Developing with IPC

The workflow for implementing a new feature into the user interface generally goes like this.


  1. Expose a feature of Pyblish through pyblish-rpc
  2. Make use of the exposed feature through pyblish-qml

In this article we will have a look at how to do just that.


As we saw in A primer on IPC, to be able to use functionality in one process from another process, we register a function with the RPC server.

Aside from registering a plain function, we are also able to register an instance of a class, called "service". We use it in place of freestanding functions in order to share memory state amongst one or more calls, such as maintaining reference to discovered plug-ins and instances.

The default service is located in pyblish-rpc/ and looks like this.

class RpcService(object):
    def __init__(self):
        self._context = None
        self._plugins = None
        self._provider = None


    def ping(self):
        """Used to check connectivity"""
        return {
            "message": "Hello, whomever you are"

This service is registered by default when launching any integrated host. If you have Pyblish installed, try launching any host and type this in.

import xmlrpclib
proxy = xmlrpclib.ServerProxy("")
# {"message": "Hello, whomever you are"}

Note the additional path /pyblish.

You can find the exposed port number of your host via the environment variable PYBLISH_CLIENT_PORT which in this case is 9001. If the variable isn't available, make sure the Pyblish has been installed correctly and that the integration is loaded.

Exposing a feature for RPC

Now that we know how features are exposed, let's try exposing something we can use with pyblish-qml.

class RpcService(object):
    def add_two_numbers(self, a, b):
        return a + b

With the method added, let's have a look at how we can use this in pyblish-qml.

Using a feature

Before diving into the QML side, let's stick to a standalone Python interpreter.

QML instantiates Proxy from pyblish-rpc/ and uses it for any and all communication with Pyblish within a host. Let's try doing that in a plain Python interpreter.

from pyblish_rpc.client import Proxy
proxy = Proxy(9001)

The port is normally passed to QML from the host during the point at which a user triggers its visibility; e.g. by pressing "Publish" from the file-menu or by calling .show() from it's integration, such as

For this example, we will assume a host is running at 9001.

Now we can call any function exposed via this service.

# Retrieve a list of plugins
plugins =

# Retrieve the current context
context = proxy.context()

Including the one we just implemented.

print(proxy.add_two_numbers(1, 2))
# 3

Using a feature in the controller

In a typical model/view/controller application, the controller is the part where logic and event handling occurs. This is where we will make use of our feature.

Open up pyblish-qml/ and add the calling method to the Controller class.

# pyblish-qml/pyblish_qml/
class Controller(QtCore.QObject):
    def add_two_numbers(self, a, b):
        result =, b)
        print("The result is: %s" % result)
        return result
    ... represents the currently active host and is an instance of the Proxy class we saw above.

Whenever a host connects/shows the GUI, Proxy is reinstantiated to reflect the change.

Using a feature in QML

To call upon a feature from QML directly, all we need to do is decorate the method we just made.

# pyblish-qml/pyblish_qml/
class Controller(QtCore.QObject):
    @QtCore.pyqtSlot(int, int, result=int)
    def add_two_numbers(self, a, b):
        result =, b)
        print("The result is: %s" % result)
        return result

We are then able to call this function from within QML.

// pyblish-qml/pyblish_qml/qml/Overview.qml
Footer {
    id: footer

    mode: overview.state == "publishing" ? 1 : overview.state == "finished" ? 2 : 0

    width: parent.width
    anchors.bottom: parent.bottom

    onReset: {
        // Run our function at the touch of the reset-button
        app.add_two_numbers(2, 4)

Notice how the controller is exposed as a global variable in QML called app. Try launching the GUI and pressing the Reset button, and keep an eye out in the output.

The result is: 6
Mon Jul 18 2016 19:42:18 GMT+0000 (UTC)

results matching ""

    No results matching ""