MEanModel/Pop Workflow: Introduce dataSelection on the SimulationSet level

[Yuri05]

This is already the case in the excel template: we have dataSelection on the Output level and on the SimulationSet level.
When generating the code from excel, those 2 dataSelections are concatenated and set as dataSelection of the Output.

Because of this, Outputs are duplicated in the generated code (SimulationSet1.Output1, SimulationSet2.Output2, etc.)

With this change, it will be possible to reduce the complexity of the generated R-Code (each Output is exported only once).
It will be also possible to assign some properties to an Output (e.g. colours) which will be used in ALL plots for this output.

[pchelle]

@Yuri05
What should be the default selection for the simulation set level ? Select all the data in the dataset or select none of the data in the dataset ?
Currently, default selection for output was none of the data was selected