Julia plugin

⚠️ Starting with DSS 15, this capability is not available anymore in DSS.

Summary

This plugin adds Julia support to DSS.

After installation, DSS users can:

• create Julia code recipes
• create Julia notebooks
• interact with DSS datasets and folders from Julia code

Installation

Install the plugin in DSS, then run the Julia integration script as a DSS administrator:

$DSS_HOME/bin/dssadmin install-julia-integration

This script creates the Julia code environment, downloads the required libraries, precompiles them, and registers the Jupyter kernel used by DSS.

Requirements

• DSS 8.0 or newer
• Julia installed on the DSS machine

Usage

After the installation is complete, Julia appears as an available language for code recipes and notebooks.

Creation of a Julia recipe from the actions panel:

Creation of Julia notebooks:

Inside recipes and notebooks, use the Dataiku.jl package to interact with DSS. It is a Julia wrapper around the DSS Public API and provides helpers to read and write datasets and folders.

Example of a Julia recipe:

Code environment and limitations

This plugin currently uses a single shared Julia environment located at:

$DSS_HOME/code-envs/julia

It is not possible to configure multiple Julia code environments with this plugin.

To install or remove Julia packages in that environment, use Julia's built-in package manager either:

• from a Julia notebook in DSS
• from the command line by running julia with JULIA_DEPOT_PATH=$DSS_HOME/code-envs/julia

Example of a Julia notebook used to manage the code environment: