Jetbrains DataSpell looks almost awesome, but I'm having an issue with the managed servers that makes it just alright. I'm using a conda environment (which I've called pymc3_env for obvious reasons) as my interpreter. I've configured both the workspace and the project to use pymc3_env. When I run code in a notebook connected to a DataSpell-managed Jupyter server I get errors (specifically some quite fundamental Clang compilation errors from theano: ld: library not found for -lSystem). The same code in the same notebook runs fine if, instead of using a DataSpell-managed server, I activate the conda environment in a terminal, launch a Jupyter notebook server, and give DataSpell the URL. What gives?

0

There are 0 best solutions below