Graphical User Interface with Python
Python is inherently a programming language ideal for data analysis, and it works well by simply entering commands and seeing the results immediately. For longer programs, or ones that are used frequently, the code is usually written into a file and then executed from the command line or as an element of a more complex set of scripted operations. However many users prefer a graphical user interface or "GUI" where mouse clicks, drop-down menus, check boxes, images, and plots access with the mouse and sometimes the keyboard. Python has that too, through a selection of tools that can be added to your code.
The most developed core component for this purpose is the Tk/Tcl library. In its most recent versions it provides a very direct way of running your programs within a GUI that is attractive, matches the native machine windows in appearance, and is highly responsive. While this is not the only one, it is a solid starting point and may be all you need. Later in this tutorial we will show you how to use Python running in background behind a graphical interface in a browser window which is another technique that is growing in capability as the browsers themselves gain computing power. However, the Tk library will allow you to build free-standing graphical applications that will run on all platforms on which Python runs.
A Tk Tutorial
An excellent starting point is the Tk documentation tutorial which will take you step by step through how to use this library in Python and in other languages. Work through this one first, and then return here and we will illustrated it with a few examples to show you how to use it in Physics and Astronomy applications.
Building a Program
In the tutorial you will see examples of code segments, and a few complete programs. Here, step by step, is how to make one that you can use as a template for something useful. First, you may see lines such as these at the beginning of a Tk program:
from tkinter import * from tkinter import ttk
which add the needed modules to your Python system. When you use "from xxx import *" that takes all the parts of xxx and adds them with their intrinsic names. Usually this is not an issue for such major parts of Python as tk, but it could cause a problem by overwriting or changing the meaning of something you are already using. Also, even if that is not the case, you cannot immediately tell if a function or constant is from tk, ttk or the core Python. An example would be Tk itself. The next lines of a simple program started this way may be
root = Tk() mainframe = ttk.Frame(root, padding=3 3 12 12")
and if we wanted to use a string inside the tkinter program we would have
mystring = StringVar()
somewhere in the code. The function "StringVar()" is not in the core Python. It comes from tkinter. Without the leading "from tkinter import *" the "StringVar()" is not defined.
For long term maintenance, a better approach is to load the module and give it a short name this way:
import tkinter as tk from tkinter import ttk
Now we have to use "tk" along with the component we are taking from tkinter, but the resulting code is much clearer. Let's complete a program that will convert feet to meters, following an example from the Tk website.
def calculate(*args): try: value = float(feet.get()) meters.set((0.3048 * value * 10000.0 + 0.5)/10000.0) except ValueError: pass root = tk.Tk() root.title("Feet to Meters")
mainframe = tk.ttk.Frame(root, padding="3 3 12 12") mainframe.grid(column=0, row=0, sticky=(tk.N, tk.W, tk.E, tk.S)) mainframe.columnconfigure(0, weight=1) mainframe.rowconfigure(0, weight=1)
feet = tk.StringVar() meters = tk.StringVar()
This way, all the tk-specific terms are clear. Notice the constants too, here "tk.W" for example, come from tkinter. Here's the remainder of the program that wil create a calculator window and run the conversion:
feet_entry = tk.ttk.Entry(mainframe, width=7, textvariable=feet) feet_entry.grid(column=2, row=1, sticky=(tk.W, tk.E)) tk.ttk.Label(mainframe, textvariable=meters).grid(column=2, row=2, sticky=(tk.W, tk.E)) tk.ttk.Button(mainframe, text="Calculate", command=calculate).grid(column=3, row=3, sticky=tk.W) tk.ttk.Label(mainframe, text="feet").grid(column=3, row=1, sticky=tk.W) tk.ttk.Label(mainframe, text="is equivalent to").grid(column=1, row=2, sticky=tk.E) tk.ttk.Label(mainframe, text="meters").grid(column=3, row=2, sticky=tk.W) for child in mainframe.winfo_children(): child.grid_configure(padx=5, pady=5) feet_entry.focus() root.bind('<Return>', calculate)
The result will look like this when it runs:
The elements of the program are
- import tkinter
- define any functions such as "calculate" that you want to run
- create a graphical environment "root" window
- add a frame to the window
- put a grid inside the window frame
- add widgets for your interface located within the grid
- pad the widgets with space so they look good and Tk does the rest
- start the user interface
If you start the Python interpreter in a shell and perform the operations step by step up, as soon as the root object is created, a GUI window will appear on the screen. What comes after determines the properties of that display and its contents. For example, adding a title has a obvious effect to this window.
The window contains a frame which is called "mainframe" here, and when that is added the window shrinks to fit the frame. It is quite small because the frame does not contain anything but a few pixels of padding around its border. You can, at this point grab the edges of the tiny box and stretch it on your screen. The command
mainframe.grid(column=0, row=0, sticky=(tk.N, tk.W, tk.E, tk.S))
creates a grid that is anchored to tk points N, W, E, and S, that is to the corners. Within the grid we add components that will make our interface, and each component has properties we can set too. In this program we will have a frame that will enclose the our interface, and the window will resize automatically to enclose everything. The Tk grid handles the location of the "widgets" that are the elements your interface, and at least for many applications will make creating an nice-looking GUI simple.
First, we add the frame:
mainframe.columnconfigure(0, weight=1) mainframe.rowconfigure(0, weight=1)
Then we define the strings that will hold input and the output from the box that will accept the input:
feet = tk.StringVar() feet_entry = tk.ttk.Entry(mainframe, width=7, textvariable=feet)
The object called "feet_entry" represents this box on the screen. The box is located inside the GUI by the grid that is defined in mainframe
feet_entry.grid(column=2, row=1, sticky=(tk.W, tk.E))
and we put it in column 2 and row 1. As soon as this is issued, a box appears in the GUI and is centered because there is nothing more there. The result of the calculation will be a string that we will show on the interface as a label below the input in the second row
meters = tk.StringVar() tk.ttk.Label(mainframe, textvariable=meters).grid(column=2, row=2, sticky=(tk.W, tk.E))
Notice how the string is given as a variable without quotes, so the interface will update when this variable changes. Along with this, we include fixed labels for the input units, an explanation and the output units:
tk.ttk.Label(mainframe, text="feet").grid(column=3, row=1, sticky=tk.W) tk.ttk.Label(mainframe, text="is equivalent to").grid(column=1, row=2, sticky=tk.E) tk.ttk.Label(mainframe, text="meters").grid(column=3, row=2, sticky=tk.W)
Lastly, we add a button that will start the calculation
tk.ttk.Button(mainframe, text="Calculate", command=calculate).grid(column=3, row=3, sticky=tk.W)
All of these elements are children of the frame, and to have them appear on the grid with a little padding we take each one and configure the grid for it
for child in mainframe.winfo_children(): child.grid_configure(padx=5, pady=5)
Now focus the window on the entry box to make it easy on the user and use the Return key to start a conversion (as well as the button)
feet_entry.focus() root.bind('<Return>', calculate)
The GUI is ready to use. Start it with this
You will find this program among our examples as the file "feet_to_meters_tk_explicit.example".
Interfacing an Instrument: Phidgets
The power of a graphical user interface or GUI is that it is easy to use when the actions are repetitive and the end user does not need to know the details of what is happening. It will deliver to the user the result that is needed and even provide cues on how to make that happen. As in the calculation of a feet to meters conversion, the GUI does not inform the user of how it is done, and simply does it without a fuss. In a laboratory environment that can be very useful, assuming the user understands the process, and needs the answer on a single click. An example would be if there is an instrument which is capable of making a measurement and the user needs to know the result.
The Canadian Company Phidgets manufactures a wide variety of small modules for the hobbiest, lab, and even industry that provide control of motors, relays digital signal levels, analog to digital (ADC) and digital to analog (DAC) conversion, and environmental sensors for temperature, humidty, pressure, light level, and stress. Their hardware has been traditionally designed to interface with computers using the USB bus, and recently (2017) they introduced a simpler 2-wire interface that is more versatile for some applications. They also make single board computers to run their devices under Linux. Their products are not unlike the famous Rasberry Pi or Beagle Board systems, except they are uniformly supported, consistently manufactured, and offered with supporting opensource software. Conceptually it offers an Internet of Things for the lab. Of course there are many other vendors with similar intent, National Instruments with their proprietary and mostly Windows-only LabVIEW software, Labjack with cost effective versatile interfaces and solid cross-platform Python support, and Measurement Computing with cutting edge interfaces supported in Linux and Python worthy of consideration. Whatever you pick for your lab experiments, if it has a Python library or module, you will have access to the device from within the Python program. Therefore, it may be useful to have an example of what can be done to interface a user with a device through a Tk GUI.
The instrument we will use is an infrared thermometer from Phidgets because it uses interesting basic physics, and it is so easily interfaced the device technology does not intrude on showing the programming elements. The device is very small, and is powered from the USB interface. It consists of an infrared sensing pyrometer that is internally a Peltier array of thermocouples, and the electronics to connect it to the computer through the standard bus. The black anodized aluminum cylinder is baffled such that infrared light from a spot 10 degrees in diameter (1 meter at a distance of 11.4 meters) illuminates the sensor. The electronics make a measurement of the IR radiation by determining the voltage from the Peltier array, which is to say the temperature it reaches in equilibrium with the incident radiation. It also measures the ambient temperature. The signal it gets is fundamentally the blackbody temperature of the 10 degree spot, and the device functions as a non-contact thermometer that can measure from -70 to +380 celsius with a temperature resolution of 0.02 C and an error of <math>\pm</math> 4 C.