R Shiny: Update with multiple sources of input - r

Question

I am creating a shiny app where I would like to offer users two ways to give two input values: Either directly via two input sliders, or via two data tables on which the app performs some computation with two values as output.
The output values depend on each other, and should sum to one.
I get both these methods working separately: Slider 1 is updated when the value of slider 2 is changed and vice versa, and I managed to compute the output values using the data (See MWE).
Now I would like to combine them such that each slider is updated when a.) the other slider is updated manually; or b.) the data table is changed.
I would like to use these updated slider values for computations in another function.
Any suggestions to change the value of updateSliderInput() when two sources of input are available?
MWE:
library(shiny)
library(rhandsontable)
#### Initialization ####
# Initial data
datA <- datB <- data.frame(success=c(1,1,2), failure=c(1,1,2), total=c(2,2,4))
colnames(datA) <- colnames(datB) <- c(expression(x[T*","*2]==1), expression(x[T*","*2]==0)," " )
rownames(datA) <- rownames(datB) <- c(expression(x[T*","*1]==1), expression(x[T*","*1]==0)," " )
# Dummy function to perform some computation on input data
ComputeWeights <- function(datA,datB){
c(sum(datA)/sum(datA,datB),sum(datB)/sum(datA,datB))
}
#### Server ####
server <- shinyServer(function(input, output, session) {
# Allow users to provide slider input via data table
prevA <- reactive({datA})
prevB <- reactive({datB})
changeA <- reactive({
if(is.null(input$hotA)){return(prevA())}
else{
datA <- as.data.frame(hot_to_r(input$hotA))
datA[c(1,2),3] <- rowSums(datA[c(1,2),c(1,2)])
datA[3,c(1,2)] <- colSums(datA[c(1,2),c(1,2)])
datA[3,3] <- sum(datA[c(1,2),3])
datA
}
})
changeB <- reactive({
if(is.null(input$hotB)){return(prevB())}
else{
datB <- as.data.frame(hot_to_r(input$hotB))
datB[c(1,2),3] <- rowSums(datB[c(1,2),c(1,2)])
datB[3,c(1,2)] <- colSums(datB[c(1,2),c(1,2)])
datB[3,3] <- sum(datB[c(1,2),3])
datB
}
})
output$hotA <- renderRHandsontable({rhandsontable(changeA(), width=375,
rowHeaders=c(expression(x[T*","*2]==1), expression(x[T*","*2]==0)," " ),
colHeaders=c(expression(x[T*","*1]==1), expression(x[T*","*1]==0)," " )) %>%
hot_table(rowHeaderWidth=50) %>%
hot_cols(colWidths = c(50,50,50), format="0", allowInvalid=FALSE)%>%
hot_col(3, readOnly=TRUE)%>%
hot_row(3, readOnly=TRUE)})
output$hotB <- renderRHandsontable({rhandsontable(changeB(), width=375,
rowHeaders=c(expression(x[T*","*2]==1), expression(x[T*","*2]==0)," " ),
colHeaders=c(expression(x[T*","*1]==1), expression(x[T*","*1]==0)," " )) %>%
hot_table(rowHeaderWidth=50) %>%
hot_cols(colWidths = c(50,50,50), format="0", allowInvalid=FALSE)%>%
hot_col(3, readOnly=TRUE)%>%
hot_row(3, readOnly=TRUE)})
# Slider updated with input from slider
observe({
updateSliderInput(session, "w2", value = 1-input$w1,
min = 0, max = 1, step=0.01)
})
observe({
updateSliderInput(session, "w1", value = 1-input$w2,
min = 0, max = 1, step=0.01)
})
output$weights <- renderPrint({
ComputeWeights(changeA(),changeB())
})
})
#### UI ####
ui <- shinyUI(fluidPage(
# First input source: sliders
sidebarLayout(sliderInput("w1", "W1",
min = 0, max = 1, value = 0.5, step=0.01),
sliderInput("w2", "W2",
min = 0, max = 1, value = 0.5, step=0.01)),
mainPanel(
# Second input source: data
rHandsontableOutput("hotA", width="100%"),
rHandsontableOutput("hotB", width="100%"),
textOutput("weights")))
)
shinyApp(ui=ui, server=server)

Answer 1

Would it be possible to just add an observe for changeA and changeB which are reactive?
observe({
wts <- ComputeWeights(changeA(),changeB())
updateSliderInput(session, "w1", value = wts[1], min = 0, max = 1, step=0.01)
updateSliderInput(session, "w2", value = wts[2], min = 0, max = 1, step=0.01)
})
Then you could set the slider input to each value of your returned vector from ComputeWeights.
Let me know if this is what you had in mind.