Python packages

It's difficult for the HPCC to host a vast volume of Python packages, and to resolve conflicts between Python versions. Users are encouraged to install and manage packages on their own, through conda or virtual environments.

Using conda

Users install their own version of Python through Anaconda in the home or research space. This gives them full control on their preferred versions of Python and packages. Isolated virtual environments can be created and controlled by conda environments.

Using virtualenv

By default, Python version 3.6.4 (compiled by GCC/6.4.0) is loaded and ready to use after logging into HPCC dev-nodes. Other versions can be found by running module spider Python. Usually, pre-installed packages are available for each Python version. If users need to install new ones, they can install them in the home or research space through virtual environments.