Make file that contains service_id for every train/datetime

[brechtvdv]

Now service_ids are auto-incremented.

To keep the same service_id for every train-datetime pair persistent over different datasets, we need to make a file that makes the link between train-datetimepair -> service_id