README.md

scriptgloss

Reconstruct static code from shiny apps

scriptgloss exposes functionality for building static code to recreate outputs built in a shiny context. Construct something in shiny and generate the code needed to produce it without needing to trudge through the app code.

Installation

For now, due to a dependency on the graph package in Bioconductor by way of CodeDepends, this package will fail to install via the typical devtools mechanisms. Instead, please use BiocManager and follow any prompts to install the package, pulling from Bioconductor when necessary. I'll be working to reduce this impedence.

install.packages("BiocManager") # R (>3.5.0)
BiocManager::install("dgkf/scriptgloss")

Developer Note This was my first foray into advising installation via BiocManager. If you run into issues, please report them and I'll work to make the installation seamless until this dependency requirement has been resolved.

Getting Started

Start by building a shiny app

Let's say we're building a shiny app to explore the mtcars dataset. We might start with something like this:

library(shiny)

ui <- fluidPage(
  selectInput('x', 'x axis', choices = names(mtcars)),
  selectInput('y', 'y axis', choices = names(mtcars)),
  plotOutput('plot'))

srv <- function(input, output, session) {
  output$plot <- renderPlot({
    plot(x = mtcars[[input$x]], y = mtcars[[input$y]], 
      xlab = input$x, ylab = input$y)
  })
}

shinyApp(ui, srv)

If you run the code above, you'll see that it's quite simple; just two drop down menus to pick an x and y variable which will update a plot to compare the two.

Adding code export elements

In order to add static code output, there are just a few small steps that need to be taken.

  1. The JavaScript dependencies need to be added to the webpage header. This can be done easily by adding the UI element, scriptglossJS().
  2. A UI button needs to be added to export the code. For this, scriptgloss provides the functions showCodeButton() and clipCodeButton() for showing code in a modal window or copying to clipboard respectively.
  3. An observer of the UI button needs to be added to prepare the code.

Adding a "Show Code" button

A minimal example showing how you would use the showCodeButton() UI element. Not the most interesting code in the world, but operational!

library(shiny)
library(scriptgloss)

ui <- fluidPage(
  scriptglossJS(),
  showCodeButton("show_code"))

srv <- function(input, output, session) {
  # ui observer to display a modal code window when the button is pressed
  observeEvent(input$show_code, show_code_modal(srv))
}

shinyApp(ui, srv)

Adding a clipboard button to copy code

A minimal example of how you would add a button to copy static code to the clipboard. Again, nothing too exciting, but hopefully enough to build it into your own work.

library(shiny)
library(scriptgloss)

ui <- fluidPage(
  scriptglossJS(),
  uiOutput('copy_code_btn'))

srv <- function(input, output, session) {
  # ui observer to copy code to the user's clipboard
  output$copy_code_btn <- renderUI(clipCodeButton(srv))
}

shinyApp(ui, srv)

Reworking our shiny app to produce static code

Just by following the steps above, we can quickly add a button to show off our code!

library(shiny)
library(scriptgloss)

ui <- fluidPage(
  selectInput('x', 'x axis', choices = names(mtcars)),
  selectInput('y', 'y axis', choices = names(mtcars)),
  plotOutput('plot'),
  # add our UI elements ... 
  scriptglossJS(),              # <-- don't forget the JavaScript part!
  showCodeButton("show_code"),  # button to show code as a pop-up
  uiOutput("clip_code_btn"))    # button to copy to clipboard

srv <- function(input, output, session) {
  output$plot <- renderPlot({
    plot(x = mtcars[[input$x]], y = mtcars[[input$y]], 
      xlab = input$x, ylab = input$y)
  })

  # observer for our modal button
  observeEvent(input$show_code, show_code_modal(srv, "plot"))

  # button renderer for our copy-to-clipboard button
  output$clip_code_btn <- renderUI(clipCodeButton(text = get_code(srv, "plot")))
}

shinyApp(ui, srv)

Acknowledgements

Early prototypes developed at Genentech. Many thanks to my employer for their encouragement and accommodation in allowing me to release this work publicly.



dgkf/scriptgloss documentation built on June 8, 2019, 8:43 p.m.