Jupyter Notebook Kernels

Things on this page are fragmentary and immature notes/thoughts of the author. Please read with your own judgement!

Jupyter Kernels

By default, Python kernels are installed to /usr/local/share/jupyter/kernels and BeakerX kernels are installed to /usr/share/jupyter/kernels.

Python¶

ipykernel¶

IPython Kernel for Jupyter/Lab, which is the default Python kernel for Jupyter/Lab currently.

xeus-python¶

xeus-python is a Jupyter kernel for Python based on the native implementation of the Jupyter protocol xeus.

A new Python kernel for Jupyter

SQL Kernels¶

There are multiple ways to run SQL in a Jupyter notebook.

• using a dedicated SQL kernel

• using cell magics

• using a Python library (in a notebook with a Python kernel)

I personally prefer running SQL through a Python library.

• No need to install dedicated SQL kernels or kernels support cells magics for SQL.

• We often need to do data transformation and visualization after querying data. This is convenient to do in Python.

Spark Kernel¶

There are dedicated kernels for Spark, e.g.,

1. toree (a good one）

2. sparkmagic (seems like a good choice)

3. spylon-kernel

However, such a Spark kernel isn’t really needed no matter you use Spark/Scala or PySpark (preferred) .

JVM Language Kernels¶

• JJava

• rapaio-jupyter-kernel

• Ganymede

  • works well for Java

  • print doesn’t work for other JVM languages based on cell magics

• kotlin-jupyter

  • Kotlin 2 is not supported yet

• almond

  • works well

Remote Kernels¶

IPython is able to run remtoe kernels.

https://github.com/ipython/ipython/wiki/Cookbook:-Connecting-to-a-remote-kernel-via-ssh

https://stackoverflow.com/questions/29037211/how-do-i-add-a-kernel-on-a-remote-machine-in-ipython-jupyter-notebook

jupyter/notebook#1487