If you have been coding in python, even for a while, you must be familiar with the concept of ‘pip’ by now. It is a package management system used to install and manage software packages/libraries written in Python.
Then one may ask that where are all these packages/libraries stored? It is obvious that there must be a large “on-line repository” which stores all this code. And the answer to this question is Python Package Index (PyPI).
PyPI is the official third-party software repository for Python. At the time of writing this article, PyPI was already hosting 95971 packages!
pip uses PyPI as the default source for packages and their dependencies. So whenever you type:
pip install package_name
pip will look for that package on PyPI and if found, it will download and install the package on your local system.
In this article, I will demonstrate how you can publish your own Python package on PyPI so that it is one-line installable and easily available to all other python users online! I will take the example of a sample package and show you the complete process. The example package is hosted on Github.
Step 1: Get the python scripts ready
First step is, of course, getting your python program (which you want to publish on PyPI) ready!
It could be any python script. Here I am using my own python script which I have named ‘locator.py‘ (I am using this name just for reference purpose. Feel free to save your python scripts by any name.) This file is available here.
Step 2: Getting the package-directory structure ready
This is the most important step. Now, we have to follow some pre-defined structure for our package’s directory.
As a reference, feel free to check out the Github repository of the sample project which is used in this tutorial. You can clone this repository and make some modifications to create your own package.
The directory structure has to be like this :
Ok let’s discuss what all these files will contain.
- setup.py : It is the most important file. It’s the file where various aspects of your project are configured. The primary feature of setup.py is that it contains a global setup() function. The keyword arguments to this function are how specific details of your project are defined.
You will need to install this setuptools library using pip:
pip install setuptools
Here is how my setup.py looks like:
# reading long description from file
# specify requirements of your package here
# some more details
'Development Status :: 4 - Beta'
'Intended Audience :: Developers'
'Topic :: Internet'
'License :: OSI Approved :: MIT License'
'Programming Language :: Python'
'Programming Language :: Python :: 2'
'Programming Language :: Python :: 2.6'
'Programming Language :: Python :: 2.7'
'Programming Language :: Python :: 3'
'Programming Language :: Python :: 3.3'
'Programming Language :: Python :: 3.4'
'Programming Language :: Python :: 3.5'
# calling the setup function
'A small wrapper around google maps api'
'Nikhil Kumar Singh'
'maps location address'
Let us see what different arguments of setup function do:
- name: It is the name of your project. Your package will be listed by this name on PyPI.
- version: It is a string in which you can specify the current version of your project.
It is totally your choice how you want to set the scheme of the series of versions (You may use ‘1.0’ or ‘0.1’ or even ‘0.0.1’).
This version is displayed on PyPI for each release if you publish your project. Every-time you upload a new version, you will have to change this argument as well.
- description: A short description about the package. You can use long_description argument to write long descriptions.
- long_description: We can use rich text for better description of our files. The default file format is reStructuredText . You can have a look at DESCRIPTION.txt to get a feel of the syntax.
- url: A homepage URL for your project. This makes it easier for people to follow or contribute to your project.
- author, author_email: Details about the author
- license: specify the type of license you are using.
- classifiers: It is a list of strings in which we can specify more details about our project like its development status, topic, license and supported python versions for your project. You can see more classifiers here.
- install_requires: It can be used to specify what third-party libraries your package needs to run. These dependencies will be installed by pip when someone installs your package.
- keywords: List keywords to describe your project.
- DESCRIPTION.txt : This file contains the long description about our package to show on the PyPI page. We use reStructuredText file format here. Check the file used in our package here.
- LICENSE.txt: It is a good practice to set a license for the usage of your project. You can use any of the freely available templates. Most commonly used is the MIT license.
The license I am using for this project is available here.(You may replace the Author’s name for using this license in your projects)
- README.md: This file has got nothing to do with our PyPI package. It contains description to be shown on the Github page. You can use it for PyPI page too but it will need some more modifications to our code. For now, lets keep it simple.
- __init__.py: The primary use of __init__.py is to initialize a python package.
The inclusion of this file in a directory indicates to the Python interpreter that the directory should be treated like a Python package.
You can leave this file empty.
Step 3: Create your accounts
Now, its time to create an account on PyPI and Test PyPI. Test PyPI is just a testing site where we will upload our code first to see if everything works correctly or not.
Once accounts have been made, create this .pypirc file in home directory of your system and enter the account details.
[distutils] index-servers = pypi pypitest [pypi] repository=https://pypi.python.org/pypi username= your_username password= your_password [pypitest] repository=https://testpypi.python.org/pypi username= your_username password= your_password
Note: If you are on a windows system, just type echo %USERPROFILE% in command prompt to know the home directory of your PC. Put the .pypirc file there.
Step 4: Upload the package
Finally, we are ready to upload our package on PyPI!
- First of all, we will check if our package installs correctly on Test PyPI.
Open command prompt/terminal in the root directory of your package.
Run this in terminal:
python setup.py register -r pypitest
This will attempt to register your package against PyPI’s test server, just to make sure you’ve set up everything correctly.
Now, run this:
python setup.py sdist upload -r pypitest
You should get no errors, and should also now be able to see your library in the test PyPI repository.
- Once you’ve successfully uploaded to PyPI Test, perform the same steps but point to the live PyPI server instead.
To register on PyPI, run:
python setup.py register -r pypi
python setup.py sdist upload -r pypi
And you’re all done! Your package is now publicly available on PyPI and could be easily installed by a simple pip command!
The package we created using this tutorial is available here.
Just type on terminal,
pip install your_package_name
to check if installation process is getting completed successfully.
This article is contributed by Nikhil Kumar. If you like GeeksforGeeks and would like to contribute, you can also write an article using contribute.geeksforgeeks.org or mail your article to firstname.lastname@example.org. See your article appearing on the GeeksforGeeks main page and help other Geeks.
Please write comments if you find anything incorrect, or you want to share more information about the topic discussed above.
- Command Line Scripts | Python Packaging
- Issues with using C code in Python | Set 2
- Python code to convert SOP to POS
- Issues with using C code in Python | Set 1
- Code introspection in Python
- How to run Python code on Google Colaboratory
- Execute a String of Code in Python
- Optimization Tips for Python Code
- Debugging Python code using breakpoint() and pdb
- Different Python IDEs and Code Editors
- Morse Code Translator In Python
- Python | Generate QR Code using pyqrcode module
- Understanding Code Reuse and Modularity in Python 3
- Python | Accessing variable value from code scope
- Performing Google Search using Python code