Settings for easy selection of multiple kernels in Jupyter

Introduction

Recently, I started to create a virtual environment in Anaconda and code it in an environment that suits my purpose. Among them, for some reason only one of the PCs I own was able to select the kernel from Jupyter Notebook without any settings, but on other PCs I select the kernel without settings. I couldn't. (Probably a problem with the version of Anaconda.) So, I did a lot of research on the net so that I could select the kernel from Jupyter Notebook. However, in my environment (because it is Windows), including Additional packages did not work, or the command information was out of date, so I used it for myself. I will summarize it in. With this setting, you can change the kernel in an instant as follows!

Execution environment

• Windows 10（64bit）
• Anaconda 4.3.0（64bit）

Setup steps

Creating a virtual environment

You can easily create a virtual environment by using conda. This time, it is assumed that jupyter will be used, so install jupyter at the same time when creating this virtual environment.

create_env

#Create virtual environment, install jupyter
$ conda create --name=test python=python3.6 jupyter
#Check if the virtual environment is created, if it is displayed in the list, it is successful
$ conda info -e

Enter the appropriate python version and --name = according to your environment.

Add virtual environment kernel to jupyter

At this stage, I don't think the kernel has been added yet. Now try running the following command:

check_kernel

#View available kernels
$ jupyter kernelspec list

Add the kernel of the virtual environment you added earlier here. Execute the following command with the virtual environment activated.

install_kernel

#Activate virtual environment
$ activate test
#Add kernel
$ ipython kernel install --user --name=test --display-name=test

Now you can switch and select kernels on Jupyter Notebook.

Finally, I will summarize the main options when adding a kernel.

How to remove the added kernel

The added kernel information is located in ~ \ AppData \ Roaming \ jupyter \ kernels with the --user option. You can remove the added kernel by deleting the folder with the name specified by --name here.

You can also remove the kernel with the following command.

uninstall_kernel

#Remove kernel
$ jupyter kernelspec uninstall test

reference

1. stack overflow - Using both Python 2.x and Python 3.x in IPython Notebook
2. Coexistence of Anaconda 2 and Anaconda 3 in Jupyter + Bonus (Julia)
3. Easy way to switch virtual environment created by Conda on Jupyter