plot all figures of a loop simultaneously on one page (all plots of a loop)

[shamimashrafiyan]

I have a render plotly function in which I get some plots of a loop, how can I plot all of them at the same time?
please see a part of my code:
please notice this code is inside of a render plotly

plot_list_final = list()
            for (i in 1: length(s)) {
            Final_fig_box <- subplot(plot_list1[[i]], plot_list2[[i]])%>%
              layout(title = "Inferred Celltype Activity")
            plot_list_final[[i]] = Final_fig_box 
            return(plot_list_final[[i]]) 
            }

I would like to have all plots on a page at the same time, but the code shows only the last plot (last i).
Thanks in advance

[nirgrahamuk]

Seems you have two issues.
Not disposing of all but the last plot that you collate
Then generating a ui capable of displaying all the kept plots.
Ill help you directly with the first part, but the second i recommend you read up on dynamic UI and make a reprex if you want further help with.

plot_list_final = list()
            for (i in 1: length(s)) {
            Final_fig_box <- subplot(plot_list1[[i]], plot_list2[[i]])%>%
              layout(title = "Inferred Celltype Activity")
            plot_list_final[[i]] = Final_fig_box 
    }
            return(plot_list_final) 
            }

[shamimashrafiyan]

Yes, I have realized that I should have separate plot functions in UI section to show all figures together, but my question is: for example if I have 5 plots (i is 5 in the above code), how can I write to add 5 plot functions in UI?
I don't know the syntax!
Is it possible to automatically add UI elements based on the server part's output?

[Matthias]

Is the number of plots fixed or variable?
You could use "insertUI" to insert some elements, e.g. the plots that you could label "plot1", "plot2", "plot3" etc, however I think this isn't very elegant.

The easiest strategy would be to merge all plots into one figure and just show this!?
Did you consider using facet_wrap?
Also the cowplot-library allows to arrange multiple plots into one, here the plot_grid function might be useful when the plots are already stored in a list.

[nirgrahamuk]

library(shiny)
library(plotly)
library(tidyverse) # includes purrr::map*
library(rlang) # for list2()

ui <- fluidPage(
  actionButton("btn","add a plot"),
  uiOutput("myui")
)

server <- function(input, output, session) {
  
  my_plots <- reactiveVal(NULL)
  
  observeEvent(input$btn,{
    existing <- my_plots()
    new_plot_name <- paste0("plot",input$btn)
    new_plot <- plot_ly(economics,x=~date,
                        y=as.formula(paste0("~",sample(c("pce" ,   "pop" ,"psavert", "uempmed" ,"unemploy"),
                                                       1))),
                        type="scatter",mode="markers") %>% layout(title=new_plot_name)
    new_named_plot <- list2(!!sym(new_plot_name) := new_plot)
    new_plot_list <- append(existing,new_named_plot)
    my_plots(new_plot_list)
    output[[new_plot_name]] <- renderPlotly({
      new_plot
    })
  })
  
  
  output$myui <- renderUI({
    mp <- req(my_plots())
    
    imap(mp,
         ~{plotlyOutput(.y)})
  })
  
  
  
}

shinyApp(ui, server)

[nirgrahamuk]

Resources for dynamic UI are available both here:
Shiny - Build a dynamic UI that reacts to user input (rstudio.com)
and here : Chapter 10 Dynamic UI | Mastering Shiny (mastering-shiny.org)