use a function inside shiny server with options being reactive values - r

Question

Sorry if my question is a bit silly but I can't find a way of making to work a custom function that use reactiveValues as options.
I created several functions to do some "heavy processing" that I have put in global.R. These functions are something like this
estimateDEG <- function(variables = NULL, design = NULL, ...){
# do some processing for example
design <- model.matrix(variables$group[,1]
d <- estimateDisp(d, design))
suppressMessages(fit <- glmQLFit(d, design))
suppressMessages(out <- glmQLFTest(fit, coef = 2))
p <- out$table$PValue
p[is.na(p)] <- 1
variables$stat$p.value <<- p
variables$stat$rank <<- rank(p)
variables$stat$q.value <<- p.adjust(p, method = "BH")
variables$stat$logFC <<- out$table$logFC
... # more coding
}
Then I want to use this function in server.R
server.R
shinyServer(function(input, output, session) {
variables <- reactiveValues(
group = NULL,
stat = list()
)
# for example, I have a button that when it is clicked store some
# information in `variables$group` that I want to use in the function `estimateDEG`.
observeEvent(input$buttonList, {
group <- fread(input$groupSelectViaText, header = FALSE) # a TextAreaInput from ui.R
variables$group <- lapply(unique(group$V2), function(x) {
group[group$V2 == x, ]$V1
})
names(variables$group) <- unique(group$V2)
})
# and now I would like to use the estimateDEG function in another observeEvent
observeEvent(input$runButton, {
deg <- reactive(estimateDE(variables = variables,
test.method = input$testMethod, # another input from ui.R
FDR = input$fdr # another input f
))
})
However, when I run this code the reactiveValues are not updated, i.e, after running estimateDEG the variables$stat value is NULL. Is there any way of using a function inside server.r that use reactiveValues as options and update another values inside these reactiveValues? I would expect variables$stat to be populated with p.value, rank, q.value and so on
Many thanks in advance

Answer 1

Any time you're assigning to global in a shiny app, especially with <<- you've likely gone astray.
Reactive values are functions, and any argument passed is the new value of the function. To change the value of a reactive, you cannot use assign aka <-. This will just reassign the object from a reactive object to a static object in memory. You instead use the new value as the argument for the reactive function object. See below:
If you have the reactive value x, then want to assign its value to 3, you would use:
x <- reactiveVal(0)
# then
x(3)
print(x())
## 3
In your case, you need to pass the reactive object as a function to be called inside your custom made function. If x is a reactive value equal to 0, and I have a function that updates it:
f <- function(rval, newval) {
rval(newval)
}
f(x, 3)
x()
##3
To see that in action, the below is a quick demo app.
library(shiny)
# function that takes a reactive element as a function
# and returns that element's value plus some
# you can define this in global, under the /R directory
# or even in server.
add_some <- function(r_val, some) {
x <- r_val()
x <- x + some
r_val(x)
}
ui <- fluidPage(
sidebarLayout(
sidebarPanel(
actionButton("button", "Button"),
numericInput("some", "Some", 1)
),
mainPanel(
verbatimTextOutput("console")
)
)
)
server <- function(input, output, session) {
init <- reactiveVal(0)
observeEvent(input$button, ignoreInit = TRUE, {
add_some(init, input$some)
})
output$console <- renderPrint({
init()
})
}
shiny::runApp(list(ui = ui, server = server))
Hope that helps.