Data partitioning

## Background
Sometimes the data coming in into Enceladus are not well partitioned.

## Goal
Make Enceladus to output data files in effective sizes to well leverage partioning

## Expected Task List
A list of expected issues that will be needed to achieve this Epic
- investigate the problematic data
- create the library
- enhance dataset data model with `recordsPerPartition` value
- optional dataset setting
- heuristic based on anylysis to be saved
- implement partitioning by record count
- warning if partition size is over limit
- warning if record count is off