Error CoordTern requires the following missing aesthetics (tlr->xy) : z when using ggtern in Shiny

[pre1ryp]

I am trying to write a Shiny app that reads a file with the composition of
three component systems and plots the ternary phase diagram using the ggtern package.

I can do this in basic R but when trying to implement it as part of Shiny I get and error which I am not sure how to fix:

Error:CoordTern requires the following missing aesthetics (tlr->xy) : z

The app consists of fileInput widget and three selectInput widgets part of a renderUI functions so that the user can select the variable for each corner of the ternary plot from the file.

I am using the latest library versions:
ggplot2 3.3.0 and ggtern 3.3.0

This is the content of the ‘.csv’ file as a data frame:

df <- data.frame("Comp1" = c(0.3, 0.5, 0.6, 0.75, 0.8),
                 "Comp2" = c(0.3, 0.25, 0.15, 0.15, 0.1),
                 "Comp3" = c(0.4, 0.25, 0.25, 0.1, 0.1),
                 "Value" = c(300, 500, 1200, 2500, 4500))

^{Created on 2020-04-13 by the reprex package (v0.2.1)}

Below is the reprex of the shiny code:

#
# Ternary Phase Plot
#

library(shiny)
library(ggplot2)
library(ggtern)
#> --
#> Remember to cite, run citation(package = 'ggtern') for further info.
#> --
#> 
#> Attaching package: 'ggtern'
#> The following objects are masked from 'package:ggplot2':
#> 
#>     aes, annotate, ggplot, ggplot_build, ggplot_gtable,
#>     ggplotGrob, ggsave, layer_data, theme_bw, theme_classic,
#>     theme_dark, theme_gray, theme_light, theme_linedraw,
#>     theme_minimal, theme_void
library(DT)
#> 
#> Attaching package: 'DT'
#> The following objects are masked from 'package:shiny':
#> 
#>     dataTableOutput, renderDataTable

ui <- fluidPage(
  titlePanel("Ternary Phase Plot"),
  sidebarLayout(
    sidebarPanel(
      
      fileInput("file","Upload '.csv' file"),
      br(),
      uiOutput("vx"), # vx is coming from renderUI in server
      br(),
      uiOutput("vy"), # vy is coming from renderUI in server
      br(),
      uiOutput("vz") # vy is coming from renderUI in server
    ),
    
    mainPanel(
      plotOutput("plot", height = "800px"),
      DT::dataTableOutput("table")
    )
  )
)


server <- function(input, output, session) {
  # get the names of the variables from the data file and used them in the selectInput part of the renderUI function
  vars <- reactive({
  req(input$file)
  names(read.csv(input$file$datapath,
                 header = TRUE,
                 sep = ","))
  })
  # read the data file
  df <- reactive({
    req(input$file)
    read.csv(input$file$datapath,
                   header = TRUE,
                   sep = ",")
  })
  
  output$vx <- renderUI({
    selectInput("varx", "Select the 1st (L) component", choices = vars())
  })
  
  output$vy <- renderUI({
    selectInput("vary", "Select the 2nd (T) component", choices = vars())
  })
  
  output$vz <- renderUI({
    selectInput("varz", "Select the 3rd (R) component", choices = vars())
  })
  
  # render plot
  output$plot <- renderPlot({

    ggtern(df(), aes_string(x=input$varx, y=input$vary, z=input$varz)) + 
      geom_point(aes(fill=Value), color="black",shape=21, size=6) +
      labs(fill="Value")

    
  })
  
  # print dataframe in table to confirm it is read properly
  output$table <- DT::renderDataTable(
    df()
  )

}

# Run the application 
shinyApp(ui = ui, server = server)

[pre1ryp]

I found the answer to my problem from an stack overflow question (https://stackoverflow.com/a/55102305/11234102). It worked although the answer was provided to resolve a different problem. The ggtern function needs to be put in a print() command.