cannot apply check_model title with patchwork::plot_annotation
qdread opened this issue · 4 comments
I recently tried to run some code from 2022 again, where I applied a title to a check_model() plot using patchwork::plot_annotation(). It returns an error which I've unsuccessfully tried to debug. This reprex should be equivalent to the code I'm trying to run. Thanks very much for your help!
Reprex:
library(easystats)
library(patchwork)
m <- lm(mpg ~ factor(cyl) + disp + hp, data = mtcars)
pp <- check_model(m)
print(plot(pp) + plot_annotation(title = 'mtcars model'))
Error:
Error in p + plot_annotation(title = "mtcars model") :
non-numeric argument to binary operator
Other potentially useful info:
> class(pp)
[1] "check_model" "see_check_model"
> class(plot(pp))
[1] "list"
Session info:
R version 4.3.3 (2024-02-29 ucrt)
Platform: x86_64-w64-mingw32/x64 (64-bit)
Running under: Windows 11 x64 (build 22631)
Matrix products: default
locale:
[1] LC_COLLATE=English_United States.utf8 LC_CTYPE=English_United States.utf8 LC_MONETARY=English_United States.utf8
[4] LC_NUMERIC=C LC_TIME=English_United States.utf8
time zone: America/New_York
tzcode source: internal
attached base packages:
[1] stats graphics grDevices utils datasets methods base
other attached packages:
[1] patchwork_1.2.0 see_0.8.3 report_0.5.8 parameters_0.21.6 performance_0.11.0 modelbased_0.8.7 insight_0.19.10
[8] effectsize_0.8.7 datawizard_0.10.0 correlation_0.8.4 bayestestR_0.13.2 easystats_0.7.1
loaded via a namespace (and not attached):
[1] gtable_0.3.4 tensorA_0.36.2.1 xfun_0.43 ggplot2_3.5.0 twosamples_2.0.1
[6] caTools_1.18.2 processx_3.8.4 lattice_0.22-5 vctrs_0.6.5 tools_4.3.3
[11] ps_1.7.6 bitops_1.0-7 generics_0.1.3 parallel_4.3.3 sandwich_3.1-0
[16] pbmcapply_1.5.1 tibble_3.2.1 fansi_1.0.6 DEoptimR_1.1-3 cmdstanr_0.7.1
[21] pkgconfig_2.0.3 Matrix_1.6-5 checkmate_2.3.1 distributional_0.4.0 lifecycle_1.0.4
[26] farver_2.1.1 compiler_4.3.3 munsell_0.5.1 codetools_0.2-19 pracma_2.4.4
[31] pillar_1.9.0 MASS_7.3-60.0.1 iterators_1.0.14 abind_1.4-5 multcomp_1.4-25
[36] foreach_1.5.2 nlme_3.1-164 robustbase_0.99-2 posterior_1.5.0 tidyselect_1.2.1
[41] mvtnorm_1.2-4 dplyr_1.1.4 labeling_0.4.3 splines_4.3.3 grid_4.3.3
[46] colorspace_2.1-0 cli_3.6.2 qqconf_1.3.2 magrittr_2.0.3 survival_3.5-8
[51] utf8_1.2.4 TH.data_1.1-2 withr_3.0.0 scales_1.3.0 backports_1.4.1
[56] opdisDownsampling_1.0.1 estimability_1.5 emmeans_1.10.1 zoo_1.8-12 coda_0.19-4.1
[61] qqplotr_0.0.6 knitr_1.46 doParallel_1.0.17 mgcv_1.9-1 rlang_1.1.3
[66] Rcpp_1.0.12 glue_1.7.0 rstudioapi_0.16.0 R6_2.5.1
plot() returns a list of ggplot-objects, not an object where you can add layers from patchwork. However, is you use see::plots() on a list of ggplot-objects, that will return a patchwork-object:
library(easystats)
library(patchwork)
m <- lm(mpg ~ factor(cyl) + disp + hp, data = mtcars)
pp <- check_model(m)
x <- plot(pp)
plots(x, n_columns = 2) + plot_annotation(title = 'mtcars model')
Created on 2024-04-17 with reprex v2.1.0
@strengejacke i had thought plot() returned the patchwork object. Did we change that at some point? I don't seem much value in retuning a list vs the final thing that is shown by default--should we change it?
Maybe we can just add an else here?
https://github.com/easystats/see/blob/7af4340caa87077c50d4baa83ef35f6ff712f7d4/R/plot.check_model.R#L276
Yes I am fairly sure that plot() returned the patchwork object at some point, and now returns a list.