Fixing aggregation and some aesthetics in automatic postprocessing

[saraloo]

WIP: Addressing some issues that arose during the disparities round related to the postprocessing and aggregation functions used in the classical flepi R inference.

Fixes:

• [x] change automatic postprocessing to show the fitting data and the simulated output in the same aggregation format used within the likelihood algorithm (ie using the getStats() function)
• [x] fix the date filtering in aggregation
• [x] test to confirm these fix the problems

This addresses #272