korniichuk / rk

The remote Jupyter kernel/kernels administration utility
http://www.korniichuk.com
The Unlicense
54 stars 18 forks source link

Using conda environment on remote #6

Open mick-d opened 8 years ago

mick-d commented 8 years ago

First of all, great package, that's exactly what I was looking for. I still have an issue with using the remote conda environment: I am not sure the conda environment on the remote is used properly.

Ideally I would like to source it but the only option I found was to indicate the path of the remote conda environment python binary in kernel spec.

So when importing sys in the local notebook using the remote kernel, i get the correct environment python version (3.5):

sys.version
`3.5.1`

But clearly, the environment variables are not right (conda environment was clearly not sourced):

!python --version
Python 2.7.11

Is there a way to properly source the whole remote conda environment instead of just using its associated python binary?

mick-d commented 8 years ago

I just found out that no environment variables are loaded from remote, how to do it? Any help would be greatly appreciated @korniichuk

helger commented 5 years ago

@mick-d Have a look at dotenv .

mick-d commented 5 years ago

I couldn't even remember this issue from 3 years ago :) I'll reread what it was about, thanks