Websocket server using PostgreSQL as a message transport. Uses SocketIO, Greenlet, Flask.
SocketIO PostgreSQL PubSub Websocket Server
What Is This?
This is a websocket server that uses PostgreSQL as its message transport system.
If you have an application that already uses postgres, you can easily send and receive asynchronous events via your existing database. Check out the postgres documentation on LISTEN/NOTIFY for more details.
How Do I Use This?
First configure a database connection string in
You can create a
# local.cfg SQLALCHEMY_DATABASE_URI = 'postgresql:///mydatabase DEBUG = True
DATABASE_URL in your environment. If you use Heroku and have a database attached, this will be set already.
Note: python 3.6 or higher is required.
pip install -r requirements.txt
- Debug server mode:
DEBUG=1 python socketio_pg/server.py
- In gunicorn (production):
gunicorn --worker-class eventlet -w 1 socketio_pg.server:app
- If using Heroku, a Procfile is already set up for you.
On the client side, you simply connect a socket.io client to begin sending and receiving events. There's a simple demo HTML page at socketio_pg/static/test.html that you can access from the dev server at http://localhost:3030/static/test.html.
Why Use This?
If your application already uses PostgreSQL, you can start sending and receiving asynchronous events right away. It makes an excellent transport for messages (keep them small though, under 8000 bytes), and you can simply issue queries to do it. No additional infrastructure needed, besides this websocket server. If you aren't using PostgreSQL, maybe you should be.
One neat trick is to set up triggers that emit
NOTIFY queries when rows on certain tables are inserted or updated. This allows messages to be delivered to clients notifying them of updates without any application code at all. Some demos and slides from a talk can be found here.
How Is This Built?
This server uses the following technologies:
|Filename, size||File type||Python version||Upload date||Hashes|
|Filename, size socketio_pg-0.13.tar.gz (7.7 kB)||File type Source||Python version None||Upload date||Hashes View|