bluemax

bluemax is a python rpc server using websockets.

The concept is that you can write python functions and expose them in the __all__ property of the module.

The complexity comes when you want to do more than this. And that is what I am working on. The most important pattern to me is being able to insert, update and delete in the database and be able to broadcast the changes to all users after the actor gets their response. I call it broadcast_on_success.

    actor   crud    broadcast
      |       |         |
      |------>|         |
      |       |         |
      |<------|         |
      |       |-------->|
      |       |         |
      |<------|---------|
      |       |         |

The errors from crud are local to the actor. The actor will receive identity on success and then like every other user hear about crud via broadcast.

Using annotations allows clients to infer function. See the sample.add function.

For deployment of multiple servers see:

http://www.tornadoweb.org/en/stable/guide/running.html#running-behind-a-load-balancer

To use:

pip install bluemax

To install:

python3.7 -m venv venv
. venv/bin/activate
pip install -e git+https://bitbucket.org/blueshed/bluemax.git#egg=bluemax

To run you can choose either with workers using redis or standalone.

To run standalone:

bluemax run.server -m bluemax.tests.foo

To run with redis:

pip install bluemax[redis]
docker run -p 6379:6379 -d redis:2.8

Then create a conf.yml file containing:

---
procedures: tests.foo
services: tests.foo.services
REDIS_URL: redis://localhost

Then run a worker in one terminal and a server in the another.

bluemax run.worker -c conf.yml

and

bluemax run.services -c conf.yml

and

bluemax run.server -c conf.yml

Now go to http://localhost:8080 and add some numbers.

To build your own project try:

bluemax sidney olive

You can use any name for your module. It will create extension points for logs, settings, urls and a base procedures modules. bluemax will look for your procedures.all in the module you pass in.