Cognite Data Fusion: Move Magic Datapoint Granularity Selection Algorithm From Direct Query to the API

I have been playing around with the excellent DirectQuery preview functionality and really enjoyed that I can simply give it a timestamp range and it figures out how many points I need and what the granularity will be with the assumption that I will use the queried data for some visualization.  Even if someone has an 8K monitor with no border around their chart, with one pixel per point the most I could possibly need would be 7680 datapoints.  Reality is that we need much less since we show a datapoint as more than a pixel width.  But I don’t have to care about that complexity.  Cognite calculates optimal granularity for me.

Can this feature also be available via the API for “Retrieve data points”?
https://docs.cognite.com/api/v1/#tag/Time-series/operation/getMultiTimeSeriesDatapoints

Maybe we would be able to set granularity to “auto” then the magic happens?