MaMMoS Webinar Series, WEB 2 скачать в хорошем качестве

MaMMoS Webinar Series, WEB 2

"Making research outputs FAIRer with reproducible software installation (and pixi)" Speaker: Prof. Hans Fangohr, Max Planck Institute for the Structure and Dynamics of Matter, Germany Software is used in many stages of research activity: data capture, data processing and analysis, simulation, creation of tables and figures for publications. To make research outputs reproducible and re-usable we must archive the data but also the relevant software: for example, a Python program to create a figure from a data file. In this presentation, we discuss the importance of recording additional software dependencies (such as which version of Python, which other dependences such as matplotlib and numpy are needed). We also discuss the importance of being able to actually install the required dependencies. Following an overview of software packaging and installation tools (such as pip, conda, spack, Docker), we introduce pixi [1] in more detail. We have found pixi to be a useful in our day-to-day work (and think it will be for many others), and by using it, we have made the software installations we create more reproducible and re-usable in the future.