Why does robyn_input transforms negative context variable into positive?

[pedroaranah]

Hi all,

Was troubleshooting an MMM I'm currently building and found out that Robyn, when transforming inputs in function robyn_input, transforms negative context variables into positive in this line

Robyn/R/R/inputs.R

Line 206 in 5915aa9

dt_input <- check_allneg(dt_input)

This created an unexpected issue for me as I've built a train-test split validation process apart from Robyn one where this variable is still negative vs the one transformed in robyn_input.

I know it's easy to bypass from my side, but is there any reason for this?