What is __ doc __ in Python?

In Python, __doc__ is a special attribute that stores a string defining the documentation for a module, class, method, or function. It’s typically derived from a docstring β€” the first statement in the defined block, enclosed in triple quotes β€” providing a convenient way to access descriptive texts about the code’s functionality.

You can define the docstring using a string surrounded by triple quotes as shown in the example:

class Dog:
    """Your best friend."""

    def do_nothing(self):
        pass


print(Dog.__doc__)
# Your best friend. 

This Python code defines a class Dog with a docstring “Your best friend.” and a method do_nothing that does nothing (using the pass statement). The print(Dog.__doc__) statement then prints the docstring of the Dog class.

  • class Dog: This line starts the definition of a class named Dog.
  • """Your best friend.""": This is a docstring, a special string literal used to document the Dog class. It describes what the class represents or does.
  • def do_nothing(self): This defines a method named do_nothing for the Dog class. The self parameter refers to the instance of the class on which the method is called.
  • pass: This is a placeholder statement that does nothing. It’s used here because Python syntax requires at least one statement in the method body, and do_nothing is intended to be an empty method.
  • print(Dog.__doc__): This line prints the docstring of the Dog class. In Python, __doc__ is a special attribute that holds the docstring of a class, function, method, or module. In this case, it prints "Your best friend."

The code demonstrates how to define a simple class with a docstring and access that docstring using the __doc__ attribute. The output of this code will be: Your best friend.

Try it yourself in the interactive Python shell:

If you already learned something from this tutorial, why not joining my free Python training program? I call it the Finxter Email Computer Science Academy—and it’s just that: a free, easy-to-use email academy that teaches you Python in small daily doses for beginners and pros alike!

πŸ”— Recommended: Learn Python and join the Free Finxter Email Computer Science Academy. It’s fun!

Let’s go back to the docstring. As everything is an object in Python (even functions), you can also define a docstring on functions:

def bark():
    """Wuff"""
    pass


print(bark.__doc__)
# Wuff

Note that if you don’t define the docstring, the return value is None.

def bark():
    pass


print(bark.__doc__)
# None

Why Use Docstrings?

A great advantage of having defined docstrings in your code is that you now can create your own great-looking documentation programmatically. With tools such as Sphinx, it’s super easy to create something like this for your own code projects–only using the __doc__ values defined in the code.

In particular, you can use docstrings in Python to:

  • Document functions, classes, and modules directly in the code.
  • Enhance code readability and maintainability.
  • Standardize documentation for easy access through tools like Sphinx and IDEs.
  • Enable runtime introspection of documentation via the help() function.
  • Facilitate testing by integrating with testing frameworks for descriptive test cases.

Best Practices Docstring

There are a couple of best-practices called Docstring Conventions as defined in the official PEP standard. Adhere to them when defining your docstrings. Here are the 7 most important docstring conventions:

  1. All modules, function, methods, and classes should have docstrings.
  2. Always use """triple double quotes""" around your docstrings for consistency reasons.
  3. Use triple quotes even if the docstring fits into a single line. This allows for easy expansion later.
  4. No blank line before or after the docstring—except for classes where you should add one line after the docstring.
  5. Use a phrase that describes what your code is doing such as """Do X and return Y.""" ending in a period. Don’t use a description such as """Does X and returns Y.""".
  6. Multi-line docstrings start with a summary line (like the one-liner docstring), followed by a blank line, followed by a closer description such as argument --- name of the person (string) to describe one of the arguments of the function or method. For example, you can use one line per argument.
  7. Start a multi-line docstring immediately in the same line of the opening """triple double strings... rather than starting the text in a new line.

If you’re a perfectionist or beyond intermediate code level, check out the official documentation that contains more examples.