Jupyter Notebooks¶

Jupyter Notebooks offer a large quality of life and are becoming more important with machine learning etc. Natively PandaRoot is a collection of C++ libraries which usually are collected and called through ROOT, mostly by macros.

ROOT by itself features python bindings as well as an own C++ Kernel. With the %%cpp magic it is also possible to mix C++ code into python code. See the introduction to PyROOT.

Using the Kernels without installation¶

It it is enough to do the following:

source <build>/config.sh -p
export JUPYTER_CONFIG_DIR=$SIMPATH/etc/root/notebook
export JUPYTER_PATH=$SIMPATH/etc/root/notebook
jupyter notebook

You now can use ROOT and PandaRoot classes in the Python3 kernel or the RootC++ kernel. Here are two example notebooks to start off with.

Examples¶

PandaRoot Notebook Python style

Install the Kernels¶

In some cases it is beneficial to install the Kernels to automatically be used in the system. A simple ‘jupyter notebook’ command in a fresh shell would allow them to be used. Also VSCode with the jupyter extension would be able to spawn such notebooks inside the integrated environment.

The downside is that the Kernels are pointing to one specific installation of PandaRoot and its dependencies and managing multiple installations may become cumbersome.

As a prerequisit one has to install the metakernel package in python (pip3 install metakernel). To install the Kernels simply copy the folders from <build>/jupyter/kernels to the appropriate place:

~/.local/share/jupyter/kernels (Linux)
~/Library/Jupyter/kernels (Mac)
%APPDATA%\jupyter\kernels (Win)