A primer on IPC

Developing for the user interface of Pyblish requires an understanding of inter-process communication (IPC), so the following is a quick-start for those unfamiliar with it.

pyblish-qml runs as an individual process on the local machine and communicates with a host via Remote Procedure Calls (RPC) over standard TCP, encapsulated into the pyblish-rpc module.

The UI generally makes requests to the host, and the host replies with data. Any host interested in providing their user with a user interface to Pyblish must first:

1. Start listening at a given port number
2. Send this number to pyblish-qml - the client

The client has a server of its own, listening for incoming requests in the same manner as the host. The address of this server is fixed at 9090.

Both servers uses the Python Standard Library XML-RPC module which works like this.

from SimpleXMLRPCServer import SimpleXMLRPCServer as Server

# Implement functionality
def function(a, b):
  return a + b

# Expose functionality through server
ip, port = "127.0.0.1", 10000
server = Server((ip, port))
server.register_function(function)
server.serve_forever()

This will start an infinite loop, listening for calls to function(a, b), with which we can communicate from a separate process like this.

import xmlrpclib
proxy = xmlrpclib.ServerProxy("http://127.0.0.1:10000")
print proxy.function(5, 10)
# 15

And that's all there is to it.

Take a break. Have a Kit-kat

Take a moment to really absorb the above information, experiment with it, try to break it. This technique lays the foundation not only for how pyblish-qml communicates with the core, but for how inter-process communication works everywhere. There are many variations to this approach that you should explore to really get a sense of the possibilities of this technique.

Here are some well-known variations for you to Google about.

• Remote Procedure Calls
• Message Queues
• RESTful

And here is some recommended reading of particular interest.

• The ZeroMQ Guide
• RESTful in Practice
• Enterprise Application Integration
• Enterprise Integration Patterns
• Service Design Patterns

Break over