Force shiny to render plot in loop

问题

I have a shiny app that runs a simulation. The goal is to show the user the calculation steps in between as a plot.

How do I force shiny to update the plot?

An MWE would look like this

library(shiny)

server <- function(input, output, session) {
  # base plot as a placeholder
  output$myplot <- renderPlot(plot(1:1, main = "Placeholder"))

  # wait until the button is triggered
  observeEvent(input$run, {
    print("Do some calculations in 3 steps")
    for (i in seq_len(3)) {
      print("Do some calculations")
      # ...
      x <- seq_len(i * 100)
      y <- (x + 1)^2 - 1 # this will do for now

      print("Plot the data ")

      # ISSUE HERE!
      # this should render the current step of the simulation, instead it 
      # renders only after the whole code is run (i.e., after step 3)
      output$myplot <- renderPlot(plot(x, y, main = sprintf("Round %i", i), type = "l"))

      print("Wait for 1 second for the user to appreciate the plot...")
      Sys.sleep(1)
    }
  })
}

ui <- fluidPage(
  actionButton("run", "START"),
  plotOutput("myplot")
)

shinyApp(ui = ui, server = server)

The issue is, that shiny runs the code and produces one plot at the end of the simulation, however, I want to get a plot at each simulation step (displayed for at least one second).

Any help/hint is greatly appreciated.

Appendix

I have looked at this post, but replacing the text with a plot/renderPlot doesn't yield the correct results.

回答1:

You could nest an observer into an observeEvent to make it work. Based on Jeff Allen's code from the SO topic you linked.

Crucial part:

observeEvent(input$run, {
    rv$i <- 0
    observe({
      isolate({
        rv$i <- rv$i + 1
      })

      if (isolate(rv$i) < maxIter){
        invalidateLater(2000, session)
      }
    })
  })

Full code:

library(shiny)

server <- function(input, output, session) {
  rv <- reactiveValues(i = 0)
  maxIter <- 3

  output$myplot <- renderPlot( {
    if(rv$i > 0) {
      x <- seq_len(rv$i * 100)
      y <- (x + 1)^2 - 1 # this will do for now
      plot(x, y, main = sprintf("Round %i", rv$i), type = "l") 
    } else {
      plot(1:1, main = "Placeholder")
    }
  })

  observeEvent(input$run, {
    rv$i <- 0
    observe({
      isolate({
        rv$i <- rv$i + 1
      })

      if (isolate(rv$i) < maxIter){
        invalidateLater(2000, session)
      }
    })
  })

}

ui <- fluidPage(
  actionButton("run", "START"),
  plotOutput("myplot")
)

shinyApp(ui = ui, server = server)

来源：https://stackoverflow.com/questions/47270540/force-shiny-to-render-plot-in-loop

标签

plot

Shiny