Run R script after input in Shiny - r

Question

Good morning everyone,
I have a Shiny application that collects 5 inputs from the users and stores them into variables.
Then, I would be able to use another R script that would run based on the information provided by the user.
Here is a sample of my Shiny App :
jscode <- "shinyjs.closeWindow = function() { window.close(); }"
#Define UI for application
ui <- pageWithSidebar(
#App title
headerPanel("Filters applied for Powerpoints"),
#Panel to display the filters
sidebarPanel(
#Select dates
dateInput(inputId = "startDate", label = "Start date : ", value = "2018-12-01", format = "yyyy/mm/dd"),
dateInput(inputId = "endDate", label = "End date : ", value = "2018-12-31", format = "yyyy/mm/dd"),
#Select brand template
selectInput("Brand", label = "Select brand : ", choices = list("Carat" = "Carat", "Amplifi" = "Amplifi", "iProspect" = "iProspect", "Isobar" = "Isobar")),
#Select medium type
selectInput("Medium", label = "Select medium type : ", choices = list("Social Post" = "Social Post", "Display" = "Display", "Programmatic" = "Programmatic", "SEA" = "SEA")),
#Enter the plan ID of your campaign
textInput("Camp", label = "Enter the plan ID of your campaign : ", value = ""),
#Button to close the window, then run script
useShinyjs(),
extendShinyjs(text = jscode, functions = c("closeWindow")),
actionButton("close", "Close and run")
),
mainPanel()
)
#Define server logic
server <- function(input, output, session){
observe({
startDate <<- input$startDate
endDate <<- input$endDate
brand <<- input$Brand
medium <<- input$Medium
campaign <<- input$Camp
})
observeEvent(input$close, {
js$closeWindow()
stopApp()
})
source("C:/Users/RPeete01/Desktop/Automated powerpoints/Datorama R/Datorama reporting R/DatoramaSocial.R")
}
#Run the application
shinyApp(ui = ui, server = server)
I've used the source function but it doesn't work.
If someone has an idea, please let me know.
Thanks a lot,
Rémi

Answer 1

You should take advantage of built in onStop functions in shiny to execute some functions before the stopApp() call
library(shiny)
if (interactive()) {
# Open this application in multiple browsers, then close the browsers.
shinyApp(
ui = basicPage("onStop demo",actionButton("close", "Close and run")),
server = function(input, output, session) {
onStop(function() cat("Session stopped\n"))
observeEvent(input$close, {
stopApp()
})
},
onStart = function() {
cat("Doing application setup\n")
onStop(function() {
cat("Doing application cleanup, your functions go here\n")
})
}
)
}

Answer 2

Instead of creating a function to replace your script, you can source your script by supplying an environment to the local option. This environment must contain the objects needed by your script. Something like that:
mylist <- reactiveVal() # we will store the inputs in a reactive list
observe({ # create the list
mylist(list(
startDate = input$startDate,
endDate = input$endDate,
brand = input$Brand,
medium = input$Medium,
campaign = input$Camp))
})
observeEvent(input$runScript, { # "runScript" is an action button
source("myscript.R", local = list2env(mylist()))
})
EDIT
Here is a full example.
library(shiny)
ui <- fluidPage(
textInput("text", "Enter text", value = "test"),
actionButton("runScript", "Run")
)
server <- function(input, output, session) {
mylist <- reactiveVal() # we will store the inputs in a reactive list
observe({ # create the list
mylist(list(
text = input$text))
})
observeEvent(input$runScript, { # "runScript" is an action button
source("myscript.R", local = list2env(mylist()))
})
}
shinyApp(ui, server)
File myscript.R:
writeLines(text, "output.txt")
When I run the app and click on the button, the file output.txt is correctly created (i.e. the script is correctly sourced).

Answer 3

Your script DatoramaSocial.R should be formulated as a function that takes your 5 input values as arguments. As to the return value, well you haven't told us what you want to do with it. By formulating it as a function I mean wrap everything in DatoramaSocial.R in a function (or several subfunctions). The code for that function can easily reside in the external script file or be pasted before the ui and server statements in your shiny app. If the former, simply include the definitions by calling source('DatoramaSocial.R') before your ui and server statements.
Now, in your server function, you can simply call it as a reaction to changes in the input:
observe({
DatoramaSocial(input$startDate, input$endDate, input$Brand, input$Medium, input$Camp)
})
Although in this case, I recommend inserting an actionbuttonInput and having the user click that when they have selected all their inputs. In which case, update to:
observeEvent(input$actionbutton, ignoreInit=TRUE, {
DatoramaSocial(input$startDate, input$endDate, input$Brand, input$Medium, input$Camp)
})
where actionbutton is the actionbutton's inputId.