Closed fcollonval closed 3 years ago
You are correct @jamestwebber.
This was not related.
nb_conda_kernels
is not a dependency of this project (any more - to be precise).
If you are using JupyterLab from an environment, you need to install nb_conda_kernels
in that environment too to see the others environment.
If you are using JupyterLab from an environment, you need to install nb_conda_kernels in that environment too to see the others environment.
Am I understanding this right--I need to install nb_conda_kernels
in both the jupyter
env (that is, the one which I run the server in) as well as the kernel env? That seems strange. In the past, I would just install it in the server env, and then installing ipykernel
in another env would make it visible.
But at least for now, using ipython kernel install
is easy enough. It was a little more convenient when it could auto-detect new kernels but it's not that critical. What would be much cooler is if I didn't need to install anything in the "kernel" env, but I can't imagine that being simple to implement.
edit: Or perhaps you are just saying I need nb_conda_kernels
because I'm running jupyter in an environment. It doesn't appear to be necessary, as long as I register the other kernels.
edit: Or perhaps you are just saying I need
nb_conda_kernels
because I'm running jupyter in an environment.
Yes this is what I meant. So if you are installing and running jupyter from your environment my-second-env
, you need to install nb_conda_kernels
in my-second-env
to see the other environments.
It doesn't appear to be necessary, as long as I register the other kernels.
This actually depends on your settings. When you are installing a kernel explicitly with ipython kernel install
, a file is created to tell jupyter about how to start that kernel. Depending on the folder in which that configuration file will be created, you will see or not that kernel in jupyter. You are seeing it because by default the kernel is installed in your user settings.
usage: ipython-kernel-install [-h] [--user] [--name NAME] [--display-name DISPLAY_NAME] [--profile PROFILE]
[--prefix PREFIX] [--sys-prefix] [--env ENV VALUE]
Install the IPython kernel spec.
optional arguments:
-h, --help show this help message and exit
--user Install for the current user instead of system-wide
--name NAME Specify a name for the kernelspec. This is needed to have multiple IPython kernels at
the same time.
--display-name DISPLAY_NAME
Specify the display name for the kernelspec. This is helpful when you have multiple
IPython kernels.
--profile PROFILE Specify an IPython profile to load. This can be used to create custom versions of the
kernel.
--prefix PREFIX Specify an install prefix for the kernelspec. This is needed to install into a non-
default location, such as a conda/virtual-env.
--sys-prefix Install to Python's sys.prefix. Shorthand for --prefix='/home/fcollonval/miniconda3'.
For use in conda/virtual-envs.
--env ENV VALUE Set environment variables for the kernel.
nb_conda_kernels
on the other hand is loading dynamically the environment as kernel inside of jupyter (without writing files - expect if you use non-default configuration). This is the reason you need to install it in every environment you are running jupyter from.
Thanks for the clarification! I suppose this issue doesn't have much to do with gator
but it was useful
I have a (maybe?) related issue. I just set up a new JupyterLab 3 environment and installed
gator
to manage conda envs (I previously usednb_conda_kernels
but it doesn't look like it works anymore?).The gator tab shows all my existing kernels but I can't start a notebook in anything but the base environment. Is there some additional step I need to take to make them discoverable?
edit: I guess my issue is different and just required
ipython kernel install --user --name=<name>
. Somehow in the past that wasn't necessary for me to see new kernels.Originally posted by @jamestwebber in https://github.com/mamba-org/gator/issues/147#issuecomment-864445995