R/jagsParallel.R

In isoWater: Discovery, Retrieval, and Analysis of Water Isotope Data

###Code adapted from R2jags, https://github.com/suyusung/R2jags
###Includes fix for namespace bug proposed by user wlandu 
###Accessed 3/7/2021

jags.parallel <- function (data, inits, parameters.to.save, model.file,
                           n.chains , n.iter, n.burnin,
                           n.thin, n.cluster = n.chains)
{
  #Updated to use doParallel
  registerDoParallel(n.cluster)
  res = foreach(i = 1:n.cluster) %dopar% {
    jags(data = data, parameters.to.save = parameters.to.save, 
         model.file = model.file, n.chains = 1, n.iter = n.iter, 
         n.burnin = n.burnin, n.thin = n.thin)
  }
  stopImplicitCluster()

  ###Removed
#  .runjags <- function() {
#    requireNamespace("coda")
#    jagsfit <- R2jags::jags(data = eval(expression(data)), inits = jags.inits,
#                    parameters.to.save = eval(expression(jags.params)),
#                    model.file = eval(expression(jags.model)), n.chains = 1,
#                    n.iter = eval(expression(n.iter)), n.burnin = eval(expression(n.burnin)),
#                    n.thin = eval(expression(n.thin)), DIC = eval(expression(DIC)),
#                    working.directory = eval(expression(working.directory)),
#                    jags.seed = eval(expression(jags.seed)), progress.bar = "none",
#                    digits = eval(expression(digits)), RNGname = eval(expression(RNGname)),
#                    jags.module = eval(expression(jags.module)) )
#    return(jagsfit)
#  }
#  cl <- makeCluster(n.cluster, methods = FALSE)
#  clusterExport(cl, c(names(data), export_obj_names ), envir = envir)
#  clusterSetRNGStream(cl, jags.seed)
#  tryCatch(res <- clusterCall(cl, .runjags), finally = stopCluster(cl))

  result <- NULL
  model <- NULL
  for (ch in 1:n.chains) {
    result <- abind(result, res[[ch]]$BUGSoutput$sims.array,
                    along = 2)
    model[[ch]] <- res[[ch]]$model
  }
  if (is.function(model.file)) {
    model.file <- substitute(model.file)
  }
  result <- as.bugs.array2(result, model.file = model.file,
                           program = "jags", DIC = TRUE, n.iter = n.iter, 
                           n.burnin = n.burnin, n.thin = n.thin)
  out <- list(model = model, BUGSoutput = result, parameters.to.save = parameters.to.save,
              model.file = model.file, n.iter = n.iter, DIC = TRUE)
  class(out) <- c("rjags.parallel", "rjags")
  return(out)
}

as.bugs.array2 <- function(sims.array, model.file=NULL, program="jags",
                           DIC=FALSE, DICOutput=NULL, n.iter=NULL, n.burnin=0, n.thin=1)
{
  ## 'sims.array' is supposed to be a 3-way array with
  # n.sims*n.chains*n.parameters simulations, and
  # the 3rd component of dimnames(x) should have the parameter names.
  
  ## From Andrew Gelman's bugs.r function
  ## a couple of lines commented out by Eduardo Leoni (see comment below)
  #require("R2WinBUGS")
  d <- dim(sims.array)
  n.keep       <- d[1]
  n.chains     <- d[2]
  n.parameters <- d[3]
  n.sims       <- n.keep*n.chains
  if (is.null(n.iter)){
    n.iter <- (n.keep + n.burnin)*n.thin
  }
  #
  parameter.names <- dimnames(sims.array)[[3]]
  if (is.null(parameter.names)) {
    parameter.names <- paste("P", 1:n.parameters, sep="")
    dimnames(sims.array)[[3]] <- parameter.names
  }
  parameters.to.save <- unique(sapply(strsplit(parameter.names, "\\["), "[", 1))
  #
  sims <- matrix(NA, n.sims, n.parameters)
  root.long <- character(n.parameters)
  indexes.long <- vector(n.parameters, mode = "list")
  for (i in 1:n.parameters) {
    temp <- decode.parameter.name(parameter.names[i])
    root.long[i] <- temp$root
    indexes.long[[i]] <- temp$indexes
  }
  n.roots <- length(parameters.to.save)
  left.bracket.short <- as.vector(regexpr("[[]", parameters.to.save))
  right.bracket.short <- as.vector(regexpr("[]]", parameters.to.save))
  root.short <- ifelse(left.bracket.short == -1, parameters.to.save,
                       substring(parameters.to.save, 1, left.bracket.short -1))
  dimension.short <- rep(0, n.roots)
  indexes.short <- vector(n.roots, mode = "list")
  n.indexes.short <- vector(n.roots, mode = "list")
  long.short <- vector(n.roots, mode = "list")
  length.short <- numeric(n.roots)
  for (j in 1:n.roots) {
    long.short[[j]] <- (1:n.parameters)[root.long == root.short[j]]
    length.short[j] <- length(long.short[[j]])
    if (length.short[j] == 0){
      stop(paste("parameter", root.short[[j]], "is not in the model"))
    }
    else if (length.short[j] > 1) {
      dimension.short[j] <- length(indexes.long[[long.short[[j]][1]]])
      n.indexes.short[[j]] <- numeric(dimension.short[j])
      for (k in 1:dimension.short[j]){
        n.indexes.short[[j]][k] <- length(unique(unlist(lapply(indexes.long[long.short[[j]]], .subset, k))))
      }
      length.short[j] <- prod(n.indexes.short[[j]])
      ## Modified by Eduardo Leoni
      ## this check fails if you take out a part of the simulations
      ## (for example, you don't want the array to have some of the
      ## parameters) so I took them out.
      
      ## if (length(long.short[[j]]) != length.short[j])
      ##   stop(paste("error in parameter", root.short[[j]],
      ##   "in parameters.to.save"))
      indexes.short[[j]] <- as.list(numeric(length.short[j]))
      for (k in 1:length.short[j]){
        indexes.short[[j]][[k]] <- indexes.long[[long.short[[j]][k]]]
      }
    }
  }
  rank.long <- unlist(long.short)
  # -----
  # yes, it's inefficient to do this, but for now I'm just letting this be as it is:
  for (k in 1:n.parameters) {
    sims[,k] <- as.vector(sims.array[,,k])
  }
  # ----
  dimnames(sims) <- list(NULL, parameter.names)
  summary <- monitor(sims.array, n.chains, keep.all = TRUE)
  last.values <- as.list(numeric(n.chains))
  
  for(nn in 1:length(n.indexes.short)){
    if(is.null(n.indexes.short[[nn]])){
      n.indexes.short[[nn]] <- 1
    }
  }
  
  
  for (i in 1:n.chains) {
    n.roots.0 <- if (DIC){
      n.roots - 1
    }
    else{
      n.roots
    }
    last.values[[i]] <- as.list(numeric(n.roots.0))
    names(last.values[[i]]) <- root.short[1:n.roots.0]
    for (j in 1:n.roots.0) {
      if (dimension.short[j] <= 1) {
        last.values[[i]][[j]] <- sims.array[n.keep, i, long.short[[j]]]
        names(last.values[[i]][[j]]) <- NULL
      }
      if (program=="jags"){
        last.values[[i]][[j]] <- array(sims.array[n.keep, i, long.short[[j]]], n.indexes.short[[j]])
      }
      ## only winbugs have to permute the array.
      else {
        last.values[[i]][[j]] <- aperm(array(sims.array[n.keep, i, long.short[[j]]], rev(n.indexes.short[[j]])), dimension.short[j]:1)
      }
    }
  }
  sims <- sims[sample(n.sims), , drop = FALSE]
  sims.list <- summary.mean <- summary.sd <- summary.median <- summary.025 <-  summary.975 <- vector(n.roots, mode = "list")
  names(sims.list) <- names(summary.mean) <- names(summary.sd) <- names(summary.median) <- names(summary.025) <- names(summary.975) <- root.short
  for (j in 1:n.roots) {
    if (length.short[j] == 1) {
      sims.list[[j]] <- sims[, long.short[[j]]]
      summary.mean[[j]] <- summary[long.short[[j]], "mean"]
      summary.sd[[j]] <- summary[long.short[[j]], "sd"]
      summary.median[[j]] <- summary[long.short[[j]], "50%"]
      ##ell: added 025 and 975
      summary.025[[j]] <- summary[long.short[[j]], "2.5%"]
      summary.975[[j]] <- summary[long.short[[j]], "97.5%"]
    }
    if (program=="bugs") {
      temp2 <- dimension.short[j]:1
      sims.list[[j]] <- aperm(array(sims[, long.short[[j]]], c(n.sims, rev(n.indexes.short[[j]]))), c(1, (dimension.short[j] + 1):2))
      summary.mean[[j]] <- aperm(array(summary[long.short[[j]], "mean"], rev(n.indexes.short[[j]])), temp2)
      summary.sd[[j]] <- aperm(array(summary[long.short[[j]], "sd"], rev(n.indexes.short[[j]])), temp2)
      summary.median[[j]] <- aperm(array(summary[long.short[[j]], "50%"], rev(n.indexes.short[[j]])), temp2)
      ##ell: added 025 and 975
      #      summary.025[[j]] <- aperm(array(summary[long.short[[j]], "2.5%"], rev(n.indexes.short[[j]])), temp2)
      #      summary.975[[j]] <- aperm(array(summary[long.short[[j]], "97.5%"], rev(n.indexes.short[[j]])), temp2)
    }
    if (program=="jags") {
      ##fix this list
      #sims.list[[j]] <- aperm(array(sims[, long.short[[j]]], c(n.sims, rev(n.indexes.short[[j]]))), c(1, (dimension.short[j] + 1):2))
      sims.list[[j]] <- array(sims[, long.short[[j]]], c(n.sims, n.indexes.short[[j]]))
      #sims.list[[j]] <- sims[, long.short[[j]]]
      summary.mean[[j]] <- array(summary[long.short[[j]],"mean"],n.indexes.short[[j]])
      summary.sd[[j]] <- array(summary[long.short[[j]],"sd"],n.indexes.short[[j]])
      summary.median[[j]] <- array(summary[long.short[[j]],"50%"],n.indexes.short[[j]])
      ##ell: added 025 and 975
      #      summary.025[[j]] <- array(summary[long.short[[j]],"2.5%"],n.indexes.short[[j]])
      #      summary.975[[j]] <- array(summary[long.short[[j]],"97.5%"],n.indexes.short[[j]])
    }
  }
  summary <- summary[rank.long, ]
  #all <- list(n.chains = n.chains, n.iter = n.iter, n.burnin = n.burnin,
  #        n.thin = n.thin, n.keep = n.keep, n.sims = n.sims,
  #        sims.array = sims.array[, , rank.long, drop = FALSE], sims.list = sims.list,
  #        sims.matrix = sims[, rank.long], summary = summary, mean = summary.mean,
  #        sd = summary.sd, median = summary.median, root.short = root.short,
  #        long.short = long.short, dimension.short = dimension.short,
  #        indexes.short = indexes.short, last.values = last.values,
  #        is.DIC=DIC, p02.5=summary.025, p97.5=summary.975)
  
  all <- list(n.chains = n.chains, n.iter = n.iter, n.burnin = n.burnin,
              n.thin = n.thin, n.keep = n.keep, n.sims = n.sims,
              sims.array = sims.array[,,rank.long,drop = FALSE], sims.list = sims.list,
              sims.matrix = sims[, rank.long], summary = summary, mean = summary.mean,
              sd = summary.sd, median = summary.median, root.short = root.short,
              long.short = long.short, dimension.short = dimension.short,
              indexes.short = indexes.short, last.values = last.values, program=program,
              model.file=model.file)
  
  if(DIC && is.null(DICOutput)) { ## calculate DIC from deviance
    deviance <- all$sims.array[, , "deviance", drop = FALSE]
    dim(deviance) <- dim(deviance)[1:2]
    pD <- numeric(n.chains)
    DIC <- numeric(n.chains)
    for (i in 1:n.chains) {
      pD[i] <- var(deviance[, i])/2
      DIC[i] <- mean(deviance[, i]) + pD[i]
    }
    all <- c(all, list(isDIC=TRUE, DICbyR=TRUE,  pD=mean(pD), DIC=mean(DIC)))
  }
  else if(DIC && !is.null(DICOutput)) { ## use DIC from BUGS
    all <- c(all, list(isDIC=TRUE, DICbyR=FALSE,
                       pD=DICOutput[nrow(DICOutput),4],
                       DIC=DICOutput[nrow(DICOutput),3]))
  }
  else {
    all <- c(all, isDIC=FALSE)
  }
  class(all) <- "bugs"
  all
}

###Internal function from package R2WinBUGS by Uwe Ligges:
###https://cran.r-project.org/package=R2WinBUGS, accessed 3/7/2021
decode.parameter.name <-
  function (a){
    #
    # Decodes Bugs parameter names
    #   (e.g., "beta[3,14]" becomes "beta" with 2 indexes:  3 and 14)
    # for use by the bugs.sim() function
    #
    left.bracket <- regexpr ("[[]", a)
    if (left.bracket==-1){
      root <- a
      dimension <- 0
      indexes <- NA
    }
    else {
      root <- substring (a, 1, left.bracket-1)
      right.bracket <- regexpr ("[]]", a)
      a <- substring (a, left.bracket+1, right.bracket-1)
      indexes <- as.numeric(unlist(strsplit(a, ",")))
      dimension <- length(indexes)
    }
    list(root=root, dimension=dimension, indexes=indexes)
  }