Python package mgmt is aweful between the virtual env you need, pip that install different things on different OS, sometime the exec is no even in your path, I don't recall how many time I had to install awscli because it was broken, on top of that you add all the problem with python 2/3 ( pip3 ect .. ). Pip installing things as root or in your home folder ...
From all the mgmt I had to use, Python is by far the worst.
Every time I have something in Python to use I pray that pip will install it correctly which is roughly 50% of the cases.
Yeah, the current state of Python is the worst. Especially in machine learning, it's common that a project only supports one of conda or pip installation, which are often but not always compatible. The idea of storing all dependencies in "requirements.txt" is a good idea but has not standardized across the Python community. IMO all the 2 -> 3 upgrade difficulties have slowed down the development of the Python tooling ecosystem.
I take it you've never used npm for a large project that's spanned multiple years of different developers while supporting multiple versions of your application while also staying up to date with dependencies.
From all the mgmt I had to use, Python is by far the worst.
Every time I have something in Python to use I pray that pip will install it correctly which is roughly 50% of the cases.