In addition to the builtin mechanisms to create and manage code environments (ie, virtualenv for Python and custom mechanism for R), you can choose to use Conda.
Conda is a packages management system.
There are several advantages to using Conda:
- Conda provides a lot of packages in prebuilt format, avoiding hassle with compiling packages
- Conda can take care of installing versions of Python and R themselves so you don’t need to have them preinstalled
You need to install miniconda or anaconda. The “conda” binary should be in the PATH of DSS.
Using Conda environments¶
To use Conda for an environment instead of the builtin mechanisms, check the “Use conda” checkbox when creating the environment.
When managing the packages of a Conda-based environment, you actually need to manage two lists of packages:
- The list of Conda packages
- The “regular” list of language packages (either pip requirements or R packages)
This is due to the fact that not all packages are available through Conda. For packages not available through Conda, you need to put them in the “regular” list.
It is recommended not to put the same packages in both lists.