Abstract paths for dataset server template

[antru6]

Currently, the implementor of a dataset server template should read the dump file from the directory /kobe/dataset/$DATASET_NAME/dump and should create a copy of the database in some custom directory inside /kobe/dataset/$DATASET_NAME.

As an improvement, the api should be extended such as these directories should be chosen by the author of the template (as already done in the federator templates).