Mixed Models Module
- mcfanda@gmail.com
- Posts: 463
- Joined: Thu Mar 23, 2017 9:24 pm
Re: Mixed Models Module
weired things happening here .
After checking all versions (that seem to be ok), here is what happens: If I open jamovi, load a file, run any standard module that uses car (say ANOVA), then run one of my module analyses (say GLM), everything works fine.
If I open jamovi, load the same file, and run GLM as the first analysis, it crashes at the line where car::Anova is executed and exit with exit code -11
On the other workstation of mine (ubuntu 16), everything works fine
After checking all versions (that seem to be ok), here is what happens: If I open jamovi, load a file, run any standard module that uses car (say ANOVA), then run one of my module analyses (say GLM), everything works fine.
If I open jamovi, load the same file, and run GLM as the first analysis, it crashes at the line where car::Anova is executed and exit with exit code -11
On the other workstation of mine (ubuntu 16), everything works fine
- mcfanda@gmail.com
- Posts: 463
- Joined: Thu Mar 23, 2017 9:24 pm
Re: Mixed Models Module
and the same goes with "lsmeans" package. If I run any other modules that uses it (ANOVA post-hoc) first, it works on my module too. If I run my module first, it crashes at the line when lsmeans is executed.
Re: Mixed Models Module
at a guess, this would be a compiler mismatch issue. jamovi for linux (and it's native packages) need to be built with gcc 5.4. (i didn't consider this angle when prepping for artful)
the easiest way to get up and running here, would be to copy the required native R packages from /usr/lib/jamovi/Resources/modules/jmv/R, into your modules build/R
i can't think of a good solution for this off the top of my head (there's likely not a good one).
jonathon
the easiest way to get up and running here, would be to copy the required native R packages from /usr/lib/jamovi/Resources/modules/jmv/R, into your modules build/R
i can't think of a good solution for this off the top of my head (there's likely not a good one).
jonathon
Re: Mixed Models Module
actually, possibly an R version mismatch ...
i'll be taking a look at this today.
cheers
i'll be taking a look at this today.
cheers
- mcfanda@gmail.com
- Posts: 463
- Joined: Thu Mar 23, 2017 9:24 pm
Re: Mixed Models Module
Hi
the "coping strategy" from /usr/lib/jamovi/Resources/modules/jmv/R, does not work
the "coping strategy" from /usr/lib/jamovi/Resources/modules/jmv/R, does not work
Re: Mixed Models Module
righto. just working on this now.
- mcfanda@gmail.com
- Posts: 463
- Joined: Thu Mar 23, 2017 9:24 pm
Re: Mixed Models Module
In the meantime I finished the multinomial and fixed the estimated marginal means and post-hoc for all models. As soon as I can compile it all push it on github. BTW, lsmeans package rocks!!
Re: Mixed Models Module
incidentally, lsmeans is being discontinued. you should check out emmeans.
http://cran.r-project.org/package=emmeans
i'll forward you the email i received
jonathon
http://cran.r-project.org/package=emmeans
i'll forward you the email i received
jonathon
Re: Mixed Models Module
hey, sorry this took so long.
if you update jamovi and jmvtools you should be gtg.
however, you'll have the same issue with viridis and viridisLite, so copy viridisLite from /usr/lib/jamovi/Resources/modules/jmv/R into gamlj/build/R before building.
cheers
jonathon
if you update jamovi and jmvtools you should be gtg.
however, you'll have the same issue with viridis and viridisLite, so copy viridisLite from /usr/lib/jamovi/Resources/modules/jmv/R into gamlj/build/R before building.
cheers
jonathon
- mcfanda@gmail.com
- Posts: 463
- Joined: Thu Mar 23, 2017 9:24 pm
Re: Mixed Models Module
and I'm indeed go to go. thanks a lot
cheers
marcello
cheers
marcello