How to write a Python module/package?

A module is a file containing Python definitions and statements. The file name is the module name with the suffix .py

create hello.py then write the following function as its content:

def helloworld():   print "hello"

Then you can import hello:

>>> import hello>>> hello.helloworld()'hello'>>>

To group many .py files put them in a folder. Any folder with an __init__.py is considered a module by python and you can call them a package

|-HelloModule  |_ __init__.py  |_ hellomodule.py

You can go about with the import statement on your module the usual way.

For more information, see 6.4. Packages.

Python 3 - UPDATED 18th November 2015

Found the accepted answer useful, yet wished to expand on several points for the benefit of others based on my own experiences.

Module: A module is a file containing Python definitions and statements. The file name is the module name with the suffix .py appended.

Module Example: Assume we have a single python script in the current directory, here I am calling it mymodule.py

The file mymodule.py contains the following code:

def myfunc():    print("Hello!")

If we run the python3 interpreter from the current directory, we can import and run the function myfunc in the following different ways (you would typically just choose one of the following):

>>> import mymodule>>> mymodule.myfunc()Hello!>>> from mymodule import myfunc>>> myfunc()Hello!>>> from mymodule import *>>> myfunc()Hello!

Ok, so that was easy enough.

Now assume you have the need to put this module into its own dedicated folder to provide a module namespace, instead of just running it ad-hoc from the current working directory. This is where it is worth explaining the concept of a package.

Package: Packages are a way of structuring Python’s module namespace by using “dotted module names”. For example, the module name A.B designates a submodule named B in a package named A. Just like the use of modules saves the authors of different modules from having to worry about each other’s global variable names, the use of dotted module names saves the authors of multi-module packages like NumPy or the Python Imaging Library from having to worry about each other’s module names.

Package Example: Let's now assume we have the following folder and files. Here, mymodule.py is identical to before, and __init__.py is an empty file:

.└── mypackage    ├── __init__.py    └── mymodule.py

The __init__.py files are required to make Python treat the directories as containing packages. For further information, please see the Modules documentation link provided later on.

Our current working directory is one level above the ordinary folder called mypackage

$ lsmypackage

If we run the python3 interpreter now, we can import and run the module mymodule.py containing the required function myfunc in the following different ways (you would typically just choose one of the following):

>>> import mypackage>>> from mypackage import mymodule>>> mymodule.myfunc()Hello!>>> import mypackage.mymodule>>> mypackage.mymodule.myfunc()Hello!>>> from mypackage import mymodule>>> mymodule.myfunc()Hello!>>> from mypackage.mymodule import myfunc>>> myfunc()Hello!>>> from mypackage.mymodule import *>>> myfunc()Hello!

Assuming Python 3, there is excellent documentation at: Modules

In terms of naming conventions for packages and modules, the general guidelines are given in PEP-0008 - please see Package and Module Names

Modules should have short, all-lowercase names. Underscores can be used in the module name if it improves readability. Python packages should also have short, all-lowercase names, although the use of underscores is discouraged.

Since nobody did cover this question of the OP yet:

What I wanted to do:

Make a python module install-able with "pip install ..."

Here is an absolute minimal example, showing the basic steps of preparing and uploading your package to PyPI using setuptools and twine.

This is by no means a substitute for reading at least the tutorial, there is much more to it than covered in this very basic example.

Creating the package itself is already covered by other answers here, so let us assume we have that step covered and our project structure like this:

.└── hellostackoverflow/    ├── __init__.py    └── hellostackoverflow.py

In order to use setuptools for packaging, we need to add a file setup.py, this goes into the root folder of our project:

.├── setup.py└── hellostackoverflow/    ├── __init__.py    └── hellostackoverflow.py

At the minimum, we specify the metadata for our package, our setup.py would look like this:

from setuptools import setupsetup(    name='hellostackoverflow',    version='0.0.1',    description='a pip-installable package example',    license='MIT',    packages=['hellostackoverflow'],    author='Benjamin Gerfelder',    author_email='benjamin.gerfelder@gmail.com',    keywords=['example'],    url='https://github.com/bgse/hellostackoverflow')

Since we have set license='MIT', we include a copy in our project as LICENCE.txt, alongside a readme file in reStructuredText as README.rst:

.├── LICENCE.txt├── README.rst├── setup.py└── hellostackoverflow/    ├── __init__.py    └── hellostackoverflow.py

At this point, we are ready to go to start packaging using setuptools, if we do not have it already installed, we can install it with pip:

pip install setuptools

In order to do that and create a source distribution, at our project root folder we call our setup.py from the command line, specifying we want sdist:

python setup.py sdist

This will create our distribution package and egg-info, and result in a folder structure like this, with our package in dist:

.├── dist/├── hellostackoverflow.egg-info/├── LICENCE.txt├── README.rst├── setup.py└── hellostackoverflow/    ├── __init__.py    └── hellostackoverflow.py

At this point, we have a package we can install using pip, so from our project root (assuming you have all the naming like in this example):

pip install ./dist/hellostackoverflow-0.0.1.tar.gz

If all goes well, we can now open a Python interpreter, I would say somewhere outside our project directory to avoid any confusion, and try to use our shiny new package:

Python 3.5.2 (default, Sep 14 2017, 22:51:06) [GCC 5.4.0 20160609] on linuxType "help", "copyright", "credits" or "license" for more information.>>> from hellostackoverflow import hellostackoverflow>>> hellostackoverflow.greeting()'Hello Stack Overflow!'

Now that we have confirmed the package installs and works, we can upload it to PyPI.

Since we do not want to pollute the live repository with our experiments, we create an account for the testing repository, and install twine for the upload process:

pip install twine

Now we're almost there, with our account created we simply tell twine to upload our package, it will ask for our credentials and upload our package to the specified repository:

twine upload --repository-url https://test.pypi.org/legacy/ dist/*

We can now log into our account on the PyPI test repository and marvel at our freshly uploaded package for a while, and then grab it using pip:

pip install --index-url https://test.pypi.org/simple/ hellostackoverflow

As we can see, the basic process is not very complicated. As I said earlier, there is a lot more to it than covered here, so go ahead and read the tutorial for more in-depth explanation.