Implement PROFILE MEMORY clause / command

[gitbuda]

In the similar way how the PROFILE works it would be very interesting how much each clause consumes memory.

E.g.

MATCH (n) // all below are just illustrations of the complexity -> double-check for the exact numbers
WITH n // copies of accessors - 48B each
WITH n.age // copies of ints - 8B each
WITH n.string // copies of string - arbitrary in size, maybe smaller compared to node, maybe much bigger