Hacker News Reader

Comment by ShorsHammer | Hacker News Reader

Given how fragmented virtual environments are in python is which is the most popular currently? Last I looked there was venv, pyenv, pipenv, pipx, poetry and pipsi.

Haven't done any python work in a while but am curious to know what people use. Starting right now, I'd probably just stick with venv, though guessing the others do offer some extra benefits.

View on HN

The virtual environment part isn't really very fragmented at all -- I believe all of those use virtualenv/venv under the hood for that and just add layers of additional features for package and/or interpreter management. (that part is where the fragmentation happens, especially packages.) The virtualenv/venv decision pretty much boils down to are you py3 only? Use the venv module in the standard library. Do you still support py2? Then virtualenv.

I see conda a lot more for new python coders in the machine learning space, having a GUI helps newer programmers. Check out Anaconda/Miniconda. Code is as simple as

  conda create -n myenv python==3.8
  conda activate myenv

Do all your things

  conda env export > myenv.yml

On your coworkers machine

  conda env create -f myenv.yml

conda can also manage packages that aren’t part of python such as R, which is why I mainly use it.

Reply on news.ycombinator.com