app.R

library(shiny)
library(bslib)
library(leaflet)
library(plotly)
library(dplyr)
library(ghcn)  # Your custom package
library(DT)
library(httr)
library(shinyalert)
library(promises)
library(future)
library(shinyjs)

# Enable parallel processing
future::plan(multisession)

ui <- bslib::page_sidebar(
  title = shiny::tags$span(style = "color: white; font-weight: bold;", "iSci GHCN-Daily Data Portal"),
  theme = bslib::bs_theme(
    bg = "#FFFFFF",
    fg = "#000000",
    primary = "#3398cb"
  ),
  shinyjs::useShinyjs(),
  # Add a container for the loading message
  div(id = "loading-content",
      style = "position: fixed; top: 50%; left: 50%; transform: translate(-50%, -50%); 
                 z-index: 9999; background-color: #f0f0f0; padding: 20px; 
                 border-radius: 5px; text-align: center;
                 border: 2px solid #4a4a4a;",
      h3("Checking NOAA CDO API Status. Please Wait...", 
         style = "color: #3398cb; margin-bottom: 15px;"),
      tags$img(src = "isci_logo.png", height = "50px", style = "margin-bottom: 15px;"),
      style = "display: none;"
  ),
  sidebar = bslib::sidebar(
    width = 300,
    
    # Sidebar inputs
    shiny::numericInput("radius", "Search Radius (km)", value = 10, min = 1, max = 500),
    shiny::numericInput("limit", "Max Number of Stations", value = 10, min = 1, max = 50),
    shiny::actionButton("search_stations", "Search Stations", class = "btn-custom"),
    shiny::br(),
    shiny::br(),
    shiny::dateRangeInput("date_range", "Date Range", 
                          start = Sys.Date() - 365, end = Sys.Date()),
    shiny::selectizeInput("variables", "Select Variables", 
                          choices = NULL,
                          multiple = TRUE,
                          options = list(maxItems = 5)),
    shiny::actionButton("fetch_data", "Fetch Data", class = "btn-custom")
  ),
  
  # Main content
  bslib::layout_column_wrap(
    width = 1,
    heights_equal = "row",
    # Map and Station Info Card
    bslib::layout_columns(
      col_widths = c(8, 4),
      bslib::card(
        full_screen = TRUE,
        bslib::card_header("Station Map: Click to Establish Search Radius"),
        leaflet::leafletOutput("station_map", height = 800)
      ),
      bslib::card(
        full_screen = TRUE,
        bslib::card_header("Station Information"),
        shiny::uiOutput("station_info")
      )
    ),
    # Plots and Data
    bslib::navset_card_tab(
      full_screen = TRUE,
      bslib::nav_panel("Plots", plotly::plotlyOutput("data_plot", height = 800)),
      bslib::nav_panel("Data", DT::dataTableOutput("data_table"))
    )
  ),
  
  # Custom CSS
  tags$style(HTML("
    .bslib-page-sidebar > .navbar { background-color: #3398cb !important; }
    .bslib-sidebar-layout > .sidebar { background-color: #3398cb !important; }
    .bslib-sidebar-layout > .sidebar { color: white !important; }
    .bslib-sidebar-layout > .sidebar .form-control { color: black !important; background-color: white !important; }
    .bslib-sidebar-layout > .sidebar .selectize-input { color: black !important; background-color: white !important; }
    .btn-custom { 
      background-color: #ed5535 !important; 
      border-color: #ed5535 !important; 
      color: white !important;
    }
    .btn-custom:hover {
      background-color: #d64a2e !important;
      border-color: #d64a2e !important;
    }
  "))
)

server <- function(input, output, session) {
  # Show loading message
  shinyjs::show("loading-content")
  
  # Perform API check
  tryCatch({
    base_url <- "https://www.ncei.noaa.gov/cdo-web/api/v2"
    response <- httr::GET(
      url = file.path(base_url, "stations"),
      query = list(
        datasetid = "GHCND",
        limit = 5
      ),
      httr::add_headers("token" = "jNgLeoBbPesCsXNbybBUwfTKPuXktBGS"),
      httr::timeout(20)
    )
    
    if (httr::status_code(response) != 200) {
      stop(paste("API request failed with status code:", httr::status_code(response)))
    }
  }, error = function(e) {
    error_message <- if (inherits(e, "timeout")) {
      "The NOAA CDO server was unresponsive and timed out after 20 seconds."
    } else {
      paste("The NOAA CDO server failed with the following error:", e$message)
    }
    shinyalert::shinyalert(
      title = "NOAA CDO API Error",
      text = error_message,
      type = "error"
    )
  }, finally = {
    # Hide loading message when done (success or failure)
    shinyjs::hide("loading-content")
  })
  
  # Reactive values for storing the clicked point and selected station
  clicked_point <- shiny::reactiveVal(NULL)
  selected_station <- shiny::reactiveVal(NULL)
  
  # Initialize the map
  output$station_map <- leaflet::renderLeaflet({
    leaflet::leaflet() %>%
      leaflet::addTiles() %>%
      leaflet::addAwesomeMarkers(lng = -98.5795, lat = 39.8283, 
                                 icon = leaflet::makeAwesomeIcon(icon = 'flag', markerColor = 'red', iconColor = 'white'),
                                 label = "Click to select a location") %>%
      leaflet::setView(lng = -98.5795, lat = 39.8283, zoom = 4)  # Center on US
  })
  
  # Update clicked point when map is clicked
  shiny::observeEvent(input$station_map_click, {
    click <- input$station_map_click
    clicked_point(c(lat = click$lat, lng = click$lng))
    
    leaflet::leafletProxy("station_map") %>%
      leaflet::clearMarkers() %>%
      leaflet::addAwesomeMarkers(lng = click$lng, lat = click$lat, 
                                 icon = leaflet::makeAwesomeIcon(icon = 'flag', markerColor = 'red', iconColor = 'white'),
                                 label = "Selected location")
  })
  
  # Reactive expression for getting stations
  stations <- shiny::eventReactive(input$search_stations, {
    point <- clicked_point()
    
    shiny::req(point, input$radius, input$limit)
    
    result <- ghcn::get_ghcn_daily_stations(
      lat = point["lat"],
      lon = point["lng"],
      radius = input$radius,
      limit = input$limit,
      token = "jNgLeoBbPesCsXNbybBUwfTKPuXktBGS"
    )
    
    # Sort stations by distance
    result %>%
      dplyr::mutate(distance = sqrt((latitude - point["lat"])^2 + (longitude - point["lng"])^2)) %>%
      dplyr::arrange(distance)
  })
  
  # Update map with station markers
  shiny::observeEvent(stations(), {
    req(stations())
    leaflet::leafletProxy("station_map") %>%
      leaflet::clearMarkers() %>%
      leaflet::addAwesomeMarkers(data = stations(),
                                 ~longitude, ~latitude,
                                 icon = leaflet::makeAwesomeIcon(icon = 'info-sign', markerColor = 'blue', iconColor = 'white'),
                                 popup = ~paste(name, "<br>", id),
                                 label = ~name,
                                 layerId = ~id) %>%
      leaflet::addAwesomeMarkers(lng = clicked_point()["lng"], lat = clicked_point()["lat"], 
                                 icon = leaflet::makeAwesomeIcon(icon = 'flag', markerColor = 'red', iconColor = 'white'),
                                 label = "Selected location")
  })
  
  # Update selected station when a marker is clicked
  shiny::observeEvent(input$station_map_marker_click, {
    click <- input$station_map_marker_click
    if (!is.null(click$id)) {
      selected_station(stations() %>% dplyr::filter(id == click$id))
      # Update the variables selectInput
      available_vars <- unlist(strsplit(selected_station()$available_vars, ", "))
      shiny::updateSelectizeInput(session, "variables", choices = available_vars, selected = character(0))
    }
  })
  
  # Render station information card
  output$station_info <- shiny::renderUI({
    station <- selected_station()
    if (is.null(station)) {
      return(shiny::p("'Search Stations' after selecting a map location. Then click on a station marker to view its information."))
    }
    
    # Split available variables
    available_vars <- strsplit(station$available_vars, ", ")[[1]]
    
    shiny::tagList(
      shiny::h5(station$name),
      shiny::tags$div(
        style = "line-height: 1.2; margin-bottom: 5px;",
        shiny::p(shiny::strong("ID: "), station$id),
        shiny::p(shiny::strong("Date Range: "), paste(station$mindate, "to", station$maxdate)),
        shiny::p(shiny::strong("Distance: "), sprintf("%.2f km", station$distance * 111)),  # Approximate conversion to km
        shiny::p(shiny::strong("Latitude: "), station$latitude),
        shiny::p(shiny::strong("Longitude: "), station$longitude),
        shiny::p(shiny::strong("Elevation: "), station$elevation, " m")
      ),
      shiny::h6("Available Variables:"),
      shiny::tags$ul(
        style = "padding-left: 20px; margin-top: 2px;",
        lapply(strsplit(station$available_vars, ", ")[[1]], function(var) {
          description <- ghcn::ghcn_daily_datatypes$description[ghcn::ghcn_daily_datatypes$datatype == var]
          shiny::tags$li(
            style = "margin-bottom: 2px;",
            shiny::strong(var), ": ", description
          )
        })
      )
    )
  })
  
  # Reactive expression for getting data
  station_data <- shiny::eventReactive(input$fetch_data, {
    shiny::req(selected_station(), input$variables, input$date_range)
    station <- selected_station()
    
    # Check if selected date range is within the station's data range
    station_start <- as.Date(station$mindate)
    station_end <- as.Date(station$maxdate)
    selected_start <- input$date_range[1]
    selected_end <- input$date_range[2]
    
    if (selected_start < station_start || selected_end > station_end) {
      shiny::showModal(shiny::modalDialog(
        title = "Date Range Error",
        "The selected date range is outside the available data range for this station. 
                Please adjust your date selection or choose a different station.",
        easyClose = TRUE,
        footer = NULL
      ))
      return(NULL)
    }
    
    ghcn::get_ghcn_daily_data(station_id = station$id,
                              start_date = selected_start,
                              end_date = selected_end,
                              token = "jNgLeoBbPesCsXNbybBUwfTKPuXktBGS",
                              datatype = input$variables)
  })
  
  # Render the data plot
  output$data_plot <- plotly::renderPlotly({
    shiny::req(station_data(), selected_station())
    p <- ghcn::visualize_ghcn_daily_data(station_data(), selected_station()$name)
    plotly::ggplotly(p)
  })
  
  # Render the data table
  output$data_table <- DT::renderDataTable({
    shiny::req(station_data())
    DT::datatable(station_data(), 
                  options = list(pageLength = 25,
                                 order = list(list(0, 'desc')),  # Sort by first column (date) in descending order
                                 columnDefs = list(list(
                                   targets = "_all",
                                   render = DT::JS(
                                     "function(data, type, row, meta) {",
                                     "return type === 'display' && data != null && data !== ''",
                                     "  ? '<span style=\"font-size:0.7em;\">' + data + '</span>'",
                                     "  : data;",
                                     "}"
                                   )
                                 ))))
  })
}

shiny::shinyApp(ui = ui, server = server)