Best way to have a globally accessible experimental env

[foxale]

My current setup is:

• pyenv for multiple python versions, but all clean (with no packages installed)
• poetry for project environments managed via pipx

This is a perfect setup for working on multiple projects, as I have separate python versions, environments, etc.

However, many times I have a situation that I'm in a random directory and need to quickly pull up python interactive prompt to verify some random though. Usually a bare python is not enough, I need it to have a bunch of most often used packages pre-installed (e.g. pandas, numpy).

In other words, I need to have a globally accessible python environment for experimenting with random stuff, not related to any particular project.

What's the best way to achieve this, without cluttering the pyenv python interpreters with random packages?

[neersighted]

I highly suggest use of pipx to manage Poetry and other Python CLI tools that are used in a global and not project-specific context: https://pypa.github.io/pipx/

[foxale]

I use pipx to manage Poetry, but how is that related to the question?

[foxale]

I edited the post to make it more clear what is my goal :)

[neersighted]

Ah, I see you are referring to importable modules and not tools with entrypoints. I think the answer to that question is the same as any other tool -- install things into one or more sandbox/pet virtual envs you keep in a common location.

e.g.

python3 -m venv ~/.virtualenvs/sandbox
source ~/.virtualenvs/sandbox/activate

There are common tools to support this workflow, like pyenv-virtualenv and virtualenvwrapper.