Hi guys. I got a taste for functional programming in shiny, but given my level of expertise I would like to get a bit of feedback on whether or not it is valid way of doing stuff. I wrote a simple function which returns a function that can be used in collapsing time series using tsibble. When plotting longer time series, it is useful to have a function which automatically collapses the data when a longer time series is plotted but provides uncollapsed time series for shorter ones.

```
library(plotly)
library(dplyr)
library(shiny)
library(shinyWidgets)
library(lubridate)
library(tsibble)
ui = fluidPage(
dateRangeInput("dr", label = "f", start = today() - 60, end = today()),
plotlyOutput("plot")
)
server = function(input, output, session){
output$plot = renderPlotly({
get_fct =
function(x){
if(x <= 30){
function(y){
ymd(y)
}
}else if(x > 30){
function(y){
yearmonth(y)
}
}
}
collapsing_function = get_fct(x = as.numeric(input$dr[2] - input$dr[1]))
df = tibble(date = seq(input$dr[1], input$dr[2], by = "days"), value = rnorm(as.numeric(input$dr[2] - input$dr[1]) + 1), key = 1:(as.numeric(input$dr[2] - input$dr[1]) + 1)) #dateRangeInput generates data - think of it as if filtering SQL database using input$dr
df %>%
as_tsibble(index = date, key = key) %>%
index_by(Period = collapsing_function(date)) %>%
summarise(value = sum(value)) %>%
plot_ly(x = ~Period, y = ~value, type = "bar")
})
}
shinyApp(ui, server)
```