Merge pull request #157 from jacob-long/master

Integrate recent package updates

.Rbuildignore

@@ -19,11 +19,11 @@
 ^CONTRIBUTING.md$
 ^\.github$
 ^\.vs$
-^doc$
 ^Meta$
 ^make_hex\.R$
 ^hex\.png$
 ^LICENSE\.md$
 ^paper\.md$
 ^pkgdown$
 ^CRAN-SUBMISSION$
+^doc$

.github/workflows/recheck.yml

@@ -0,0 +1,18 @@
+on:
+    workflow_dispatch:
+      inputs:
+        which:
+          type: choice
+          description: Which dependents to check
+          options:
+          - strong
+          - most
+
+name: Reverse dependency check
+
+jobs:
+  revdep_check:
+    name: Reverse check ${{ inputs.which }} dependents
+    uses: r-devel/recheck/.github/workflows/recheck.yml@v1
+    with:
+        which: ${{ inputs.which }}

DESCRIPTION

@@ -15,27 +15,27 @@ License: GPL (>= 3)
 Encoding: UTF-8
 LazyData: true
 Imports: 
-    crayon,
+    cli,
     generics,
+    broom,
+    broom.mixed,
     ggplot2 (>= 3.4.0),
     magrittr,
     pander,
     pkgconfig,
-    rlang (>= 0.3.0),
+    rlang,
+    sandwich,
     tibble
 Suggests:
   boot,
-  broom,
-  broom.mixed,
-  huxtable (>= 3.0.0),
+  huxtable,
   kableExtra,
   lme4,
   lmerTest,
   MASS,
   methods,
   pbkrtest,
   RColorBrewer,
-  sandwich,
   scales,
   survey,
   weights,
@@ -47,8 +47,8 @@ Enhances:
   brms,
   quantreg,
   rstanarm
-RoxygenNote: 7.2.3
+RoxygenNote: 7.3.2.9000
 VignetteBuilder: knitr
 Roxygen: list(markdown = TRUE)
 Depends: 
-    R (>= 3.5.0)
+    R (>= 3.6.0)

NAMESPACE

@@ -20,90 +20,6 @@ if (!nzchar(system.file(package = "interactions"))) {
 if (!nzchar(system.file(package = "interactions"))) {
   export(sim_slopes)
 }
-
-if (getRversion() >= "3.6.0") {
-  S3method(generics::tidy, glht)
-} else {
-  export(tidy.glht)
-}
-
-if (getRversion() >= "3.6.0") {
-  S3method(generics::tidy, summary.glht)
-} else {
-  export(tidy.summary.glht)
-}
-
-if (getRversion() >= "3.6.0") {
-  S3method(knitr::knit_print, summ.glm)
-} else {
-  export(knit_print.summ.glm)
-}
-
-if (getRversion() >= "3.6.0") {
-  S3method(knitr::knit_print, summ.lm)
-} else {
-  export(knit_print.summ.lm)
-}
-
-if (getRversion() >= "3.6.0") {
-  S3method(knitr::knit_print, summ.merMod)
-} else {
-  export(knit_print.summ.merMod)
-}
-
-if (getRversion() >= "3.6.0") {
-  S3method(knitr::knit_print, summ.rq)
-} else {
-  export(knit_print.summ.rq)
-}
-
-if (getRversion() >= "3.6.0") {
-  S3method(knitr::knit_print, summ.svyglm)
-} else {
-  export(knit_print.summ.svyglm)
-}
-
-if (getRversion() >= "3.6.0") {
- S3method(generics::glance, summ.lm)
-} else {
- export(glance.summ.lm)
-}
-
-if (getRversion() >= "3.6.0") {
- S3method(generics::glance, summ.glm)
-} else {
- export(glance.summ.glm)
-}
-
-if (getRversion() >= "3.6.0") {
- S3method(generics::glance, summ.svyglm)
-} else {
- export(glance.summ.svyglm)
-}
-
-if (getRversion() >= "3.6.0") {
- S3method(generics::glance, summ.merMod)
-} else {
- export(glance.summ.merMod)
-}
-
-if (getRversion() >= "3.6.0") {
- S3method(generics::glance, summ.rq)
-} else {
- export(glance.summ.rq)
-}
-
-if (getRversion() >= "3.6.0") {
- S3method(generics::tidy, summ)
-} else {
- export(tidy.summ)
-}
-
-if (getRversion() >= "3.6.0") {
- S3method(generics::tidy, summ.merMod)
-} else {
- export(tidy.summ.merMod)
-}
 S3method("%just%",data.frame)
 S3method("%just%",default)
 S3method("%just%",list)
@@ -126,8 +42,15 @@ S3method(get_formula,brmsfit)
 S3method(get_formula,default)
 S3method(get_formula,panelmodel)
 S3method(glance,summ.glm)
+S3method(glance,summ.lm)
 S3method(glance,summ.merMod)
+S3method(glance,summ.rq)
 S3method(glance,summ.svyglm)
+S3method(knitr::knit_print,summ.glm)
+S3method(knitr::knit_print,summ.lm)
+S3method(knitr::knit_print,summ.merMod)
+S3method(knitr::knit_print,summ.rq)
+S3method(knitr::knit_print,summ.svyglm)
 S3method(make_predictions,brmsfit)
 S3method(make_predictions,default)
 S3method(make_predictions,merMod)
@@ -159,6 +82,9 @@ S3method(summ,lm)
 S3method(summ,merMod)
 S3method(summ,rq)
 S3method(summ,svyglm)
+S3method(tidy,summ)
+S3method(tidy,summ.merMod)
+S3method(tidy,summary.glht)
 export("%just%")
 export("%just%<-")
 export("%nin%")
@@ -212,13 +138,11 @@ export(weights_tests)
 export(wgttest)
 export(wrap_str)
 export(wtd.sd)
+import(broom)
+import(broom.mixed)
+import(cli)
 import(ggplot2)
 import(rlang)
-importFrom(crayon,cyan)
-importFrom(crayon,inverse)
-importFrom(crayon,italic)
-importFrom(crayon,red)
-importFrom(crayon,underline)
 importFrom(generics,glance)
 importFrom(generics,tidy)
 importFrom(ggplot2,aes_string)
@@ -234,6 +158,8 @@ importFrom(grDevices,rgb)
 importFrom(magrittr,"%<>%")
 importFrom(magrittr,"%>%")
 importFrom(pander,pandoc.table.return)
+importFrom(rlang,check_installed)
+importFrom(rlang,is_installed)
 importFrom(stats,"contrasts<-")
 importFrom(stats,AIC)
 importFrom(stats,BIC)

NEWS.md

@@ -1,16 +1,52 @@
 # jtools 2.2.3.9999
 
-Bug fix:
+Bug fixes:
 * `predict_merMod()` was miscalculating predictions under a specific set of
 conditions: When the model was fit with `lme4::glmer()`, the link was not
 identity, standard errors were simultaneously being calculated, and the 
 user included random effects. This error has been corrected 
 ([#144](https://github.com/jacob-long/jtools/issues/144))
+* When printing `summ.glm()` results with `knitr`, the *p* value for the 
+omnibus chi-squared test is now included in the model statistics, like it 
+already is when using the function in the console. ([#138](https://github.com/jacob-long/jtools/issues/138)
+* `effect_plot()` no longer fails with an error when models weights are 
+specified without a variable name. 
+([#156](https://github.com/jacob-long/jtools/issues/156))
 
-Enhancement:
+Enhancements:
 
 * `summ()` will now produce model fit statistics for `glm.nb` models 
 ([#142](https://github.com/jacob-long/jtools/issues/142)).
+* `effect_plot()` now includes the argument `facet.by`. If you want to plot
+your prediction variable at multiple values of some other variable, you can 
+pass the name of that variable to `facet.by`. It will, by default, create a 
+separate plot for each unique level of that variable using 
+`ggplot2::facet_wrap()`. If you only want specific levels of the `facet.by`
+variable, you can specify them by giving a vector of values to the `at` argument.
+More than most kinds of plots, you may need to do some further customization
+to the results of `effect_plot()` since the plots may not fit the space well,
+or you may want to rearrange into different row/column configurations, etc. I
+expect this to be most useful for cases when you have a multilevel model and 
+there is a random slope specified for the `pred` variable. A version of this
+feature was requested by Github user *5tokesy* (#147). 
+* `plot_coefs()` now handles `fixest_multi` objects from the `fixest` package
+out of the box. (#123)
+* The `omit.coefs` and `coefs` arguments to  `plot_coefs()` can now be modified
+using the new argument, `coefs.match`. You may use regular expressions to more
+efficiently match coefficient names, which will be most useful when you have 
+models with many coefficients with predictable naming schemes. To have your
+arguments interepreted this way, set `coefs.match = "regex"`. 
+([#122](https://github.com/jacob-long/jtools/issues/122))
+
+Other changes:
+
+* Removed `tidy.glht()` method because the `broom` package now implements one.
+(#139)
+* Made `broom` and `broom.mixed` hard dependencies to help avoid user confusion
+and a few errors that could occur when one was missing. (#149) 
+* `cli` is now used to format (colorizing, italicizing, etc.) console output 
+rather than `crayon`. Please report any issues if formatting doesn't work 
+right for you; there shouldn't be any changes from past releases.
 
 # jtools 2.2.2