Skip to main content

Installing additional Jupyter kernels

Python, R, Julia, and Stata kernels on Code Ocean.

Updated over 6 years ago

The Python 3 kernel:

  • is already included when you install Jupyter.

  • To install Jupyter, If you have Conda or pip installers available in your environment, you can just add package Jupyter . 

  • If you do not, you can add python3-pip , python3-setuptools , and python3-wheel  via apt-get, and then add jupyter via the pip3 installer (which will appear automatically when you've added python3-pip ).

The R kernel:

  • requires the package IRkernel: available either through CRAN, or, if you have the Conda installer available, you can add package r-irkernel and then add the channel r  to Conda's list of available channels.

  • You'll also need to run, in the postInstall script, the command Rscript -e "IRkernel::installspec()" .

The Julia kernel: 

  • requires the Julia package IJulia  (https://github.com/JuliaLang/IJulia.jl). 

  • You could theoretically install this in any environment, but it's probably best to start with a Python 3.7 (with Conda) base image. 

  • First, add julia and jupyter  as conda packages, and then add the channel conda-forge   to Conda's list of available channels.

  • Second, in the postInstall script, run julia -e 'using Pkg; Pkg.add("IJulia"); Pkg.update();' .

The Stata kernel:

  • Start your capsule from an environment with Stata available.

  • Via the apt-get installer, add the packages python3-pip , python3-setuptools, and python3-wheel .

  • Via the pip3  installer that dynamically appears, add  ipystata , pandas, and jupyter.

  • In the postInstall script, run the following:

ipython -c "import ipystata; \
  from ipystata.config import config_stata; \
  config_stata('/usr/local/stata/stata-mp')"

 What if I need a different (or older) Kernel? 

Please write to us at support@codeocean.com or via live chat, and we'll be happy to help (and to amend this article accordingly). 

Did this answer your question?