The following steps show how to make an IPython kernel (even if it is in a conda/virtualenv environment) available to an existing jupyter installation:
-
Enable the environment whose IPython kernel you want to use from Jupyter. For a conda environment, you need to do:
activate <env_name>
-
Install the IPython Kernel for Jupyter package (a.k.a.
ipykernel
):pip install ipykernel
-
Create your kernel. Assign it a
<name>
and, optionally, a<display_name>
to be shown in the Jupyter Notebook menu:python -m ipykernel install --user --name <name> --display_name <display_name>
Note: Since we have passed the
--user
option, this will create a user-specific kernelspec directory describing the kernel so that it is locatable by every Jupyter installation. Non-user kernels can be also available. You can list the kernels available to any Jupyter executable as follows:jupyter kernelspec list
The specific steps will depend on the particular kernel. Check the corresponding documentation for details.