Python module - working with modules in Python (2024)

last modified January 29, 2024

In this article we work with Python modules. Several examples show how to createand use Python modules.

A module is a file containing Python code. Python modules have the.py extension.

Python code can be managed using:

functions
classes
modules
packages

Python modules are used to organize Python code. For example, database relatedcode is placed inside a database module, security code in a security module etc.Smaller Python scripts can have one module. But larger programs are split intoseveral modules. Modules are grouped together to form packages.

The .pyc file

Python caches the compiled content of modules in .pyc files tospeed up loading modules. Python compiles the program source code into bytecode. To improve performance, it caches the byte code on the file systemwhenever the source file has changes.

This caching makes loading of Python modules much faster because the compilationphase can be bypassed. Python caches the compiled version of each module in the__pycache__ directory under the namemodule.version.pyc.

Python checks the modification date of the source against the compiled versionto see if it's out of date and needs to be recompiled

main.py

#!/usr/bin/pythonimport compileallcompileall.compile_dir('lib/', force=True)

The compileall module can be used to programtically compile Pythonmodules.

Python module names

A module name is the file name with the .py extension. When wehave a file called empty.py, empty is the module name. The__name__ is a variable that holds the name of the module beingreferenced. The current module, the module being executed (called also the mainmodule) has a special name: '__main__'. With this name it can bereferenced from the Python code.

We have two files in the current working directory: empty.py andtest_empty.py. The second module is the main module, which isexecuted. It imports the first module. Modules are imported using theimport keyword.

empty.py

"""An empty module"""

This is empty.py module.

test_empty.py

#!/usr/bin/pythonimport emptyimport sysprint(__name__)print(empty.__name__)print(sys.__name__)

In this code example we import two modules: the built-in module sysand one custom module empty. We print the names of modules to theconsole.

$ ./test_empty.py__main__emptysys

The name of the module, which is being executed is always '__main__'.Other modules are named after the file name.Modules can be imported into other modules using the import keyword.

Python locating modules

When a module is imported the interpreter first searches for a built-in modulewith that name. If not found, it then searches in a list of directories given bythe variable sys.path. The sys.path is a list ofstrings that specifies the search path for modules. It consists of the currentworking directory, directory names specified in the PYTHONPATHenvironment variable, and some additional installation dependent directories. Ifthe module is not found, an ImportError is raised.

locating_modules.py

#!/usr/bin/pythonimport sysimport textwrapsp = sorted(sys.path)dnames = ', '.join(sp)print(textwrap.fill(dnames))

The script prints all directories from sys.path variable.

Python import keyword

The import keyword can be used in several ways.

from module import *

This construct will import all Python definitions into the namespace of anothermodule. There is one exception. Objects beginning with underscore character_ are not imported. They are expected to be used only internally bythe module being imported. This way of importing modules is not recommended.

everything.py

#!/usr/bin/pythonfrom math import *print(cos(3))print(pi)

This import construct has imported all definitions from the built-inmath module. We can call the math functions directly, withoutreferencing the math module.

$ ./everything.py-0.98999249660044543.141592653589793

The use of this import construct may result in namespace pollution. We may haveseveral objects of the same name and their definitions can be overridden.

pollution.py

#!/usr/bin/pythonfrom math import *pi = 3.14print(cos(3))print(pi)

The example will print 3.14 to the console. Which may not be what we wanted. Thenamespace pollution may become critical in larger projects.

Python objects that are not imported

The following example shows definitions that are not beingimported using this import construct.

names.py

#!/usr/bin/python"""names is a test module"""_version = 1.0names = ["Paul", "Frank", "Jessica", "Thomas", "Katherine"]def show_names(): for i in names: print(i)def _show_version(): print(_version)

test_names.py

#!/usr/bin/pythonfrom names import *print(locals())show_names()

The _version variable and the _show_version functionare not imported into the test_names module. We do not see them inthe namespace. The locals function give us all the definitionsavailable in the module.

Importing specific objects

With the from and import keywords,it is possible to import only some objects.

from module import fun, var

This import construct imports only specific objects from a module. This way weimport only definitions that we need.

import_specific.py

#!/usr/bin/pythonfrom math import sin, piprint(sin(3))print(pi)

We import two objects from the math module. There is no way how wecould reference other definitions such as a cos function.

imnames.py

#!/usr/bin/pythonfrom names import _version, _show_versionprint(_version)_show_version()

We could also import definitions beginning with an underscore. But this is a badpractice.

Python import module

The last construct is most widely used.

import module

It prevents the namespace pollution and enables to access all definitions from amodule.

impmod.py

#!/usr/bin/pythonimport mathpi = 3.14print(math.cos(3))print(math.pi)print(math.sin(3))print(pi)

In this case, we reference the definitions via the module name. As we can see,we are able to use both pi variables. Our definition and the one from themath module.

$ ./impmod.py-0.98999249660044543.1415926535897930.14112000805986723.14

Python aliasing modules

We can create an alias for the module with the as keyword.

importas.py

#!/usr/bin/python# importas.pyimport math as mprint(m.pi)print(m.cos(3))

We can change the name through which we can reference the module. To do this, weuse the as keyword.

$ ./importas.py3.14159265359-0.9899924966

ImportError

An ImportError is raised if a module cannot be imported.

importerror.py

#!/usr/bin/pythontry: import empty2except ImportError as e: print('Failed to import:', e)

We have not created an empty2 module. Therefore an exception israised.

$ ./importerror.pyFailed to import: No module named empty2

Executing Python modules

Modules can be imported into other modules or they can be also executed.Module authors often create a testing suiteto test the module. Only if the module is executed as a script,the __name__ attribute equals to '__main__'.

We will demonstrate this on a fibonacci module. Fibonacci numbers isa sequence of numbers, where each is the sumof its two immediate predecessors.

fibonacci.py

#!/usr/bin/python"""A module containing the fibonaccifunction."""def fib(n): a, b = 0, 1 while b < n: print(b, end=" ") (a, b) = (b, a + b)# testingif __name__ == '__main__': fib(500)

The module can be normally imported as usual. The module can be also executed.

$ ./fibonacci.py1 1 2 3 5 8 13 21 34 55 89 144 233 377

If we do import the fibonacci module, the test is not executed automatically.

>>> import fibonacci as fib>>> fib.fib(500)1 1 2 3 5 8 13 21 34 55 89 144 233 377

The fibonacci module is imported and the fib function is executed.

Python dir function

The built-in dir function gives a sorted list of stringscontaining the names defined by a module.

dirfun.py

#!/usr/bin/python"""This is dirfun module"""import math, sysversion = 1.0names = ["Paul", "Frank", "Jessica", "Thomas", "Katherine"]def show_names(): for i in names: print(i)print(dir())

In this module, we import two system modules. We define a variable, a list anda function.

print(dir())

The dir function returns all the names available in thecurrent namespace of the module.

$ ./dirfun.py['__builtins__', '__cached__', '__doc__', '__file__', '__loader__', '__name__','__package__', '__spec__', 'math', 'names', 'show_names', 'sys', 'version']

We can see some built-in names like '__file__' or'__name__' and all the others that we have defined and imported.

Python globals function

The globals function returns a dictionary that represents thecurrent global namespace. It is a dictionary of global names and their values.It is the dictionary of the current module.

globalsfun.py

#!/usr/bin/pythonimport textwrapversion = 1.0def myfun(): passgl = globals()gnames = ', '.join(gl)print(textwrap.fill(gnames))

We use the globals function to print all the global names of thecurrent module.

$ ./globalsfun.pytextwrap, __package__, version, __builtins__, __name__, __spec__,__doc__, gl, __cached__, myfun, __loader__, __file__

These are the global names of the current module.

Python module attribute

The __module__ class attribute has the name of the module in whichthe class is defined.

animals.py

"""module animals"""class Cat: passclass Dog: pass

This are the contents of the animals.py file. We have two classes.

mclass.py

#!/usr/bin/pythonfrom animals import Catclass Being: passb = Being()print(b.__module__)c = Cat()print(c.__module__)

In this code we use the __module__ attribute.

from animals import Cat

From the animals module, we import the Cat class.

class Being: pass

In the current module, we define a class Being.

b = Being()print(b.__module__)

An instance of the Being class is created. We print the nameof its module.

c = Cat()print(c.__module__)

We create an object from the Cat class. We also print the modulewhere it was defined.

$ ./mclass.py__main__animals

The current module's name is '__main__'. And the Cat's module name isanimals.

Source

Python modules - language reference

In this article we covered Python modules.

Author

My name is Jan Bodnar and I am a passionate programmer with many years ofprogramming experience. I have been writing programming articles since 2007. Sofar, I have written over 1400 articles and 8 e-books. I have over eight years ofexperience in teaching programming.

List all Python tutorials.