Copy entire input-path folder to VM for pipelines
Problem to solve
As a user I want to be able to select one specific file, a folder or eventually several data files as input for my experiments. The current limitation is that the input-path MUST be a folder. When, for example, selecting a file for pre-processing or training, this does not work as the experiment pipeline requires the input path to be a folder.
Also, a user script might require not only one input-file but more (e.g. containing lable information, a test file, etc).
The potential solution is that the entire repository is cloned as the first step during pipeline execution. The input-path than can be either a file or folder and all other files are also available for use.