A framework for implementing real-time APIs in Django web applications
Django-RT is a framework for implementing real-time APIs in Django web applications. This framework follows the traditional publish/subscribe methodology, using Redis as the pub/sub provider.
Django-RT consists of two major components:
- Application API: This provides the necessary functionality to allow you to easily add a real-time API to your Django application.
- “Courier” HTTP server: A lightweight, asynchronous server designed solely to deliver real-time messages to web clients.
WARNING: Django-RT is currently in pre-alpha stages of development, and is not ready for public consumption. Expect major API changes and, potentially, security risks in pre-alpha releases. Use in production at your own risk!
- Web chat
- User notifications
- Real-time charts (server monitoring, stock tickers, etc)
- Simple to integrate with any existing Django project
- Asynchronous server design to handle many long-running client connections from a single thread
- Decoupled server architecture allows for the use of blocking code (database access, cache access, etc) in callbacks, without compromising the operation of the delivery server
- Deployable on existing Heroku apps, using uWSGI
- Django >= 1.7
- Python >= 3.4 (2.7+ support planned)
- A Redis server (TODO: minimum Redis version)
There are currently 2 production servers to choose from:
- aiohttp/asyncio (recommended): Python >= 3.4
- gevent WSGI: Python >= 3.4
SSE (Server-Sent Events) is the only delivery protocol currently supported.
- Only one-way delivery is currently supported (server to client). Messages originating from the client should use a conventional AJAX request. This should be sufficient for the majority of applications.
- Real-time functionality can only be added to Django Class-Based Views (CBVs). Older function-style views are not supported.
How it works