WIP: Split object event data model

For experimenting with a new event model. Relations (production/end vertex for a particle as well as in/out-going particles for vertices) have been removed from the EDM objects (which are placed in Gaussino for convenience for now). Instead of writing the four different container to augment the basic particles when zipping, one object is temporarily written to the TES that stores particle-index and vertex-index pairs together with the information whether it is an ingoing or outgoing relation.

The idea here is that the augmenting objects can afterwards be created from these linking information by a different algorithm when the particles or vertices are constant or utilised in the packer. This is obviously open for suggestions.

I'll add another algorithm that will imitate the two-stage conversion of the HepMC information asap.