WIP: use CaloFuture code in TupleToolCaloHypo

This is an attempt to resurrect TupleToolCaloHypo since it's broken in master after the clean-up of old calo code in Rec!1998 (merged) and it computes old (wrong) neutral PID variables in v31r0.

I naively thought it might just require changing all Calo to CaloFuture interfaces but it's proven more complicated since few things have diverged between the old and new calo code. Thanks to @sponce I've been able to advance a bit. It needs LHCb!2565 (closed) and Rec!2067 (closed) but still doesn't work. I will add comments on the code directly where I am stuck.

General questions: @rmatev @sstahl @graven

• what's the policy for Analysis code in master to work? Do we want to maintain it as much as possible? This one requires quite some changes and it might have been easier to directly start from scratch once Rec is more stable
• I guess we don't have any test for this TupleTool since we didn't realise Rec!1998 (merged) was breaking it. If we agree to resurrect it, we should add one to prevent this from happening again
• I am afraid the last might apply to other parts of the code, so if the policy is to keep all code in master functioning, more work will be needed to identify and fix other broken pieces