To use virtual environments with jupyter notebook, best practice is to create an independent virtual environment and add it to jupyter notebook as a kernel. This provides the portability to use the same venv on other IDEs or commandline as well as remove the venv from jupyter notebook incase it is no longer required and delete the venv. Let’s dive into creating and setting up a new ven on jupyter notebook in windoes system. Recommended to create and add new venvs for each python project so that each project’s dependencies can be safely isolated from others to avoid conflicts.

Jupyter notebook handles venvs as kernels. These kernels can be added to existing jupyter notebook by installing ipykernel library on venv and installing the venv as new kernel to existing jupyter notebook installation. Follow the steps below by updating the command format as per your OS and commandline.

OBJECTIVE

  1. Create & activate New Virtual Environment on Windows local system using virtualenv wrapper
  2. Add Virtual Environment Kernel to jupyter notebook
  3. Activate New virtual environment Kernel in jupyterbook using ‘Change Kernel’
  4. Remove/Delete a kernel from jupyter notebook

1. CREATE & ACTIVATE NEW VIRTUAL ENVIRONMENT ON WINDOWS SYSTEM

NOTE: All commands are based on Windows PowerShell 2. Please follow the steps by changing command formats accordingly for your commandline tool and OS.

cd Replace_this_with\path_to_directory\for_new_venv\

Create new folder where environment will be saved such as venv_project_name

mkdir venv_NEW_DIRECTORY_NAME

cd .\venv_NEW_DIRECTORY_NAME\

Create New Virtual Environment Using Virtualenv Wrapper

  • Choose Any Suitable Virtual Environment Name Such As virtual_env_name
  • Providing path to python installation is optional. Use any one command below: 
    virtualenv virtual_env_name
virtualenv virtual_env_name -p C:/ProgramData/Anaconda3/python.exe

Activate New Virtual Environment for windows

./virtual_env_name/Scripts/activate

Alternatively Activate New Virtual Environment for macOS

source ./virtual_env_name/bin/activate

2. ADD VIRTUAL ENVIRONMENT KERNEL TO JUPYTER NOTEBOOK

Intsall ipykernel in New Virtual Environment

pip install ipykernel

Add New Venv as ipykernel to Jupyter Notebook & list all ipykernels

python -m ipykernel install --name=virtual_env_name
jupyter kernelspec list

3. ACTIVATE NEW VIRUAL ENVIRONMENT OR KERNEL IN JUPYTER NOTEBOOK

  • Launch jupyter notebook through commandline using jupyter notebook command
  • Open a new notebook
  • Alternatively, if jupyter notebook is already running then refresh/reload browser
  • In the Menu bar navigate to Kernel –> Change Kernel –> Select virtual_env_name (name of the newly created and added virtual environment)
  • Observe Kernel name on the top right change from Python 3 to virtual_env_name

Congrats! New Virtual Environment Kernel is now actiavted in Jupyter notebook

  • For using an existing kernel already added to jupyter notebook only navigate to Kernel –> Change Kernel –> Select virtual_env_name
  • There is no need to activate virtual environment in command prompt or terminal to use the kernel in jupyter notebook. The terminal for jupyter notebook will shutdown the running Python 3 Kernel and activated the selected virtual_env_name kernel
  • To install pip packages activate virtual environment on commandline and install packages using pip install package_name as usual on commandline

4. DELETE KERNEL FROM JUPYTER NOTEBOOK

List all kernels to copy name of kernel to be removed

jupyter kernelspec list

Remove kernel using unistall

jupyter kernelspec uninstall unwanted_kernel_name