brisk_python.Rmd

---
jupyter:
  jupytext:
    notebook_metadata_filter: all,-language_info
    split_at_heading: true
    text_representation:
      extension: .Rmd
      format_name: rmarkdown
      format_version: '1.2'
      jupytext_version: 1.13.7
  kernelspec:
    display_name: Python 3
    language: python
    name: python3
---

# Brisk introduction to Python

This is an introduction designed for those of us who already know a [dynamic
programming
language](https://en.wikipedia.org/wiki/Dynamic_programming_language) fairly
well. MATLAB and the R language are [examples of dynamic programming
languages](https://en.wikipedia.org/wiki/Dynamic_programming_language#Examples_of_Dynamic_Programming_Languages).

For an alternative introduction at a slightly slower pace, see [Introducing
Python](./introducing_python.Rmd).

## Numbers

There are two types of numbers in Python: integer and floating point.

You may remember that an integer is a whole number - a number without anything
after the decimal point.  The *counting numbers* are integers — e.g. 0, 1, 2,
3, ....  But integers also include negative whole numbers — e.g. -1, -2, -3
...

In Python, an integer is an *object* of type `int`, and a float is an object
of type `float`.

```{python}
a = 99
type(a)
```

```{python}
b = 99.0
type(b)
```

You can create ints and floats by using `int` and `float` like this:

```{python}
float('1')
```

```{python}
float(1)
```

```{python}
int('1')
```

```{python}
int(1)
```

`+`,  `-`, `\*` or `/` on a mix of floats and ints, give floats:

```{python}
a + b
a * b
```

Dividing an int by an int also gives a float:

```{python}
1 / 2
```

If you only want the integer part of the division, use `//`

```{python}
1 // 2
```

```{python}
1.0 // 2.0
```

Python has built-in function called `round`:

```{python}
round(5.0 / 2.0)
```

Built-in means it is always available in Python, without you having to load an optional module (library).  You will see modules soon.

The `%` operator on numbers gives you the remainder of integer division
(also known as the modulus):

```{python}
5 % 2
```

```{python}
5.0 % 2.0
```

(true-and-false)=

## True and False

`True` and `False` are special objects in Python.  They are of type `bool`
(for Boolean).

```{python}
type(True)
```

```{python}
type(False)
```

To show several results from one cell, we can use the `print` function, that
displays the value.  As you will soon see, we use a function by giving its
name (here: `print`) followed by round brackets (parentheses) that contain the
thing or things we want the function to work on.  For example, to display the
value `False` I can write:

```{python}
print(False)
```

This is similar to what I would get by just putting `False` on the last line of the cell, in the way we have been doing up to now, to show values.  One advantage of `print` is that I can display multiple values from one cell:

```{python}
print(True == False)
print(True == True)
print(False == False)
```

## None

`None` is also a special object in Python.  By convention, Python often uses
`None` to mean that no valid value resulted from an operation, or to signal
that we don’t have a value for a parameter.

```{python}
type(None)
```

Unlike most other values in Python, the default display output from None, is
nothing:

```{python}
None
```

## Equals

As for MATLAB and R, `=` is for assignment, `==` is for testing equality.

```{python}
# a gets the value 1.  Notice the single =
a = 1
a
```

`==` is a test to ask if the left hand side value is equal to the right hand side value:

```{python}
a == 1
```

Notice that Python returns True in this case because `a` is equal to 1.  On the other hand:

```{python}
a == 2
```

Like R, Python uses `!=` for testing that objects are *not* equal. This is
different from MATLAB, which uses `~=`:

```{python}
a != 1
```

(comparison-operators)=

## Comparison operators

You have just seen the `==` *operator* in action, as well as the `!=`
operator.

These are *operators* because they operate on values.  Here they operate on the values to their left and right.

`+` and `-` and so on, are also operators.  You can read this:

```{python}
3 + 4
```

as "Apply the addition operator to 3 and 4, returning the result (here, 7).

Similarly, you can read:

```{python}
a == 2
```

as "Apply the equality operator to the value in `a` and 2, returning the
result (here, `True`).

The equality operator `==` and the inequality operator `!=` are examples of
*comparison* operators.  These are operators that apply a *comparison*
question to the values to their left and right.  They always return True or False.

Here are Python's comparison operators:

| Operator | Name | Example | Result for example |
| --- | --- | --- | --- |
| `==` | equal to | `3 == 2` | `False` |
| `!=` | not equal to | `3 != 2` | `False` |
| `<` | less than | `2 < 3` | `True` |
| `>` | greater than | `2 > 3` | `False` |
| `<=` | less than or equal to | `2 <= 2` | `True` |
| `>=` | greater than or equal to | `2 >= 2` | `True` |

Here are the examples:

```{python}
print('3 == 2 gives', 3 == 2)
print('3 != 2 gives', 3 != 2)
print('2 < 3 gives', 2 < 3)
print('2 > 3 gives', 2 > 3)
print('2 <= 2 gives', 2 <= 2)
print('2 >= 2 gives', 2 >= 2)
```

## Logical operators

Logical operators are like comparison operators, but they ask *logical* questions.

For example, in logic, by definition, the *and* operator asks the question:
*Are both the left and right values True?* — like this:

```{python}
print('True and True:', True and True)
print('True and False:', True and False)
print('False and True:', False and True)
print('False and False:', False and False)
```

Similarly, the logical operator *or* asks the question: "Is either of the left or the right values True?".  The answer is True to the question if either the left value is True, or the right, or both:

```{python}
print('True or True:', True or True)
print('True or False:', True or False)
print('False or True:', False or True)
print('False or False:', False or False)
```

The operator *not* only works on the value to its left, and it flips a True
value to False, or a False value to True.

```{python}
print('not True:', not True)
print('not False:', not False)
```

In fact, the logical operators will first force their arguments to be True or False before they give their answer.  So, in the case of `and` or `or`, they force force their left and right arguents to be `bool` values, before they calculate the answer.  So, in fact, you can use things other than exact True and False on either side of the `and` or `or`, as long as applying `bool(value)` to the thing to the left and right will produce a True or False value.  See {doc}`truthiness` for more detail.

## “If” statements, blocks and indention

A conditional statement in Python looks like this:

```{python}
my_var = 10
if my_var == 10:
    print("The conditional is True!")
    print("my_var does equal 10")
```

The first line of the conditional statement, that contains the conditional
test, ends in a colon.  Call this the *if test*.  There follow some lines
*indented* relative to the `if` test.  Call these indented lines the *if
block*.  Python executes the statements in the `if` block only when the
`if` test evaluates to True.  For example, in this case, the `if` test
evaluates to False, and the block does not execute:

```{python}
my_var = 11
# This time the conditional evaluates to False
if my_var == 10:  # the "if test"
    # The indented lines are the "if block"
    print("The conditional is True!")
    print("my_var does equal 10")
```

The first line that returns to the same level of indentation as the `if`
test line, closes the `if` block.

Unless the `if` block has a further indented block (for example, another
`if` block), then all the lines in the block must have the same indentation.

See note  for equivalent `if` statements in R
and MATLAB.

The `if` block may be followed by another block where the conditional is
`else:`. This block will only run if the initial conditional test evaluates
to False.

```{python}
my_var = 11
if my_var == 10:
    print("The conditional is True!")
    print("my_var does equal 10")
else:
    print("The conditional is False!")
    print("my_var does not equal 10")
```

There may be other conditional tests, with associated conditional blocks.
These tests use the contraction `elif conditional_test`, where `elif` is a
contraction for `else if`:

```{python}
my_var = 12
if my_var == 10:
    print("The conditional is True!")
    print("my_var does equal 10")
elif my_var == 11:
    print("The second conditional is True!")
    print("my_var does equal 11")
elif my_var == 12:
    print("The third conditional is True!")
    print("my_var does equal 12")
else:
    print("All conditionals are False!")
    print("my_var does not equal 10, 11 or 12")
```

## “While” statements

`while` statements are another example with an initial test followed by an
indented block.   Here’s an example where we find the largest [Fibonacci
number](https://en.wikipedia.org/wiki/Fibonacci_number) less than 1000:

```{python}
last_but_1 = 0
fibonacci = 1
while fibonacci < 1000:
    last_but_2 = last_but_1
    last_but_1 = fibonacci
    fibonacci = last_but_2 + last_but_1

print("Largest Fibonacci < 1000 is", last_but_1)
```

Notice the initial *while test*: `while fibonacci < 1000:`, followed by the
indented *while block*.  Unlike the `if` statement, Python will continue to
run the statements in the `while` block until the conditional in the
`while` test evaluates to False.

## Lists

Make a list like this:

```{python}
my_list = [9, 4, 7, 0, 8]
my_list
```

```{python}
type(my_list)
```

A list element can be any type of object, including another list:

```{python}
mixed_list = [9, 3.0, True, my_list]
mixed_list
```

```{python}
type(mixed_list)
```

A Python list is like a cell array in MATLAB, or a `list` in R.

## “for” loops and iteration

We can *iterate* over a list.  To iterate, means to fetch one element after
another from some container, such as a list.  We can use a `for` loop to
iterate over a list:

```{python}
for e in my_list:
    print(e)
```

The `for` loop has the same form as `if` statements and `while` loops,
with a first line ending in a colon, followed by an indented block.

The first line in the `for` loop is of form: `for loop_variable in
container:`.  The *container* is the container from which we will fetch the
elements.  At each iteration of the `for` loop, Python gets a new element
from the container to put into the *loop variable*.  For each element in the
container, Python executes the *for block*.

Note  shows equivalent `for` loops in Python, R and
MATLAB.

See
[`range`](https://docs.python.org/3/library/functions.html#func-range)
for a common way of writing a `for` loop that iterates over a sequence of
integers.

## Lists are sequences

A [sequence](https://docs.python.org/3/library/stdtypes.html#sequence-types-list-tuple-range)
is a category of Python objects that have a defined element order, have a
length, are iterable, can be indexed with integers, and *sliced* (see below).
If object `s` is a sequence, then:

* `s` has a length that can be found with `len(s)`;

* we can iterate over the elements in `s` with `for element in s: # do
  something with element`;

* we can return the element at position `n` with `s[n]`;

* we can get another sequence by *slicing* `s`.  For example, `s[0:n]`
  will give a new sequence containing the first `n` elements of `s`.

```{python}
# Has a length
len(my_list)
```

```{python}
# Is iterable
for e in my_list:
    print(e)
```

```{python}
# Can be indexed
my_list[1]
```

```{python}
# Can be sliced
my_list[0:2]
```

## Python indices are 0-based

Indices for Python sequences start at 0.  For Python, the first element is at
index 0, the second element is at index 1, and so on:

```{python}
# the first element
my_list[0]
```

```{python}
# the second element
my_list[1]
```

## Negative indices

Negative numbers as indices count back from the end of the list. For
example, use index `-1` to return the last element in the list:

```{python}
print(my_list)

# the last element
my_list[-1]
```

This is the third from last element:

```{python}
my_list[-3]
```

## Lists are mutable

A list is a *mutable* object. Mutable means, that we can change the elements
in the list, without creating a new list.

```{python}
my_list[1] = 99
my_list
```

In Python, variable names point to an object.

When you do `another_variable = a_variable`, you are telling the name
`another_variable` to point to the same object as the name
`a_variable`. When objects are mutable, this can be confusing:

```{python}
another_list = my_list
another_list
```

`my_list` points to a list object in memory. When you do
`another_list = my_list`, it tells Python that `another_list` points
to *the same object*. So, if we modify the list, pointed to by
`my_list`, we also modify the value of `another_list`, because `my_list`
and `another_list` point at the same list.

```{python}
my_list[1] = 101
another_list
```

## Adding lists

Adding two lists with `+` returns a new list that is the concatenation of
the two lists:

```{python}
new_list = my_list + [False, 1, 2]
new_list
```

## Appending and removing elements

You can append elements with the `append` method.

A method is a function attached to the object.  See Functions for more
on functions in Python.

We can see that `append` is a method by displaying the value of
`my_list.append`:

```{python}
my_list.append
```

To call the method, we add parentheses, surrounding any arguments we want to
pass into the method.  In this case we want to pass in the element to append:

```{python}
my_list.append(20)
my_list
```

Note that the `append` method does *not* return the list, it just changes
the list in-place. Python returns `None` from the `append` method:

```{python}
result = my_list.append(42)
result == None
```

This is also true for some other methods that modify the list in-place, such
as the `sort` method:

```{python}
new_list = [10, 1, 3]
result = new_list.sort()
# Return value is None
result == None
# But the original list now in ascending order from sort
new_list
```

You can remove elements from the list with the `pop` method:

```{python}
# Remove and return the last element of the list
my_list.pop()
my_list
```

```{python}
# Remove and return the third element of the list
my_list.pop(2)
my_list
```

## Slicing

You can return slices from any sequence, including lists, by putting a slice
specifier in square brackets. For example, this returns the first 3 elements
of the list:

```{python}
my_list[0:3]
```

The first number after the square bracket and before the colon is the *start*
index. In this case we start at the first element (element at index 0). The
second number, after the colon, is the *stop* index. This is the end index
*plus one*.  So we return elements at index 0, 1 and 2. That is, elements *up
to, but not including* 3.

If you omit the first number (the start index) Python assumes 0:

```{python}
my_list[:3]
```

If you omit the second number, Python assumes the length of the list as
the stop index.

```{python}
my_list[2:]
```

```{python}
my_list[2:len(my_list)]
```

You can omit both numbers, in which case you return all the elements of the
list. This can be useful if you want to make a new list that contains the same
elements as the first:

```{python}
another_list = my_list[:]
another_list
```

Because this is a new list object, you can change the original list without
changing the new list:

```{python}
my_list[1] = 999
another_list
```

You can also specify a second colon, and a third number. This third
number is the *step size*. For example, to get every second element of
the list:

```{python}
my_list[0:len(my_list):2]
```

```{python}
# Length of list assumed as stop index if omitted
my_list[0::2]
```

You can use negative numbers for the `start` and `stop` indices. As for
indexing, negative `start` and `stop` values count back from the end of
the list:

```{python}
print(my_list)
my_list[-4:-2]
```

Negative numbers for the `step` count backwards from the `start` to the
`stop` index:

```{python}
my_list[4:1:-1]
```

If you have a negative step size, and you don’t specify the start index, then
the start index defaults to the last element in the list. If you don’t specify
the stop index, it defaults to one prior to index 0:

```{python}
print(my_list)
my_list[-1:1:-1]
```

```{python}
my_list[:1:-1]
```

```{python}
my_list[-2::-1]
```

One consequence that is worth remembering is that the following idiom gives
you a reversed copy of the list:

```{python}
my_list[::-1]
```

(tuples)=

## Tuples

Tuples are almost the same as lists, except they are not mutable. That
is, you cannot change the elements of a tuple, or change the number of
elements.

```{python}
my_tuple = (9, 4, 7, 0, 8)
my_tuple
```

```{python tags=c("raises-exception")}
# This raises a TypeError
my_tuple[1] = 99
```

```{python tags=c("raises-exception")}
# This raises an AttributeError, because tuples have no append method
my_tuple.append(20)
```

Here’s an empty tuple:

```{python}
empty_tuple = ()
empty_tuple
```

A tuple with two elements:

```{python}
two_tuple = (1, 5)
two_tuple
```

As with lists, you can add tuples, forming the concatenation of the tuples:

```{python}
(1, 2) + (5, 6)
```

There is a little complication when making a tuple with one element:

```{python}
not_a_tuple = (1)
not_a_tuple
```

This is because Python can’t tell that you meant this to be a tuple,
rather than an expression with parentheses round it.  See {doc}`length_one_tuples` for an explanation.

To tell Python that you mean this to be a length-one tuple, add a comma after
the element, and before the closing parenthesis:

```{python}
length_one_tuple = (1,)
length_one_tuple
```

## Strings

Make a string like this:

```{python}
my_string = 'interesting text'
my_string
```

You can use single quotes or double quotes for your string, the two strings
are the same:

```{python}
another_string = "interesting text"
another_string
my_string == another_string
```

Convert other objects to strings using `str`:

```{python}
# Convert integer to string
str(9)
```

```{python}
# Convert floating point value to string
str(1.2)
```

## Strings are sequences

Like lists, strings are sequences (have length, can be iterated, can index,
can slice).

```{python}
# Length
len(my_string)
```

```{python}
# Iterable
for c in my_string:
    print(c)
```

```{python}
# Can index
my_string[1]
```

```{python}
# Can slice
my_string[1:5]
```

```{python}
# Can slice
my_string[::-1]
```

## Strings are immutable

Unlike lists, strings are immutable. You cannot change the characters within a
string:

```{python tags=c("raises-exception")}
# Raises a TypeError
my_string[1] = 'N'
```

## Adding strings

```{python}
my_string + ' with added insight'
```

## String methods

Strings have lots of interesting methods. In IPython, try tab-complete on a
string variable name, followed by a period – e.g. type `my_string.`,
followed by the tab key.  See also the [list of string methods in the Python
docs](http://docs.python.org/library/stdtypes.html#string-methods).

One interesting method is `replace`. It returns a new string that is a copy
of the input, but replacing instances of one string with another:

```{python}
another_string = my_string.replace('interesting', 'extraordinary')
another_string
```

Notice that the original string has not changed (it’s immutable):

```{python}
my_string
```

Use the `split` method to break a string into a list of strings.  By
default, `split` will split the string at any white space (spaces, tab
characters or line breaks):

```{python}
my_string.split()
```

Pass a character to `split` to split the string at that character:

```{python}
another_example = 'one:two:three'
another_example.split(":")
```

The `strip` method returns a new string with spaces, tabs and end of line
characters removed from the beginning and end of the string:

```{python}
# A string with a newline character at the end
my_string = ' a string\n'
my_string
my_string.strip()
```

## Inserting values into strings

We often want to insert values into strings.  This is called *string interpolation*.

For example, let us say we are running a shepherding business.  The shepherds rotate, some days Mary is on, sometimes Joseph, sometimes their son James.

Today, Mary is on.

```{python}
shepherd_name = "Mary"
```

She is herding 92 sheep.

```{python}
flock_size = 92
```

We may want to send out an announcement, say to the LED message board in front
of our shepherding business, that tells people which shepherd is on duty, and how many sheep they have.

So, if the number of sheep is 92, the message could be any of these three:

* "Shepherd Mary is on duty with 92 sheep." or
* "Shepherd Joseph is on duty with 92 sheep." or
* "Shepherd James is on duty with 92 sheep."

depending on the value in the `shepherd_name` variable.  And, of course, the `flock_size` could be almost any number.  So there are a huge number of potential sentences, that depend on the `shepherd_name` variable, and the `flock_size` variable.

Usually the best way to do this is using something called
[f-strings](https://datagy.io/python-f-strings).  These are strings with an `f` before the opening quote.  The `f` tells Python you may want to insert a variable into the string.  You specify variables to insert by putting them inside curly braces (`{}`) in the string, like this:

```{python}
# Notice the f before the first quote to tell Python there may
# be variables inside this string.
f"Shepherd {shepherd_name} is on duty with {flock_size} sheep."
```

There are many ways you can tell Python how to format the values you insert, and there are other, less common and useful ways to do this string interpolation.

For more details and more options, see: [Inserting values into strings](./string_formatting.Rmd).

## Range

`range` returns a *range object*.  It is a sequence, and so it is rather like
a list .  When you use `range` with one argument, the argument value is the
`stop` index.  For example, to make a `range` object generating the numbers
from 0 up to *but not including* 5:

```{python}
my_range = range(5)
my_range
```

You can make a range object into a list by using `list`:

```{python}
list(range(5))
```

A `range` object is a sequence:

```{python}
# Has a length
print('Length', len(my_range))
# Is iterable
for e in my_range:
    print('Value', e)
# Can be indexed
print('Value at position 1', my_range[1])
# Can be sliced
print('Slice 0:2', my_range[0:2])
```

Set the start element for `range` by passing two arguments:

```{python}
my_range = range(1, 7)
print(my_range)
print(list(my_range))
```

Set the step size with a third argument:

```{python}
my_range = range(1, 7, 2)
list(my_range)
```

One common use of `range` is to iterate over a sequence of numbers in a `for`
loop:

```{python}
for i in range(5):
    print(i)
```

## Sets

Sets are collections of unique elements, with no defined order.  Python
reserves the right to order set elements in any way it chooses:

```{python}
# Only unique elements collected in the set
my_set = set((5, 3, 1, 3))
my_set
```

Because there is no defined order, you cannot index into a set:

```{python tags=c("raises-exception")}
my_set[1]
```

You can add elements to a set with the `add` method:

```{python}
my_set.add(10)
my_set
```

Because set elements must be unique, if you add an element already in the set,
this does not change the set:

```{python}
my_set.add(5)
my_set
```

You can iterate over a set, but the order of elements is arbitrary.  You
cannot rely on the same order in any two runs of your program:

```{python}
# The order of elements is arbitrary
for element in my_set:
    print(element)
```

Look at the [methods of the set
object](https://docs.python.org/3/library/stdtypes.html#set-types-set-frozenset)
for interesting operations such as `difference`, `union`, `intersection` etc.

## Sets, lists and tuples are containers

A *container* is a Python object for which you can test an element for
membership.  So, if an object `c` is a container then we can test if an
element is in the container with `true_or_false = element in c`.

Be careful – the word `in` has different meanings in `for element in
c:` and `true_or_false = element in c`.  With  `for element in c:`,
`in` is a part of the `for` loop syntax.  With `true_or_false = element
in c`, `in` triggers a test of membership, returning True or False.

```{python}
5 in my_set
```

```{python}
11 in my_set
```

You can use `not in` to test if an element is not in a container:

```{python}
11 not in my_set
```

Lists and tuples are also containers:

```{python}
9 in [9, 4, 7, 0, 8]
```

```{python}
3 in (1, 3, 5)
```

## Dictionaries

A dictionary is collection of key / value pairs. The *key* is something that
identifies the element, and the *value* is the value corresponding to the
particular key.

For nearly all the dictionaries you will use in practice, the keys will be strings, but when you get more advanced, you will find that many things can be used as keys.

```{python}
# This is an empty dictionary
software = {}
```

Here we insert a new key / value mapping into the dictionary. The key is a
string — `'Python'` — and the corresponding value is an integer 50:

```{python}
software['Python'] = 100
software
```

Now we insert another key / value mapping:

```{python}
software['MATLAB'] = 50
software
```

Get the value corresponding to a key by indexing the dictionary with the key:

```{python}
software['Python']
```

We can iterate over the keys in the dictionary, but the order of the
keys depends on the order you put the keys into the dictionary — and it easy
to lose track of that order. As as result, we generally find we don't rely much on the particular order of the keys, when we iterate over them.

For example, here you see that the first key is 'Python' (the key for the first key/value pair we put in), even though 'Python' is after 'MATLAB' in an alphabetical sort, for example.

```{python}
for key in software.keys():
    print(key)
```

We can also iterate over the values, with the same constraint, that the order
depends on the order we put in the values:

```{python}
for value in software.values():
    print(value)
```

We can use the `items` method to iterate over the key / value pairs. In this
case each element is a tuple of length two, where the first element is the key
and the second element is the value:

```{python}
for key_value in software.items():
    print(key_value)
```

One way to construct a dictionary is with curly brackets, using colons to
separate the key and value, and commas to separate the pairs:

```{python}
software = {'MATLAB': 50, 'Python': 100}
software
```

Keys must be unique. A later key / value pair will overwrite an earlier key /
value pair that had the same key:

```{python}
software = {'MATLAB': 50, 'Python': 100, 'MATLAB': 45}
software  
```

## Dictionaries are containers

Dictionaries are also containers.  Python takes the elements in the container
to be the dictionary keys.  This is a convenient way to test if you already
have a key in a dictionary:

```{python}
'MATLAB' in software
```

```{python}
'happiness' in software
```

## “for”, “while”, “continue” and “break”

`for` statements and `while` statement are *loops*, because Python keeps
executing the `for` or `while` block until the `for` runs out of
elements or the `while` condition is False.  You can break out of a loop
using the `break` statement:

```{python}
for i in range(10):
    if i == 6:
        break
    print(i)
```

The `continue` statement short-circuits execution of the current iteration
of the `for` or `while` block, to continue with the next iteration:

```{python}
for i in range(10):
    if i == 6:
        continue
    print(i)
```

See “for” and “while”, “break” and “else:” for more on loops and `break`.

## Functions

Here we define our first function in Python:

```{python}
def my_function(an_argument):
    return an_argument + 1
```

The function definition begins with the `def` keyword followed by a space.
There follows the name of the function `my_function`. Next we have an open
parenthesis, followed by a specification of the arguments that the function
expects to be passed to it. In this case, the function expects a single
argument. In our case, the value of the input argument will be attached to the
name `an_argument` when the function starts to execute.  Last, we have an
indented block, with code that will run when the function is called. We can
return a value from the function using the `return` statement.

```{python}
my_function(10)
```

We called `my_function` by appending the opening parenthesis, and the
arguments, followed by the closing parenthesis. The function began to execute
with the variable `an_argument` set to 10. It returned 10 + 1 = 11.

A function need not accept any arguments:

```{python}
def my_second_function():
    return 42

my_second_function()
```

A function does not need to have a `return` statement.  If there is no
return statement, the function returns `None`:

```{python}
def function_with_no_return():
    # Function with no return statement
    a = 1

function_with_no_return() == None
```

A function can have more than one argument:

```{python}
def my_third_function(first_argument, second_argument):
    return first_argument + second_argument

my_third_function(10, 42)
```

## Default values for function arguments

The function definition can give a default value for a function argument:

```{python}
def my_fourth_function(first_argument, extra_argument=101):
    return first_argument + extra_argument
```

This function, like `my_third_function`, has two arguments, and we can call
it the same way that we call `my_third_function`:

```{python}
my_fourth_function(10, 42)
```

But, we can also omit the second argument, because it has a default value.  In
that case the argument will get its default value:

```{python}
my_fourth_function(10)  # Pass one argument, get default for second
```

So far we have passed in arguments by position, the first argument in our call
becoming the first argument in the function, and so on.  We can also pass in
arguments by name.  For example, we could pass in `extra_argument` by giving
the parameter name and value, like this:

```{python}
my_fourth_function(10, extra_argument=202)
```

Passing arguments this way can make the code easier to read, because the name
of the argument often gives a good clue as to its purpose in the function.  It
can also be useful with functions having many parameters with default values;
in that case using the argument name makes it easier to pass in one or few
values that are different from the defaults.

## Functions are objects too

Remember that everything in Python is an object. The function is itself an
object, where the name of the function is a variable, that refers to the
function:

```{python}
my_fourth_function
```

```{python}
type(my_fourth_function)
```

We call the function by adding the open parenthesis followed by the arguments
and the close parenthesis:

```{python}
my_fourth_function(10)
```

We can make a new name to point to this same function as easily as we can
could with any other Python variable:

```{python}
another_reference_to_func4 = my_fourth_function
print('Type of new variable', type(another_reference_to_func4))
# We call this function using the new name
another_reference_to_func4(10)
```

## Sorting

The Python function `sorted` returns a sorted list from something that
Python can iterate over:

```{python}
sorted('adcea')
```

```{python}
sorted((1, 5, 3, 2))
```

In order to do the sorting, Python compares the elements with
`one_element < another_element`. For example, to do the sort above,
Python needed results like:

```{python}
3 < 5
```

Sometimes you want to order the objects in some other way than simply
comparing the elements. If so, then you can define a *sort function*.  A sort
function is a function that accepts an element as its argument, and returns a
*sort value* for that element. Python does the sorting, not on the elements
themselves, but on the returned sort value for each element.

For example, let’s say we have first and last names stored as tuples:

```{python}
people = [('JB', 'Poline'), ('Matthew', 'Brett'), ('Mark', 'DEsposito')]
```

By default, Python compares tuples by comparing the first value first, then
the second value, and so on. This means for our case that we are sorting on
the first name:

```{python}
('Matthew', 'Brett') > ('Mark', 'DEsposito')
```

```{python}
sorted(people)
```

That may not be what you want.  You might want to sort by the last name, which
is the second value in the tuple.  In that case you can make a sort function,
that accepts the element as an input (the tuple in this case), and returns a
value:

```{python}
def get_last_name(person):
    return person[1]  # The last name
```

Remember everything in Python is an object. The function we have just defined
is also an object, with name `get_last_name`:

```{python}
get_last_name
```

We can pass this value to the `sorted` function as a sort function.  We will
pass this in using the sort function parameter name, which is `key`:

```{python}
sorted(people, key=get_last_name)
```

## Files

You can open a file in several different *modes*.  The mode specifies whether
you want to read or write the file, and whether the data in the file is, or
will be, text data (string) or binary data (bytes).  For example, here we open
a file for Writing Text (`wt`):

```{python}
my_file = open("a_text_file.txt", "wt")
```

If we had wanted to write binary (byte) data, we would have used `wb` for
the mode (Write Binary).

As usual, you can explore this new file object in IPython by appending the
object name with a period, and pressing the tab key to get a list of
attributes and methods – e.g. `myfile.` followed by tab.

To write to a file, use the `write` method.

```{python}
# Write a line of text with a newline character at the end
# The method returns the number of characters written
my_file.write("MATLAB is good for matrices\n")
# Another line
my_file.write("Python is good for coding\n")
```

You should close the file when you’ve finished with it:

```{python}
my_file.close()
```

To read a file, open the file in read mode:

```{python}
# Open file in Read Text mode
my_file2 = open("a_text_file.txt", "rt")
```

You can read all the contents in one shot by calling the `read` method
without arguments:

```{python}
contents = my_file2.read()
print(contents)
```

Remember to close the file afterwards:

```{python}
my_file2.close()
```

An open text file object is also *iterable*, meaning, that you can ask the
file object to return its contents line by line, in a `for` loop. Let’s open
the file again to show this in action:

```{python}
my_file2 = open("a_text_file.txt", "rt")
for line in my_file2:  # iterating over the file object
    print("Line is:", line)

my_file2.close()
```

We can delete files by using routines from various *modules*.

## Modules

Modules are libraries of data and functions.  Many of these come pre-installed with Python.

One common module is the `os` module, where `os` is short for [operating system](https://en.wikipedia.org/wiki/Operating_system).  The module has various functions that allow us to look at the system, and work with files.

In order to use the functions and data in modules, we have to *import* the module.  For example, imagine I forgot that fact, and tried to use the `unlink` function from the `os` module:

```{python tags=c("raises-exception")}
# Try to delete the file.
# Oops - I haven't imported 'os' yet!
os.unlink('a_text_file.txt')
```

This reminds me I need to *import* the module so I can use it:

```{python}
# Get the module ready to use.
import os
```

Now I can use the function as I intended:

```{python tags=c("raises-exception")}
# Actually delete the file.
os.unlink('a_text_file.txt')
```