Using DataSpell with Nb Conda Kernels
Is Dataspell compatible with Nb Conda Kernels?
I have some existing Jupyter Notebooks that use Conda virtual environments created using Nb Conda Kernels. I've tried loading those notebooks into DataSpell:
1. If I set the Python interpreter to be the same as the kernel (i.e. NOT the base Conda environment), then Dataspell asks me to install Jupyter. If I ignore that message and run a cell, I get a message saying that the SSL module cannot be found (although I have checked the location it is supposed to be missing from and it is present.).
2. If I set the Python interpreter to be the base Conda environment, then Dataspell shows invalid references for any module not in the base environment (although I can run the notebook, presumably because the kernel is configured correctly.).
Should this work? If so, then how do I configure DataSpell? And if it does not work, is there a workaround? I'd prefer not to have to install Jupyter in every Conda environment, that is one of the benefits of Nb Conda Kernels.
Please sign in to leave a comment.
On Mac M1 I removed old miniforge installation and installed Miniconda 3 from scratch. Also DataSpell was reinstalled completely. There are very good videos on YouTube authored by Jeff Heaton explaining the installation process of Miniconda3 and Jupyter notebooks.
Thank you Sven for your response. I have Anaconda and Dataspell working together fine, that is not the problem. The issue is specifically about the compatibility of conda virtual environments created using the Nb Conda Kernels module with DataSpell.
In this setup, the Jupyter notebook itself runs inside the Anaconda base venv, whilst your Jupyter notebook code runs inside the Nb Conda Kernel venv that you select within the notebook. Thus, the latter venvs do not need to have Jupyter installed, which seems to cause a problem when you use those within Dataspell.
Make sure that your computer meets the minimum system requirements to run DataSpell. Check the official documentation for specific requirements. Make sure you have the latest version of DataSpell installed. Updates often include fixes and improvements that can resolve problems. If you encounter any error messages, Minesweeper or warnings, read them carefully as they may provide clues about the issue. You can search for the error message online or consult the DataSpell community for assistance.
A year later, I stumbled across exactly the same thing. Is there a workaround for this in the meantime?
It would be great if it is possible to install jupyter only once and still have access to other conda environments using nb_conda_kernels.
The two points from the original post are still the same today.