Question about deprecation of spend_exposure_fitting.png #780
arowley-ai
started this conversation in
General
Replies: 1 comment
-
Sorry for the late reply because the discussions here are not as well monitored as the issues. I guess it doesn't hurt to export thos too. Will consider it in our next update. Will let you know. |
Beta Was this translation helpful? Give feedback.
0 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
-
Hello,
Can anyone please explain why the code to generate spend_exposure_fitting.png in the current version has been commented out?
I understand that the fitting for spend/exposure is done earlier, when a RobynInput is updated with hyperparameters, but as far as I can tell the directory for outputting plots isn't made until a RobynOutput is created. Having the plots generated without being saved, as is done in demo.R, is not quite the same because a lot of users (myself included) are going to want to run Robyn in a container and export plots out later.
So my argument is actually that the code is in the right place and does useful stuff, could it be reintegrated back in? I haven't tested but it looks like it would work exactly as is. Personally I would be happy with simply uncommenting the affected lines and maybe amending or removing the 'message' that is displayed if
any(InputCollect$exposure_selector) == False
Could make a PR if this was agreeable with all?
Edit: Meanwhile, I have also noticed that pareto_media_transform_matrix.csv no longer seems to outputs rawExposure values correctly (instead, they are the same as rawMedia values). I don't think this was intended? Could be fixed separately or at the same time
Edit again: Scratch that, it doesn't work as is, but I did get it working like this:
if(length(InputCollect$exposure_vars)>0){ pSpendExposure <- wrap_plots( InputCollect$modNLS$plots, ncol = ifelse(length(InputCollect$modNLS$plots) <= 3, length(InputCollect$modNLS$plots), 3) ) + plot_annotation( title = "Spend-exposure fitting with Michaelis-Menten model", theme = theme(plot.title = element_text(hjust = 0.5)) ) ggsave( paste0(OutputCollect$plot_folder, "spend_exposure_fitting.png"), plot = pSpendExposure, dpi = 600, width = 12, limitsize = FALSE, height = ceiling(length(InputCollect$modNLS$plots) / 3) * 7 ) }
Thanks
Beta Was this translation helpful? Give feedback.
All reactions