New depedencies to be handled by L1 simulation

Fix for ATR-26993. Would be nice to have @pavol commenting. This would be easily handled if I made my old promise of transferring all the configuration to LAr code (from L1Calo). To be done soon. In the mean time, this fixes the issue found in the referenced JIRA. For me, it is important to have it. It is a bit weird that dependency through the algorithm to the tool. I did not know that this could be the case (that tool dependencies go to the algorithm they pertain to). Anyway, the calib line (as in the MR !60680 (merged)) introduced is not necessary for other uses of the BS converter code, hence, the AllowEmpty and, of course, the resetting of the property in the python side.