Python Dependency Management
I really enjoy Python. It is one of my go to languages, especially when I want to prototype something quickly. The standard library is quite mature, the third party library selection is massive, I am familiar with the syntax, and I find The Zen of Python (import this) to be a relatively strong set of ideals to guide development.
That said, I consider managing dependencies in Python to be a less than ideal experience. From my perspective Python lacks some of the quality of life improvements that many other languages have implemented with regards to managing dependencies: Rust has the fantastic Cargo, Node has the effective NPM and Java has a smattering of tools my favorite being Gradle. I preferred all of these tools to the way I have handled Python virtual environments in the past.
My Current Knowledge
I have personally worked with multiple python virtual environment tools, virtualenv, venv, and pyenv to create isolated virtual environments. I would then use Pip to load third party libraries into the virtual environments. I have also used the impressive conda package and environment management tool. Conda is great, but is specifically geared towards data science. You can certainly use conda outside of data science, but I want a tool that isn’t flavored in any way. I use Python as a general purpose programming language and regularly perform file manipulation, web request & response parsing, web application development, email management, text message management, and more so to use a data science geared package management tool seems like a miscommunication to my future self, or other developers.
This leads me to continue using virtualenv and pip as my virtual environment and package management tool-chain.
Something Better?
I have heard promising things about Python Poetry, but haven’t dived into using that tool yet. This project is simple enough that I can try out Python Poetry to assess it as an alternative to my current workflow.
Worst case scenario: I add Python Poetry to the list of virtual environment management tools I’ve tried, but I continue to use virtualenv and pip.
Best case scenario: I deem Python Poetry an improvement over my current workflow and I become a more effective Python developer!
Before we explore this new tool let’s take a look at how I currently use virtualenv and pip.