Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Failure showing fitted models on julia-1.6.0-DEV.855 and later #363

Closed
dmbates opened this issue Sep 7, 2020 · 4 comments · Fixed by #364
Closed

Failure showing fitted models on julia-1.6.0-DEV.855 and later #363

dmbates opened this issue Sep 7, 2020 · 4 comments · Fixed by #364

Comments

@dmbates
Copy link
Collaborator

dmbates commented Sep 7, 2020

As I reported in JuliaGraphics/Showoff.jl#24 the re-write of Printf in JuliaLang/julia#32859 removed Base.Grisu, which has broken the Showoff package.

I have felt for a while that the use of Showoff is more ornamental than useful so maybe now is the time to bite the bullet and strip the use of Showoff.showoff from MixedModels.

One fewer dependency sounds like a good idea to me. Opinions?

@palday
Copy link
Member

palday commented Sep 7, 2020

One fewer dependency seems fine with me. Grep'ing the sources, we're only using it in three lines and two of them are essentially the same from a printing perspective (sigmas and variances).

(I wasn't even aware we were using Showoff!)

Without actually trying it out, I would guess we can write a replacement in < 10 lines of code. 😀

@dmbates
Copy link
Collaborator Author

dmbates commented Sep 7, 2020

My recollection is that the Ryu package, which has been part of Julia's stdlib since 1.4, is reasonably accessible.

My sleep schedule is kind of messed up right now so if I find myself awake in the middle of the night I will take a run at it.

@dmbates
Copy link
Collaborator Author

dmbates commented Sep 7, 2020

It appears that we use Ryu.writefixed(obj, 4) for the deviance in the show method for a GeneralizedLinearMixedModel so we could probably do the same for the show method for a LinearMixedModel. The 4 is four digits after the decimal point, which is more than enough for those values. (Recall that changes in the deviance are compared to a chi-square distribution.)

@dmbates
Copy link
Collaborator Author

dmbates commented Sep 7, 2020

Seems like the changes to Printf have also broken the ProgressMeter code so I am going to punt for the time being on getting MixedModels to work with the development version of Julia. I did remove the dependency on Showoff in #364 but we probably don't want to commit that to master until after the tutorial this week.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging a pull request may close this issue.

2 participants