Feature Request: Copy single file for a dataset

[jbstettner]

It would be great and ensure better book-keeping if one could define a pre-dataset-hook, i.e. a job running once before the actual tasks of the datasets are queued.

Use-case: Copy a config-file/readme or so with the settings of that specific dataset. This would enable users to have all the infos they need right away at the location of the produced files.