Resetting plotly event data in quarto dashboard

I'm using a quarto dashboard with some shiny for the interactivity functionality.
At the moment I have a side bar with some filters that loads the data and after the data is loaded a scatter plot (via plotly) is created. Using the plotly_selected event data I'm able to filter other datatables and plots in the dashboard.

The issue is that once I go back to the side bar, change the filters and load the data, if it's a subset of the first data - which was selected before - the selection doesn't reset.
An example, if in the first data I have 20 data points and select 3, after I use the filters and 2 data points coincide with the ones before, now I have 2 selected instead of just showing all data points from that subset.

I tried solving this using this solution: Reseting click-events - #2 by helpmeplot - Plotly R - Plotly Community Forum
But it seems that the useShinyjs() only works inside a fluidPage() which I actually don't use since in quarto dashboards we can just use the R cards for the UI part (as shown here: Quarto – Dashboards with Shiny for R)

My question is: Is there a way to make this useShinyjs() work in a quarto dashboard or is there some other way to reset the event data that might work for the dashboard?

This is an example of my code at the moment:

---
title: "Test"
format: dashboard
server: shiny
---

```{r}
#| context: setup
library(ggplot2)
library(shinyjs)
dataset <- diamonds

{.sidebar}

numericRangeInput(
  "carat_filter", 
  "carat",
  c(0, 1),
  separator = "to", 
  min = 0, 
  max = 1)

div(style="float: right", actionButton("sidebar_ok", "OK"))

Plot

useShinyjs()
extendShinyjs(text = "shinyjs.resetClick = function() { Shiny.onInputChange('plotly_selected-diamonds_plot', 'null'); }")
plotlyOutput('plot')
tableOutput('data')
#| context: server

dataset_filtered <- reactiveVal(NULL)
observeEvent(input$sidebar_ok, {
  data <- dataset %>%
  filter(carat >= carat_filter[1] & carat<= carat_filter[2])
  if (!is.null(filtered_df)) {
    dataset_filtered(data)
    js$resetClick()
  }
})
 
output$plot <- renderPlotly({
  
  p <- ggplot(
    dataset_filtered(), 
    aes_string(x=input$x, y=input$y)) + geom_point()
  
  return(ggplotly(p, source = "diamonds_plot"))
  
})

observe({

  selected_data <- event_data("plotly_selected", source = "diamonds_plot")

  datapoints_selected <- dataset_filtered()

  if (!is.null(selected_data)) {
    data <- 
      dataset_filtered()[dataset_filtered()[, input$x_variable] %in% selected_data$x &
                         dataset_filtered()[, input$y_variable] %in% selected_data$y,]
    
    datapoints_selected <- if(nrow(data) > 0) data else dataset_filtered()
  }


  output$data <- renderTable({
    datapoints_selected
  })
})