Function to set non-detects in residue series

6 files changed, 187 insertions, 4 deletions

diff --git a/ChangeLog b/ChangeLog
index 1f47ac8..17ffa8a 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -1,3 +1,9 @@
+commit 9f848a9518aabf162723271bafba244221ee83ed
+Author: Johannes Ranke <johannes.ranke@jrwb.de>
+Date:   2019-09-27 12:55:03 +0200
+
+    Built windows binary for drat
+
 commit 435e07a6f7fd2599d783fd306ee6d9e08acc0c6b
 Author: Johannes Ranke <jranke@uni-bremen.de>
 Date:   2019-09-27 10:00:15 +0200
diff --git a/DESCRIPTION b/DESCRIPTION
index 1589b14..73c561d 100644
--- a/DESCRIPTION
+++ b/DESCRIPTION
@@ -2,7 +2,7 @@ Package: pfm
 Type: Package
 Title: Utilities for Pesticide Fate Modelling
 Version: 0.5.7
-Date: 2019-09-27
+Date: 2019-10-09
 Authors@R: person("Johannes Ranke", email = "jranke@uni-bremen.de",
                   role = c("aut", "cre", "cph"),
                   comment = c(ORCID = "0000-0003-4371-6538"))
diff --git a/NAMESPACE b/NAMESPACE
index 59e8040..8eb6863 100644
--- a/NAMESPACE
+++ b/NAMESPACE
@@ -36,6 +36,8 @@ export(perc_runoff_reduction_exposit)
 export(pfm_degradation)
 export(read.TOXSWA_cwa)
 export(sawtooth)
+export(set_nd)
+export(set_nd_focus)
 export(soil_DT50)
 export(soil_Kfoc)
 export(soil_N)
diff --git a/R/process_residues.R b/R/process_residues.R
new file mode 100644
index 0000000..35d3a4e
--- /dev/null
+++ b/R/process_residues.R
@@ -0,0 +1,107 @@
+#' Set non-detects in residue series without replicates
+
+#' Sets non-detects directly before or directly after detects to NA. Values between
+#' lod and loq are set to their mean value if an loq is specified.
+#' If 'time_zero' is set to TRUE, the residue series is assumed to start with time
+#' zero, and non-detects at time zero are set to 'time_zero_nd_value'. For the
+#' set_nd_focus variant, this is zero, otherwise this argument has NA as default
+#' value.
+#' If stopping after the first non-detection is requested, as in in the FOCUS
+#' variant of the function, an loq has to be specified in order to decide
+#' if any later detections are above the loq.
+
+#' @param r A character vector of sequential residues without replicates, with
+#' non-detects specified as 'nd' and unquantified values above the limit of
+#' detection specified as 'nq', otherwise coercible to numeric
+#' @param lod Limit of detection (numeric)
+#' @param loq Limit of quantification(numeric). Must be specified if the FOCUS rule to
+#' stop after the first non-detection is to be applied
+#' @param stop_after_first_nondetect Should we really stop after the first non-detection?
+#' @references FOCUS (2014) p. 75, 76, 131, 132
+#' @export
+#' @examples
+#' parent_1 <- c(.12, .09, .05, .03, "nd", "nd", "nd", "nd", "nd", "nd")
+#' set_nd(parent_1, 0.02)
+#' parent_2 <- c(.12, .09, .05, .03, "nd", "nd", .03, "nd", "nd", "nd")
+#' set_nd(parent_2, 0.02)
+#' set_nd_focus(parent_2, 0.02, loq = 0.05)
+#' parent_3 <- c(.12, .09, .05, .03, "nd", "nd", .06, "nd", "nd", "nd")
+#' set_nd(parent_3, 0.02)
+#' set_nd_focus(parent_3, 0.02, loq = 0.05)
+#' metabolite <- c("nd", "nd", "nd", 0.03, 0.06, 0.10, 0.11, 0.10, 0.09, 0.05, 0.03, "nd", "nd")
+#' set_nd(metabolite, 0.02)
+set_nd <- function(r, lod, loq = NA,
+  time_zero = TRUE, time_zero_nd_value = NA, stop_after_first_nondetect = FALSE)
+{
+
+  if (stop_after_first_nondetect & is.na(loq)) {
+    stop("You need to specify an loq to decide if the curve should be cut off after the first non-detect")
+  }
+
+  result <- r
+
+  # Handle nq values
+  if (!missing(loq)) {
+    nq = 0.5 * (lod + loq)
+    result[r == "nq"] <- nq
+  } else {
+    if (any(r == "nq")) stop("You need to specify lod and loq")
+  }
+
+  #  Handle nd values
+  if (time_zero) {
+    if (r[1] == "nd") {
+      residues_present = FALSE
+      result[1] <- time_zero_nd_value
+    } else {
+      residues_present = TRUE
+    }
+    start_i <- 2
+  } else {
+    residues_present <- if (r[1] == "nd") FALSE else TRUE
+    start_i <- 1
+  }
+
+  for (i in start_i:length(r)) {
+
+    # residues_in_next
+    if (i < length(r)) {
+      next_value <- r[i + 1]
+      if (next_value == "nd") residues_in_next = FALSE
+      else residues_in_next = TRUE
+    } else {
+      residues_in_next = FALSE
+    }
+
+    if (r[i] == "nd") {
+      if (residues_present | residues_in_next) {
+        result[i] <- 0.5 * lod
+      } else {
+        result[i] <- NA
+      }
+
+      if (stop_after_first_nondetect) {
+        if (residues_present & !residues_in_next) {
+          remaining <- (i + 1):length(r)
+          if (!any(suppressWarnings(as.numeric(r[remaining])) > loq, na.rm = TRUE)) {
+            result[remaining] <- NA
+            return(as.numeric(result))
+          }
+        }
+
+      }
+      if (!residues_in_next) residues_present <- FALSE
+      else residues_present <- TRUE
+    }
+
+  }
+  return(as.numeric(result))
+}
+
+#' @describeIn set_nd Set non-detects in residues series according to FOCUS rules
+#' @export
+set_nd_focus <- function(res, lod, loq = NA, time_zero = TRUE) {
+  result <- set_nd(res, lod, loq = loq, time_zero = time_zero,
+    time_zero_nd_value = 0, stop_after_first_nondetect = TRUE)
+  return(result)
+}
diff --git a/build.log b/build.log
index 6048d87..e001d03 100644
--- a/build.log
+++ b/build.log
@@ -1,9 +1,12 @@
-* checking for file './DESCRIPTION' ... OK
-* preparing 'pfm':
+* checking for file ‘./DESCRIPTION’ ... OK
+* preparing ‘pfm’:
 * checking DESCRIPTION meta-information ... OK
 * checking for LF line-endings in source and make files and shell scripts
 * checking for empty or unneeded directories
-* looking to see if a 'data/datalist' file should be added
+Removed empty directory ‘pfm/inst/testdata/SwashProjects/Project_1/MACRO’
+Removed empty directory ‘pfm/inst/testdata/SwashProjects/Project_1’
+Removed empty directory ‘pfm/inst/testdata/SwashProjects’
+* looking to see if a ‘data/datalist’ file should be added
   NB: this package now depends on R (>= 3.5.0)
   WARNING: Added dependency on R >= 3.5.0 because serialized objects in  serialize/load version 3 cannot be read in older versions of R.  File(s) containing such objects:  'pfm/data/EFSA_GW_interception_2014.RData'  WARNING: Added dependency on R >= 3.5.0 because serialized objects in  serialize/load version 3 cannot be read in older versions of R.  File(s) containing such objects: 'pfm/data/EFSA_washoff_2017.RData'
 * building 'pfm_0.5.7.tar.gz'
diff --git a/man/set_nd.Rd b/man/set_nd.Rd
new file mode 100644
index 0000000..8c766ac
--- /dev/null
+++ b/man/set_nd.Rd
@@ -0,0 +1,65 @@
+% Generated by roxygen2: do not edit by hand
+% Please edit documentation in R/process_residues.R
+\name{set_nd}
+\alias{set_nd}
+\alias{set_nd_focus}
+\title{Set non-detects in residue series without replicates
+Sets non-detects directly before or directly after detects to NA. Values between
+lod and loq are set to their mean value if an loq is specified.
+If 'time_zero' is set to TRUE, the residue series is assumed to start with time
+zero, and non-detects at time zero are set to 'time_zero_nd_value'. For the
+set_nd_focus variant, this is zero, otherwise this argument has NA as default
+value.
+If stopping after the first non-detection is requested, as in in the FOCUS
+variant of the function, an loq has to be specified in order to decide
+if any later detections are above the loq.}
+\usage{
+set_nd(r, lod, loq = NA, time_zero = TRUE, time_zero_nd_value = NA,
+  stop_after_first_nondetect = FALSE)
+
+set_nd_focus(res, lod, loq = NA, time_zero = TRUE)
+}
+\arguments{
+\item{r}{A character vector of sequential residues without replicates, with
+non-detects specified as 'nd' and unquantified values above the limit of
+detection specified as 'nq', otherwise coercible to numeric}
+
+\item{lod}{Limit of detection (numeric)}
+
+\item{loq}{Limit of quantification(numeric). Must be specified if the FOCUS rule to
+stop after the first non-detection is to be applied}
+
+\item{stop_after_first_nondetect}{Should we really stop after the first non-detection?}
+}
+\description{
+Set non-detects in residue series without replicates
+Sets non-detects directly before or directly after detects to NA. Values between
+lod and loq are set to their mean value if an loq is specified.
+If 'time_zero' is set to TRUE, the residue series is assumed to start with time
+zero, and non-detects at time zero are set to 'time_zero_nd_value'. For the
+set_nd_focus variant, this is zero, otherwise this argument has NA as default
+value.
+If stopping after the first non-detection is requested, as in in the FOCUS
+variant of the function, an loq has to be specified in order to decide
+if any later detections are above the loq.
+}
+\section{Functions}{
+\itemize{
+\item \code{set_nd_focus}: Set non-detects in residues series according to FOCUS rules
+}}
+
+\examples{
+parent_1 <- c(.12, .09, .05, .03, "nd", "nd", "nd", "nd", "nd", "nd")
+set_nd(parent_1, 0.02)
+parent_2 <- c(.12, .09, .05, .03, "nd", "nd", .03, "nd", "nd", "nd")
+set_nd(parent_2, 0.02)
+set_nd_focus(parent_2, 0.02, loq = 0.05)
+parent_3 <- c(.12, .09, .05, .03, "nd", "nd", .06, "nd", "nd", "nd")
+set_nd(parent_3, 0.02)
+set_nd_focus(parent_3, 0.02, loq = 0.05)
+metabolite <- c("nd", "nd", "nd", 0.03, 0.06, 0.10, 0.11, 0.10, 0.09, 0.05, 0.03, "nd", "nd")
+set_nd(metabolite, 0.02)
+}
+\references{
+FOCUS (2014) p. 75, 76, 131, 132
+}