Display output on tm_data_table, tm_variable_browser, tm_missing_data

[llrs-roche]

Pull Request

Fixes insightsengineering/teal#1452

This PR works on the edge cases of datanames = "all" and NULL for each three modules, in addition I also revised the decision on tm_front_page to show all modules by default, and reverted back to not showing the panels by default, as per previous behavior and @vedhav compelling comment.

Code used to check this (besides the examples)

library("teal")
devtools::load_all("../teal.modules.general")

data <- teal_data()
data <- within(data, {
  library(dplyr)
  IRIS <- iris
  MTCARS <- mtcars
  # ADSL <- teal.data::rADSL
})

datanames <- "all"

app <- init(
  data = data,
  modules = modules(
    example_module(),
    tm_front_page(datanames = datanames),
    tm_data_table(datanames = datanames),
    tm_variable_browser(datanames = datanames),
    tm_missing_data(datanames = datanames)
  )
)

shinyApp(app$ui, app$server)


datanames <- NULL

app <- init(
  data = data,
  modules = modules(
    example_module(),
    tm_front_page(datanames = datanames),
    tm_data_table(datanames = datanames),
    tm_variable_browser(datanames = datanames),
    tm_missing_data(datanames = datanames)
  )
)

shinyApp(app$ui, app$server)

module	"all"	NULL
tm_data_table()	✅	✅ 1
tm_variable_browser()	✅	✅ 2
tm_missing_data()	✅	✅ 2

1: No right-panel but the table uses all datasets that are data.frames. I was thinking if variables_selected names should be used when datanames is NULL ? So that something like this should only show IRIS dataset and no left panel.

teal.modules.general::tm_data_table(datanames = NULL,
                                      variables_selected = list(IRIS = colnames(data$IRIS)[1:3])),

But this gets complicated if one wants to keep the standard 6 columns of the datasests as assertion fails:

teal.modules.general::tm_data_table(datanames = NULL,
                                      variables_selected = list(IRIS = colnames(data$IRIS)[1:3], MTCARS = NULL)),
Assertion on 'variables_selected' failed: May only contain the following types: {character}, but element 2 has type 'NULL'.

2: No right-panel but all the datasets that are data.frames are shown.

---

I also tested whether transformators worked with tm_missing_data (the only one of these three modules with a transformator argument) because the datanames documentation implies it will be merged with it: currently it doesn't and I'm not sure how to fix it.

If I create a transformator and use datanames not equal to all by default the new dataname won't be used.
I need to explore better the interaction between teal::module and the transformator.

reprex with transformator

library("teal")

data <- teal_data()
data <- within(data, {
  library(dplyr)
  IRIS <- iris
  MTCARS <- mtcars
  # ADSL <- teal.data::rADSL
})


transf <- list(teal_transform_module(
  label = "Transformator",
  ui = function(id) {
    ns <- NS(id)
    tags$div(
      numericInput(ns("n_rows"), "Number of rows to subset", value = 32, min = 1, max = 150, step = 1)
    )
  },
  server = function(id, data) {
    moduleServer(id, function(input, output, session) {
      reactive({
        within(data(),
               {
                 iris <- head(IRIS, num_rows)
               },
               num_rows = input$n_rows
        )
      })
    })
  },
  datanames = "all"
))


datanames <- "IRIS"

app <- init(
  data = data,
  modules = modules(
    example_module(),
    tm_front_page(datanames = datanames),
    tm_data_table(datanames = datanames),
    tm_variable_browser(datanames = datanames),
    tm_missing_data(datanames = datanames, transformators = transf)
  )
)

shinyApp(app$ui, app$server)

[m7pr]

@llrs-roche about transformators and datanames. Have you tried to unify names?
Looks like teal_data() only contains IRIS and MTCARS, however module uses only IRIS BUT your transformator tries to change iris