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