5. Functions

  • Script/program files are a nice way to organize many statements

  • However, eventually you will find yourself writing the same series of statements over and over

    • Or copying and pasting in your editor

  • This is troubling since

    • It is tedious

    • You can introduce errors

    • Your code can become harder to maintain

5.1. Functions

  • We like to group together sequences of statements that serve a coherent purpose

  • In Python, and many other programming languages, we do this with functions

5.1.1. Defining a Function

  • Once you’ve defined a function, you can call it exactly the same way you call built-in functions like print

1def celsius_to_fahrenheit(temp_in_celsius):
2    partial_conversion = temp_in_celsius * 9/5
3    temp_in_fahrenheit = partial_conversion + 32
4    return temp_in_fahrenheit

  • Above is an example definition of a function that converts temperature units from Celsius to Fahrenheit

  • The details on all the parts will be discussed as we move through this topic

  • You can probably already get a sense that

    • It is a group of statements that serves a single purpose

    • It can be reused every time we want to do the conversion without having to write the whole conversion formula out

5.1.2. Calling a Function

  • Given the function definition of celsius_to_fahrenheit, we can now call the function

    • celsius_to_fahrenheit(0) gives us 32.0

    • celsius_to_fahrenheit(-40) gives us -40.0

    • celsius_to_fahrenheit(32) gives us 89.6

  • It’s always good to test your functions a little to verify that they are doing what you expect

  • When we call a function, Python executes the statements within the function in the order that they appear

  • Functions allow for

    • Reuse — Just call it any time you need it

    • Maintainability — If there is a bug in the function, you only need to change the function once

    • Readability — celsius_to_fahrenheit is a lot easier to recognize when compared to something like (C * 9/5) + 32

  • More importantly they facilitate abstraction

    • More on this later

Quick Activity

Write your own function to do something with math. Honestly, whatever you want.

5.1.3. Function Parameters

  • Notice how the function celsius_to_fahrenheit has a parameter called temp_in_celsius

  • When defining functions, we want them to be very general

    • For example, it would be rather silly to write a function twenty_degrees_celsius_to_fahrenheit that was only capable of converting 20 degrees Celsius to Fahrenheit

      • What happens if we want to calculate 30 degrees Celsius in Fahrenheit, write another function called thirty_degrees_celsius_to_fahrenheit?

    • Instead, we wrote the function such that the temperature in Celsius is a parameter that we can specify when calling the function

      • celsius_to_fahrenheit(20)

      • celsius_to_fahrenheit(30)

1# This is rather silly and useless
2def twenty_degrees_celsius_to_fahrenheit():
3    partial_conversion = 20 * 9/5
4    temp_in_fahrenheit = partial_conversion + 32
5    return temp_in_fahrenheit

  • If it helps, just think of the parameters as variables that belong to the function

  • You have already been specifying the parameter in the print function

    • The value you want printed out is the value you are setting for the parameter that print takes

  • In reality, you have been using this idea in math class for years

    \(f(x) = x * 9/5 + 32\)

  • This is the definition of a function called \(f\) that takes a parameter \(x\)

  • If I asked you what \(f(20)\) is, you can calculate the result

    \(f(20) = 20 * 9/5 + 32\)

    \(f(20) = 36 + 32\)

    \(f(20) = 68\)

  • This is the same idea we used in celsius_to_fahrenheit, but in Python instead of our typical math syntax

5.1.4. Execution of a Function

  • Below is a simple and arbitrary function (square_of_sum) that takes two parameters (a and b) and calculates what the square of their sum is

1def square_of_sum(a, b):
2    c = a + b
3    d = c * c
4    return d

  • If I were to call this function with square_of_sum(2, 3), Python handles the execution like this

    1. Python will check to see if it knows about a function called square_of_sum

    2. Python takes the values supplied to it when called (2 and 3) and assigns them to their respective parameters

      • a = 2 and b = 3

    3. The sum of a and b is calculated and put into a variable c

    4. The variable c is multiplied with itself (effectively squaring it) and the result is assigned to d

    5. The function returns the value associated with d

  • If you provide too few or too many parameters, Python will raise an error like the following examples

    • Missing one parameter — TypeError: square_of_sum() missing 1 required positional argument: 'b'

    • Too many parameters — TypeError: square_of_sum() takes 2 positional arguments but 3 were given



5.2. Return

  • Every function you call returns a value

  • Notice that we wrote return at the end of the previous functions

  • This allows me to specify the value being returned by the function

  • Consider the following example

1temperature = celsius_to_fahrenheit(20)
2print(temperature)

  • Here, the function call celsius_to_fahrenheit(20) will execute the function

  • When the function finishes executing, it returns the value 68

  • The value 68 is stored in the variable temperature

  • The value of temperature (68) is printed out

  • If you do not write a return statement in a function, the function will still return a value, but the value will be None

  • None is a special type and value that means nothing

Activity

Write a function print_celsius_to_fahrenheit that is identical to celsius_to_fahrenheit, except instead of returning the final value, it just prints the value out within the function (just replace the final line with print(temp_in_fahrenheit)).

Run the following and see if you can figure out why it is doing what it is doing:

  • print(print_celsius_to_fahrenheit(20))

  • print(celsius_to_fahrenheit(20))

  • print_celsius_to_fahrenheit(20)

  • celsius_to_fahrenheit(20) NOTE: This one is lying to you

Warning

Colab is misleading you when you call celsius_to_fahrenheit(20). Colab will make it seem as if celsius_to_fahrenheit(20) is printing out the result, but it is not — Colab is being “nice” and just displaying any values produced on the last line of code, regardless of if you printed it out. To demonstrate this to yourself, run the following:

1celsius_to_fahrenheit(20)
2print("Now you do not see any value from celsius_to_fahrenheit")

In the above example, if you want to keep track of the value returned by celsius_to_fahrenheit, simply assign it to a variable for later some_variable = celsius_to_fahrenheit(20).

Activity

Write a function called euclidean_distance(x1, y1, x2, y2) that calculates and returns the Euclidean distance between two points. Remember, Euclidean distance is defined as \(\sqrt{(x1 - x2)^{2} + (y1 - y2)^{2}}\)

Does Python have a square root function? How do you calculate the square of a value in Python? How would I find out?

5.3. Execution Flow

  • Python executes one statement at a time

  • In a program, the statements get executed in the order in which they appear, top to bottom

  • However, statements within a function only get executed when that function is called

 1def celsius_to_fahrenheit(temp_in_celsius):
 2    partial_conversion = temp_in_celsius * 9/5
 3    temp_in_fahrenheit = partial_conversion + 32
 4    return temp_in_fahrenheit
 5
 6celsius = 24
 7fahrenheit = celsius_to_fahrenheit(celsius)
 8print(fahrenheit)
 9
10celsius = 32
11fahrenheit = celsius_to_fahrenheit(celsius)
12print(fahrenheit)

  • In the above example, the program starts running at line 1, however Python notes that this is a function definition

    • It is not called yet — it does not run yet

  • Python takes note of the function and knows that it exists

  • The first line to get executed in this program is line 6 where the value of 24 is assigned to celsius

  • Line 7 makes a call to the function celsius_to_fahrenheit, and so the execution jumps to line 1

  • The program will run the whole function (lines 1 – 4) and return the value to where it was called (line 7) and the value is stored in the variable fahrenheit

    • Functions end when there are no more lines to execute, or a return statement is hit

  • Line 8 prints out the value of fahrenheit

  • Line 10 assigns a value to a variable

  • Like 11 calls the function celsius_to_fahrenheit again, which means our execution jumps to line 1 again

  • Once the function is complete (lines 1 – 4), the value is returned to like 11 and the returned value is assigned to fahrenheit

  • Line 12 prints out the value of fahrenheit

  • The program is now complete

5.4. Abstraction

Activity

Write down a “program” to make spaghetti (not in python, like on paper). You can only use the following statements:

  • locate [object]

  • grasp [limb]

  • release [limb]

  • move_limb_to [location]

  • wait [time in seconds]

Assume you start from a clean, empty, kitchen.

Activity

Write down a “program” to make spaghetti (not in python, like on paper). You can use plain English prose and assume you are addressing a human being who is familiar with a kitchen and making pasta.

Assume you start from a clean, empty, kitchen.

  • You have now created two different programs for making spaghetti at two different levels of abstraction, which version was easier?

5.4.1. Making Use of Abstraction

  • You have been making use of the print function every time you needed to display something

  • Fortunately, you did not need to worry about setting individual pixels on your display to show the characters

  • print has collected all the complex information and instructions needed to print

  • Because of this, we can think about print every time we need to print instead of worrying about the underlying workings of how to set pixels on a display

5.4.2. Creating Abstraction

  • Without being able to organize things into levels of abstraction, writing complex software would be prohibitively difficult

    • The same is true for your every day life — learning to think of things in terms of levels of abstraction is very important

    • For example, when driving a car, do you think about the pistons firing?

    • Or, do you need to think about neurons firing and ion pumps to move your arm?

  • You are already experts at this in real life

  • Unfortunately, however, it is not simple to just start creating abstraction in your code

  • Knowing how and where to create levels of abstraction requires a deep understanding of the problem you are trying to address

Warning

It needs to be emphasized that how, when, and where to create levels of abstraction can be very difficult. This is not something you can learn by just reading about it; this is a skill that you will be developing throughout your careers. Over time you will get better and better at it.

Often there isn’t even a right way to do it. Some may be better than others, but that does not mean that the others are not “correct”.

The point is, however, that no one is expecting you to be an expert in this at this stage. Expect this to not be trivial.

5.5. Format of a Function

  • The format for defining a function is as follows

1def function_name(parameter_1, parameter_2, parameter_3, parameter_4, ... parameter_n):
2    statement 1
3    statement 2
4    ...
5    statement m

  • function_name is what the function is named

  • parameter_1, parameter_2 , etc. are called the parameters, you can have as many as you like and call them almost whatever you want

    • You can think of these like variables

  • You tell Python which statements make up the body of the function by using indentation

    • Some languages use other syntax, like begin and end or braces ({ })

Activity

Write a function concatenate_strings that takes two strings as parameters and then returns the concatenation of the strings. For example, if I call concatenate_strings("Hello","World!") it will return HelloWorld!. Do not add a space between the two strings being concatenated.

Activity

Warning: This one is tricky. If you’re still stuck after lecture, be sure to take your time to figure this out. There’s a YouTube video to help.

Now write a function criss_cross_concatenation that will take four strings and return the concatenation of the first, third, second, and fourth, in that order. BUT your function isn’t allowed to directly use the + to concatenate strings. You can, however, use the concatenate_strings function.

5.6. Composition

  • We can compose functions of other functions

    • Like in the criss_cross_concatenation function that makes use of concatenate_strings

  • Since functions often return a useful value, we can nest function calls in statements

    • square_of_sum(2, 2) + square_of_sum(3, 3)

    • 16 + 36

    • 52

  • Similarly, we can nest function calls as parameters to other functions

    • square_of_sum(square_of_sum(2, 2), square_of_sum(3, 3))

    • square_of_sum(16, 36)

    • 2704

  • It is a good exercise to work these out by hand to help with your understanding

  • If you get confused tracking what is happening in the above example

    1. Slow down; there is one trick — follow the code

    2. Functions get evaluated and turned into values

    3. Find a function you can evaluate and evaluate it

    4. Cross out the function and replace it with the returned value

    5. Keep going

Activity

Figure out the value of square_of_sum(square_of_sum(1, 1), (square_of_sum(1, 1) + square_of_sum(0, 1))) using only pencil and paper — no computers!

5.7. Variable scope

  • You may have already noticed that variables you create within a function are not accessible outside the function

  • These variables within the function are local to that function

  • Nothing can access a function’s local variables

1def celsius_to_fahrenheit(temp_in_celsius):
2    partial_conversion = temp_in_celsius * 9/5
3    temp_in_fahrenheit = partial_conversion + 32
4    return temp_in_fahrenheit

  • With the celsius_to_fahrenheit function, the variables temp_in_celsius, partial_conversion, and temp_in_fahrenheit are local

  • If I were to call the function and then later try to access the partial_conversion variable, I would have a problem

1print(celsius_to_fahrenheit(28))
2print(temp_in_fahrenheit)

  • Trying to run the above code would result in the error NameError: name 'temp_in_fahrenheit' is not defined

  • Where this can get tricky is when you have two variables with the same name, but in different scope

1def celsius_to_fahrenheit(temp_in_celsius):
2    partial_conversion = temp_in_celsius * 9/5
3    temp_in_fahrenheit = partial_conversion + 32
4    return temp_in_fahrenheit
5
6temp_in_fahrenheit = "Hello, world!"
7print(celsius_to_fahrenheit(28))
8print(temp_in_fahrenheit)

  • In the above example, print(temp_in_fahrenheit) prints out Hello, world!

  • Although variables with the name temp_in_fahrenheit exists within the function celsius_to_fahrenheit and outside the function, they are actually different variables

    • The variable temp_in_fahrenheit outside the function has no knowledge of the one inside the function

    • The variable temp_in_fahrenheit inside the function has no knowledge of the one outside the function

Note

There is a way to set a variable to have global scope, which allows the variable to be accessed everywhere. This, however is generally bad practice and something we will not do.

5.7.1. Import

  • Scope does not only apply to variables

  • Sometimes you need to access complex functions that already exist out there and you don’t want to write

  • For example, you may have already found that you wanted to make use of the square root (sqrt) function

  • Fortunately the sqrt function exists in Python

  • Unfortunately, you don’t get it by default when you start up Python as it is not in scope

  • Fortunately, in spite of this, there is a rather simple way to access the function

  • We import the math module, which is where the function is stored

1import math
2
3root_of_two = math.sqrt(2)
4print(root_of_two)

  • In the above example, the math module is imported

  • We then access the sqrt function by prefacing it with the module name math

    • math.sqrt

  • You can think of it as, from the math module, call the square root function

5.8. Comments

  • You can add comments to your code in Python with #

  • As soon as Python sees # it ignores the rest of the current line

1# Calculate the Euclidean distance between two points
2d = ((x1 - x2)**2 + (y1 - y2)**2)**0.5

  • We want our code to be written in such a way that it is correct, but also understandable

  • However, sometimes we may have some code that is rather complex and not immediately clear

    • This becomes particularly important if you are working with others that need to look at your code

  • When situations like these arise, we add comments to our code to explain what’s going on

  • It’s not about explaining everything, but explaining what is likely to be unclear

5.8.1. Docstring

  • We will find that, as we write bigger and bigger programs, we will be making use of functions a lot

    • Not only those that already exist like print, but functions we write

  • Since functions tend to be some coherent set of statements that serve a purpose, we write docstrings to describe what the function does

 1def celsius_to_fahrenheit(temp_in_celsius):
 2    """
 3    Convert a temperature from Celsius units to Fahrenheit units.
 4
 5    :param temp_in_celsius: The temperature in Celsius to be converted.
 6    :return: The temperature in Fahrenheit.
 7    """
 8    partial_conversion = temp_in_celsius * 9/5
 9    temp_in_fahrenheit = partial_conversion + 32
10    return temp_in_fahrenheit

  • The stuff between the """ is the docstring and should appear immediately after the def line

  • It explains what the function does in plane English

  • It explains what each parameter is

    • Mind the use of :s

  • If the function returns something, then explain that too

    • Mind the use of :s here too

  • This may feel like a lot of work, especially with such a simple function in the above example

  • But having these describing the functions makes it easier for anyone looking at your code

  • Trust me when I say, there will be a time in your life where you regret not writing comments/docstrings

5.9. For Next Class