Jupyter Notebooks

Contents

• Installation
  • Installing the C++ Kernel
• Getting Started
• Converting Notebooks

Installation

Make sure to use an updated version of pip.

pip install --user --upgrade pip

Install jupyterlab which includes the kernel for running Python code.

pip install --user --upgrade jupyterlab

Installing the C++ Kernel

In order to run C++ code, the cling interactive C++ interpreter must be available. It can be downloaded from root.cern.ch for selected Linux distributions. After extracting the archive, the PATH variable must be adjusted to include the path to cling's binaries.

export PATH=$PATH:${YOUR_CLING_ROOT}/bin

To make the change persistent, add the above line to your `~/.profile'. Afterwards, prepare and install the corresponding Jupyter kernel. You may select any of C++11, C++14 and C++17. In the instructions below, the C++17 kernel is made available.

cd ${YOUR_CLING_ROOT}/share/cling/Jupyter/kernel/
pip install -e .
jupyter-kernelspec install --user cling-cpp17

Once this is done, language support for C++ will be available in future jupyterlab sessions.

Getting Started

After completing the installation, you can either start the classic jupyter-notebook or the more recent jupyter-lab. Once started, the built-in help is available via the Help menu. While using jupyter-notebook, the shortcuts can be displayed using the h key and an interactive tour is available in Help->User Interface Tour.

Converting Notebooks

It is possible to convert notebook files to a variety of formats. For example, you can extract the contained python code into a proper python script via

jupyter-nbconvert --to python yournotebook.ipynb

Multiple other formats are available as well (including $\LaTeX$). Consult

jupyter-nbconvert --help

more information.