CYBER WEEK SALE: Get 50% off your Pro annual membership using code [CYBER23](https://www.codecademy.com/checkout?periods=year&plan_id=proGoldAnnualV2&discountCode=CYBER23)


Dive into data types and learn to create your own Python Classes.

Introduction to Classes

Attributes can be added to user-defined objects after instantiation, so it's possible for an object to have some attributes that are not explicitly defined in an object's constructor. We can use the `dir()` function to investigate an object's attributes at runtime. `dir()` is short for _directory_ and offers an organized presentation of object attributes.

```py
class FakeDict:
  pass

fake_dict = FakeDict()
fake_dict.attribute = "Cool"

print(dir(fake_dict))
# Prints ['__class__', '__delattr__', '__dict__', '__dir__', '__doc__', '__eq__', '__format__', '__ge__', '__getattribute__', '__gt__', '__hash__', '__init__()', '__init_subclass__', '__le__', '__lt__', '__module__', '__ne__', '__new__', '__reduce__', '__reduce_ex__', '__repr__', '__setattr__', '__sizeof__', '__str__', '__subclasshook__', '__weakref__', 'attribute']
```

That's certainly a lot more attributes than we defined! Python automatically adds a number of attributes to all objects that get created. These internal attributes are usually indicated by double-underscores. But sure enough, `attribute` is in that list.

Do you remember being able to use `type()` on Python's native data types? This is because they are also objects in Python. Their classes are `int`, `float`, `str`, `list`, and `dict`. These Python classes have special syntax for their instantiation, `1`, `1.0`, `"hello"`, `[]`, and `{}` specifically. But these instances are still full-blown objects to Python.

```py
fun_list = [10, "string", {'abc': True}]

type(fun_list)
# Prints <class 'list'>

print(dir(fun_list))
# Prints ['__add__', '__class__', [...], 'append', 'clear', 'copy', 'count', 'extend', 'index', 'insert', 'pop', 'remove', 'reverse', 'sort']
```

Above we define a new list. We check its type and see that's an instantiation of class `list`. We use `dir()` to explore its attributes, and it gives us a large number of internal Python dunder attributes, but afterward, we get the usual list methods.

Everything is an Object

Instance variables and class variables are both accessed similarly in Python. This is no mistake, they are both considered _attributes_ of an object. If we attempt to access an attribute that is neither a class variable nor an instance variable of the object Python will throw an `AttributeError`.

```py
class NoCustomAttributes:
  pass

attributeless = NoCustomAttributes()

try:
  attributeless.fake_attribute
except AttributeError:
  print("This text gets printed!")

# prints "This text gets printed!"
```

What if we aren't sure if an object has an attribute or not? [`hasattr()`](https://www.codecademy.com/resources/docs/python/built-in-functions/hasattr?page_ref=catalog) will return `True` if an object has a given attribute and `False` otherwise. If we want to get the actual value of the attribute, [`getattr()`](https://www.codecademy.com/resources/docs/python/built-in-functions/getattr?page_ref=catalog) is a Python function that will return the value of a given object and attribute. In this function, we can also supply a third argument that will be the default if the object does not have the given attribute. 

The syntax and parameters for these functions look like this:

`hasattr(object, “attribute”)` has two parameters:
- _object_ : the object we are testing to see if it has a certain attribute
- _attribute_ : name of attribute we want to see if it exists

`getattr(object, “attribute”, default)` has three parameters (one of which is optional):
- _object_ : the object whose attribute we want to evaluate
- _attribute_ : name of attribute we want to evaluate
- _default_ : the value that is returned if the attribute does not exist (note: this parameter is **optional**)


Calling those functions looks like this:

```py
hasattr(attributeless, "fake_attribute")
# returns False

getattr(attributeless, "other_fake_attribute", 800)
# returns 800, the default value
```

Above, we checked if the `attributeless` object has the attribute `.fake_attribute`. Since it does not, `hasattr()` returned `False`. After that, we used `getattr` to attempt to retrieve `.other_fake_attribute`. Since `.other_fake_attribute` isn't a real attribute on `attributeless`, our call to `getattr()` returned the supplied default value `800`, instead of throwing an `AttributeError`.


Attribute Functions

_Methods_ are functions that are defined as part of a class. The first argument in a method is always the object that is calling the method. Convention recommends that we name this first argument `self`. Methods always have at least this one argument. 

We define methods similarly to functions, except that they are indented to be part of the class.

```py
class Dog:
  dog_time_dilation = 7

  def time_explanation(self):
    print("Dogs experience {} years for every 1 human year.".format(self.dog_time_dilation))

pipi_pitbull = Dog()
pipi_pitbull.time_explanation()
# Prints "Dogs experience 7 years for every 1 human year."
```

Above we created a `Dog` class with a `.time_explanation()` method that takes one argument, `self`, which refers to the object calling the function. We created a `Dog` named `pipi_pitbull` and called the `.time_explanation()` method on our new object for Pipi.

Notice we didn't pass any arguments when we called `.time_explanation()`, but were able to refer to `self` in the function body. When you call a method it automatically passes the object calling the method as the first argument.


Methods

When we want the same data to be available to every instance of a class we use a _class variable_. A class variable is a variable that's the same for every instance of the class. 

You can define a class variable by including it in the indented part of your class definition, and you can access all of an object's class variables with `object.variable` syntax.

```py
class Musician:
  title = "Rockstar"

drummer = Musician()
print(drummer.title)
# prints "Rockstar"
```

Above we defined the class `Musician`, then instantiated `drummer` to be an object of type `Musician`. We then printed out the drummer's `.title` attribute, which is a class variable that we defined as the string "Rockstar". 

If we defined another musician, like `guitarist = Musician()` they would have the same `.title` attribute.

**NOTE**: Class variables are often referenced with a leading period, like `.title` above. This is done to quickly show that the variable belongs to a class and must be accessed with dot notation, like `drummer.title`.

Class Variables

There are several methods that we can define in a Python class that have special behavior. These methods are sometimes called "magic," because they behave differently from regular methods. Another popular term is [_dunder methods_](https://www.codecademy.com/resources/docs/python/dunder-methods?page_ref=catalog), so-named because they have two underscores (double-underscore abbreviated to "dunder") on either side of them.

The first dunder method we're going to use is the [`__init__()`](https://www.codecademy.com/resources/docs/python/dunder-methods/init?page_ref=catalog) method (note the two underscores before and after the word "init"). This method is used to _initialize_ a newly created object. It is called every time the class is instantiated.

Methods that are used to prepare an object being instantiated are called _constructors_. The word "constructor" is used to describe similar features in other object-oriented programming languages, but programmers who refer to a constructor in Python are usually talking about the `__init__()` method.

```py
class Shouter:
  def __init__(self):
    print("HELLO?!")

shout1 = Shouter()
# prints "HELLO?!"

shout2 = Shouter()
# prints "HELLO?!"
```

Above, we created a class called `Shouter` and every time we create an instance of `Shouter` the program prints out a shout. Don't worry, this doesn't hurt the computer at all.

Pay careful attention to the instantiation syntax we use. `Shouter()` looks a lot like a function call, doesn't it? If it's a function, can we pass parameters to it? We absolutely can, and those parameters will be received by the `__init__()` method.

```py
class Shouter:
  def __init__(self, phrase):
    # make sure phrase is a string
    if type(phrase) == str:

      # then shout it out
      print(phrase.upper())

shout1 = Shouter("shout")
# prints "SHOUT"

shout2 = Shouter("shout")
# prints "SHOUT"

shout3 = Shouter("let it all out")
# prints "LET IT ALL OUT"
```

Above, we've updated our `Shouter` class to take the additional parameter `phrase`. When we created each of our objects, we passed an argument to the constructor. The constructor takes the argument `phrase` and, if it's a string, prints out the all-caps version of `phrase`.


Constructors

A class doesn't accomplish anything simply by being defined. A class must be _instantiated_. In other words, we must create an _instance_ of the class, in order to breathe life into the schematic.  

Instantiating a class looks a lot like calling a function.  We would be able to create an instance of our defined `CoolClass` as follows:

```py
cool_instance = CoolClass()
```

Above, we created an object by adding parentheses to the name of the class. We then assigned that new instance to the variable `cool_instance` for safe-keeping so we can access our instance of `CoolClass` at a later time. 

Instantiation

Python equips us with many different ways to store data. A `float` is a different kind of number from an `int`, and we store different data in a `list` than we do in a `dict`. These are known as different _types_. We can check the type of a Python variable using the [`type()`](https://www.codecademy.com/resources/docs/python/built-in-functions/type?page_ref=catalog) function.

```py
a_string = "Cool String"
an_int = 12

print(type(a_string))
# prints "<class 'str'>"

print(type(an_int))
# prints "<class 'int'>"
```
Above, we defined two variables, and checked the `type` of these two variables. A variable's type determines what you can do with it and how you can use it. You can't `.get()` something from an integer, just as you can't add two dictionaries together using `+`. This is because those operations are defined at the `type` level.


Types

A [_class_](https://www.codecademy.com/resources/docs/python/classes?page_ref=catalog) is a template for a data type. It describes the kinds of information that class will hold and how a programmer will interact with that data. Define a class using the [`class`](https://www.codecademy.com/resources/docs/python/keywords/class?page_ref=catalog) keyword. [PEP 8 Style Guide for Python Code](https://www.python.org/dev/peps/pep-0008/#class-names) recommends capitalizing the names of classes to make them easier to identify.

```py
class CoolClass:
  pass
```

In the above example we created a class and named it `CoolClass`. We used the `pass` keyword in Python to indicate that the body of the class was intentionally left blank so we don't cause an `IndentationError`. We’ll learn about all the things we can put in the body of a class in the next few exercises.

Class

Since we can already use dictionaries to store key-value pairs, using objects for that purpose is not really useful. Instance variables are more powerful when you can guarantee a rigidity to the data the object is holding. 

This convenience is most apparent when the constructor creates the instance variables using the arguments passed into it. If we were creating a search engine and wanted to create a class to hold each search entry, we could do so like this:

```py
class SearchEngineEntry:
  def __init__(self, url):
    self.url = url

codecademy = SearchEngineEntry("www.codecademy.com")
wikipedia = SearchEngineEntry("www.wikipedia.org")

print(codecademy.url)
# prints "www.codecademy.com"

print(wikipedia.url)
# prints "www.wikipedia.org"
```

In the preceding code sample, we define a `SearchEngineEntry` class, which contains a constructor with two parameters, `self` and `url`. Inside the constructor body, we create an instance variable named `self.url` and assign it the value of the `url` parameter that is passed into the constructor. 

We then create the `codecademy` instance of `SearchEngineEntry` by passing the URL as an argument to the constructor. After `codecademy` is defined, printing `codecademy.url` shows that the URL has been assigned to the `url` instance variable of `codecademy`. Similarly, `wikipedia.url` holds the value that was passed into the constructor when `wikipedia` was defined.

Since the `self` keyword refers to the object and not the class being called, we can define a `.secure()` method on the `SearchEngineEntry` class that returns the secure link to an entry.

```py
class SearchEngineEntry:
  secure_prefix = "https://"
  def __init__(self, url):
    self.url = url

  def secure(self):
    return "{prefix}{site}".format(prefix=self.secure_prefix, site=self.url)

codecademy = SearchEngineEntry("www.codecademy.com")
wikipedia = SearchEngineEntry("www.wikipedia.org")

print(codecademy.secure())
# prints "https://www.codecademy.com"

print(wikipedia.secure())
# prints "https://www.wikipedia.org"
```

Above, we define our `.secure()` method to take just the one required argument, `self`. We access both the class variable `self.secure_prefix` and the instance variable `self.url` to return a secure URL. 

This is the strength of writing object-oriented programs. We can write our classes to structure the data that we need and write methods that will interact with that data in a meaningful way.

Self

Methods can also take more arguments than just `self`:

```py
class DistanceConverter:
  kms_in_a_mile = 1.609
  def how_many_kms(self, miles):
    return miles * self.kms_in_a_mile

converter = DistanceConverter()
kms_in_5_miles = converter.how_many_kms(5)
print(kms_in_5_miles)
# prints "8.045"
```

Above we defined a `DistanceConverter` class, instantiated it, and used it to convert 5 miles into kilometers. Notice again that even though `.how_many_kms()` takes two arguments in its definition, we only pass `miles`, because `self` is implicitly passed (and refers to the object `converter`).

Methods with Arguments

One of the first things we learn as programmers is how to print out information that we need for debugging. Unfortunately, when we print out an object we get a default representation that seems fairly useless.

```py
class Employee():
  def __init__(self, name):
    self.name = name

argus = Employee("Argus Filch")
print(argus)
# prints "<__main__.Employee object at 0x104e88390>"
```

This default string representation gives us some information, like where the class is defined and our computer's memory address where this object is stored, but is usually not useful information to have when we are trying to debug our code.

We learned about the dunder method `__init__()`. Now, we will learn another dunder method called  [`__repr__()`](https://www.codecademy.com/resources/docs/python/dunder-methods/repr?page_ref=catalog). This is a method we can use to tell Python what we want the _string representation_ of the class to be. `__repr__()` can only have one parameter, `self`, and must return a string.

In our `Employee` class above, we have an instance variable called `.name` that should be unique enough to be useful when we're printing out an instance of the  `Employee` class.

```py
class Employee():
  def __init__(self, name):
    self.name = name

  def __repr__(self):
    return self.name

argus = Employee("Argus Filch")
print(argus)
# prints "Argus Filch"
```
We implemented the `__repr__()` method and had it return the `.name` attribute of the object. When we printed the object out it simply printed the `.name` of the object! Cool!

String Representation

A class instance is also called an _object_. The pattern of defining classes and creating objects to represent the responsibilities of a program is known as [_Object Oriented Programming_](https://www.codecademy.com/resources/docs/general/programming-paradigms/object-oriented-programming?page_ref=catalog) or OOP.

Instantiation takes a class and turns it into an object, the `type()` function does the opposite of that. When called with an object, it returns the class that the object is an instance of.

```py
print(type(cool_instance))
# prints "<class '__main__.CoolClass'>"
```
We then print out the `type() of cool_instance` and it shows us that this object is of type `__main__.CoolClass`. 

In Python `__main__`  means "this current file that we're running" and so one could read the output from `type()` to mean  "the class `CoolClass` that was defined here, in the script you're currently running."

Object-Oriented Programming

So far we've covered what a data type actually is in Python. We explored what the functionality of Python's built-in types (also referred to as _primitives_) are. We learned how to create our own data types using the `class` keyword.

We explored the relationship between a class and an object &mdash; we create objects when we instantiate a class, we find the class when we check the `type()` of an object. We learned the difference between class variables (the same for all objects of a class) and instance variables (unique for each object).

We learned about how to define an object's functionality with methods. We created multiple objects from the same class, all with similar functionality, but with different internal data. They all had the same methods, but produced different output because they were different instances.

Take a moment to congratulate yourself, object-oriented programming is a complicated concept.

Review

We've learned so far that a class is a schematic for a data type and an object is an instance of a class, but why is there such a strong need to differentiate the two if each object can only have the methods and class variables the class has? This is because each instance of a class can hold different kinds of data. 

The data held by an object is referred to as an _instance variable_. Instance variables aren't shared by all instances of a class &mdash; they are variables that are specific to the object they are attached to.

Let's say that we have the following class definition:

```py
class FakeDict:
  pass
```

We can instantiate two different objects from this class, `fake_dict1` and `fake_dict2`, and assign instance variables to these objects using the same attribute notation that was used for accessing class variables.

```py
fake_dict1 = FakeDict()
fake_dict2 = FakeDict()

fake_dict1.fake_key = "This works!"
fake_dict2.fake_key = "This too!"

# Let's join the two strings together!
working_string = "{} {}".format(fake_dict1.fake_key, fake_dict2.fake_key)
print(working_string)
# prints "This works! This too!"
```