Migrate specific checks into vignette

courtiol · Mar 7, 2021 · 00a65be · 00a65be
1 parent 9d32ed0
commit 00a65be
Show file tree

Hide file tree

Showing 2 changed files with 169 additions and 27 deletions.
diff --git a/TODO_spaMM.md b/TODO_spaMM.md
@@ -22,34 +22,17 @@ This document describes what should be done so that spaMM becomes supported by b
 See https://www.tidymodels.org/learn/develop/broom/ and the broom.mixed vignette for detailed instructions
 
 ### Improve the main methods
-TO COME
+See the dedicated spaMM vignette in this package
 
-### Check that it works for all common random effect structures
-- [ ] check with no random effect
-- [ ] check one and two random effects
-- [ ] check one and two random slopes
-- [ ] check Matern(1|x, y)
-- [ ] check AR1(1|time)
-- [ ] check corrMatrix(1|group)
+### Checks / Tests
+Make sure it works for:
 
-### Check that it works with the 3 fitting functions
-- [ ] check output from fitme()
-- [ ] check output from HLfit()
-- [ ] check output from corrHLfit()
+- [ ] all common random effect structures
+- [ ] all common families and links
+- [ ] the 3 fitting functions (`fitme()`, `HLfit()`, `corrHLfit()`)
+- [ ] multivariate models?
 
-### Check that it works with common families and links
-- [ ] check gaussian (log + identity)
-- [ ] check Gamma (log + identity)
-- [ ] check Binomial (logit + probit)
-- [ ] check Poisson (log + identity)
-- [ ] check truncated Poisson (log)
-- [ ] check NegBin (log)
-- [ ] check truncated NegBin (log)
-
-### Handle multivariate models?
-- [ ] tidy for HLfitlist?
-- [ ] augment for HLfitlist?
-- [ ] glance for HLfitlist?
+See the dedicated spaMM vignette in this package
 
 ### Polish
 - [ ] cleanup test file
@@ -58,6 +41,7 @@ TO COME
 - [ ] store fitted models as rda in inst/extdata (produced in run_example.R)
 - [ ] update documentation (see https://roxygen2.r-lib.org/articles/markdown.html)
 - [ ] update examples
+- [ ] remove special spaMM vignette
 - [ ] update vignette
 - [ ] update NEWS
 - [ ] delete this file

diff --git a/vignettes/spaMM.rmd b/vignettes/spaMM.rmd
@@ -5,6 +5,7 @@ date: "`r Sys.Date()`"
 output:
   rmarkdown::html_vignette:
     toc: yes
+    toc_depth: 4
 vignette: >
   %\VignetteIndexEntry{Testing broom.mixed for spaMM}
   %\VignetteEngine{knitr::rmarkdown}
@@ -54,7 +55,7 @@ glance(lm1_spaMM_fitme)
 glance(lm1_spaMM_fitme, npar.details = TRUE)
 ```
 
-#### Notes
+#### FIXME?
 Some statistics produced by `glance.lm()` are missing from `glance.HLfit()`:
 - `r-squared` \& `adj.r.squared` statistics $\rightarrow$ unlikely to work in general unless new function appear in spaMM, so perhaps it makes sense to skip it
 - `p.value` \& `df` $\rightarrow$ this corresponds to a test with a null model $\rightarrow$  unlikely to work in general so perhaps it makes sense to skip it
@@ -102,5 +103,162 @@ glance(lm3_spaMM_fitme)
 glance(lm4_spaMM_fitme)
 ```
 
-#### Notes
+#### FIXME?
 These cases do not seem to be handled properly $\rightarrow$ the question is whether we should deal with them and detect such cases and produce a warning
+
+## Poisson GLM
+
+### Outputs from `tidy()`
+TO COME
+
+### Outputs from `augment()`
+TO COME
+
+### Outputs from `glance()`
+TO COME
+
+
+## Binomial GLM
+
+### Outputs from `tidy()`
+TO COME
+
+### Outputs from `augment()`
+TO COME
+
+### Outputs from `glance()`
+TO COME
+
+## Negative Binomial GLM
+
+### Outputs from `tidy()`
+TO COME
+
+### Outputs from `augment()`
+TO COME
+
+### Outputs from `glance()`
+TO COME
+
+## COM-Poisson GLM
+
+### Outputs from `tidy()`
+TO COME
+
+### Outputs from `augment()`
+TO COME
+
+### Outputs from `glance()`
+TO COME
+
+## Zero truncated GLM
+
+### Outputs from `tidy()`
+TO COME
+
+### Outputs from `augment()`
+TO COME
+
+### Outputs from `glance()`
+TO COME
+
+## LMM with 1 simple random effect
+
+### Outputs from `tidy()`
+TO COME
+
+### Outputs from `augment()`
+TO COME
+
+### Outputs from `glance()`
+TO COME
+
+## LMM with 2 simple random effect
+
+### Outputs from `tidy()`
+TO COME
+
+### Outputs from `augment()`
+TO COME
+
+### Outputs from `glance()`
+TO COME
+
+## LMM with random slopes
+
+### Outputs from `tidy()`
+TO COME
+
+### Outputs from `augment()`
+TO COME
+
+### Outputs from `glance()`
+TO COME
+
+## LMM with autocorrelation
+
+### Outputs from `tidy()`
+TO COME
+
+### Outputs from `augment()`
+TO COME
+
+### Outputs from `glance()`
+TO COME
+
+## LMM with correlation matrix
+
+### Outputs from `tidy()`
+TO COME
+
+### Outputs from `augment()`
+TO COME
+
+### Outputs from `glance()`
+TO COME
+
+## DHLM
+
+### Outputs from `tidy()`
+TO COME
+
+### Outputs from `augment()`
+TO COME
+
+### Outputs from `glance()`
+TO COME
+
+
+## GLMM
+
+### Outputs from `tidy()`
+TO COME
+
+### Outputs from `augment()`
+TO COME
+
+### Outputs from `glance()`
+TO COME
+
+## DHGLM
+
+### Outputs from `tidy()`
+TO COME
+
+### Outputs from `augment()`
+TO COME
+
+### Outputs from `glance()`
+TO COME
+
+
+## Multivariate fits
+
+### Outputs from `tidy()`
+TO COME
+
+### Outputs from `augment()`
+TO COME
+
+### Outputs from `glance()`
+TO COME