nmode's Git Repositories - Rnaught/blob - R/server.R

   1 #' @importFrom methods is
   2 #' @importFrom utils read.csv write.csv
   3 server <- function(input, output) {
   4     # Hide the sidebar if the 'About' tab is active.
   5     shiny::observeEvent(input$tabset, {
   6         if (input$tabset == "About") {
   7             shinyjs::hideElement(selector="#sidebar")
   8             shinyjs::removeCssClass("main", "col-sm-8")
   9             shinyjs::addCssClass("main", "col-sm-12")
  10         } else {
  11             shinyjs::showElement(selector="#sidebar")
  12             shinyjs::removeCssClass("main", "col-sm-12")
  13             shinyjs::addCssClass("main", "col-sm-8")
  14         }
  15     })
  16
  17     reactive <- shiny::reactiveValues(
  18         data_table=data.frame(Name=character(0), `Reporting Frequency`=character(0), `Case Counts`=numeric(0), check.names=FALSE),
  19         est_table=data.frame(Dataset=character(0)),
  20         estimators=list()
  21     )
  22
  23     # Validate and add datasets when button is clicked.
  24     # Also evaluate the new datasets on existing estimators.
  25     shiny::observeEvent(input$addData, {
  26         # Option 1: Manual entry.
  27         if (input$dataInputMethod == 1) {
  28             checks_passed <- TRUE
  29
  30             # Ensure the dataset name is not blank.
  31             if (grepl("^\\s*$", input$dataName)) {
  32                 output$dataNameWarn <- shiny::renderText("Error: The dataset name cannot be blank.")
  33                 checks_passed <- FALSE
  34             }
  35             # Ensure the dataset name is not a duplicate.
  36             else if (input$dataName %in% reactive$data_table[,1]) {
  37                 output$dataNameWarn <- shiny::renderText("Error: There is already a dataset with the specified name.")
  38                 checks_passed <- FALSE
  39             }
  40             else
  41                 output$dataNameWarn <- shiny::renderText("")
  42
  43             # Ensure the case counts consist only of non-negative integers, separated by commas.
  44             counts <- as.numeric(unlist(strsplit(input$dataCounts, split=",")))
  45             if (any(is.na(counts)) || any(counts <= 0) || any(counts %% 1 != 0)) {
  46                 output$dataCountsWarn <- shiny::renderText("Error: The list of case counts should only contain non-negative integers, separated by commas.")
  47                 checks_passed <- FALSE
  48             }
  49             # Ensure the case counts contain at least two entries.
  50             else if (length(counts) < 2) {
  51                 output$dataCountsWarn <- shiny::renderText("Error: The list of case counts should contain at least two entries.")
  52                 checks_passed <- FALSE
  53             }
  54             else
  55                 output$dataCountsWarn <- shiny::renderText("")
  56
  57             if (checks_passed)
  58                 d <- data.frame(input$dataName, input$dataUnits, t(counts))
  59         }
  60
  61         else {
  62             checks_passed <- FALSE
  63
  64             # Option 2: Upload .csv
  65             if (input$dataInputMethod == 2)
  66                 d <- try(read.csv(input$dataUpload$datapath, header=FALSE))
  67             # Option 3: Paste .csv
  68             else
  69                 d <- try(read.csv(text=input$dataPaste, header=FALSE))
  70
  71             if (is(d, "try-error"))
  72                 output$dataCSVWarn <- shiny::renderText("Error reading file.")
  73             else if (ncol(d) < 4 || anyNA(d[,1]) || anyNA(sapply(d[,3:4], as.numeric)) || !all(trimws(d[,2]) %in% c("Daily", "Weekly")))
  74                 output$dataCSVWarn <- shiny::renderText("Error: The provided .csv file does not match the required format.")
  75             else if (length(intersect(reactive$data_table[,1], d[,1])) > 0)
  76                 output$dataCSVWarn <- shiny::renderText("Error: The provided .csv file contains dataset names which already exist.")
  77             else if (length(unique(d[,1])) != length(d[,1]))
  78                 output$dataCSVWarn <- shiny::renderText("Error: The provided .csv file contains duplicate dataset names.")
  79             else {
  80                 output$dataCSVWarn <- shiny::renderText("")
  81                 checks_passed <- TRUE
  82             }
  83         }
  84
  85         if (checks_passed) {
  86             d[,3:ncol(d)] <- apply(d[,3:ncol(d)], 2, as.numeric)
  87             d[,3] <- data.frame(I(lapply(split(d[,3:ncol(d)], 1:nrow(d)), function(x) x[!is.na(x)])))
  88             d <- d[,1:3]
  89             d[,2] <- trimws(d[,2])
  90             colnames(d) <- c("Name", "Reporting Frequency", "Case Counts")
  91             reactive$data_table <- rbind(reactive$data_table, d)
  92             reactive$est_table <- update_est_row(input, output, d, reactive$estimators, reactive$est_table)
  93         }
  94     })
  95
  96     output$dataTable <- shiny::renderDataTable(reactive$data_table, escape=FALSE)
  97     output$estTable <- shiny::renderDataTable(reactive$est_table, escape=FALSE)
  98
  99     # Download table of estimates as a .csv file.
 100     output$downloadEst <- shiny::downloadHandler(
 101         filename=function() { paste0("Rnaught-", Sys.Date(), ".csv") },
 102         content=function(file) { write.csv(reactive$est_table, file) }
 103     )
 104
 105     shiny::observeEvent(input$addWP, {
 106         if (input$serialWPKnown == 1) {
 107             serial <- validate_serial(input, output, "serialWPInput", "serialWPWarn")
 108             if (!is.na(serial)) {
 109                 reactive$estimators[[length(reactive$estimators)+1]] <- list(method="WP", mu=serial, search=list(B=100, shape.max=10, scale.max=10), mu_units=input$serialWPUnits)
 110                 reactive$est_table <- update_est_col(input, output, reactive$data_table, reactive$estimators[[length(reactive$estimators)]], reactive$est_table)
 111             }
 112         }
 113         else {
 114             checks_passed <- TRUE
 115
 116             grid_length <- as.numeric(input$gridLengthInput)
 117             max_shape <- as.numeric(input$gridShapeInput)
 118             max_scale <- as.numeric(input$gridScaleInput)
 119
 120             if (is.na(grid_length) || grid_length <= 0 || grid_length %% 1 != 0) {
 121                 output$gridLengthWarn <- shiny::renderText("Error: The grid size must be a positive integer.")
 122                 output$gridShapeWarn <- shiny::renderText("")
 123                 output$gridScaleWarn <- shiny::renderText("")
 124                 checks_passed <- FALSE
 125             }
 126             else {
 127                 output$gridLengthWarn <- shiny::renderText("")
 128
 129             if (is.na(max_shape) || max_shape < 1 / grid_length) {
 130                 output$gridShapeWarn <- shiny::renderText("Error: The maximum shape must be at least the reciprocal of the grid length.")
 131                 checks_passed <- FALSE
 132             }
 133             else
 134                 output$gridShapeWarn <- shiny::renderText("")
 135
 136             if (is.na(max_scale) || max_scale < 1 / grid_length) {
 137                 output$gridShapeWarn <- shiny::renderText("Error: The maximum scale must be at least the reciprocal of the grid length.")
 138                 checks_passed <- FALSE
 139             }
 140             else
 141                 output$gridScaleWarn <- shiny::renderText("")
 142             }
 143
 144             if (checks_passed) {
 145                 reactive$estimators[[length(reactive$estimators)+1]] <- list(method="WP", mu=NA, search=list(B=grid_length, shape.max=max_shape, scale.max=max_scale), mu_units=input$serialWPUnits)
 146                 reactive$est_table <- update_est_col(input, output, reactive$data_table, reactive$estimators[[length(reactive$estimators)]], reactive$est_table)
 147             }
 148         }
 149     })
 150
 151     shiny::observeEvent(input$addseqB, {
 152         serial <- validate_serial(input, output, "serialseqBInput", "serialseqBWarn")
 153         checks_passed <- !is.na(serial)
 154
 155         kappa <- as.numeric(input$kappaInput)
 156         if (is.na(kappa) || kappa <= 0) {
 157             output$kappaWarn <- shiny::renderText("Error: The maximum value must be a positive number.")
 158             checks_passed <- FALSE
 159         }
 160         else
 161             output$kappaWarn <- shiny::renderText("")
 162
 163         if (checks_passed) {
 164             reactive$estimators[[length(reactive$estimators)+1]] <- list(method="seqB", mu=serial, kappa=kappa, mu_units=input$serialseqBUnits)
 165             reactive$est_table <- update_est_col(input, output, reactive$data_table, reactive$estimators[[length(reactive$estimators)]], reactive$est_table)
 166         }
 167     })
 168
 169     shiny::observeEvent(input$addID, {
 170         serial <- validate_serial(input, output, "serialIDInput", "serialIDWarn")
 171         if (!is.na(serial)) {
 172             reactive$estimators[[length(reactive$estimators)+1]] <- list(method="ID", mu=serial, mu_units=input$serialIDUnits)
 173             reactive$est_table <- update_est_col(input, output, reactive$data_table, reactive$estimators[[length(reactive$estimators)]], reactive$est_table)
 174         }
 175     })
 176
 177     shiny::observeEvent(input$addIDEA, {
 178         serial <- validate_serial(input, output, "serialIDEAInput", "serialIDEAWarn")
 179         if (!is.na(serial)) {
 180             reactive$estimators[[length(reactive$estimators)+1]] <- list(method="IDEA", mu=serial, mu_units=input$serialIDEAUnits)
 181             reactive$est_table <- update_est_col(input, output, reactive$data_table, reactive$estimators[[length(reactive$estimators)]], reactive$est_table)
 182         }
 183     })
 184 }
 185
 186 validate_serial <- function(input, output, serialInputId, serialWarnId) {
 187     serial <- as.numeric(input[[serialInputId]])
 188     if (is.na(serial) || serial <= 0) {
 189         output[[serialWarnId]] <- shiny::renderText("Error: The mean serial interval should be a non-negative number.")
 190         serial <- NA
 191     }
 192     else
 193         output[[serialWarnId]] <- shiny::renderText("") # Clear warning text.
 194
 195     return(serial)
 196 }
 197
 198 update_est_col <- function(input, output, datasets, estimator, est_table) {
 199     if (nrow(datasets) == 0)
 200         new_est_table <- data.frame(matrix(nrow=0, ncol=ncol(est_table)+1))
 201     else {
 202         estimates <- rep(NA, nrow(datasets))
 203
 204     for (row in 1:nrow(datasets))
 205         estimates[row] <- eval_estimator(input, output, estimator, datasets[row,])
 206
 207         if (nrow(est_table) == 0)
 208             new_est_table <- cbind(datasets[,1], estimates)
 209         else
 210             new_est_table <- cbind(est_table, estimates)
 211     }
 212
 213     colnames(new_est_table) <- c(colnames(est_table), shiny::HTML(paste0(estimator$method, "<br/>(&mu; = ", estimator$mu, " ", tolower(estimator$mu_units), ")")))
 214     return(new_est_table)
 215 }
 216
 217 update_est_row <- function(input, output, datasets, estimators, est_table) {
 218     if (length(estimators) == 0) {
 219         if (nrow(est_table) == 0)
 220             new_est_table <- data.frame(datasets[,1])
 221         else
 222             new_est_table <- data.frame(c(est_table[,1], datasets[,1]))
 223
 224         colnames(new_est_table) <- colnames(est_table)
 225     }
 226     else {
 227         new_est_table <- data.frame(matrix(nrow=nrow(datasets), ncol=length(estimators)))
 228
 229         for (row in 1:nrow(datasets))
 230             for (col in 1:length(estimators))
 231                 new_est_table[row, col] <- eval_estimator(input, output, estimators[[col]], datasets[row,])
 232
 233         new_est_table <- cbind(datasets[,1], new_est_table)
 234         colnames(new_est_table) <- colnames(est_table)
 235         new_est_table <- rbind(est_table, new_est_table)
 236     }
 237
 238     return(new_est_table)
 239 }
 240
 241 eval_estimator <- function(input, output, estimator, dataset) {
 242     # Adjust serial interval to match time unit of case counts.
 243     serial <- estimator$mu
 244     if (estimator$mu_units == "Days" && dataset[2] == "Weekly")
 245         serial <- serial / 7
 246     else if (estimator$mu_units == "Weeks" && dataset[2] == "Daily")
 247         serial <- serial * 7
 248
 249     # White and Panago
 250     if (estimator$method == "WP") {
 251         estimate <- WP(unlist(dataset[3]), mu=serial, search=estimator$search)
 252
 253         if (!is.na(estimator$mu))
 254             estimate <- round(estimate$Rhat, 2)
 255         # Display the estimated mean of the serial distribution if mu was not specified.
 256         else {
 257             if (dataset[2] == "Daily")
 258                 mu_units <- "days"
 259             else
 260                 mu_units <- "weeks"
 261             MSI <- sum(estimate$SD$supp * estimate$SD$pmf)
 262             estimate <- shiny::HTML(paste0(round(estimate$Rhat, 2), "<br/>(&mu; = ", round(MSI, 2), " ", mu_units, ")"))
 263         }
 264     }
 265     # Sequential Bayes
 266     else if (estimator$method == "seqB")
 267         estimate <- round(seqB(unlist(dataset[3]), mu=serial, kappa=estimator$kappa)$Rhat, 2)
 268     # Incidence Decay
 269     else if (estimator$method == "ID")
 270         estimate <- round(ID(unlist(dataset[3]), mu=serial), 2)
 271     # Incidence Decay with Exponential Adjustement
 272         else if (estimator$method == "IDEA")
 273         estimate <- round(IDEA(unlist(dataset[3]), mu=serial), 2)
 274
 275     return(estimate)
 276 }