Django Suspense is small package to easily display a fallback in templates until children have finished loading.
To get started, install the package from pypi:
pip install django-suspense
Now you can add suspense
to your django project. Change your INSTALLED_APPS
setting like this:
INSTALLED_APPS = [
...,
"suspense",
]
Optionally, you can specify suspense
as builtins and this will be available in any of your templates without additionally specifying {% load suspense %}
:
TEMPLATES = [
{
"BACKEND": "django.template.backends.django.DjangoTemplates",
"DIRS": [BASE_DIR / "templates"],
"APP_DIRS": True,
"OPTIONS": {
"context_processors": [
# ... list of default processors
],
"builtins": ["suspense.templatetags.suspense"], # <--
},
},
]
If you choose not use it as a built-in, you will need to add {% load suspense %}
to the top of your template whenever you want to use suspense.
Because django executes database queries lazily, they may sometimes not work as expected. Let's try to create a very slow but lazy object and write a view function:
from suspense.shortcuts import render
# app/views.py
def view(request):
def obj():
import time
time.sleep(1)
return range(10)
return render(request, 'template.html', {'obj': obj})
Let's now add the output of the received data to the template. At this point, we still haven't made a database query, so we can easily and quickly show the template right away.
{% load suspense %}
<ul>
{% suspense %}
{% fallback %}
<li class="skeleton">Loading ... </li>
{% endfallback %}
{% for data in obj %}
<li>{{ data }}</li>
{% endfor %}
{% endsuspense %}
</ul>
Once obj is ready for use, we will show it. But until it is ready, fallback works. While we are waiting for the data to be displayed, a request is made on the client side.
Suspense does not add additional DOM elements after rendering the final result. That's why syntactically the code above will be valid.
On Safari if your webpage is very light/simple, you may experience a delay in rendering.
Ex: the page renders only after some django-suspense or all content is downloaded.
WebKit has an issue with streaming responses requiring a certain amount of visible content before to actually start rendering.
If you are experiencing this issue, you can use the additional {% webkit_extra_invisible_bytes %}
template tag to add a few extra invisible bytes in Safari.
{% load suspense %}
{% webkit_extra_invisible_bytes %}
By default the webkit_extra_invisible_bytes
adds 200 bytes but you can specify a different amount:
{% webkit_extra_invisible_bytes 300 %}
If you are using a Content Security Policy (CSP) with nonce
and strict-dynamic
, you may need to add the nonce
attribute to the script tag.
You can override the suspense/replacer.html
template and add the nonce
attribute to the script tag.
With django-csp:
{% extends "suspense/replacer.html" %}
{% block script_attributes %}nonce="{{request.csp_nonce}}"{% endblock %}
Async views with an ASGI server is also supported.
import asyncio
from suspense.shortcuts import async_render
# app/views.py
async def view(request):
async def obj():
await asyncio.sleep(1)
return range(10)
return async_render(request, 'template.html', {'obj': obj()})
Suspense will wait for any awaitable object to finish before rendering the suspense tags.
If you have multiple suspense blocks with different awaitable, you can specify which awaitable to wait for or each suspense block will await everything.
Ex: {% suspense obj obj2 %}
{% load suspense %}
<ul>
{% suspense obj %}
{% fallback %}
<li class="skeleton">Loading ... </li>
{% endfallback %}
{% for data in obj %}
<li>{{ data }}</li>
{% endfor %}
{% endsuspense %}
{% suspense obj2 %}
{% fallback %}
<li class="skeleton">Loading 2... </li>
{% endfallback %}
{% for data in obj2 %}
<li>{{ data }}</li>
{% endfor %}
{% endsuspense %}
</ul>
Important: If your async context variable is used by more than one suspense block, or you did not specify any variables on the tags, make sure to wrap your coroutines in tasks so they can be awaited multiple times.
Ex: asyncio.create_task(obj())
import asyncio
from suspense.shortcuts import async_render
# app/views.py
async def view(request):
async def obj():
await asyncio.sleep(1)
return range(10)
task_obj = asyncio.create_task(obj())
return async_render(request, 'template.html', {'obj': task_obj})
- synchronous streaming response with AGSI will wait for the full render before sending the response to the client.
If you would like to suggest a new feature, you can create an issue on the GitHub repository for this project. Also you can fork the repository and submit a pull request with your changes.
This project is licensed under the MIT License. See the LICENSE file for more information.