textplot_wordcloud() with comparison = TRUE split not max.words evenly by groups

[p-dre]

Describe the bug

I would like to create a wordcloud in which the most used words are displayed by groups. Unfortunately, the maximum number of words is not evenly split between the groups like it is writen at the docs: "The maximum frequency will be split evenly across categories when comparison = TRUE.". I couldn't quite figure out what the split is based on, but it seems to have something to do with the relative importance of the words within each group.

Reproducible code

Dataset: https://dataverse.harvard.edu/dataset.xhtml?persistentId=doi:10.7910/DVN/L4OAKN

Bundestagganz <- readRDS("Corp_Bundestag_V2.rds")
datenneu1 <- subset (Bundestagganz, date <"2018-01-01")

datenneu1 %>% 
  corpus %>%  
  tokens(remove_punct = TRUE, remove_numbers = TRUE, remove_symbols = TRUE) %>% 
  dfm(verbose = FALSE) %>%  
  dfm_group(groups = party) %>% 
  quanteda.textplots::textplot_wordcloud(comparison = TRUE, max.words = 10,title.size = 1)

datenneu1 %>% 
  corpus %>% 
  tokens(remove_punct = TRUE, remove_numbers = TRUE, remove_symbols = TRUE) %>% 
  dfm(verbose = FALSE) %>%
  dfm_group(groups = party) %>% 
  quanteda.textplots::textplot_wordcloud(comparison = TRUE, max_words = 1000 , min_size = 0.1, max_size = 1)  

Expected behavior

My expectation is that max.words describes either the maximlae number of words per group or the maximum number of words in total but then split evenly between the groups.

System information

sessionInfo()
R version 4.2.2 (2022-10-31 ucrt)
Platform: x86_64-w64-mingw32/x64 (64-bit)
Running under: Windows 10 x64 (build 19044)

Matrix products: default

locale:
[1] LC_COLLATE=German_Germany.utf8  LC_CTYPE=German_Germany.utf8    LC_MONETARY=German_Germany.utf8 LC_NUMERIC=C                    LC_TIME=German_Germany.utf8    

attached base packages:
[1] stats     graphics  grDevices utils     datasets  methods   base     

other attached packages:
 [1] word2vec_0.3.4            udpipe_0.8.9              wordcloud_2.6             RColorBrewer_1.1-3        reshape2_1.4.4            quanteda.textplots_0.94.1 forcats_0.5.1             stringr_1.4.0            
 [9] purrr_0.3.4               readr_2.1.2               tidyr_1.2.0               tibble_3.1.7              ggplot2_3.3.6             tidyverse_1.3.2           tidytext_0.3.4            dplyr_1.0.10             
[17] ROCR_1.0-11               quanteda_3.2.1           

loaded via a namespace (and not attached):
 [1] httr_1.4.3          jsonlite_1.8.0      modelr_0.1.8        RcppParallel_5.1.5  assertthat_0.2.1    googlesheets4_1.0.0 cellranger_1.1.0    yaml_2.3.5          pillar_1.7.0        backports_1.4.1     lattice_0.20-45    
[12] glue_1.6.2          digest_0.6.29       rvest_1.0.2         colorspace_2.0-3    htmltools_0.5.2     Matrix_1.5-1        plyr_1.8.7          pkgconfig_2.0.3     broom_0.8.0         haven_2.5.0         scales_1.2.0       
[23] tzdb_0.3.0          googledrive_2.0.0   generics_0.1.2      ellipsis_0.3.2      withr_2.5.0         cli_3.3.0           magrittr_2.0.3      crayon_1.5.1        readxl_1.4.0        evaluate_0.15       stopwords_2.3      
[34] tokenizers_0.2.1    janeaustenr_0.1.5   fs_1.5.2            fansi_1.0.3         SnowballC_0.7.0     xml2_1.3.3          data.table_1.14.2   tools_4.2.2         hms_1.1.1           gargle_1.2.0        lifecycle_1.0.1    
[45] munsell_0.5.0       reprex_2.0.1        compiler_4.2.2      rlang_1.0.2         grid_4.2.2          rstudioapi_0.13     rmarkdown_2.13      gtable_0.3.0        DBI_1.1.2           R6_2.5.1            lubridate_1.8.0    
[56] knitr_1.38          fastmap_1.1.0       utf8_1.2.2          fastmatch_1.1-3     stringi_1.7.6       Rcpp_1.0.8.3        vctrs_0.4.1         dbplyr_2.1.1        tidyselect_1.1.2    xfun_0.30       

[stuckel]

I had exactly the same issue and would also appreciate a solution.