radar2/server.R

server <- function(input, output, session) {
  
  hideTab(inputId = 'RadaRtabs', target = 'Data')
  
  output$disclaimerText = renderText('you must confirm understanding the disclaimer')
  
  # define data selection ---------------------------------------------------
  
  data_select <- reactive({if (input$box2.6_first != 365) {
    radar_data %>% 
      filter(specialty_shiny %in% input$specialtyInput & 
               department %in% input$departmentInput &
               Materiaal %in% input$materialInput) %>%
      filter_first_isolate(col_patient_id = "patientid", episode_days = input$box2.6_first) %>% 
      mutate(mo = as.mo(mo, Becker = TRUE))
  } else {
    radar_data_first %>% 
      filter(specialty_shiny %in% input$specialtyInput & 
               department %in% input$departmentInput &
               Materiaal %in% input$materialInput)
  }
  })
  
  # update specialties & departments --------------------------------------------------------------
  
  observe({
    x <- input$allInput
    if (x == TRUE) {
      x <- sort(unique(radar_data$specialty_shiny))
    }
    else {
      x <- character(0)
    }
    
    updateCheckboxGroupInput(
      session,
      "specialtyInput",
      label = NULL,
      choices = sort(unique(radar_data$specialty_shiny)),
      selected = x
    )
  })
  
  observe({
    x <- input$specialtyInput
    if (input$allDepartments) {
      if (!is.null(x)) {
        update_departments <-
          radar_data %>% filter(specialty_shiny %in% input$specialtyInput)
        x <- sort(unique(update_departments$department))
      }
      else {
        x <- character(0)
      }
    } else {
      x <- character(0)
    }
    updateCheckboxGroupInput(
      session,
      inputId = "departmentInput",
      label = NULL,
      choices = sort(unique(radar_data$department)),
      selected = x
    )
  })
  
  observeEvent(input$btnDisclaimerUnderstood, {
    if (input$cbDisclaimerRead == T) {
      showTab(inputId = 'RadaRtabs', target = 'Data', select = T)
    }
  })
  
  
  # sidebar hover -----------------------------------------------------------
  
  onevent("mouseenter", "sidebarCollapsed", shinyjs::removeCssClass(selector = "body", class = "sidebar-collapse"))
  onevent("mouseleave", "sidebarCollapsed", shinyjs::addCssClass(selector = "body", class = "sidebar-collapse"))
  
  # BOX top left -----------------------------------------------------------
  
  output$box2 <- renderUI({
    div(
      style = "position: relative",
      tabBox(
        id = "box2",
        width = NULL,
        height = 500,
        tabPanel(
          title = trnslt("Positive & negative cultures"),
          div(
            style = "position: absolute; left: 0.5em; bottom: 0.5em;",
            dropdown(
              radioGroupButtons(
                inputId = "box2.1_group",
                label = trnslt("Select group"),
                choiceNames = c(trnslt("All"), trnslt("Year"), trnslt("Gender"), trnslt("Department"), trnslt("Specialty"), trnslt("Specialty code"), trnslt("ICU status"), trnslt("Clinical status"), trnslt("Outward status")),
                choiceValues = c("group_all", "year", "gender", "department", "specialty", "specialism", "is_icu", "is_clinical", "is_outward"),
                selected = "group_all",
                direction = "vertical",
              ),
              size = "xs",
              icon = icon("gear", class = "opt"),
              up = TRUE
            )
          ),
          div(
            style = "position: absolute; right: 3.5em; bottom: 0.5em;",
            dropdown(
              downloadButton(outputId = "down_box_pos_neg", label = trnslt("Download plot")),
              size = "xs",
              icon = icon("download", class = "opt"),
              up = TRUE
            )
          ),
          withSpinner(
            girafeOutput("pos_neg_plot", height = 400),
            type = 4,
            color = "#d33724",
            size = 0.7
          )
        ),
        tabPanel(
          title = "Isolates (and pathogens) detected",
          div(
            style = "position: absolute; left: 0.5em; bottom: 0.5em;",
            dropdown(
              radioGroupButtons(
                inputId = "box2.3_group",
                label = trnslt("Select group"),
                choiceNames = c(trnslt("All"), trnslt("Year"), trnslt("Gender"), trnslt("Department"), trnslt("Specialty"), trnslt("Specialty code"), trnslt("ICU status"), trnslt("Clinical status"), trnslt("Outward status")),
                choiceValues = c("group_all", "year", "gender", "department", "specialty", "specialism", "is_icu", "is_clinical", "is_outward"),
                selected = "group_all",
                direction = "vertical"
              ),
              size = "xs",
              icon = icon("gear", class = "opt"),
              up = TRUE
            )
          ),
          div(
            style = "position: absolute; left: 4.6em; bottom: 0.5em;",
            dropdown(
              sliderTextInput(
                inputId = "box2.4_top",
                label = trnslt("Use slider to select by count"),
                choices = seq(0, n_distinct(radar_data$mo), 10),
                selected = c(0, 10)
              ),
              size = "xs",
              label = trnslt("Select top ..."),
              up = TRUE)
          ),
          div(
            style = "position: absolute; left: 12em; bottom: 0.5em;",
            dropdown(
              radioGroupButtons(
                inputId = "box2.3_pathogen",
                label = trnslt("Select pathogen group"),
                choiceNames = c("all", "definite", "probable", "improbable"),
                choiceValues = c(0, 1, 2, 3),
                selected = 0
              ),
              size = "xs",
              label = trnslt("Pathogens"),
              up = TRUE
            )),
          div(
            style = "position: absolute; left: 18.7em; bottom: 0.5em;",
            dropdown(
              selectInput(
                inputId = "box2.5_search",
                label = trnslt("Search and select isolates"),
                choices = mo_fullname(unique(radar_data$mo)[which(!is.na(unique(radar_data$mo)))]),
                
                multiple = TRUE
              ),
              size = "xs",
              label = trnslt("Search isolates"),
              up = TRUE
            )),
          div(
            style = "position: absolute; left: 27em; bottom: 0.5em;",
            dropdown(
              sliderTextInput(
                inputId = "box2.6_first",
                label = "Select range in days (use standard = 365 if not specially requested otherwise)",
                choices = c(15, 30, 60, 90, 365),
                selected = 365
              ),
              size = "xs",
              label = trnslt("Define first isolate guidelines"),
              up = TRUE
            )),
          div(
            style = "position: absolute; right: 3.5em; bottom: 0.5em;",
            dropdown(
              downloadButton(outputId = "down_box_patho", label = trnslt("Download plot")),
              size = "xs",
              icon = icon("download", class = "opt"),
              up = TRUE
            )
          ),
          withSpinner(
            girafeOutput("patho_plot", height = 400),
            type = 4,
            color = "#d33724",
            size = 0.7
          )
        ),
        tabPanel(
          title = trnslt("First isolates per episode"),
          div(
            style = "position: absolute; left: 0.5em; bottom: 0.5em;",
            dropdown(
              radioGroupButtons(
                inputId = "box2.2_group",
                label = trnslt("Select group"),
                choiceNames = c(trnslt("All"), trnslt("Year"), trnslt("Gender"), trnslt("Department"), trnslt("Department type"), trnslt("Specialty"), trnslt("ICU status"), trnslt("Clinical status"), trnslt("Outward status")),
                choiceValues = c("group_all", "year", "gender", "department", "type_dept", "specialism", "is_icu", "is_clinical", "is_outward"),
                selected = "group_all",
                direction = "vertical"
              ),
              size = "xs",
              icon = icon("gear", class = "opt"),
              up = TRUE
            )
          ),
          withSpinner(
            plotOutput("test_plot", height = 400),
            type = 4,
            color = "#d33724",
            size = 0.7
          )
        ),
        div(
          style = "position:absolute;right:0.5em;bottom: 0.5em;",
          conditionalPanel(
            "input.box2 == 'Positive & negative cultures'",
            actionBttn(
              inputId = "pos_neg_plus",
              icon = icon("search-plus", class = "opt"),
              style = "fill",
              color = "danger",
              size = "xs"
            )
          )
        ),
        div(
          style = "position:absolute;right:0.5em;bottom: 0.5em;",
          conditionalPanel(
            "input.box2 == 'First isolates per episode'",
            actionBttn(
              inputId = "test_plus",
              icon = icon("search-plus", class = "opt"),
              style = "fill",
              color = "danger",
              size = "xs"
            )
          )
        ),
        div(
          style = "position:absolute;right:0.5em;bottom: 0.5em;",
          conditionalPanel(
            "input.box2 == 'Isolates (and pathogens) detected'",
            actionBttn(
              inputId = "patho_plus",
              icon = icon("search-plus", class = "opt"),
              style = "fill",
              color = "danger",
              size = "xs"
            )
          )
        )
      )
    )
  })
  
  observeEvent((input$test_plus), {
    showModal(modalDialog(
      renderPlot({
        test_plot() + theme(
          text = element_text(family = "Arial"),
          axis.title = element_text(size = 20),
          text = element_text(size = 20),
          plot.title = element_text(size = 26)
        )
      }, height = 600),
      easyClose = TRUE,
      size = "l",
      footer = NULL
    ))
  })
  
  observeEvent((input$pos_neg_plus), {
    showModal(modalDialog(
      renderPlot({
        pos_neg_plot() + theme(
          axis.title = element_text(size = 20),
          text = element_text(size = 20),
          plot.title = element_text(size = 26)
        )
      }, height = 600),
      easyClose = TRUE,
      size = "l",
      footer = NULL
    ))
  })
  
  observeEvent((input$patho_plus), {
    showModal(modalDialog(
      renderPlot({
        patho_plot() + theme(
          text = element_text(family = "Arial"),
          axis.title = element_text(size = 20),
          plot.title = element_text(size = 26)
        )
      }, height = 600),
      easyClose = TRUE,
      size = "l",
      footer = NULL
    ))
  })
  
  
  # positive negative plot --------------------------------------------------
  
  pos_neg_plot <- reactive({
    
    pos_neg_plot <- radar_data %>% 
      filter(specialty_shiny %in% input$specialtyInput & 
               department %in% input$departmentInput &
               Materiaal %in% input$materialInput) %>% 
      group_by_at(input$box2.1_group) %>% 
      summarise(patients = n_distinct(patientid),
                total = n_distinct(sampleid),
                positive = n_distinct(sampleid[!is.na(mo)]),
                negative = total - positive) %>% 
      pivot_longer(cols = c("total", "positive", "negative")) %>% 
      mutate(name = factor(name, levels = c("total", "negative", "positive")))
    
    pos_neg_plot %>% 
      ggplot(aes(name, value, fill = name, tooltip = value)) +
      geom_col_interactive(colour = "black") +
      scale_fill_manual(values = c("white", "lightgrey", "darkred")) +
      labs(x = "", y = trnslt("Count"), fill = "", title = trnslt("Number of positive & negative blood culture tests")) +
      theme_minimal() +
      theme(text = element_text(family = "Arial"),
            legend.title = element_blank(),
            legend.text = element_text(margin = margin(l = 5)),
            axis.title.y = element_blank(),
            axis.text = element_text(size = 12),
            plot.title = element_text(face = "bold", size = 14),
            plot.caption = element_text(colour = "grey")) +
      {if (input$box2.1_group != "group_all") {
        facet_wrap(input$box2.1_group)
      }} +
      NULL
    
  })
  
  output$pos_neg_plot <- renderGirafe({
    
    ggiraph(ggobj = pos_neg_plot(),
            height_svg = 6,
            width_svg = 10,
            selection_type = "none")
  })
  
  output$down_box_pos_neg <- download_box("positive_negative_plot", pos_neg_plot())
  
  # pathogen distribution plot ---------------------------------------------------
  
  patho_plot <- reactive({
    
    if (!is.null(input$box2.5_search)) {
      mos <- as.mo(input$box2.5_search)
    } else {
      mos <- character(0)
    }
    mo_search <-  c("none", mos)
    
    top_10 <- data_select() %>% 
      filter(!is.na(mo)) %>% 
      {if(input$box2.3_pathogen != 0) {
        filter(., pathogen_group == input$box2.3_pathogen | mo %in% mo_search) 
      } else {
        .
      }} %>% 
      count(mo) %>% 
      mutate(rank = rank(-n, ties.method = "first")) %>% 
      filter(rank >= min(input$box2.4_top) & rank <= max(input$box2.4_top))
    
    patho_plot <- data_select() %>%
      {if(input$box2.3_pathogen != 0) {
        filter(., pathogen_group == input$box2.3_pathogen | mo %in% mo_search) 
      } else {
        .
      }} %>% 
      filter(mo %in% top_10$mo | mo %in% mo_search) %>% 
      mutate(show = if_else(!is.na(mo_search) & mo %in% mo_search, TRUE, FALSE))
    
    patho_plot %>% 
      ggplot(aes(fct_rev(fct_infreq(mo_name(mo))), 
                 tooltip = ..count..,
                 data_id = mo_name(mo),
                 fill = show)) +
      geom_bar_interactive(
        colour = "black"
      ) +
      scale_fill_manual(breaks = c(TRUE, FALSE), values = c("red", "darkgrey")) +
      scale_y_continuous(limits = if (max(patho_plot %>% count(mo) %>% pull(n)) < 25) {
        c(0, 25)
      } else {
        c(0, max(patho_plot %>% count(mo) %>% pull(n)))
      }) +
      labs(x = "", y = trnslt("Count"), title = trnslt("Number of first isolates detected")) +
      coord_flip() +
      theme_minimal() +
      theme(text = element_text(family = "Arial"),
            legend.title = element_blank(),
            legend.position = "none", 
            legend.text = element_text(margin = margin(l = 5)),
            axis.title.y = element_blank(),
            axis.title.x = element_blank(),
            axis.text = element_text(size = 12),
            axis.text.y = element_text(face = "italic"),
            plot.title = element_text(face = "bold", size = 14),
            plot.caption = element_text(colour = "grey")) +
      {if (input$box2.3_group != "group_all") {
        facet_wrap(input$box2.3_group)
      }} +
      NULL
  }) 
  
  output$patho_plot <- renderGirafe({
    ggiraph(ggobj = patho_plot(),
            height_svg = 6,
            width_svg = 10,
            selection_type = "single")
    
  })
  
  output$down_box_patho <- download_box("pathogen_distribution", patho_plot())
  
  
  
  
  # episode plot ---------------------------------------------------------------
  
  test_plot <- reactive({
    
    radar_data %>% 
      mutate(first_14 = first_isolate(col_mo = "mo", col_date = "date", col_patient_id = "patientid",
                                      episode_days = 14, info = FALSE), 
             first_30 = first_isolate(col_mo = "mo", col_date = "date", col_patient_id = "patientid",
                                      episode_days = 30, info = FALSE),
             first_60 = first_isolate(col_mo = "mo", col_date = "date", col_patient_id = "patientid",
                                      episode_days = 60, info = FALSE)) %>% 
      group_by_at(input$box2.2_group) %>% 
      summarise("14 days" = sum(first_14, na.rm = TRUE),
                "30 days" = sum(first_30, na.rm = TRUE),
                "60 days" = sum(first_60, na.rm = TRUE)) %>%
      pivot_longer(cols = c("14 days", "30 days", "60 days"))
    
  })
  
  output$test_plot <- renderPlot({
    test_plot() %>% 
      ggplot(aes(value, name)) +
      geom_col(colour = "black", fill = "lightgrey") +
      labs(x = trnslt("Count"), y = trnslt("Episode"), title = trnslt("Number of first isolates per episode")) +
      theme_minimal() +
      theme(text = element_text(family = "Arial"),
            legend.title = element_blank(),
            legend.text = element_text(margin = margin(l = 5)),
            axis.title.y = element_blank(),
            axis.text = element_text(size = 12),
            plot.title = element_text(face = "bold", size = 14),
            plot.caption = element_text(colour = "grey")) +
      {if (input$box2.2_group != "group_all") {
        facet_wrap(input$box2.2_group)
      }} +
      NULL
  })
  
  
  
  
  
  
  
  
  
  # BOX top right ------------------------------------------------------
  
  output$box1 <- renderUI({
    div(
      style = "position: relative;",
      tabBox(
        id = "box1",
        width = NULL,
        height = 500,
        tabPanel(
          title = trnslt("Resistance profile"),
          div(
            style = "position: absolute; left: 0.5em; bottom: 0.5em;",
            dropdown(
              radioGroupButtons(
                inputId = "box1.2_group",
                label = trnslt("Select group"),
                choiceNames = c(trnslt("All"), trnslt("Year"), trnslt("Gender"), trnslt("Department"), trnslt("Specialty"), trnslt("Specialty code"), trnslt("ICU status"), trnslt("Clinical status"), trnslt("Outward status")),
                choiceValues = c("mo", "year", "gender", "department", "specialty", "specialism", "is_icu", "is_clinical", "is_outward"),
                selected = "mo",
                direction = "vertical"
              ),
              size = "xs",
              icon = icon("gear", class = "opt"),
              up = TRUE
            )
          ), 
          div(
            style = "position: absolute; right: 3.5em; bottom: 0.5em;",
            dropdown(
              downloadButton(outputId = "down_box_res_prop", label = trnslt("Download plot")),
              size = "xs",
              icon = icon("download", class = "opt"),
              up = TRUE,
              right = TRUE
            )
          ),
          withSpinner(
            girafeOutput("isolate_prop_plot", height = 400),
            type = 4,
            color = "#d33724",
            size = 0.7
          ),
          div(
            style = "position:absolute;right:0.5em;bottom: 0.5em;",
            conditionalPanel(
              "input.box1 == 'Resistance profile'",
              actionBttn(
                inputId = "res_prop_plus",
                icon = icon("search-plus", class = "opt"),
                style = "fill",
                color = "danger",
                size = "xs"
              )
            )
          )
        ),
        tabPanel(
          title = trnslt("Combination therapy"),
          withSpinner(
            girafeOutput("comb_plot", height = 400),
            type = 4,
            color = "#d33724",
            size = 0.7
          ),
          div(
            style = "position: absolute; left: 0.5em; bottom: 0.5em;",
            dropdown(
              label = "HANDLE WITH CARE",
              radioGroupButtons(
                inputId = "comb", 
                label = HTML(trnslt("Please read the documentation before interpreting results and modifying underlying algorithm"), 
                             as.character(actionLink(inputId = "action_link", label = trnslt("(click here)"), onclick = 'window.open("https://msberends.github.io/AMR/reference/proportion.html#combination-therapy")'))), 
                justified = TRUE, 
                width = 150,
                choiceNames = c(trnslt("Only all tested"), trnslt("All")),
                choiceValues = c(TRUE, FALSE), 
                size = "xs", 
                direction = "vertical"
              ), 
              icon = icon("info-circle"), 
              size = "xs",
              up = TRUE
            )
          ),
          div(
            style = "position: absolute; right: 3.5em; bottom: 0.5em;",
            dropdown(
              downloadButton(outputId = "down_box_comb_prop", label = trnslt("Download plot")),
              size = "xs",
              icon = icon("download", class = "opt"),
              up = TRUE,
              right = TRUE
            )
          ),
          div(
            style = "position:absolute;right:0.5em;bottom: 0.5em;",
            actionBttn(
              inputId = "res_comb_plus",
              icon = icon("search-plus", class = "opt"),
              style = "fill",
              color = "danger",
              size = "xs"
            )
          )
        )
      )
    )
  })
  
  observeEvent((input$res_prop_plus), {
    showModal(modalDialog(
      renderPlot({
        isolate_prop_plot() + theme(
          text = element_text(family = "Arial"),
          axis.title = element_text(size = 20),
          text = element_text(size = 20),
          plot.title = element_text(size = 26)
        )
      }, height = 600),
      easyClose = TRUE,
      size = "l",
      footer = NULL
    ))
  })
  
  observeEvent((input$res_comb_plus), {
    showModal(modalDialog(
      renderPlot({
        comb_plot() + theme(
          text = element_text(family = "Arial"),
          axis.title = element_text(size = 20),
          text = element_text(size = 20),
          plot.title = element_text(size = 26)
        )
      }, height = 600),
      easyClose = TRUE,
      size = "l",
      footer = NULL
    ))
  })
  
  
  # plot resistance proportion ---------------------------------------------------
  
  isolate_prop_data <- reactive({
    
    # mo_selected <- input$box1_mo
    mo_selected <- input$patho_plot_selected
    
    e_coli_ab <- c("Amoxicillin",
                   "Amoxicillin/clavulanic acid",
                   "Piperacillin/tazobactam",
                   "Cefuroxim",
                   "Ceftriaxone",
                   "Ceftazidime",
                   "Meropenem",
                   "Ciprofloxacin",
                   "Gentamicin",
                   "Tobramycin",
                   "Fosfomycin",
                   "Trimethoprim",
                   "Co-trimoxazole",
                   "Nitrofurantoine")
    
    e_cloacae_ab <- c("Ciprofloxacin",
                      "Gentamicin",
                      "Tobramycin",
                      "Co-trimoxazole",
                      "Meropenem")
    
    p_aeruginosa_ab <- c("Piperacillin/tazobactam",
                         "Ceftazidime",
                         "Meropenem",
                         "Ciprofloxacin",
                         "Gentamicin",
                         "Tobramycin")
    
    s_aureus_ab <- c("Flucloxacillin",
      "Penicillin",
      "Ciprofloxacin",
      "Gentamicin",
      "Erythromycin",
      "Clindamycin",
      "Doxycycline",
      "Linezolid",
      "Co-trimoxazol",
      "Rifampicin")
    
    cons_ab <- c(#"Flucloxacillin",
      "Ciprofloxacin",
      "Gentamicin",
      "Erythromycin",
      "Clindamycin",
      "Doxycycline",
      "Linezolid",
      "Co-trimoxazol",
      "Rifampicin")
    
    e_faecalis_ab <- c("Amoxicillin",
                       "Vancomycin")
    
    get_resistance_df <- function(.data, mo_selected, abx, language = "en") {
      
      out <- .data %>%
        filter(mo == as.mo(mo_selected))
      
      print(paste("rijen voor ", mo_selected, ":", nrow(out)))
      abx <- gsub("Co-trimoxazole", "Trimetoprim / sufamethoxazole", abx)
                          
      if (nrow(out) > 0) {
        out %>% 
          select(mo, as.character(as.ab(abx))) %>%
          count_df(., translate_ab = "name", language = language)
      } else {
        out
      }

      # .data %>%
      #   filter(mo == as.mo(mo_selected)) %>% 
      #   {if (nrow(.) != 0) {
      #     group_by_at(., input$box1.2_group) %>% 
      #       select(group_vars(.), as.character(as.ab(abx))) %>% 
      #       count_df(., translate_ab = "name", language = language)
      #   }}
    }
    
    cat('Implicit join in get_resistance_df(..) @server.R:692\n')
    
    get_resistance_df(data_select(),
                      mo_selected = mo_selected,
                      abx = if (mo_genus(mo_selected) %in% mo_genus(c("E. coli",
                                                                      "K. pneumoniae",
                                                                      "P. mirabilis")))  {
                        e_coli_ab
                      } else if (mo_genus(mo_selected) %in% mo_genus(c("E. cloacae"))) {
                        e_cloacae_ab
                      } else if (mo_genus(mo_selected) %in% mo_genus(c("P. aeruginosa"))) {
                        p_aeruginosa_ab
                      } else if (as.mo(mo_selected) == as.mo("S. aureus")) {
                        s_aureus_ab
                      } else if (as.mo(mo_selected, Becker = TRUE) == as.mo("CoNS")) {
                        cons_ab
                      } else if (mo_genus(mo_selected) %in% mo_genus(c("E. faecalis"))) {
                        e_faecalis_ab
                      } else {
                        radar_data %>% select_if(is.rsi) %>% names()
                      }
    ) %>% 
      pivot_wider(names_from = interpretation, values_from = value) %>% 
      mutate(value_r = R,
             value_si = SI,
             all = R+SI,
             percent_r = percent(R/all, 0.1),
             percent_si = percent(SI/all, 0.1)) %>% 
      pivot_longer(cols = c("SI", "R"), names_to = "interpretation", values_to = "value") %>% 
      mutate(mo = as.mo(input$patho_plot_selected)) %>% 
      mutate(flag = "")
      # left_join(intrinsic_info) %>% 
      # mutate(flag = if_else(intrinsic_r == TRUE, "(intrinsic resistance)", NA_character_))
    
  })
  
  isolate_prop_plot <- reactive({
    
    if (class(try(isolate_prop_data())) == "try-error") {
      ggplot(data.frame(), aes(x = 1, y = 1, label = trnslt("No tests available"))) +
        geom_text(size = 5, colour = "red", fontface = "bold", family = "Arial") +
        theme_void()
    } else {
      
      ggplot(isolate_prop_data(), 
             aes(value, reorder(antibiotic, value_r), 
                 fill = factor(interpretation, levels = c("SI", "R")),
                 tooltip = paste0("R = ", value_r, " (", percent_r, ")\n", 
                                  "SI = ", value_si, " (", percent_si, ")")
             )) + 
        
        geom_col_interactive(colour = "black", position = "fill") +
        scale_fill_manual(limits = c("R", "SI"),
                          breaks = c("SI", "R"),
                          values = c("#FDE725FF", "#440154FF"),
                          na.value = "lightgrey") +
        scale_x_continuous(labels = percent) +
        # flag not used at the moment - for intrinsic resistance
        geom_text(data = isolate_prop_data(), aes(0.5, reorder(antibiotic, value_r), label = flag), colour = "red") +
        ggtitle(label = paste0("Selected isolates: *", mo_fullname(as.mo(input$patho_plot_selected)), "*"),
                subtitle = if (any(isolate_prop_data()$all < 30)) {
                  "Minimum number of valid tests per agent not reached (n < 30)"
                } else {
                  ""
                }) +
        {if (input$box1.2_group != "mo") {
          facet_wrap(input$box1.2_group)
        }} +
        theme_minimal() +
        theme(text = element_text(family = "Arial"),
              legend.title = element_blank(),
              legend.text = element_text(margin = margin(l = 5)),
              axis.title.y = element_blank(),
              axis.title.x = element_blank(),
              axis.text = element_text(size = 12),
              plot.title = ggtext::element_markdown(face = "bold", size = 14),
              plot.subtitle = element_text(colour = "red", face = "bold"))
    }
    
  })
  
  output$isolate_prop_plot <- renderGirafe({
    validate(
      need(input$patho_plot_selected, 'Please select from the tab: Isolates (and pathogens) detected'))
    ggiraph(ggobj = isolate_prop_plot(),
            height_svg = 6,
            width_svg = 10,
            selection_type = "single")
  })
  
  output$down_box_res_prop <- download_box("resistance_prop", isolate_prop_plot())
  
  
  # combination therapy -----------------------------------------------------
  
  comb_data <- reactive({
    
    d <- data_select() %>% 
      filter(mo == as.mo(input$patho_plot_selected))
    
    # resistance proportion
    
    amc_tob <- d %>%
      resistance(AMC, TOB, only_all_tested = as.logical(input$comb), minimum = 0)
    amc_gen <- d %>%
      resistance(AMC, GEN, only_all_tested = as.logical(input$comb), minimum = 0)
    
    cxm_tob <- d %>%
      resistance(CXM, TOB, only_all_tested = as.logical(input$comb), minimum = 0)
    cxm_gen <- d %>%
      resistance(CXM, GEN, only_all_tested = as.logical(input$comb), minimum = 0)
    
    cro_tob <- d %>%
      resistance(CRO, TOB, only_all_tested = as.logical(input$comb), minimum = 0)
    cro_gen <- d %>%
      resistance(CRO, GEN,  only_all_tested = as.logical(input$comb), minimum = 0)
    
    pip_taz_tob <- d %>%
      resistance(TZP, TOB, only_all_tested = as.logical(input$comb), minimum = 0)
    pip_taz_gen <- d %>%
      resistance(TZP, GEN, only_all_tested = as.logical(input$comb), minimum = 0)
    
    
    prop <- tibble(
      "Amoxicillin/clavulanic acid &\nTobramycin" = amc_tob,
      "Amoxicillin/clavulanic acid &\nGentamicin" = amc_gen,
      "Cefuroxime &\nTobramycin" = cxm_tob,
      "Cefuroxime & \nGentamicin" = cxm_gen,
      "Ceftriaxone &\nTobramycin" = cro_tob,
      "Ceftriaxone &\nGentamicin" = cro_gen,
      "Piperacillin/tazobactam &\nTobramycin" = pip_taz_tob,
      "Piperacillin/tazobactam &\nGentamicin" = pip_taz_gen) %>%
      bind_rows(summarise_all(., ~.)) %>%
      bind_rows(summarise_all(., ~1 - .)) %>%
      bind_cols(interpretation = c("R", "R_max", "SI", "SI_max")) %>%
      pivot_longer(cols = c("Amoxicillin/clavulanic acid &\nTobramycin":"Piperacillin/tazobactam &\nGentamicin")) %>%
      pivot_wider(names_from = interpretation, values_from = value) %>%
      pivot_longer(cols = c("R", "SI"), names_to = "interpretation")
    
    
    amc_tob_c <- d %>%
      count_R(AMC, TOB, only_all_tested = as.logical(input$comb), minimum = 0)
    amc_gen_c <- d %>%
      count_R(AMC, GEN, only_all_tested = as.logical(input$comb), minimum = 0)
    
    cxm_tob_c <- d %>%
      count_R(CXM, TOB, only_all_tested = as.logical(input$comb), minimum = 0)
    cxm_gen_c <- d %>%
      count_R(CXM, GEN, only_all_tested = as.logical(input$comb), minimum = 0)
    
    cro_tob_c <- d %>%
      count_R(CRO, TOB, only_all_tested = as.logical(input$comb), minimum = 0)
    cro_gen_c <- d %>%
      count_R(CRO, GEN,  only_all_tested = as.logical(input$comb), minimum = 0)
    
    pip_taz_tob_c <- d %>%
      count_R(TZP, TOB, only_all_tested = as.logical(input$comb), minimum = 0)
    pip_taz_gen_c <- d %>%
      count_R(TZP, GEN, only_all_tested = as.logical(input$comb), minimum = 0)
    
    # SI
    amc_tob_c_si <- d %>%
      count_SI(AMC, TOB, only_all_tested = as.logical(input$comb), minimum = 0)
    amc_gen_c_si <- d %>%
      count_SI(AMC, GEN, only_all_tested = as.logical(input$comb), minimum = 0)
    
    cxm_tob_c_si <- d %>%
      count_SI(CXM, TOB, only_all_tested = as.logical(input$comb), minimum = 0)
    cxm_gen_c_si <- d %>%
      count_SI(CXM, GEN, only_all_tested = as.logical(input$comb), minimum = 0)
    
    cro_tob_c_si <- d %>%
      count_SI(CRO, TOB, only_all_tested = as.logical(input$comb), minimum = 0)
    cro_gen_c_si <- d %>%
      count_SI(CRO, GEN,  only_all_tested = as.logical(input$comb), minimum = 0)
    
    pip_taz_tob_c_si <- d %>%
      count_SI(TZP, TOB, only_all_tested = as.logical(input$comb), minimum = 0)
    pip_taz_gen_c_si <- d %>%
      count_SI(TZP, GEN, only_all_tested = as.logical(input$comb), minimum = 0)
    
    count_r <- tibble(
      "Amoxicillin/clavulanic acid &\nTobramycin" = amc_tob_c,
      "Amoxicillin/clavulanic acid &\nGentamicin" = amc_gen_c,
      "Cefuroxime &\nTobramycin" = cxm_tob_c,
      "Cefuroxime & \nGentamicin" = cxm_gen_c,
      "Ceftriaxone &\nTobramycin" = cro_tob_c,
      "Ceftriaxone &\nGentamicin" = cro_gen_c,
      "Piperacillin/tazobactam &\nTobramycin" = pip_taz_tob_c,
      "Piperacillin/tazobactam &\nGentamicin" = pip_taz_gen_c) %>% 
      bind_cols(interpretation = c("R")) %>%
      pivot_longer(cols = c("Amoxicillin/clavulanic acid &\nTobramycin":"Piperacillin/tazobactam &\nGentamicin")) %>% 
      pivot_wider(names_from = interpretation, values_from = value) %>% 
      pivot_longer(cols = "R", names_to = "interpretation", values_to = "count_r")
    
    count_si <- tibble(
      "Amoxicillin/clavulanic acid &\nTobramycin" = amc_tob_c_si,
      "Amoxicillin/clavulanic acid &\nGentamicin" = amc_gen_c_si,
      "Cefuroxime &\nTobramycin" = cxm_tob_c_si,
      "Cefuroxime & \nGentamicin" = cxm_gen_c_si,
      "Ceftriaxone &\nTobramycin" = cro_tob_c_si,
      "Ceftriaxone &\nGentamicin" = cro_gen_c_si,
      "Piperacillin/tazobactam &\nTobramycin" = pip_taz_tob_c_si,
      "Piperacillin/tazobactam &\nGentamicin" = pip_taz_gen_c_si) %>% 
      bind_cols(interpretation = c("SI")) %>%
      pivot_longer(cols = c("Amoxicillin/clavulanic acid &\nTobramycin":"Piperacillin/tazobactam &\nGentamicin")) %>% 
      pivot_wider(names_from = interpretation, values_from = value) %>% 
      pivot_longer(cols = "SI", names_to = "interpretation", values_to = "count_si")
    
    prop %>% 
      left_join(count_r) %>% 
      left_join(count_si) %>% 
      pivot_longer(cols = c("count_r", "count_si"), names_to = "count", values_to  = "value_count") %>% 
      filter(!is.na(value_count)) %>% 
      group_by(name) %>% 
      mutate(all = sum(value_count)) %>% 
      ungroup() %>% 
      mutate(value_count = paste0(value_count, " (", percent(value, 0.1), ")"))
    
    
  })
  
  comb_plot <- reactive({
    
    if ((!as.mo(input$patho_plot_selected) %in% as.mo(c("E. coli", "K. pneumoniae")))) {
      ggplot(data.frame(), aes(x = 1, y = 1, label = "No tests available\nor not applicable for selected isolate\n(results only available for E. coli and K. pneumoniae)")) +
        geom_text(size = 5, colour = "red", fontface = "bold", family = "Arial") +
        theme_void()
    } else {
      
      ggplot(comb_data(), aes(value, name, 
                              fill = factor(interpretation, levels = c("SI", "R")), 
                              tooltip = paste0("R = ", percent(R_max, 0.01))),
             data_id = name) +
        geom_col_interactive(colour = "black", position = "fill") +
        scale_fill_manual(limits = c("R", "SI"),
                          breaks = c("SI", "R"),
                          values = c("#FDE725FF", "#440154FF"),
                          na.value = "lightgrey") +
        scale_x_continuous(labels = percent) +
        ggtitle(label = paste("Selected isolates: ", mo_fullname(as.mo(input$patho_plot_selected))),
                subtitle = if (any(isolate_prop_data()$all < 30)) {
                  "Minimum number of valid tests per agent not reached (n < 30)"
                } else {
                  ""
                }) +
        theme_minimal() +
        theme(text = element_text(family = "Arial"),
              legend.title = element_blank(),
              legend.text = element_text(margin = margin(l = 5)),
              axis.title.y = element_blank(),
              axis.title.x = element_blank(),
              plot.title = element_text(face = "bold", size = 14),
              axis.text = element_text(size = 12),
              plot.subtitle = element_text(colour = "red", face = "bold")) +
        {if (input$box1.2_group != "mo") {
          facet_wrap(input$box1.2_group)
        }} +
        NULL
    }
    
  })
  
  output$comb_plot <- renderGirafe({
    
    validate(
      need(input$patho_plot_selected, 'Please select from the tab: Isolates (and pathogens) detected'))
    
    
    ggiraph(ggobj = comb_plot(), 
            height_svg = 6,
            width_svg = 10,
            selection_type = "single")
    
  })
  
  output$down_box_comb_prop <- download_box("resistance_combination", comb_plot())
  
  # table -------------------------------------------------------------------
  
  output$box3 <- renderUI({
    div(
      style = "position: relative",
      tabBox(
        id = "box3",
        width = NULL,
        height = 550,
        tabPanel(title = "",
                 dataTableOutput("test_table"), 
                 style = "height:500px; overflow-y: scroll;overflow-x: scroll;")))
    
  })
  
  dt_custom <- function(data) {
    datatable(data, rownames = FALSE, 
              extensions = c('Buttons', 'Scroller'),
              options = list( 
                dom = "Bftp", 
                buttons = 
                  list("copy", list(
                    extend = "collection", 
                    buttons = c("csv", "excel", "pdf"), 
                    text = trnslt("Download"))), 
                lengthMenu = list( c(10, 20, -1), c(10, 20, trnslt("All"))) , 
                pageLength = -1,
                columnDefs = list(list(className = 'dt-right', targets = "_all"))))
  }
  
  output$test_table <- renderDataTable({
    
    # isolate_prop_data() %>% 
    if (input$box2 == "Positive & negative cultures") {
      all <- radar_data %>% 
        filter(specialty_shiny %in% input$specialtyInput & 
                 department %in% input$departmentInput &
                 Materiaal %in% input$materialInput) %>% 
        rename(group = input$box2.1_group) %>% 
        filter(!is.na(group)) %>% 
        distinct(patientid, .keep_all = TRUE) %>%
        count(group) %>%
        rename(n_all = n)
      pos <- radar_data %>% 
        filter(specialty_shiny %in% input$specialtyInput & 
                 department %in% input$departmentInput &
                 Materiaal %in% input$materialInput) %>% 
        rename(group = input$box2.1_group) %>% 
        group_by(group) %>%
        # group_by_at("gender") %>%
        filter(!is.na(mo)) %>% 
        distinct(patientid) %>% 
        count(group) %>%
        # count_("gender") %>%
        left_join(all) %>% 
        mutate(n_rel = percent(n/n_all, 0.1)) %>%
        select(group, n, n_rel, n_all) %>% 
        rename("Patients (n) with at least 1 positive culture" = n,
               "Patients (%) with at least 1 positive culture" = n_rel,
               "All patients (n)" = n_all) %>%
        {if (input$box2.1_group == "group_all") {
          # select(., -group)
          mutate(., group = "All patients")
        } else {
          .
        }
        } %>%
        rename_all(str_to_sentence) %>%
        dt_custom()
    } else if (input$box2 == "First isolates per episode") {
      test_plot() %>% 
        select("Episode selected" = name, "First isolates" = value) %>% 
        dt_custom()
    } else {
      data_select() %>% 
        freq(mo) %>% 
        mutate(Name = mo_name(item),
               percent = percent(percent, accuracy = 0.1), 
               cum_percent = percent(cum_percent)) %>% 
        select(-cum_percent) %>% 
        select(Name, n = count, "%" = percent, "n (cummulative)" = cum_count) %>% 
        dt_custom()
    }
    
  })
  
  output$box4 <- renderUI({
    div(
      style = "position: relative",
      tabBox(
        id = "box4",
        width = NULL,
        height = 550,
        tabPanel(title = "",
                 dataTableOutput("res_table"), style = "height:500px; overflow-y: scroll;")))
  })
  
  output$res_table <- renderDataTable({
    
    validate(
      need(input$patho_plot_selected, 'Please select from the tab: Isolates (and pathogens) detected'))
    
    if (class(try(isolate_prop_data())) == "try-error") {
      data.frame()
    } else {
      
      if (input$box1 == "Resistance profile") { 
        isolate_prop_data() %>% 
          group_by(antibiotic) %>%
          mutate(percent = percent(value/all, 0.1),
                 join = paste0(value, " (", percent, ")")) %>%
          pivot_wider(id_cols = c(antibiotic, all), names_from = interpretation, values_from = join) %>% 
          rename(Antibiotic = antibiotic,
                 total = all) %>% 
          dt_custom()
      } else {
        if (as.mo(input$patho_plot_selected) %in% as.mo(c("E. coli", "K. pneumoniae"))) {
          comb_data() %>% 
            pivot_wider(id_cols = c(name, all), names_from = interpretation, values_from = value_count) %>%
            select(Combination = name, total = all, SI, R) %>% 
            dt_custom()
        } else {
          data.frame()
        }
      }
    }
  })
  
  
}