stephanie

venv solved some of my issues, but it is sometimes cumbersome and I wish I didnt have to use that kind of tool to manage them.

trying out pyenv on my work computer for versions i'm pretty into it so far. my personal computer is a horrific mix of multiple system installations, brew py versions, virtualenvs, venvs, and like 2 or 3 conda varieties so i've decided it's beyond saving

The key I found back in the day to solve this once and for all: a) No weird anaconda shenanigans, just plain pip b) Set up a global python virtualenv ~/env for example c) Source the env automatically in every terminal so all python is scoped to the env by default

d) In the event of gross python fuckery, delete the folder, make another one, everything is now pristine. With this workflow I've very rarely had python issues for years, even when installing stuff like tensorflow from source