Interactive web apps with Shiny
RaukR 2024 • Advanced R for Bioinformatics
21-Jun-2024
Language of the web: HTML, CSS & Javascript
Hypertext Markup Language (HTML)
- All UI code is translated to HTML
- Inspect with browser
<div class="col-sm-4">
<form class="well" role="complementary">
<span class="help-block">Sidebar Panel</span>
</form>
</div>
Reactivity
- Code doesn’t always run line-by-line (Non-linear execution)
- Code executes when dependencies change
Reactive context
Reactive values cannot be accessed outside a reactive context
Reactive dynamics
#| standalone: true
#| components: [editor, viewer]
shinyApp(
ui=fluidPage(
numericInput("num_input",label="Observations",value=50),
textOutput("text_output")),
server=function(input,output) {
output$text_output <- renderText({
mean(rnorm(input$num_input))
})
})
Reactive dynamics
#| standalone: true
#| components: [editor, viewer]
shinyApp(
ui=fluidPage(
numericInput("num_input",label="Observations",value=50),
textOutput("text_output1"), textOutput("text_output2")),
server=function(input,output) {
output$text_output1 <- renderText({
mean(rnorm(input$num_input))
})
output$text_output2 <- renderText({
mean(rnorm(input$num_input))
})
})
Reactive dynamics
#| standalone: true
#| components: [editor, viewer]
shinyApp(
ui=fluidPage(
numericInput("num_input",label="Observations",value=50),
textOutput("text_output1"), textOutput("text_output2")),
server=function(input,output) {
rand <- function(x) rnorm(x)
output$text_output1 <- renderText({
mean(rand(input$num_input))
})
output$text_output2 <- renderText({
mean(rand(input$num_input))
})
})Reactive dynamics
#| standalone: true
#| components: [editor, viewer]
shinyApp(
ui=fluidPage(
numericInput("num_input",label="Observations",value=50),
textOutput("text_output1"), textOutput("text_output2")
),
server=function(input,output) {
rand <- reactive({ rnorm(input$num_input) })
output$text_output1 <- renderText({
mean(rand())
})
output$text_output2 <- renderText({
mean(rand())
})
})
Error validation
- Shiny returns an error with missing or incorrect values
Error validation
- Errors can be handled in a controlled manner
validate()can be used to check inputvalidate()usingneed()
shinyApp(
ui=fluidPage(
selectInput("data_input",label="Select data",
choices=c("","unknown","mtcars","faithful","iris")),
tableOutput("table_output")
),
server=function(input, output) {
getdata <- reactive({
validate(need(try(input$data_input),"Please select a data set"))
get(input$data_input,'package:datasets')
})
output$table_output <- renderTable({head(getdata())})
})
Error validation
validate()using custom function
valfn <- function(x) if(is.null(x) | is.na(x) | x=="") return("Input data is incorrect.")
shinyApp(
ui=fluidPage(
selectInput("data_input",label="Select data",
choices=c("","unknown","mtcars","faithful","iris")),
tableOutput("table_output")
),
server=function(input,output) {
getdata <- reactive({
validate(valfn(try(input$data_input)))
get(input$data_input,'package:datasets')
})
output$table_output <- renderTable({head(getdata())})
})
shiny::req()checks input variable and silently stops execution
Debugging & Optimizing
print()statementsInterrupt execution and inspect environment
browser()Visualize relationships using reactlog
![]()
Debugging & Optimizing
Assess compute and RAM usage using profvis
![]()
Simple profiling using shiny.tictoc
Learning & community
Documentation
Books
Conferences
Blogs & Podcasts
Thank you! Questions?
_
platform x86_64-pc-linux-gnu
os linux-gnu
major 4
minor 3.2 2024 • SciLifeLab • NBIS • RaukR
