SugiharaLab/rEDM

Question: which version of rEDM had the syntax ...$stats$rho$rho to extract the rho from a block_lnlp call?

Closed this issue · 4 comments

Hello, I'm very sorry to bother with such a trivial question but I've been trying to figure this out for a while without success.
In April and May 2022 I used rEDM for a research project. For unforeseen reasons I now have to re-run some of the analyses from this project. I want to ensure that I use the same version of rEDM that I initially used back in 2022.

Here's what my code looks like:

simplex_output = block_lnlp(df, lib = c(1, floor(endPoint / 2)), pred = c(floor(endPoint/2)+1, endPoint), target_column = colname, 
                            tp = forecast_time, stats_only = FALSE, first_column_time = FALSE, silent = TRUE)

simplex_output_rho = simplex_output$stats$rho$rho

Considering when I wrote my code, I would presume that I used version 1.2.2 or 1.2.3, but in these versions (and later ones) the call to block_lnlp() does not work. I've tried earlier versions too, but while block_lnlp() works in these, the $stats$rho$rho syntax does not.
So my question is this: what version of rEDM could I possibly have used for the above code to work?

Thanks in advance and again, sorry to bother.

Please see Issue #64 .

Although calls to block_lnlp should work until rEDM_1.14.3.tar.gz 2023-07-07, this is not recommended with the 1.X code.

@SoftwareLiteracy Thank you for your response and sorry for taking so long to get back.
Are you by any chance aware of a version of rEDM where the syntax 'simplex_output$stats$rho$rho' would have worked in my use case? I am trying to ensure that I am using the same version of the package that I was using back in 2022 (I regret not taking note of it back then).

Hi again.

I looked at the code for the 2022 versions 1.10-1.13, and they all have the same return for block_lnlp(stats_only=FALSE) which is to return a named list with items stats and model_output where stats is a data.frame with rho from a named list. So it looks like it could have been any of those versions.

@SoftwareLiteracy thank you so much! no idea how i missed that.