Merge pull request #1234 from wadpac/garmin

Add UNIXmsec as additional timestamp format option for read.myacc.csv functionality
wadpac · Nov 25, 2024 · fed05c7 · fed05c7
2 parents 2cc32ce + 5d0b308
commit fed05c7
Show file tree

Hide file tree

Showing 7 changed files with 78 additions and 27 deletions.
diff --git a/NEWS.md b/NEWS.md
@@ -4,6 +4,8 @@
 
 - Part 2: Code revisions in preparation for expansion of functionality to better facilitate external function produced event data. #653 and #1228
 
+- Part 1: Ad-hoc csv file formats (read.myacc.csv functionality) now also excepts "UNIXmsec" as optional timestamp format. #1233
+
 - Part 5: Improved handling of partially available and non-available qwindow segments in the recording #1229
 
 # CHANGES IN GGIR VERSION 3.1-6

diff --git a/R/appendRecords.R b/R/appendRecords.R
@@ -25,27 +25,28 @@ appendRecords = function(metadatadir, desiredtz = "", idloc = 1, maxRecordingInt
         } else {
           N = overlap - 1
         }
-        NEWSHORT = df[nrow(df), ]
-        NEWSHORT[1:N, ] = NEWSHORT
-        T0 = as.POSIXct(x = NEWSHORT$timestamp[1], format = "%Y-%m-%dT%H:%M:%S%z", tz = tz) + epochSize
-        NEWSHORT$timestamp = POSIXtime2iso8601(seq(T0, T0 + (N * epochSize) - 1, by =  epochSize), tz)
-        if (pattern != "long") {
-          NEWSHORT[, grep(pattern = pattern, x = colnames(NEWSHORT), invert = TRUE, ignore.case = TRUE)] = 0
-        } else {
-          enCol = grep("en", colnames(NEWSHORT), ignore.case = TRUE)
-          NEWSHORT[, enCol] = 1
-          NEWSHORT$clippingscore = 0
-          NEWSHORT$nonwearscore = 3
-          if (any(colnames(NEWSHORT) %in% c("lightmean", "lightpeak"))) {
-            NEWSHORT$lightmean = 0
-            NEWSHORT$lightpeak = 0
-          }
-          if ("temperaturemean" %in% colnames(NEWSHORT)) {
-            NEWSHORT$temperaturemean = 0
+        if (N > 0) {
+          NEWSHORT = df[nrow(df), ]
+          NEWSHORT[1:N, ] = NEWSHORT
+          T0 = as.POSIXct(x = NEWSHORT$timestamp[1], format = "%Y-%m-%dT%H:%M:%S%z", tz = tz) + epochSize
+          NEWSHORT$timestamp = POSIXtime2iso8601(seq(T0, T0 + (N * epochSize) - 1, by =  epochSize), tz)
+          if (pattern != "long") {
+            NEWSHORT[, grep(pattern = pattern, x = colnames(NEWSHORT), invert = TRUE, ignore.case = TRUE)] = 0
+          } else {
+            enCol = grep("en", colnames(NEWSHORT), ignore.case = TRUE)
+            NEWSHORT[, enCol] = 1
+            NEWSHORT$clippingscore = 0
+            NEWSHORT$nonwearscore = 3
+            if (any(colnames(NEWSHORT) %in% c("lightmean", "lightpeak"))) {
+              NEWSHORT$lightmean = 0
+              NEWSHORT$lightpeak = 0
+            }
+            if ("temperaturemean" %in% colnames(NEWSHORT)) {
+              NEWSHORT$temperaturemean = 0
+            }
           }
+          df = rbind(df, NEWSHORT)
         }
-
-        df = rbind(df, NEWSHORT)
         return(df)
       }
       # insert missing values in metashort

diff --git a/R/g.part5.R b/R/g.part5.R
@@ -260,8 +260,14 @@ g.part5 = function(datadir = c(), metadatadir = c(), f0=c(), f1=c(),
                               desiredtz = params_general[["desiredtz"]],
                               sibDefinition = sibDef,
                               nightsi)
-          # Fix missing nights in part 4 data:
-          summarysleep_tmp2 = g.part5.fixmissingnight(summarysleep_tmp2, sleeplog = sleeplog, ID)
+          if (nrow(summarysleep_tmp2) > 0) {
+            if (!all(is.na(summarysleep_tmp$sleepparam))) {
+              # Fix missing nights in part 4 data:
+              summarysleep_tmp2 = g.part5.fixmissingnight(summarysleep_tmp2, sleeplog = sleeplog, ID)
+            } else {
+              summarysleep_tmp2 = summarysleep_tmp2[-which(is.na(summarysleep_tmp$sleepparam)), ]
+            }
+          }
           #Initialise diur variable, which will  indicate the diurnal rhythm: 0 if wake/daytime, 1 if sleep/nighttime
           ts$diur = 0
           if (nrow(summarysleep_tmp2) > 0) {

diff --git a/R/read.myacc.csv.R b/R/read.myacc.csv.R
@@ -26,8 +26,9 @@ read.myacc.csv = function(rmc.file=c(), rmc.nrow=Inf, rmc.skip=c(), rmc.dec=".",
                           configtz = NULL,
                           header = NULL) {
 
-  if (length(rmc.col.time) > 0 && !(rmc.unit.time %in% c("POSIX", "character", "UNIXsec", "ActivPAL"))) {
-    stop("\nUnrecognized rmc.col.time value. The only accepted values are \"POSIX\", \"character\", \"UNIXsec\", and \"ActivPAL\".", call. = FALSE)
+  if (length(rmc.col.time) > 0 && !(rmc.unit.time %in% c("POSIX", "character", "UNIXsec", "UNIXmsec", "ActivPAL"))) {
+    stop(paste0("\nUnrecognized rmc.col.time value. The only accepted values are \"POSIX\", ",
+                "\"character\", \"UNIXsec\", \"UNIXmsec\", and \"ActivPAL\"."), call. = FALSE)
   }
 
   if (!is.null(rmc.desiredtz) || !is.null(rmc.configtz)) {
@@ -257,8 +258,13 @@ read.myacc.csv = function(rmc.file=c(), rmc.nrow=Inf, rmc.skip=c(), rmc.dec=".",
       }
     } else if (rmc.unit.time == "character") {
       P$time = as.POSIXct(P$time, format = rmc.format.time, origin = rmc.origin, tz = configtz)
-    } else if (rmc.unit.time == "UNIXsec" && rmc.origin != "1970-01-01") {
-      P$time = as.POSIXct(P$time, origin = rmc.origin, tz = desiredtz)
+    } else if (rmc.unit.time == "UNIXsec" || rmc.unit.time == "UNIXmsec") {
+      if (rmc.unit.time == "UNIXmsec") {
+        P$time = P$time / 1000
+      }
+      if (rmc.origin != "1970-01-01") {
+        P$time = as.POSIXct(P$time, origin = rmc.origin, tz = desiredtz)
+      }
     } else if (rmc.unit.time == "ActivPAL") {
       # origin should be specified as: "1899-12-30"
       rmc.origin = "1899-12-30"

diff --git a/man/GGIR.Rd b/man/GGIR.Rd
@@ -433,7 +433,8 @@ GGIR(mode = 1:5,
 
       \item{rmc.unit.time}{
         Character (default = "POSIX").
-        Character with unit of timestamps: "POSIX", "UNIXsec" (seconds since origin, see argument \code{rmc.origin}),
+        Character with unit of timestamps: "POSIX", "UNIXsec" (seconds since origin, 
+        see argument \code{rmc.origin}), "UNIXmsec" (same as UNIXsec but in milliseconds),
         "character", or "ActivPAL" (exotic timestamp format only used in the ActivPAL
         activity monitor).}
 

diff --git a/tests/testthat/test_read.myacc.csv.R b/tests/testthat/test_read.myacc.csv.R
@@ -47,6 +47,18 @@ test_that("read.myacc.csv can handle files without header, no decimal places in
   testfile[4] = "testcsv4.csv"
   write.csv(S3, file = testfile[4], row.names = FALSE)
 
+  # UNIXsec
+  testfile[5] = "testcsv5.csv"
+  S5 = S1
+  S5$time = as.numeric(S5$time)
+  write.csv(S5, file = testfile[5], row.names = FALSE)
+
+  # UNIXmsec
+  testfile[6] = "testcsv6.csv"
+  S6 = S5
+  S6$time = S6$time * 1000
+  write.csv(S6, file = testfile[6], row.names = FALSE)
+
   # attempt to load
   D1 = read.myacc.csv(rmc.file = testfile[1], rmc.nrow = 20, rmc.dec = ".",
                       rmc.firstrow.acc = 1, rmc.firstrow.header = c(),
@@ -196,6 +208,29 @@ test_that("read.myacc.csv can handle files without header, no decimal places in
   expect_that(ncol(D4$data), equals(4))
   expect_that(D4$header, equals("no header"))
 
+  # Timestamps in UNIXsec
+  D5 = read.myacc.csv(rmc.file = testfile[5], rmc.nrow = 20, rmc.dec = ".",
+                      rmc.firstrow.acc = 1, rmc.firstrow.header = c(),
+                      rmc.col.acc = c(1,3,4), rmc.col.temp = 5, rmc.col.time = 2,
+                      rmc.unit.acc = "g", 
+                      rmc.unit.time = "UNIXsec",
+                      desiredtz = "Europe/London", rmc.sf = sf)
+
+  expect_equal(mean(D5$data$x), 0.1078671, tol = 0.0001)
+  expect_equal(mean(D5$data$y), -0.06675716, tol = 0.0001)
+  expect_equal(D5$data$time[1], 1667394076, tol = 0.1)
+
+  # Timestamps in UNIXmsec
+  D6 = read.myacc.csv(rmc.file = testfile[6], rmc.nrow = 20, rmc.dec = ".",
+                      rmc.firstrow.acc = 1, rmc.firstrow.header = c(),
+                      rmc.col.acc = c(1,3,4), rmc.col.temp = 5, rmc.col.time = 2,
+                      rmc.unit.acc = "g", 
+                      rmc.unit.time = "UNIXmsec",
+                      desiredtz = "Europe/London", rmc.sf = sf)
+  expect_equal(mean(D6$data$x), 0.1078671, tol = 0.0001)
+  expect_equal(mean(D6$data$y), -0.06675716, tol = 0.0001)
+  expect_equal(D6$data$time[1], 1667394076, tol = 0.1)
+
   for (i in 1:length(testfile)) {
     expect_true(file.exists(testfile[i]))
     if (file.exists(testfile[i])) file.remove(testfile[i])

diff --git a/vignettes/readmyacccsv.Rmd b/vignettes/readmyacccsv.Rmd
@@ -69,7 +69,7 @@ Below we present a summary of the available input arguments. Please see the [par
 ### Arguments for files including timestamps
 
 - `rmc.col.time` - Scalar with column (number) in which the timestamps are stored. Leave in default setting if timestamps are not stored.
-- `rmc.unit.time` - Character with unit of timestamps: "POSIX", "UNIXsec" (seconds since origin, see argument rmc.origin), "character", or "ActivPAL" (exotic timestamp format only used in the ActivPAL activity monitor).
+- `rmc.unit.time` - Character with unit of timestamps: "POSIX", "UNIXsec" (seconds since origin, see argument rmc.origin), "UNIXmsec" (milliseconds since origin, see argument rmc.origin), "character", or "ActivPAL" (exotic timestamp format only used in the ActivPAL activity monitor).
 - `rmc.format.time` - Character string giving a date-time format as used by \link[base]{strptime}. Only used for rmc.unit.time: character and POSIX.
 - `rmc.origin` - Origin of time when unit of time is UNIXsec, e.g. 1970-1-1.