The following is the overall structure of my typical python tkinter program.

def funA():
    def funA1():
        def funA12():
            # stuff

    def funA2():
        # stuff

def funB():
    def funB1():
        # stuff

    def funB2():
        # stuff

def funC():
    def funC1():
        # stuff

    def funC2():
        # stuff


root = tk.Tk()

button1 = tk.Button(root, command=funA)
button1.pack()
button2 = tk.Button(root, command=funB)
button2.pack()
button3 = tk.Button(root, command=funC)
button3.pack()

funA funB and funC will bring up another Toplevel windows with widgets when user click on button 1, 2, 3.

I am wondering if this is the right way to write a python tkinter program? Sure, it will work even if I write this way, but is it the best way? It sounds stupid but when I see the codes other people written, their code is not messed up with bunch of functions and mostly they have classes.

Is there any specific structure that we should follow as good practice? How should I plan before start writing a python program?

I know there is no such thing as best practice in programming and I am not asking for it either. I just want some advice and explanations to keep me on the right direction as I am learning Python by myself.

I advocate an object oriented approach. This is the template that I start out with:

# Use Tkinter for python 2, tkinter for python 3
import tkinter as tk

class MainApplication(tk.Frame):
    def __init__(self, parent, *args, **kwargs):
        tk.Frame.__init__(self, parent, *args, **kwargs)
        self.parent = parent

        <create the rest of your GUI here>

if __name__ == "__main__":
    root = tk.Tk()
    MainApplication(root).pack(side="top", fill="both", expand=True)
    root.mainloop()

The important things to notice are:

• I don’t use a wildcard import. I import the package as “tk”, which requires that I prefix all commands with tk.. This prevents global namespace pollution, plus it makes the code completely obvious when you are using Tkinter classes, ttk classes, or some of your own.

• The main application is a class. This gives you a private namespace for all of your callbacks and private functions, and just generally makes it easier to organize your code. In a procedural style you have to code top-down, defining functions before using them, etc. With this method you don’t since you don’t actually create the main window until the very last step. I prefer inheriting from tk.Frame just because I typically start by creating a frame, but it is by no means necessary.

If your app has additional toplevel windows, I recommend making each of those a separate class, inheriting from tk.Toplevel. This gives you all of the same advantages mentioned above — the windows are atomic, they have their own namespace, and the code is well organized. Plus, it makes it easy to put each into its own module once the code starts to get large.

Finally, you might want to consider using classes for every major portion of your interface. For example, if you’re creating an app with a toolbar, a navigation pane, a statusbar, and a main area, you could make each one of those classes. This makes your main code quite small and easy to understand:

class Navbar(tk.Frame): ...
class Toolbar(tk.Frame): ...
class Statusbar(tk.Frame): ...
class Main(tk.Frame): ...

class MainApplication(tk.Frame):
    def __init__(self, parent, *args, **kwargs):
        tk.Frame.__init__(self, parent, *args, **kwargs)
        self.statusbar = Statusbar(self, ...)
        self.toolbar = Toolbar(self, ...)
        self.navbar = Navbar(self, ...)
        self.main = Main(self, ...)

        self.statusbar.pack(side="bottom", fill="x")
        self.toolbar.pack(side="top", fill="x")
        self.navbar.pack(side="left", fill="y")
        self.main.pack(side="right", fill="both", expand=True)

Since all of those instances share a common parent, the parent effectively becomes the “controller” part of a model-view-controller architecture. So, for example, the main window could place something on the statusbar by calling self.parent.statusbar.set("Hello, world"). This allows you to define a simple interface between the components, helping to keep coupling to a minimun.

Putting each of your top-level windows into it’s own separate class gives you code re-use and better code organization. Any buttons and relevant methods that are present in the window should be defined inside this class. Here’s an example (taken from here):

import tkinter as tk

class Demo1:
    def __init__(self, master):
        self.master = master
        self.frame = tk.Frame(self.master)
        self.button1 = tk.Button(self.frame, text = 'New Window', width = 25, command = self.new_window)
        self.button1.pack()
        self.frame.pack()
    def new_window(self):
        self.newWindow = tk.Toplevel(self.master)
        self.app = Demo2(self.newWindow)

class Demo2:
    def __init__(self, master):
        self.master = master
        self.frame = tk.Frame(self.master)
        self.quitButton = tk.Button(self.frame, text = 'Quit', width = 25, command = self.close_windows)
        self.quitButton.pack()
        self.frame.pack()
    def close_windows(self):
        self.master.destroy()

def main(): 
    root = tk.Tk()
    app = Demo1(root)
    root.mainloop()

if __name__ == '__main__':
    main()

Also see:

• simple hello world from tkinter docs
• Tkinter example code for multiple windows, why won’t buttons load correctly?
• Tkinter: How to Show / Hide a Window

Hope that helps.

This isn’t a bad structure; it will work just fine. However, you do have to have functions in a function to do commands when someone clicks on a button or something

So what you could do is write classes for these then have methods in the class that handle commands for the button clicks and such.

Here’s an example:

import tkinter as tk

class Window1:
    def __init__(self, master):
        pass
        # Create labels, entries,buttons
    def button_click(self):
        pass
        # If button is clicked, run this method and open window 2


class Window2:
    def __init__(self, master):
        #create buttons,entries,etc

    def button_method(self):
        #run this when button click to close window
        self.master.destroy()

def main(): #run mianloop 
    root = tk.Tk()
    app = Window1(root)
    root.mainloop()

if __name__ == '__main__':
    main()

Usually tk programs with multiple windows are multiple big classes and in the __init__ all the entries, labels etc are created and then each method is to handle button click events

There isn’t really a right way to do it, whatever works for you and gets the job done as long as its readable and you can easily explain it because if you cant easily explain your program, there probably is a better way to do it.

Take a look at Thinking in Tkinter.

OOP should be the approach and frame should be a class variable instead of instance variable.

from Tkinter import *
class App:
  def __init__(self, master):
    frame = Frame(master)
    frame.pack()
    self.button = Button(frame, 
                         text="QUIT", fg="red",
                         command=frame.quit)
    self.button.pack(side=LEFT)
    self.slogan = Button(frame,
                         text="Hello",
                         command=self.write_slogan)
    self.slogan.pack(side=LEFT)
  def write_slogan(self):
    print "Tkinter is easy to use!"

root = Tk()
app = App(root)
root.mainloop()

Reference: http://www.python-course.eu/tkinter_buttons.php

Organizing your application using class make it easy to you and others who work with you to debug problems and improve the app easily.

You can easily organize your application like this:

class hello(Tk):
    def __init__(self):
        super(hello, self).__init__()
        self.btn = Button(text = "Click me", command=close)
        self.btn.pack()
    def close():
        self.destroy()

app = hello()
app.mainloop()

Probably the best way to learn how to structure your program is by reading other people’s code, especially if it’s a large program to which many people have contributed. After looking at the code of many projects, you should get an idea of what the consensus style should be.

Python, as a language, is special in that there are some strong guidelines as to how you should format your code. The first is the so-called “Zen of Python”:

• Beautiful is better than ugly.
• Explicit is better than implicit.
• Simple is better than complex.
• Complex is better than complicated.
• Flat is better than nested.
• Sparse is better than dense.
• Readability counts.
• Special cases aren’t special enough to break the rules.
• Although practicality beats purity.
• Errors should never pass silently.
• Unless explicitly silenced.
• In the face of ambiguity, refuse the temptation to guess.
• There should be one– and preferably only one –obvious way to do it.
• Although that way may not be obvious at first unless you’re Dutch.
• Now is better than never.
• Although never is often better than right now.
• If the implementation is hard to explain, it’s a bad idea.
• If the implementation is easy to explain, it may be a good idea.
• Namespaces are one honking great idea — let’s do more of those!

On a more practical level, there is PEP8, the style guide for Python.

With those in mind, I would say that your code style doesn’t really fit, particularly the nested functions. Find a way to flatten those out, either by using classes or moving them into separate modules. This will make the structure of your program much easier to understand.

I personally do not use the objected oriented approach, mostly because it a) only get in the way; b) you will never reuse that as a module.

but something that is not discussed here, is that you must use threading or multiprocessing. Always. otherwise your application will be awful.

just do a simple test: start a window, and then fetch some URL or anything else. changes are your UI will not be updated while the network request is happening. Meaning, your application window will be broken. depend on the OS you are on, but most times, it will not redraw, anything you drag over the window will be plastered on it, until the process is back to the TK mainloop.