[ML] Add information about samples per node to the tree

[valeriy42]

I propose to extend the node definition by adding information about the number of training samples that passed through the node. This information provide one of the simpler feature importance indicators (see xgboost cover). It also helps to test and prevent degenerated trees such as those described in #849 .

Sub-Tasks

• Update schema definition
• Update Node logic
• Update inference model code
• Update persist/restore schema version
• Add enhancement note