Development/Tutorials/Python introduction to signals and slots

Warning

This tutorial is work in progress.

Abstract

The signal and slot architecture is design to simplify communication between objects. It's a fact that GUI programming is mostly event driven. The traditionally approach to event driven programming is to use callbaks. Callbacks have a number of limitations that Qt tries to solve with it's signal and slot architecture. The concept is that every object can emit signals. For example when a button is clicked it emits a “clicked()” signal. Signals does not do anything alone, but when connected to a slot the code in the slot will be executed every time the signal is emitted. In python every function is a slot. It's possible to connect one signal to multiple slots

Prerequisites

General understanding of the python programming language. No prior knowledge of QT is required.

Connecting signals and slots.

This example demonstrates how to connect signals and slots.

from PyQt4.QtGui import * from PyQt4.QtCore import * import sys

1. First we create a QApplication and QPushButton

app=QApplication(sys.argv) exitButton=QPushButton("Exit")

1. Her we connect the exitButton's "clicked()" signals to the app's exit method.
2. This will have the effect that every time some one clicks the exitButton the app.exit method will execute and the application will close.

QObject.connect(exitButton,SIGNAL("clicked()"),app.exit)

exitButton.show()

1. Start the evnt loop

sys.exit(app.exec_())

Signals and slots with parameters

The signal and slots mechanism is type safe. In C++ this implies that both the number of arguments and the type of the arguments in a signal must match the arguments in the receiving slot. The term is a bit unclear when dealing with Python because of it's dynamically typed nature.

If we connect custom made signals to a Qt built in slots we have to make sure that parameters is of the correct type or that they can be automatically converted to the correct type. For example a python string will automatically be converted to a QString. If we send a object of an incompatible type we will get an runtime error. It's possible to send a python object of any type using PyQt_PyObject in the signature. This is recommended when both signal and slot is implemented in python. By using PyQt_PyObject we avoid unnecessary conversions between python objects and C++ types and it more consistent with python dynamically typed nature.

Example 1 from PyQt4.QtGui import * from PyQt4.QtCore import * import sys

def printNumber(number):

   print number

if __name__=="__main__":

   #First we create a QApplication and QPushButton
   app=QApplication(sys.argv)
   slider=QSlider(Qt.Horizontal)
   
   
   QObject.connect(slider,SIGNAL("valueChanged(int)"),printNumber)
   
   
   slider.show()
   #Start the evnt loop
   sys.exit(app.exec_())

Signals and slots and threading

To send signal across threads we have to use the Qt.QueuedConnection parameter. Without this parameter the code will be executed in the same thread. import sys from time import time from PyQt4.QtCore import *

class A (QThread):

   def __init__(self):
       QThread.__init__(self)
   
   def afunc (self):
       p("starting in a()")
       self.emit(SIGNAL("asignal"))
       p("finished in a()")
      
   def bfunc(self):
       p("starting in b()")
       self.sleep(3) 
       p("finished in b()")
               
   def run(self):
       self.exec_()
       

def p(msg): print str(int(time()-start)) + "s",msg

if __name__=="__main__":

   start=time()
   p("starting in __main__")
   app=QCoreApplication(sys.argv)
   a=A()
   a.start()
   QObject.connect(a,SIGNAL("asignal"),a.bfunc,Qt.QueuedConnection)
   a.afunc()
   p("finished in __main__")
      
   sys.exit(app.exec_())

Output: 0s starting in __main__ 0s starting in a() 0s finished in a() 0s finished in __main__ 0s starting in b() 3s finished in b()


without Qt.QueuedConnection the example will output: 0s starting in __main__ 0s starting in a() 0s starting in b() 3s finished in b() 3s finished in a() 3s finished in __main__