Merge pull request #21 from habitus-eu/issue20_redundantcolumn

Fixes: redundant column and trip duration one epoch too short

.Rbuildignore

@@ -3,3 +3,4 @@
 ^\.github
 prepareNewRelease.R
 LICENSE
+codecov.yml

DESCRIPTION

@@ -1,8 +1,8 @@
 Package: hbGPS
 Type: Package
 Title: Process GPS Data For Human Behaviour Research
-Version: 0.0.4
-Date: 2023-09-28
+Version: 0.0.6
+Date: 2024-03-20
 Authors@R: c(person("Vincent T","van Hees",role=c("aut","cre"),
                   email="[email protected]"),
              person("SDU Denmark",  role = c("cph", "fnd")))

R/deriveTrips.R

@@ -48,7 +48,7 @@ deriveTrips = function(df, tz, minTripDur, minTripDist_m) {
     s1 = as.POSIXct(Segs$time_t1[tripEnd[j]], tz = tz, origin = "1970-01-01")
     tripInd = which(df$time >= s0 & df$time <= s1)
     # trip duration in minutes
-    tripDur = (Segs$time_t1[tripEnd[j]] - Segs$time_t0[tripStart[j]])
+    tripDur = diff(as.numeric(range(df$time[c(tripInd, tripInd[length(tripInd)] + 1)])))
     tripDist_m = sum(abs(df$distance_m[tripInd]))
     maxTimeGap = max(df$deltaTime[tripInd])
     tripIncl = mean(df$inclination_deg[tripInd])

R/mergeGGIR.R

@@ -4,7 +4,7 @@ mergeGGIR = function(GGIRpath, GPSdf, ID, verbose) {
 
   # If user by accident specified GGIR output folder then attempt
   # to update path to specify the ms5.outraw subfolder inside it
-  if (length(grep(pattern = "ms5.outraw", x = GGIRpath)) != 0) {
+  if (length(grep(pattern = "ms5.outraw", x = GGIRpath)) == 0) {
     newGGIRpath = paste0(GGIRpath, "/meta/ms5.outraw")
     if (dir.exists(newGGIRpath)) {
       GGIRpath = newGGIRpath
@@ -47,7 +47,9 @@ mergeGGIR = function(GGIRpath, GPSdf, ID, verbose) {
           # Linearly interpolate acceleration
           GPSdf$GGIR_ACC = GGIRread::resample(raw = as.matrix(G$ACC), rawTime = G$timenum, time = GPSdf$time, stop = nrow(G), type = 1)
           # Nearest neigbour interpolate other GGIR output columns
-          col2impute = c("SleepPeriodTime", "invalidepoch", "guider", "window", "class_id")
+          if ("class_id" %in% colnames(G) == FALSE) stop("GGIR time series misses class_id column")
+          if ("invalidepoch" %in% colnames(G) == FALSE) stop("GGIR time series misses invalidepoch column")
+          col2impute = colnames(G)[which(colnames(G) %in% c("SleepPeriodTime", "invalidepoch", "window", "class_id") == TRUE)]
           GS = as.data.frame(GGIRread::resample(raw = as.matrix(G[, col2impute]),
                                                 rawTime = G$timenum,
                                                 time = GPSdf$time, stop = nrow(G), type = 2))

inst/NEWS.Rd

@@ -1,6 +1,13 @@
 \name{NEWS}
 \title{News for Package \pkg{GGIRread}}
 \newcommand{\cpkg}{\href{http://CRAN.R-project.org/package=#1}{\pkg{#1}}}
+\section{Changes in version 0.0.6 (release date:20-03-2024)}{
+    \itemize{
+      \item Simplify mergeGGIR to no longer expect window column, see #20
+      \item Allow mergeGGIR to handle both full path to ms5.outraw folder or 
+      only the GGIR output folder
+    }
+}
 \section{Changes in version 0.0.5 (release date:05-03-2024)}{
     \itemize{
       \item More formats for GPS satillitate info now supported, see #12

tests/testthat/test_hbGPS.R

@@ -28,9 +28,9 @@ test_that("hbGPS pipeline process file 4 correctly", {
             return_object = TRUE)
 
   expect_equal(nrow(D), 20029)
-  expect_equal(ncol(D), 44)
+  expect_equal(ncol(D), 43)
 
-  expect_equal(length(unique(D$trip)), 39)
+  expect_equal(length(unique(D$trip)), 40)
   expect_equal(mean(D$GGIR_ACC), 26.19259, tolerance = 0.0001)
   expect_equal(mean(D$snr), 256.9834, tolerance = 0.0001)
   expect_equal(mean(D$snr_ratio), 78.81412, tolerance = 0.0001)
@@ -43,11 +43,13 @@ test_that("hbGPS pipeline process file 4 correctly", {
   expect_equal(mean(D$bearing_deg), 10.62006, tolerance = 0.00001)
   expect_equal(mean(D$deltaBearing), 65.73111, tolerance = 0.00001)
 
-  expect_equal(mean(D$tripMaxTimegGap), 0.7691348, tolerance = 0.000001)
-  expect_equal(mean(D$tripEleSpeed_kmh), 0.03796242, tolerance = 0.000001)
+  expect_equal(mean(D$tripMaxTimegGap), 0.7728793, tolerance = 0.000001)
+  expect_equal(mean(D$tripEleSpeed_kmh), 0.03658415, tolerance = 0.000001)
 
   # Finally a crude global check to see whether anything has changed
-  expect_equal(sum(rowSums(D[,c(11:15, 17:41)]), na.rm = TRUE), 26828374, tolerance = 0.1)
+  num_cols = which(unlist(lapply(D, is.numeric), use.names = FALSE) == TRUE)
+  num_cols = num_cols[num_cols %in% which(colnames(D) == "timenum") == FALSE]
+  expect_equal(sum(rowSums(D[, num_cols], na.rm = TRUE)), 27142209, tolerance = 0.1)
 
   # Check that it also works with a configFile
   DC = hbGPS(gps_file = gps_file,
@@ -59,8 +61,8 @@ test_that("hbGPS pipeline process file 4 correctly", {
             configFile =  system.file("testfiles/config_hbGPS.csv", package = "hbGPS"),
             return_object = TRUE)
   expect_equal(nrow(DC), 20029)
-  expect_equal(ncol(DC), 44)
-  expect_equal(length(unique(DC$trip)), 39)
+  expect_equal(ncol(DC), 43)
+  expect_equal(length(unique(DC$trip)), 40)
   expect_equal(mean(DC$GGIR_ACC), 26.19259, tolerance = 0.0001)
   expect_equal(mean(DC$speed_ms), 0.6386111, tolerance = 0.0001)