Problem with procedural tabset and ggplot

[Pierre9344]

Description

Hello,
I'm using quarto 1.4.424 on windows 11 to create a report.

In this report, I need to create a plot and tables for multiple variables. I decided to use procedural tabset with the output: asis option to not create 50 code block. But I encounter 2 problems:

1. the time to render my document is abnormally long if the code under the summary table and statistical tests sections. I tried yesterday and it wasn't finished even after 10 hours while it take only 10 minutes if I comment the code of those sections (if I launch this code in RStudio it take about 40 minutes to run the tests and create the tables.
2. When I render the document with only the plots, they are offset by one tabset. It doesn't seem to impact text output but at least the pictures generated by ggplot.

The picture next highlight the second problem I got. The plot should be under the V3_CTS_1_DIC section, not the V3_CTS_2_DIC:

Sorry for the long code. I can't share all my data but the object res[[i]][["data"] correspond to a dataframe with the variables to test as the columns and the participants as the rows (this concern a longitudinal study with 1 to 4 timepoints).
The VarInfo variable is a dataframe that list the variables that need to be tested and their categories.

The text file I attached correspond to the output of the cat function by the code bellow.
ProceduralTabset.txt

Does someone know how I can resolve those problems?

I thank you in advance.
Pierre

```{r}
#| output: asis
#| label: ClusterComparison

load("./data/base/proposalClassification/data.RData")
VarInfo <- openxlsx::read.xlsx("./data/base/proposalClassification/var.xlsx", 1)
# VarInfo <- read.delim("./var.txt")

VarClustering <- c("PANSS_YMRS_IDSC_FUNCTIONNING_COGNITION")


  # create tab for each groups of individuals and prepare the sub-tabs for the different clustering on this group
  for (l in VarClustering) {
    cat("\n\n::: {.panel-tabset group='CATEGORY'}  \n\n")
    for (j in unique(VarInfo$CATEGORY)) {
      VarInfo2 <- subset(VarInfo, VarInfo$CATEGORY == j)
      cat("  \n\n### ", j, "  \n\n")
      cat("\n\n::: {.panel-tabset group='VARIABLE'}  \n\n")
      for (k in 1:nrow(VarInfo2)) {
        cat("\n\n#### ", toupper(VarInfo2$Variable.name[k]), "  \n\n")
        if (is.na(VarInfo2$PsyCourse.codebook.link[k])) {
          cat(paste0(VarInfo2$Variable.name[k], ": ", VarInfo2$DESCRIPTION[k], "\n\n"))
        } else {
          cat(paste0(
            VarInfo2$Variable.name[k], ": [", VarInfo2$DESCRIPTION[k], "](",
            VarInfo2$PsyCourse.codebook.link[k], ")  \n\n"
          ))
        }
        cat("\n\n::: {.panel-tabset group='ANALYSiS'}  \n\n")
        ##################################
        ############## PLOT ##############
        ##################################
        cat("  \n\n##### PLOT  \n\n")
       cat("VarInfo2$Variable.name[k]")
        if (class(as.vector(res[[i]][["data"]][VarInfo2$Variable.name[k]])[[1]]) %in% c("numeric", "integer")) {
          p <- plot_spaghetti_2(res[[i]][["data"]],
            var_name = VarInfo2$Variable.name[k],
            num_clus = l
          )
          ggplot2::ggsave(paste0("./FIG/", VarInfo2$Variable.name[k], ".pdf"), p)
          plot(p)
          if (VarInfo2$Variable.name[k] == "age") {
            # girafe(ggobj = p)
          }
        } else if (class(as.vector(res[[i]][["data"]][VarInfo2$Variable.name[k]])[[1]]) == "factor") {
          if (strsplit(VarInfo2$Variable.name[k], "_")[[1]][1] %in% c("v1", "v2", "v3", "v4", "PRS")) {
            p <- res[[i]][["data"]] %>%
              drop_na(!!sym(VarInfo2$Variable.name[k])) %>%
              droplevels() %>%
              mutate(!!l := as.factor(!!sym(l))) %>%
              ggplot(aes_string(x = l, fill = VarInfo2$Variable.name[k])) +
              geom_bar(position = "fill", colour = "black") +
              scale_y_continuous(labels = scales::percent) +
              ggtitle(paste0(
                "Barplot of ", VarInfo2$Variable.name[k],
                " proportion\n in the ", l, " clusters"
              )) +
              scale_color_brewer(palette = "Set1")
            ggplot2::ggsave(paste0("./FIG/", VarInfo2$Variable.name[k], ".pdf"), p)
            plot(p)
            # girafe(ggobj = p)
          } else {
            p <- plot_alluvial(res[[i]][["data"]],
              num_clus = l,
              var_name = VarInfo2$Variable.name[k]
            )
            ggplot2::ggsave(paste0("./FIG/", VarInfo2$Variable.name[k], ".pdf"), p)
            plot(p)
            # girafe(ggobj = p)
          }
        }

        #################################
        ######### SUMMARY TABLE #########
        #################################
        cat("\n\n##### SUMMARY  \n\n")
#        if (class(as.vector(res[[i]][["data"]][VarInfo2$Variable.name[k]])[[1]]) %in% c("numeric", #"integer")) {
#          if (strsplit(VarInfo2$Variable.name[k], "_")[[1]][1] %in% c("v1", "v2", "v3", "v4")) {
#            summariseContinuous(res[[i]][["data"]],
#              var_name = VarInfo2$Variable.name[k],
#              clustering_name = l,
#              specific_visit = substr(VarInfo2$Variable.name[k], 2, 2)
#            ) %>%
#              kbl(table.attr = "style='width:30%;'", align = c("c", "r")) %>%
#              kable_styling(bootstrap_options = c("bordered", "striped", "hover")) %>%
#              column_spec(1:6, border_left = T, border_right = T) %>%
#              print()
#            cat("  \n  \n")
#          } else {
#            summariseContinuous(res[[i]][["data"]],
#              var_name = VarInfo2$Variable.name[k],
#              clustering_name = l
#            ) %>%
#              kbl(table.attr = "style='width:30%;'", align = c("c", "r")) %>%
#              kable_styling(bootstrap_options = c("bordered", "striped", "hover")) %>%
#              column_spec(1:6, border_left = T, border_right = T) %>%
#              print()
#            cat("  \n  \n")
#          }
#        } else if (class(as.vector(res[[i]][["data"]][VarInfo2$Variable.name[k]])[[1]]) == "factor") {
#          if (strsplit(VarInfo2$Variable.name[k], "_")[[1]][1] %in% c("v1", "v2", "v3", "v4")) {
#            summariseCategorical(res[[i]][["data"]],
#              var_name = VarInfo2$Variable.name[k],
#              clustering_name = l,
#              specific_visit = substr(VarInfo2$Variable.name[k], 2, 2)
#            ) %>%
#              kbl(table.attr = "style='width:30%;'", align = c("c", "r")) %>%
#              kable_styling(bootstrap_options = c("bordered", "striped", "hover")) %>%
#              column_spec(1:4, border_left = T, border_right = T) %>%
#              print()
#            cat("  \n  \n")
#          } else {
#            summariseCategorical(res[[i]][["data"]],
#              var_name = VarInfo2$Variable.name[k],
#              clustering_name = l
#            ) %>%
#              kbl(table.attr = "style='width:30%;'", align = c("c", "r")) %>%
#              kable_styling(bootstrap_options = c("bordered", "striped", "hover")) %>%
#              column_spec(1:4, border_left = T, border_right = T) %>%
#              print()
#            cat("\n  \n")
#          }
#        }

        #########################################
        ########### STATISTICAL TESTS ###########
        #########################################
        cat("\n\n##### CLUSTER COMPARISON  \n\n")
        if (class(as.vector(res[[i]][["data"]][VarInfo2$Variable.name[k]])[[1]]) %in% c("numeric", "integer")) {
          if (strsplit(VarInfo2$Variable.name[k], "_")[[1]][1] %in% c("v1", "v2", "v3", "v4")) {
            cat("  \n<center>**ONE WAY ANOVA**</center>  \n")
            res2 <- res[[i]][["data"]] %>%
              drop_na(any_of(c(VarInfo2$Variable.name[k], l))) %>%
              filter(!!as.symbol(VarInfo2$Variable.name[k]) != -999, visit == substr(VarInfo2$Variable.name[k], 2, 2)) %>%
              anova_test(dv = all_of(VarInfo2$Variable.name[k]), wid = "ID", between = all_of(l))
            get_anova_table(res2, correction = "auto") %>%
              as_tibble() %>%
              kbl(table.attr = "style='width:30%;'", align = c("c", "r")) %>%
              kable_styling(bootstrap_options = c("bordered", "striped", "hover")) %>%
              column_spec(1:6, border_left = T, border_right = T) %>%
              print()
            cat("  \n\n  \n\n")
          } else if (i == "C" && l == "alc_5orm") {
            cat("  \n<center>**ONE WAY ANOVA**</center>  \n\nWe can't do a TWO WAY ANOVA on alc_5orm for the C group as the values for the last two visit are too similar and the function return an error so we do a ONE WAY ANOVA on the first visit values.\n\n")
            res2 <- res[[i]][["data"]] %>%
              drop_na(any_of(c(VarInfo2$Variable.name[k], l))) %>%
              filter(!!as.symbol(VarInfo2$Variable.name[k]) != -999, visit == 1) %>%
              anova_test(dv = all_of(VarInfo2$Variable.name[k]), wid = "ID", between = all_of(l))
            get_anova_table(res2, correction = "auto") %>%
              as_tibble() %>%
              kbl(table.attr = "style='width:30%;'", align = c("c", "r")) %>%
              kable_styling(bootstrap_options = c("bordered", "striped", "hover")) %>%
              column_spec(1:6, border_left = T, border_right = T) %>%
              print()
            cat("  \n\n  \n\n")
          } else if (strsplit(VarInfo2$Variable.name[k], "_")[[1]][1] %in% c("PRS")) {
            cat("  \n<center>**ONE WAY ANOVA**</center>  \n\n")
            res2 <- res[[i]][["data"]] %>%
              drop_na(any_of(c(VarInfo2$Variable.name[k], l))) %>%
              filter(!!as.symbol(VarInfo2$Variable.name[k]) != -999, visit == 1) %>%
              anova_test(dv = all_of(VarInfo2$Variable.name[k]), wid = "ID", between = all_of(l))
            get_anova_table(res2, correction = "auto") %>%
              as_tibble() %>%
              kbl(table.attr = "style='width:30%;'", align = c("c", "r")) %>%
              kable_styling(bootstrap_options = c("bordered", "striped", "hover")) %>%
              column_spec(1:6, border_left = T, border_right = T) %>%
              print()
          } else {
            cat("  \n<center>**TWO WAY ANOVA**</center>  \n")
            res2 <- res[[i]][["data"]] %>%
              drop_na(any_of(c(VarInfo2$Variable.name[k], l))) %>%
              filter(!!as.symbol(VarInfo2$Variable.name[k]) != -999) %>%
              anova_test(dv = all_of(VarInfo2$Variable.name[k]), wid = "ID", between = all_of(l), within = "visit")

            get_anova_table(res2, correction = "auto") %>%
              as_tibble() %>%
              kbl(table.attr = "style='width:30%;'", align = c("c", "r")) %>%
              kable_styling(bootstrap_options = c("bordered", "striped", "hover")) %>%
              column_spec(1:6, border_left = T, border_right = T) %>%
              print()
            cat("  \n\n  \n\n")
          }
        } else if (class(as.vector(res[[i]][["data"]][VarInfo2$Variable.name[k]])[[1]]) == "factor") {
          if (strsplit(VarInfo2$Variable.name[k], "_")[[1]][1] %in% c("v1", "v2", "v3", "v4")) {
            cat("  \n<center>**X²**</center>  \n")
            res2 <- res[[i]][["data"]] %>%
              dplyr::filter(visit == substr(VarInfo2$Variable.name[k], 2, 2)) %>%
              drop_na(!!sym(l)) %>%
              droplevels() %>%
              dplyr::mutate(!!l := as.factor(!!sym(l))) %>%
              dplyr::group_by(!!sym(l), !!sym(VarInfo2$Variable.name[k])) %>%
              summarise(n = n()) %>%
              spread(!!sym(l), n) %>%
              subset(!is.na(.[VarInfo2$Variable.name[k]])) %>%
              base::replace(is.na(.), 0) %>%
              tibble::column_to_rownames(VarInfo2$Variable.name[k]) %>%
              dplyr::select(A, B) %>%
              as.matrix() %>%
              as.table()
            dim <- list()
            dim[[l]] <- rownames(res2)
            dim[[VarInfo2$Variable.name[k]]] <- colnames(res2)
            dimnames(res2) <- dim
            print(chisq.test(res2))
            cat("  \n\n  \n\n")
          } else {
            cat("  \n<center>**X² on visit 1**</center>  \n")
            res2 <- res[[i]][["data"]] %>%
              dplyr::filter(visit == 1) %>%
              drop_na(!!sym(l)) %>%
              droplevels() %>%
              dplyr::mutate(!!l := as.factor(!!sym(l))) %>%
              dplyr::group_by(!!sym(l), !!sym(VarInfo2$Variable.name[k])) %>%
              summarise(n = n()) %>%
              spread(!!sym(l), n) %>%
              subset(!is.na(.[VarInfo2$Variable.name[k]])) %>%
              base::replace(is.na(.), 0) %>%
              tibble::column_to_rownames(VarInfo2$Variable.name[k]) %>%
              dplyr::select(A, B) %>%
              as.matrix() %>%
              as.table()
            dim <- list()
            dim[[l]] <- rownames(res2)
            dim[[VarInfo2$Variable.name[k]]] <- colnames(res2)
            dimnames(res2) <- dim
            print(chisq.test(res2))
            cat("  \n\n  \n\n")

            cat("  \n<center>**X² on visit 2**</center>  \n")
            res2 <- res[[i]][["data"]] %>%
              dplyr::filter(visit == 2) %>%
              drop_na(!!sym(l)) %>%
              droplevels() %>%
              dplyr::mutate(!!l := as.factor(!!sym(l))) %>%
              dplyr::group_by(!!sym(l), !!sym(VarInfo2$Variable.name[k])) %>%
              summarise(n = n()) %>%
              spread(!!sym(l), n) %>%
              subset(!is.na(.[VarInfo2$Variable.name[k]])) %>%
              base::replace(is.na(.), 0) %>%
              tibble::column_to_rownames(VarInfo2$Variable.name[k]) %>%
              dplyr::select(A, B) %>%
              as.matrix() %>%
              as.table()
            dim <- list()
            dim[[l]] <- rownames(res2)
            dim[[VarInfo2$Variable.name[k]]] <- colnames(res2)
            dimnames(res2) <- dim
            print(chisq.test(res2))
            cat("  \n\n  \n\n")

            cat("  \n<center>**X² on visit 3**</center>  \n")
            res2 <- res[[i]][["data"]] %>%
              dplyr::filter(visit == 3) %>%
              drop_na(!!sym(l)) %>%
              droplevels() %>%
              dplyr::mutate(!!l := as.factor(!!sym(l))) %>%
              dplyr::group_by(!!sym(l), !!sym(VarInfo2$Variable.name[k])) %>%
              summarise(n = n()) %>%
              spread(!!sym(l), n) %>%
              subset(!is.na(.[VarInfo2$Variable.name[k]])) %>%
              base::replace(is.na(.), 0) %>%
              tibble::column_to_rownames(VarInfo2$Variable.name[k]) %>%
              dplyr::select(A, B) %>%
              as.matrix() %>%
              as.table()
            dim <- list()
            dim[[l]] <- rownames(res2)
            dim[[VarInfo2$Variable.name[k]]] <- colnames(res2)
            dimnames(res2) <- dim
            print(chisq.test(res2))
            cat("  \n\n  \n\n")

            cat("  \n<center>**X² on visit 4**</center>  \n")
            res2 <- res[[i]][["data"]] %>%
              dplyr::filter(visit == 1) %>%
              drop_na(!!sym(l)) %>%
              droplevels() %>%
              dplyr::mutate(!!l := as.factor(!!sym(l))) %>%
              dplyr::group_by(!!sym(l), !!sym(VarInfo2$Variable.name[k])) %>%
              summarise(n = n()) %>%
              spread(!!sym(l), n) %>%
              subset(!is.na(.[VarInfo2$Variable.name[k]])) %>%
              base::replace(is.na(.), 0) %>%
              tibble::column_to_rownames(VarInfo2$Variable.name[k]) %>%
              dplyr::select(A, B) %>%
              as.matrix() %>%
              as.table()
            dim <- list()
            dim[[l]] <- rownames(res2)
            dim[[VarInfo2$Variable.name[k]]] <- colnames(res2)
            dimnames(res2) <- dim
            print(chisq.test(res2))
            cat("  \n\n  \n\n")

            cat("<center>**TWO WAY ANOVA**</center>  \n")
            res2 <- res[[i]][["data"]] %>%
              drop_na(any_of(c(VarInfo2$Variable.name[k], l))) %>%
              filter(!!as.symbol(VarInfo2$Variable.name[k]) != -999) %>%
              dplyr::mutate(!!VarInfo2$Variable.name[k] := as.numeric(!!sym(VarInfo2$Variable.name[k]))) %>%
              anova_test(dv = all_of(VarInfo2$Variable.name[k]), wid = "ID", between = all_of(l), within = "visit")
            get_anova_table(res2, correction = "auto") %>%
              as_tibble() %>%
              kbl(table.attr = "style='width:30%;'", align = c("c", "r")) %>%
              kable_styling(bootstrap_options = c("bordered", "striped", "hover")) %>%
              column_spec(1:6, border_left = T, border_right = T) %>%
              print()
            cat("  \n\n  \n\n")
          }
        }

        cat("\n\n:::\n\n") # end analysis tabset
      }
      cat("\n\n:::\n\n") # end variable tabset
    }
    cat("\n\n:::\n\n") # end category tabset
  }
}
```
````

[mcanouil]

Your example is way too complex in terms of code and style and is not reproducible.
Likely an error in your code, use keep-md: true and investigate what you produced.

Could you share a small self-contained "working" (reproducible) example to work with, i.e., a complete Quarto document or a Git repository? Thanks.

You can share a Quarto document using the following syntax, i.e., using more backticks than you have in your document (usually four ````).

````qmd
---
title: "Reproducible Quarto Document"
format: html
---

This is a reproducible Quarto document using `format: html`.
It is written in Markdown and contains embedded R code.
When you run the code, it will produce a plot.

```{r}
plot(cars)
```

The end.
````

[cderv]

Just a suggestion that you could try to use child document to create your tabset content dynamically. See this example: https://examples.quarto.pub/create-tabsets-panel-from-r-code/

Using cat() to output markdown mixed with complex results (like tables or plots) in a fo-loop may not work well. It is better to use knit_child() in this case so that markdown are text and computation are chunks, so that their outputs are handled correctly.

Not sure this is related to your issue, but this is best practice with knitr

[harrelfe]

Along those lines check out the maketabs function in the R qreport package. Examples are here.

[Pierre9344]

Hello,

I managed to resolve the problem of the figure being offset. It seem it was caused by a ligne with ::: that didn't close anything. For the table it is probably because of all the conditions I placed in my loop that it didn't print. I temporarily moved back to rmarkdown. I will take a look at knit_child and qreport in the future.
Thank you all for your answers.

[mcanouil]

If there is ::: not closing anything it very likely means you have issues in your code (keep-md: true).

Personal note on design, three nested levels of tabset are not really ideal to navigate through.
Regular sections and subsections with tabsets is way better for navigation purposes.

Also here if you persist with this design, you should use knit_child as suggested because here it's nearly impossible to maintain your code or read it. Additionally, no IDE/tools can check errors, etc. in your current code.