DQ Rule that calculates attribute level Data Drift

Hi,

I want to create data quality rules that calculate the Population Stability Index (PSI) on a categorical attribute to detect distribution shifts over time.
My goal is to build a flow that:

1. Uses a date column from the rule input to split the dataset into: 

• a reference window (previous N days)
• a current window (most recent N days)

2. Calculates category distributions (percentages) for both periods
3. Computes the PSI between these two distributions.

I would appreciate guidance on the best way to implement this and other kinds of drift metrics in Ataccama on both categorical and numerical columns.