9.7 Combining Tkinter and Asynchronous I/O with Threads

Credit: Jacob Hallén

9.7.1 Problem

You need to access sockets, serial ports, and do other asynchronous (but blocking) I/O while running a Tkinter-based GUI.

9.7.2 Solution

The solution is to handle a Tkinter interface on one thread and communicate to it (via Queue objects) the events on I/O channels handled by other threads:

import Tkinter
import time
import threading
import random
import Queue

class GuiPart:
    def _ _init_ _(self, master, queue, endCommand):
        self.queue = queue
        # Set up the GUI
        console = Tkinter.Button(master, text='Done', command=endCommand)
        console.pack(  )
        # Add more GUI stuff here depending on your specific needs

    def processIncoming(self):
        """Handle all messages currently in the queue, if any."""
        while self.queue.qsize(  ):
            try:
                msg = self.queue.get(0)
                # Check contents of message and do whatever is needed. As a
                # simple test, print it (in real life, you would
                # suitably update the GUI's display in a richer fashion).
                print msg
            except Queue.Empty:
                # just on general principles, although we don't
                # expect this branch to be taken in this case
                pass

class ThreadedClient:
    """
    Launch the main part of the GUI and the worker thread. periodicCall and
    endApplication could reside in the GUI part, but putting them here
    means that you have all the thread controls in a single place.
    """
    def _ _init_ _(self, master):
        """
        Start the GUI and the asynchronous threads. We are in the main
        (original) thread of the application, which will later be used by
        the GUI as well. We spawn a new thread for the worker (I/O).
        """
        self.master = master

        # Create the queue
        self.queue = Queue.Queue(  )

        # Set up the GUI part
        self.gui = GuiPart(master, self.queue, self.endApplication)

        # Set up the thread to do asynchronous I/O
        # More threads can also be created and used, if necessary
        self.running = 1
        self.thread1 = threading.Thread(target=self.workerThread1)
        self.thread1.start(  )

        # Start the periodic call in the GUI to check if the queue contains
        # anything
        self.periodicCall(  )

    def periodicCall(self):
        """
        Check every 200 ms if there is something new in the queue.
        """
        self.gui.processIncoming(  )
        if not self.running:
            # This is the brutal stop of the system. You may want to do
            # some cleanup before actually shutting it down.
            import sys
            sys.exit(1)
        self.master.after(200, self.periodicCall)

    def workerThread1(self):
        """
        This is where we handle the asynchronous I/O. For example, it may be
        a 'select(  )'. One important thing to remember is that the thread has
        to yield control pretty regularly, by select or otherwise.
        """
        while self.running:
            # To simulate asynchronous I/O, we create a random number at
            # random intervals. Replace the following two lines with the real
            # thing.
            time.sleep(rand.random(  ) * 1.5)
            msg = rand.random(  )
            self.queue.put(msg)

    def endApplication(self):
        self.running = 0

rand = random.Random(  )
root = Tkinter.Tk(  )

client = ThreadedClient(root)
root.mainloop(  )

9.7.3 Discussion

This recipe shows the easiest way of handling access to sockets, serial ports, and other asynchronous I/O ports while running a Tkinter-based GUI. Note that the recipe's principles generalize to other GUI toolkits, since most of them make it preferable to access the GUI itself from a single thread, and all offer a toolkit-dependent way to set up periodic polling as this recipe does.

Tkinter, like most other GUIs, is best used with all graphic commands in a single thread. On the other hand, it's far more efficient to make I/O channels block, then wait for something to happen, rather than using nonblocking I/O and having to poll at regular intervals. The latter approach may not even be available in some cases, since not all data sources support nonblocking I/O. Therefore, for generality as well as for efficiency, we should handle I/O with a separate thread, or more than one. The I/O threads can communicate in a safe way with the main, GUI-handling thread through one or more Queues. In this recipe, the GUI thread still has to do some polling (on the Queues), to check if something in the Queue needs to be processed. Other architectures are possible, but they are much more complex than the one in this recipe. My advice is to start with this recipe, which will handle your needs over 90% of the time, and explore the much more complex alternatives only if it turns out that this approach cannot meet your performance requirements.

This recipe lets a worker thread block in a select (simulated by random sleeps in the recipe's example worker thread). Whenever something arrives, it is received and inserted in a Queue. The main (GUI) thread polls the Queue five times per second (often enough that the end user will not notice any significant delay, but rarely enough that the computational load on the computer will be negligible�you may want to fine-tune this, depending on your exact needs) and processes all messages that have arrived since it last checked.

This recipe seems to solve a common problem, since there is a question about how to do it a few times a month in comp.lang.python. There are other solutions, involving synchronization between threads, that let you solve such problems without polling (the root.after call in the recipe). Unfortunately, such solutions are generally complicated and messy, since you tend to raise and wait for semaphores throughout your code. In any case, a GUI already has several polling mechanisms built into it (the main event loop), so adding one more won't make much difference, especially since it seldom runs. The code has been tested only under Linux, but it should work on any platform with working threads, including Windows.

9.7.4 See Also

Documentation of the standard library modules threading and Queue in the Library Reference; information about Tkinter can be obtained from a variety of sources, such as Pythonware's An Introduction to Tkinter, by Fredrik Lundh (http://www.pythonware.com/library), New Mexico Tech's Tkinter reference (http://www.nmt.edu/tcc/help/lang/python/docs.html), and various books.