shiny object reactivity to several inputs (buttons) - r

I have a shinyapp, in which a main object should be updated depending on the change other objects/inputs (buttons that perform other operations, whose result is not easily tracked, e.g. online data). That's why I had to use the input of buttons. Is there a way to update the main object without having to re-write the code for every button? In my example, I had to use observeEvent two times:
library(datasets)
library(shiny)
ui<-fluidPage(
titlePanel("Telephones by region"),
sidebarLayout(
sidebarPanel(
selectInput("region", "Region:",
choices=colnames(WorldPhones)),
helpText("Data from AT&T (1961) The World's Telephones."),
actionButton("submit",
label = "submit"), # this also has other functions
actionButton("change",
label = "change") # this also has other functions
),
mainPanel(
plotOutput("phonePlot")
)
)
)
server<-function(input, output) {
data<-reactiveValues()
observeEvent(input$submit,{
data$data<-WorldPhones[,input$region]
})
observeEvent(input$change,{
data$data<-WorldPhones[,input$region]
})
output$phonePlot <- renderPlot({
if(!is.null(data$data))
barplot(data$data*1000,
ylab="Number of Telephones",
xlab="Year")
})
}
shinyApp(ui = ui, server = server)

You simply make an expression with both buttons, for example using c():
observeEvent(c(input$submit,input$change),{
data$data<-WorldPhones[,input$region]
})

Related

R shiny code to get output which takes input from ui

I am trying to write a script in shiny, which has two inputs and stores the inputs in two different variables and runs a code using these input variables.But i am getting an error which says :Error in .getReactiveEnvironment()$currentContext() :
Operation not allowed without an active reactive context. (You tried to do something that can only be done from inside a reactive expression or observer.)
The following is my ui code:
ui <- fluidPage(
titlePanel("Network Model"),
sidebarLayout(
sidebarPanel(
selectInput(inputId = "origin",
label = "Origin:",
choices = milk_runs$Origin),
selectInput(inputId = "destination",
label = "Destination:",
choices = milk_runs$Dest),
actionButton("go", "")
),
mainPanel(
tableOutput(
"view"))
)
)
server code :
server<- function(input, output){
origin <- input$origin
destination <- input$destination
observeEvent(input$go,source("nr9.R"))
output$summary <- renderPrint({
#dataset <- datasetInput()
summary(Tnetwork)
})
Can you please tell me how to get correct results.
I think (it would help if you provided a fully reproducible example) that the error is occurring because you are trying to run input$origin without reactive(). The input$origin will not invalidate and update based on user input unless put inside reactive. Based on the example you provided:
library(shiny)
ui <- fluidPage(
titlePanel("Network Model"),
sidebarLayout(
sidebarPanel(
selectInput(inputId = "origin", label = "Origin:", choices = c("A","B","C","D","E","F")),
selectInput(inputId = "destination", label = "Destination:", choices = c("A","B","C","D","E","F")),
actionButton("go", "GO")
),
mainPanel( tableOutput( "view"))
)
)
server<- function(input, output){
origin <- reactive(input$origin)
destination<-reactive(input$destination)
observeEvent(input$go,{
cat(origin(),'nextword',destination(),sep="-")
})
output$view <- renderTable({data.frame(origin=origin(),destination=destination())})
}
shinyApp(ui, server)
should print 'origin-nextword-destination' to the console when 'go' is activated, and the table should update. I changed a few bits in your example because it was not reproducible but hopefully it helps.

How to stop restarting list in selectize input after click in Shiny?

When I filter a list of states in Shiny for: "New" I can choose only one state. After that the list is restarting and I have to put: "New" again in order to find a state contains "New" in name. I would like to filter states and choose more states at one time.
Below I added a picture and a code which present my goal.
Picture
library(shiny)
ui <- fluidPage(
sidebarLayout(
sidebarPanel(
selectizeInput(
'e2', '2. Multi-select', choices = state.name, multiple = TRUE
)
),
mainPanel(
verbatimTextOutput('ex_out')
)
)
)
server <- function(input, output) {
output$ex_out <- renderPrint({
sapply(sprintf('e%d', 2), function(id) {
input[[id]]
})
})
}
shinyApp(ui = ui, server = server)

Shiny Application actionButton click on page load

I am building a Shiny application using the navbarPage() type. I have three tabs - the initial tab has a textInput() box that has default text defined. The mainPanel() of that page has a histogram and a table. On page load those update and reflect the proper information when the application is launched based on that default text.
The second tab is supposed to present a wordcloud based on that default text. When I switch over to that tab there is an error - if I go back to the first tab and enter new text and hit the actionButton - the wordcloud will update, but it won't do so until I perform that action.
Is there a way to have the actionButton() or some sort of submit happen when the page loads so the tab with the wordcloud can update? Or maybe I just need to make a variable global or something. I'm not sure. I've spent quite a bit of time on this and have hit a wall. Any help would be greatly appreciated.
Code for the UI:
tabPanel("Word Cloud Diagram",
fluidRow(
sidebarPanel(
width = 3,
h5("The sentence input:"),
wellPanel(span(h5(textOutput(
'sent'
)), style = "color:red")),
sliderInput(
"maxWC",
h5("Maximum Number of Words:"),
min = 10,
max = 100,
value = 50
),
br(),
#actionButton("update", "Update Word Cloud"),
hr(),
helpText(h5("Help Instruction:")),
helpText(
"Please have a try to make the prediction by using
the dashboard on right side. Specifically, you can:"
),
helpText("1. Type your sentence in the text field", style =
"color:#428ee8"),
helpText(
"2. The value will be passed to the model while you are typing.",
style = "color:#428ee8"
),
helpText("3. Obtain the instant predictions below.", style =
"color:#428ee8"),
hr(),
helpText(h5("Note:")),
helpText(
"The App will be initialized at the first load.
After",
code("100% loading"),
", you will see the prediction
for the default sentence example \"Nice to meet you\"
on the right side."
)
),
mainPanel(
h3("Word Cloud Diagram"),
hr(),
h5(
"A",
code("word cloud"),
"or data cloud is a data display which uses font size and/
or color to indicate numerical values like frequency of words. Please click",
code("Update Word Cloud"),
"button and",
code("Slide Input"),
"in the side bar to update the plot for relevant prediction."
),
plotOutput("wordCloud"),
# wordcloud
br()
)
)),
Code for the server:
wordcloud_rep <- repeatable(wordcloud)
output$wordCloud <- renderPlot({
v <- terms()
wordcloud_rep(
v[, 2],
v[, 1],
max.words = input$maxWC,
scale = c(5, 1.5),
colors = brewer.pal(4, "Dark2")
)
})
Also, I am using a single file application "app.R" - not sure if this is useful information or not. Again, on the first tab, default text is presented on the first page load, I just want this to extend to the wordcloud on page load so the plot is shown immediately without having to enter and submit new text. Thanks!
Here is an example that should be close to what you want. The trick is to use a submitButton. The wordcloud will have a default plot based on initial input, but will change when you change the text and press the submit button.
library(shiny)
library(wordcloud)
ui <- shinyUI(fluidPage(
titlePanel("Old Faithful Geyser Data"),
sidebarLayout(
sidebarPanel(
textInput("text", "Input Text", "Random text random text random is no yes"),
submitButton("Submit")
),
mainPanel(
tabsetPanel(
tabPanel("Tab1",
plotOutput("hist"),
tableOutput("hist_table")),
tabPanel("Tab2",
plotOutput("wordcloud"))
)
)
)
))
server <- shinyServer(function(input, output) {
observe({
word_list = strsplit(input$text, " ")
word_table = as.data.frame(table(word_list))
output$hist = renderPlot({
barplot(table(word_list))
})
output$hist_table = renderTable({
word_table
})
output$wordcloud = renderPlot({
wordcloud(word_table[,1], word_table[,2])
})
})
})
shinyApp(ui = ui, server = server)
Since the use of submitButton() is generally discouraged in favour of the more versatile actionButton() (see here for function documentation), here is a version of the answer above that uses a combination of actionButton() and eventReactive() with ignoreNULL = FALSE so that the plots show up upon launching the app.
library(shiny)
library(wordcloud)
ui <- fluidPage(
sidebarLayout(
sidebarPanel(
textInput("text", "Input Text", "Random text random text random is no yes"),
actionButton("submit", "Submit")
),
mainPanel(
tabsetPanel(
tabPanel(
"Tab1",
plotOutput("hist"),
tableOutput("hist_table")
),
tabPanel(
"Tab2",
plotOutput("wordcloud")
)
)
)
)
)
server <- shinyServer(function(input, output) {
word_list <- eventReactive(input$submit,{
strsplit(input$text, " ")
},
ignoreNULL = FALSE
)
word_table <- reactive(
as.data.frame(table(word_list()))
)
output$hist <- renderPlot({
barplot(table(word_list()))
})
output$hist_table <- renderTable({
word_table()
})
output$wordcloud <- renderPlot({
wordcloud(word_table()[, 1], word_table()[, 2])
})
})
shinyApp(ui = ui, server = server)
The solution to making the action button run on the first load is a simple one. Just add an ifelse statement.
Original:
eventReactive(input$submit, ...
New:
eventReactive(ifelse(input$submit == 0, 1, input$submit), ...
Yes, it's just that easy!

Add user input column to Shiny

I am trying to gather user input given a data set. I want to insert a column where the user can determine whether they would want to own one of the cars in the mtdata set. This is completely subjective as opinions differ from person to person so I am not able to program this in. Is there a way to append an extra column that can be a checkbox or dropdown menu to identify cars that a user would "Want to own?
library(shiny)
shinyApp(ui = shinyUI(fluidPage(
titlePanel("Interesting Cars"),
sidebarLayout(
sidebarPanel(
helpText("This is a side bar")),
mainPanel(
tableOutput("view")
)
)
)),
server = function(input, output) {
output$view <- renderTable({
head(mtcars[, 1:4], n = 6)
})
})
How about this, you can use the DT library. By adding the filter option the user can define the different components one wants and see what cars come up.
library(shiny)
library(DT)
shinyApp(ui = shinyUI(fluidPage(
titlePanel("Interesting Cars"),
sidebarLayout(
sidebarPanel(
helpText("This is a side bar")),
mainPanel(
DT::dataTableOutput("view")
)
)
)),
server = function(input, output) {
output$view <- DT::renderDataTable({
datatable(mtcars,
filter = "top"
)
})
})
Edit
If it truly is so important to add another column indicating if it is 'interesting' there will be significantly more code to written if you intend to have users assign it on different conditions. Here is an example with just the mpg. The fundamental idea here is that you assign your data to the reactiveValues function. It can then be modified as you like. This can obviously be improved upon more (as it will continue to add columns) but it demonstrates the concept.
shinyApp(ui = shinyUI(fluidPage(
titlePanel("Interesting Cars"),
sidebarLayout(
sidebarPanel(
helpText("This is a side bar"),
uiOutput("mpg"),
actionButton("add_label", "Mark Interesting")
),
mainPanel(
DT::dataTableOutput("view")
)
)
)),
server = function(input, output) {
values <- reactiveValues(
mydata = mtcars
)
output$mpg <- renderUI({
numericInput("mpg_input", "MPG Cutoff?",
value = 15
)
})
output$view <- DT::renderDataTable({
datatable(values$mydata
)
})
observeEvent(input$add_label, {
validate(
need(!is.null(input$mpg_input), "need mpg value")
)
values$mydata <- data.frame(values$mydata,
Interesting_Flag =
ifelse(values$mydata$mpg > input$mpg_input,
"Interesting",
"Not Interesting"))
})
})

'Reset inputs' button in shiny app

I would like to implement a 'Reset inputs' button in my shiny app.
Here is an example with just two inputs where I'm using the update functions to set the values back to the default values:
library(shiny)
runApp(list(
ui = pageWithSidebar(
headerPanel("'Reset inputs' button example"),
sidebarPanel(
numericInput("mynumber", "Enter a number", 20),
textInput("mytext", "Enter a text", "test"),
tags$hr(),
actionButton("reset_input", "Reset inputs")
),
mainPanel(
h4("Summary"),
verbatimTextOutput("summary")
)
),
server = function(input, output, session) {
output$summary <- renderText({
return(paste(input$mytext, input$mynumber))
})
observe({
input$reset_input
updateNumericInput(session, "mynumber", value = 20)
updateTextInput(session, "mytext", value = "test")
})
}
))
What I would like to know is if there is also a function that sets back everything to default? That would be useful in case of multiple inputs.
Additionally, I'm not sure if my use of the observe function in order to detect when the action button was hit is the 'proper way' of handling the action buttons?
First of all, your use of the observer is correct, but there is another way that's slightly nicer. Instead of
observe({
input$reset_input
updateNumericInput(session, "mynumber", value = 20)
updateTextInput(session, "mytext", value = "test")
})
You can change it to
observeEvent(input$reset_input, {
updateNumericInput(session, "mynumber", value = 20)
updateTextInput(session, "mytext", value = "test")
})
Also note that you don't need to explicitly "return" from a renderText function, the last statement will automatically be used.
Regarding the main question: Matthew's solution is great, but there's also a way to achieve what you want without having to move all your UI into the server. I think it's better practice to keep your UI in the UI file just because separation of structure and logic is generally a good idea.
Full disclaimer: my solution involves using a package that I wrote. My package shinyjs has a reset function that allows you to reset an input or an HTML section back to its original value. Here is how to tweak your original code to your desired behaviour in a way that will scale to any number of inputs without having to add any code. All I had to do is add a call to useShinyjs() in the UI, add an "id" attribute to the form, and call reset(id) on the form.
library(shiny)
runApp(list(
ui = pageWithSidebar(
headerPanel("'Reset inputs' button example"),
sidebarPanel(
shinyjs::useShinyjs(),
id = "side-panel",
numericInput("mynumber", "Enter a number", 20),
textInput("mytext", "Enter a text", "test"),
tags$hr(),
actionButton("reset_input", "Reset inputs")
),
mainPanel(
h4("Summary"),
verbatimTextOutput("summary")
)
),
server = function(input, output, session) {
output$summary <- renderText({
return(paste(input$mytext, input$mynumber))
})
observeEvent(input$reset_input, {
shinyjs::reset("side-panel")
})
}
))
There isn't such a function in shiny, however, here's a way to accomplish this without having to essentially define your inputs twice. The trick is to use uiOutput and wrap the inputs you want to reset in a div whose id changes to something new each time the reset button is pressed.
library(shiny)
runApp(list(
ui = pageWithSidebar(
headerPanel("'Reset inputs' button example"),
sidebarPanel(
uiOutput('resetable_input'),
tags$hr(),
actionButton("reset_input", "Reset inputs")
),
mainPanel(
h4("Summary"),
verbatimTextOutput("summary")
)
),
server = function(input, output, session) {
output$summary <- renderText({
return(paste(input$mytext, input$mynumber))
})
output$resetable_input <- renderUI({
times <- input$reset_input
div(id=letters[(times %% length(letters)) + 1],
numericInput("mynumber", "Enter a number", 20),
textInput("mytext", "Enter a text", "test"))
})
}
))
Here is yet another option that works for either static or dynamic inputs, and doesn't involve re-rendering inputs entirely.
It uses:
reactiveValuesToList to get all initial input values, and (optionally) any dynamic input values that get initialized afterward.
session$sendInputMessage to update values for generic inputs. The updateXyzInput functions call this under the hood like session$sendInputMessage(inputId, list(value = x, ...).
Every Shiny input uses value for its input message, and almost all will update with their input value as-is. Only a two inputs I've found need special casing - checkboxGroupInput to not send NULL when nothing is checked, and dateRangeInput to convert its c(start, end) to a list(start = start, end = end).
It may not be a good idea to blindly reset ALL inputs (even tabs will be reset), but this can easily be adapted to reset a filtered set of inputs.
library(shiny)
ui <- pageWithSidebar(
headerPanel("'Reset inputs' button example"),
sidebarPanel(
numericInput("mynumber", "Enter a number", 20),
textInput("mytext", "Enter text", "test"),
textAreaInput("mytextarea", "Enter text", "test"),
passwordInput("mypassword", "Enter a password", "password"),
checkboxInput("mycheckbox", "Check"),
checkboxGroupInput("mycheckboxgroup", "Choose a number", choices = c(1, 2, 3)),
radioButtons("myradio", "Select a number", c(1, 2, 3)),
sliderInput("myslider", "Select a number", 1, 5, c(1,2)),
uiOutput("myselUI"),
uiOutput("mydateUI"),
tags$hr(),
actionButton("reset_input", "Reset inputs")
),
mainPanel(
h4("Summary"),
verbatimTextOutput("summary")
)
)
server <- function(input, output, session) {
initialInputs <- isolate(reactiveValuesToList(input))
observe({
# OPTIONAL - save initial values of dynamic inputs
inputValues <- reactiveValuesToList(input)
initialInputs <<- utils::modifyList(inputValues, initialInputs)
})
observeEvent(input$reset_input, {
for (id in names(initialInputs)) {
value <- initialInputs[[id]]
# For empty checkboxGroupInputs
if (is.null(value)) value <- ""
session$sendInputMessage(id, list(value = value))
}
})
output$myselUI <- renderUI({
selectInput("mysel", "Select a number", c(1, 2, 3))
})
output$mydateUI <- renderUI({
dateInput("mydate", "Enter a date")
})
output$summary <- renderText({
return(paste(input$mytext, input$mynumber))
})
}
shinyApp(ui, server)
You can also create a reset button by assigning NULL to your reactive values object.
See this RStudio Shiny article on Using Action Buttons: http://shiny.rstudio.com/articles/action-buttons.html. Specifically, read the sections titled Pattern 4 - Reset buttons and Pattern 5 - Reset on tab change. Examples (including code) are provided in the article.
The article provides solutions that don't require additional packages if that's a concern.

Resources