Unexpected behavior of response_curves_data

[AdimDrewnik]

I have noticed that the output of these two functions is very different but should be exactly the same
response_curves_data( )
response_curves_data( selected_times = (start_date, end_date) )

start date and end date are covering the whole input data set. So in theory both calls are working on the same time range.

[cpulavarthi]

Hello @AdimDrewnik,

I'd like to inform you that in the current implementation of Meridian the selected_times argument of response_curves_data is a "list of a subset of time dimensions to include". This means that providing selected_times = (start_date, end_date) will result in a list of 2 dates, not all dates between start_date and end_date.

If you provide the full list of dates (as a tuple, because it needs to be a "frozenset"), then the result will be the same as with no arguments.

Feel free to reach out for any further queries or suggestions.

Thank you

Google Meridian Support Team

[AdimDrewnik]

Thank you but this makes no sense. What was the motivation to do this that way?
This is very inconsistent as optimize function has argument with the same name that works as range, not to mention that in function output_model_results_summary start and end date are 2 separate arguments.

[cpulavarthi]

Hello @AdimDrewnik,

Thank you for pointing out the inconsistency. We are working on unifying the naming convention. Please be informed that you will find these updates documented in the Change Log once they are done.

Thank you

Google Meridian Support Team