with the response data:
Notice that in this example we are processing the response as text, which is the default content type Django will reply with. What if we wanted to retrieve JSON with Fetch, but still reply with text HTML for browser navigations? With Fetch requests, we can set an
Accept header to request the
application/json content type.
In the Django application, chances are you’d use the
is_ajax() to check for an Ajax request, and respond with JSON. This function simply inspects the
X-Requested-With header to see if the request was initiated with XMLHttpRequest. Unfortunately, this header isn’t automatically set by Fetch because it’s a convention, not a standard as described in this Github issue.
So, we need to set these headers manually by passing an options object to the call to
Another critical difference between Fetch and XMLHttpRequest is the handling of credentials. By default, Fetch doesn’t pass any Cookies along with the request, neither browser cookies nor HTTP cookies. For browser cookies, just set another
"include", and defaults to
If your application doesn’t have users, you don’t need to worry about the credentials here. But, if you have users and your application is using Django’s cookie-based sessions you’ll need to make sure the
sessionid cookie is passed to the application with the request. This is just as easy as setting headers, and we simply set the
credentials option to
That should be it! This may look like more code than you should need but there’s a lot of added benefit from using the new Fetch API. The request-response cycle has never been easier to manage using standard browser features, and response processing is dead simple as well.