Trying to nest a plot-generating module inside a second module that creates several plots.

nested_modules.R

options(stringsAsFactors = FALSE,
        scipen = 9999)

library(shiny)
library(ggplot2)
library(dplyr)

theme_set(theme_bw())




# Function to plot a single selected measurement - Sepal.Length, Sepal.Width
generateMeasurePlot <- function(input, output, session, 
                                species_table,
                                measurement) {

    measure_plot <- renderPlot({

        ggplot(species_table(), aes_string(x = measurement())) +
            geom_histogram()


    })

    return(measure_plot)

}




# Generate a plot for every measurement
generateSpeciesPlots <- function(input, output, session, 
                               species_table) {

    # Identify the measures on the given table.
    # For iris, they're always the same, but my data has a varying number of
    # measures per table.
    measureNames <- reactive({

        grep(x = names(species_table()),
             pattern = ".*Length|.*Width",
             value = TRUE)

    })

    # Render each plot
    speciesPlots <- reactive({

        lapply(measureNames(), FUN = function(this_measure_name) {
            
            # WHY DOES THIS EVEN WORK
            this_measure_fun <- reactive(this_measure_name)

            callModule(module = generateMeasurePlot,
                       id = paste0(this_measure, "Plot"),
                       species_table = reactive(species_table()),
                       measurement = this_measure_fun)
            

        })

    })


    # Create a UI element for each plot
    speciesPlotsUI <- renderUI({
        do.call(tagList, speciesPlots())
    })


    # Return
    speciesPlotsUI


}








#######################################
# User Interface
#######################################

ui <- fluidPage(
    fluidRow(
        column(width = 3,
               selectInput(inputId = "selected_species", 
                           label = "Species",
                           choices = c("setosa", "versicolor", "virginica"),
                           selected = "setosa")),
        column(width = 3,
               selectInput(inputId = "selected_measurement",
                           label = "Measurement",
                           choices = ""))
    ),
    fluidRow(
        column(6, tag("h3", "measurePlot"), plotOutput("measurePlot")),
        column(6, tag("h3", "speciesPlots"), uiOutput("speciesPlots"))
    )
)









#######################################
# Serverside
#######################################

server <- function(input, output, session) {


    # Extract the selected species.
    # In the real app, these are different tables, with potentially varying
    # schema.
    species_table <- reactive({

        iris[iris$Species %in% input$selected_species, ]

    })

    # Update the UI with the available measures
    # In this case, the measures are the same for all Species - but in the
    # real app, they vary.
    available_measures <- reactive({
        grep(x = names(species_table()),
             pattern = ".*Length|.*Width",
             value = TRUE)
    })

    observe({
        updateSelectInput(session = session,
                          inputId = "selected_measurement",
                          choices = available_measures())
    })
    
    
    
    # Generate the output
    output$selectedMeasure <- reactive(input$selected_measurement)

    output$measurePlot <- callModule(module = generateMeasurePlot,
                                    id = "onemeasure",
                                    species_table = reactive(species_table()),
                                    measurement = reactive(input$selected_measurement))

    output$speciesPlots <- callModule(module = generateSpeciesPlots,
                                    id = "allmeasures",
                                    species_table = reactive(species_table()))


}




# Run the app
shinyApp(ui, server)