add Fitbit json read functionalty #68

wadpac · Sep 27, 2024 · 207d712 · 207d712
1 parent e452535
commit 207d712
Show file tree

Hide file tree

Showing 9 changed files with 586 additions and 3 deletions.
diff --git a/DESCRIPTION b/DESCRIPTION
@@ -22,7 +22,7 @@ URL: https://github.com/wadpac/GGIRread/
 BugReports: https://github.com/wadpac/GGIRread/issues
 License: Apache License (== 2.0)
 Suggests: testthat
-Imports: matlab, bitops, Rcpp (>= 0.12.10), data.table, readxl
+Imports: matlab, bitops, Rcpp (>= 0.12.10), data.table, readxl, jsonlite
 Depends: stats, utils, R (>= 3.5.0)
 NeedsCompilation: yes
 LinkingTo: Rcpp

diff --git a/NEWS.md b/NEWS.md
@@ -5,7 +5,7 @@
 - Migrate read function for ActiGraph (csv) and Actiwatch (csv/awd) count data to GGIRread #68.
 - Add function for reading Actical (csv) count data #68.
 - Add function to read and merge Philips Health Band file pairs (xlsx) #68.
-
+- Add function for reading Fitbit (json) with sleep, steps or calories #68.
 
 # Changes in version 1.0.1 (release date:03-06-2024)
 

diff --git a/R/readFitbit.R b/R/readFitbit.R
@@ -0,0 +1,85 @@
+readFitbit = function(filename = NULL) {
+  # Assumptions made:
+  # - sleep is sampled at 30 second resolution
+  # - steps are sampled at 60 second resolution
+  # - timestamp format is always the same per data type
+
+  # Declare local functions
+  handleTimeGaps = function(df, epochSize) {
+    timeRange = range(df$dateTime)
+    startTime = timeRange[1]
+    endTime = timeRange[2]
+    timeFrame = data.frame(dateTime = seq(startTime, endTime, by = epochSize))
+    df = merge(df, timeFrame, by = c("dateTime"), all.x = TRUE)
+    return(df)
+  }
+
+  #-------------------------------------------------
+  # Main code
+
+  D = jsonlite::read_json(path = filename,
+                simplifyVector = FALSE,
+                flatten = FALSE)
+
+  # extract dataType as json structure differs between types
+  dataType = tolower(unlist(strsplit(basename(filename), "-"))[1])
+
+  if (dataType == "sleep") {
+    epochSize = 30
+    # Put all data in data.frame
+    for (i in 1:length(D)) {
+      tmp = D[[i]][15]$levels
+      data = as.data.frame(data.table::rbindlist(tmp$data, fill = TRUE))
+      data$dateTime = as.POSIXct(data$dateTime, format = "%Y-%m-%dT%H:%M:%S")
+      if (i == 1) {
+        all_data = data
+      } else {
+        all_data = rbind(all_data, data)
+      }
+      if ("shortData" %in% names(tmp)) {
+        shortData = data.table::rbindlist(tmp$shortData, fill = TRUE)
+        shortData$dateTime = as.POSIXct(shortData$dateTime, format = "%Y-%m-%dT%H:%M:%S")
+        if (i == 1) {
+          all_shortData = shortData
+        } else {
+          all_shortData = rbind(all_shortData, shortData)
+        }
+      }
+    }
+    # Expand to full time series
+    D = as.data.frame(lapply(all_data, rep, all_data$seconds/epochSize))
+    D$dateTime = seq(from = D$dateTime[1], length.out = nrow(D), by = epochSize) 
+    D$seconds = epochSize
+    D = handleTimeGaps(D, epochSize) # Handle time gaps, if any
+
+    S = as.data.frame(lapply(all_shortData, rep, all_shortData$seconds/30))
+    S$dateTime = seq(from = S$dateTime[1], length.out = nrow(S), by = 30) 
+    S$seconds = epochSize
+
+    # merge in shortData (S)
+    matching_times = which(S$dateTime %in% D$dateTime ==  TRUE)
+    non_matching_times = which(S$dateTime %in% D$dateTime == FALSE)
+    if (length(matching_times) > 0) {
+      times_to_replace = S$dateTime[matching_times]
+      D[which(D$dateTime %in% times_to_replace), ] = S[matching_times,]
+    }
+    if (length(non_matching_times) > 0) {
+      D = rbind(D, S[non_matching_times,])
+    }
+    D = handleTimeGaps(D, epochSize) # Handle new time gaps, if any
+
+    # Order time stamps
+    D = D[order(D$dateTime), ]
+    colnames(D)[2] = "sleeplevel"
+  } else if (dataType == "steps" || dataType == "calories") {
+    epochSize = 60
+    data = as.data.frame(data.table::rbindlist(D, fill = TRUE))
+    data$dateTime = as.POSIXct(data$dateTime, format = "%m/%d/%y %H:%M:%S")
+    D = handleTimeGaps(data, epochSize = 60)
+    D$value = as.numeric(D$value)
+    colnames(D)[2] = dataType
+  } else {
+    stop("File type not recognised")
+  }
+  return(D)
+}
diff --git a/README.md b/README.md
@@ -12,4 +12,6 @@ ActivInsights Ltd https://activinsights.com/ | GENEActiv Original and Sleep | .b
 Unilever Discover Ltd | Genea (no longer manufactured) | .bin | raw gravitational units | readGenea
 ActiGraph | ??? | .csv | count data | readActigraph
 Actiwatch | ??? | .csv and .awd | count data | readActiwatch
-Actical | ??? | .csv | count data | readActical
+Actical | ??? | .csv | count data | readActical.R
+Philips Health Band | ??? | .xlsx | count data | mergePHBfilePairs.R
+Fitbit | ??? | .json | sleep, steps or calories data | readFitbit.R
diff --git a/inst/testfiles/calories-1995-06-23_Fitbit.json b/inst/testfiles/calories-1995-06-23_Fitbit.json
@@ -0,0 +1,142 @@
+[{
+  "dateTime" : "06/23/95 00:00:00",
+  "value" : "1.48"
+},{
+  "dateTime" : "06/23/95 00:01:00",
+  "value" : "1.48"
+},{
+  "dateTime" : "06/23/95 00:02:00",
+  "value" : "1.48"
+},{
+  "dateTime" : "06/23/95 00:03:00",
+  "value" : "1.48"
+},{
+  "dateTime" : "06/23/95 00:04:00",
+  "value" : "1.48"
+},{
+  "dateTime" : "06/23/95 00:05:00",
+  "value" : "1.48"
+},{
+  "dateTime" : "06/23/95 00:06:00",
+  "value" : "1.48"
+},{
+  "dateTime" : "06/23/95 00:07:00",
+  "value" : "1.48"
+},{
+  "dateTime" : "06/23/95 00:08:00",
+  "value" : "1.48"
+},{
+  "dateTime" : "06/23/95 00:09:00",
+  "value" : "1.48"
+},{
+  "dateTime" : "06/23/95 00:10:00",
+  "value" : "1.48"
+},{
+  "dateTime" : "06/23/95 00:11:00",
+  "value" : "1.48"
+},{
+  "dateTime" : "06/23/95 00:12:00",
+  "value" : "1.48"
+},{
+  "dateTime" : "06/23/95 00:13:00",
+  "value" : "1.48"
+},{
+  "dateTime" : "06/23/95 00:14:00",
+  "value" : "1.48"
+},{
+  "dateTime" : "06/23/95 00:15:00",
+  "value" : "1.48"
+},{
+  "dateTime" : "06/23/95 00:16:00",
+  "value" : "1.48"
+},{
+  "dateTime" : "06/23/95 00:17:00",
+  "value" : "1.48"
+},{
+  "dateTime" : "06/23/95 00:18:00",
+  "value" : "1.48"
+},{
+  "dateTime" : "06/23/95 00:19:00",
+  "value" : "1.48"
+},{
+  "dateTime" : "06/23/95 00:20:00",
+  "value" : "1.48"
+},{
+  "dateTime" : "06/23/95 00:21:00",
+  "value" : "1.48"
+},{
+  "dateTime" : "06/23/95 00:22:00",
+  "value" : "1.48"
+},{
+  "dateTime" : "06/23/95 00:23:00",
+  "value" : "1.48"
+},{
+  "dateTime" : "06/23/95 00:24:00",
+  "value" : "1.48"
+},{
+  "dateTime" : "06/23/95 00:25:00",
+  "value" : "1.48"
+},{
+  "dateTime" : "06/23/95 00:26:00",
+  "value" : "1.48"
+},{
+  "dateTime" : "06/23/95 00:27:00",
+  "value" : "1.48"
+},{
+  "dateTime" : "06/23/95 00:28:00",
+  "value" : "1.48"
+},{
+  "dateTime" : "06/23/95 00:29:00",
+  "value" : "1.48"
+},{
+  "dateTime" : "06/23/95 00:30:00",
+  "value" : "1.48"
+},{
+  "dateTime" : "06/23/95 00:31:00",
+  "value" : "1.48"
+},{
+  "dateTime" : "06/23/95 00:32:00",
+  "value" : "1.48"
+},{
+  "dateTime" : "06/23/95 00:33:00",
+  "value" : "1.48"
+},{
+  "dateTime" : "06/23/95 00:34:00",
+  "value" : "1.48"
+},{
+  "dateTime" : "06/23/95 00:35:00",
+  "value" : "1.48"
+},{
+  "dateTime" : "06/23/95 00:36:00",
+  "value" : "1.48"
+},{
+  "dateTime" : "06/23/95 00:37:00",
+  "value" : "1.48"
+},{
+  "dateTime" : "06/23/95 00:38:00",
+  "value" : "1.48"
+},{
+  "dateTime" : "06/23/95 00:39:00",
+  "value" : "1.48"
+},{
+  "dateTime" : "06/23/95 00:40:00",
+  "value" : "1.48"
+},{
+  "dateTime" : "06/23/95 00:41:00",
+  "value" : "1.48"
+},{
+  "dateTime" : "06/23/95 00:42:00",
+  "value" : "1.48"
+},{
+  "dateTime" : "06/23/95 00:43:00",
+  "value" : "1.48"
+},{
+  "dateTime" : "06/23/95 00:44:00",
+  "value" : "1.48"
+},{
+  "dateTime" : "06/23/95 00:45:00",
+  "value" : "1.48"
+},{
+  "dateTime" : "06/23/95 00:46:00",
+  "value" : "1.48"
+}]