how to create new kernel

Hello experts, I’m wondering if anyone explains or provide me to how to create a new kernel? let me explain my issue: I have spark on my server in /usr/local/spark, I could use the built-in SparkR package in spark package in R language as making link of that to R library like: ln -s /usr/local/spark/R/SparkR/ /usr/local/R/library

as you know, to create pyhton2 kernel for jupyterhub we only need below code in /usr/local/share/jupyter/kernels/python2/kernel.json file:

{
 "display_name": "Python 2",
 "language": "python",
 "argv": [
  "/usr/bin/python2",
  "-m",
  "ipykernel_launcher",
  "-f",
  "{connection_file}"
 ]
}

now, I need to know how can I create scala and spark kernel for jupyter like above.

many thanks.

[alexarchambault]

@vhp1360 Launching the jupyter-scala script like

$ ./jupyter-scala --jupyter-path /usr/local/share/jupyter

should install jupyter-scala under /usr/local/share/jupyter/kernels/scala. It doesn’t need a spark distribution - it downloads the spark jars itself, depending on the version users ask in the session (illustrated here).

[vhp1360]

hi @alexarchambault, and thanks for your reply. your solution is good, but I tried to find a way to link Scala from Scala package also spark and pyspark from Apache Spark package to Python and then to Jupyter. it seems, I’m looking for a hard issue and should forget it. 😃