search

wbonat / mcglm

Fitting multivariate covariance generalized linear models
GNU General Public License v3.0
21 stars 5 forks source link

Error when running basic code from http://mcglm.leg.ufpr.br/articles/mcglm.html #20

Open osupplisson opened 9 months ago

osupplisson commented 9 months ago

Hi,

I recently found your package and after looking at the 2016 Applied Stat paper, I decided to give it a try. It seems to answer one of my long-standing needs ! So first of all, thanks a lot for that and your work!!! :-)

However, when I gave it a try on R with one example from the website, I got the following error (translated from French):

Error in (function (classes, fdef, mtable) : unable to find an inherited method for the 'forceSymmetric' function for the signature '"dgCMatrix", "logical" Error in model_fit$IterationCovariance : $ operator is invalid for atomic vectors'

Please, find more details about the code that produced the error below:

# Takken from http://mcglm.leg.ufpr.br/articles/mcglm.html
library(mcglm)
data(soya)
form.grain <- grain ~ block + water * pot
Z0 <- mc_id(soya)
fit_grain <- mcglm(linear_pred = c(form.grain), matrix_pred = list(Z0), data = soya)

My sessionInfo() is as follow:

R version 4.3.2 (2023-10-31 ucrt)
Platform: x86_64-w64-mingw32/x64 (64-bit)
Running under: Windows 11 x64 (build 22621)

Matrix products: default

locale:
[1] LC_COLLATE=French_France.utf8  LC_CTYPE=French_France.utf8    LC_MONETARY=French_France.utf8
[4] LC_NUMERIC=C                   LC_TIME=French_France.utf8    

time zone: Europe/Paris
tzcode source: internal

attached base packages:
[1] stats     graphics  grDevices utils     datasets  methods   base     

other attached packages:
[1] mcglm_0.6.0

loaded via a namespace (and not attached):
[1] compiler_4.3.2   assertthat_0.2.1 Matrix_1.6-3     tools_4.3.2      Rcpp_1.0.11      grid_4.3.2      
[7] lattice_0.22-5

Do you have an idea of what's going on ? My take is that the package uses a build from 020-06-08 and some things have evolved since then ? Is the package still maintained and expanded ? I generally don't have any issues with other recent packages and everything is up-to-date here.

As an additional question, at the very end of the 2016 paper, you mentioned that penalised smoothing splines might be used (this is marked as future extension in the Journal of Statistical software paper, but I couldn't find anything in the current doc, am I right to assume that this is not implemented ?

Cheers, Olivier

wbonat commented 9 months ago

The package was archived because I did not have time to fix some issues that appears after some changes in the new R version. So, the easy way to fix it is to install an old version of R and old versions of the Matrix package. In special this error was fixed in the version 0.8.0 available here https://cran.r-project.org/src/contrib/Archive/mcglm/ The error is related to a change in the Matrix package in the use of the forceSymmetric function. Thus probably if you install R version 4.0 and mcglm version 0.8.0 you will not have any errors. I should fix the issues, but no funding for the project, so I do it only on my free time which has been so rare right now.