R/AIRS_Data_Import.R

In crtahlin/AIRS_Data_Import: Importing AIRS exported CSV data into R

#' @title Import multiple AIRS files
#' @description  A function that imports multiple AIRS data files into one dataframe. 
#' 
#' @param dataFileAIRS Dataframe containing info about uploaded files via shiny::fileInput().
#' @param sensor The sensor data to import.
#' @export
AIRSImportMultiple <- function (dataFilesAIRS, sensor) {
  data <- data.frame()
  for(file in dataFilesAIRS$datapath) {
    tempData <- AIRSImportExtended(filePath=file, sensor=sensor)
    data <- rbind(data, tempData)
  }
  return(data)
}


#' @title Import AIRS data exported as CSV from an Android phone
#' @description A function to import data for a certain sensor from and AIRS 
#' generated CSV file into R.
#' 
#' @details 
#' Currently implemented specificaly for "HP" aka "heart pulse" sensor 
#' (i.e. Zephyr bluetooth). It might in principle work on other sensor data,
#'  although the columns will surely not get proper names.
#'  
#' @param filePath The path to the CSV file containing AIRS exported data.
#' Use forward slashes to separate folders (e.g. "/").
#' 
#' @param sensor The abbreviation representing the sensor. I.e "HP" for heart pulse (heart rate).
#' 
#' @export 
AIRSImport <- function( filePath, 
                        sensor) {
  # read data from CSV file
  data <- readLines(con=filePath)
  
  # read first line of AIRS expoted data for the initial date
  # temporarily set locale to "C" for date recognition
  lct <- Sys.getlocale("LC_TIME"); Sys.setlocale("LC_TIME", "C")
  initialDate <- data[1]
  
  # remove extra date information, which breaks the code
  # this extra info was added sometime in 2014?
  temp <- strsplit(initialDate, " GMT")
  temp[[1]][2] <- gsub(temp[[1]][2], pattern=":", replacement="")
  initialDate <- paste(temp[[1]][1], temp[[1]][2])
  # end of hacking section
  
  initialDate <- as.POSIXct(initialDate, format="%a %B %d %H:%M:%S %z %Y")
  Sys.setlocale("LC_TIME", lct)
  
  # find lines with patterns that match desired sensor 
  validEntries <- grep(pattern=paste0(";",sensor,";"),x=data)
  
  # read data into dataframe
  elapsedTime <- system.time({
    temp <- gsub(x = data[validEntries], pattern = "#", replacement = "")
    assign( x=paste0("AIRS_", sensor),
            value=read.table(text=temp,
                             sep=";", 
                             comment.char="@" )
    )
    rm(temp)
    rm(data)
  })
  print(elapsedTime)
  rm(elapsedTime)
  # name the columns of the HP measurements
  if (sensor=="HP") {
    colnames(AIRS_HP) <- c("MilisecsSinceOrigin","Sensor","Value")
  }
  
  # add a date column to data
  Date <- vector(length=dim(get(paste0("AIRS_", sensor)))[1], mode="numeric")
  Date <- get(paste0("AIRS_", sensor))$MilisecsSinceOrigin/1000 + initialDate
  assign( x=paste0("AIRS_", sensor),
          value=cbind(get(paste0("AIRS_", sensor)), Date)
  )
  rm(Date)
  # Why did I choose to return AIRS+sensor names specificaly? Just to complicate things?
  return(get(paste0("AIRS_", sensor)))
}


#' @title Import AIRS data exported as CSV from an Android phone and add extra information
#' @description A function to import data for a certain sensor from and AIRS 
#' generated CSV file into R. The function adds extra columns to imported data 
#' for easier data manipulation.
#' 
#' @details 
#' Currently implemented specificaly for "HP" aka "heart pulse" sensor 
#' (i.e. Zephyr bluetooth). It might in principle work on other sensor data,
#'  although the columns will surely not get proper names.
#'  
#' @param filePath The path to the CSV file containing AIRS exported data.
#' Use forward slashes to separate folders (e.g. "/").
#' 
#' @param sensor The abbreviation representing the sensor. I.e "HP" for heart pulse (heart rate).
#' 
#' @export 
AIRSImportExtended <- function(filePath, sensor) {
  data <- AIRSImport(
    filePath=filePath,
    sensor=sensor)
  data$TimeinSeconds <- as.POSIXlt(data$Date)[["sec"]] + 
    as.POSIXlt(data$Date)[["min"]]*60 +
    as.POSIXlt(data$Date)[["hour"]]*(60*60)
  
  # head(data)
  
  data$"CyclicTime" <- (data$"TimeinSeconds" / (60*60*24)) %% 1
  data$"Weekday" <- c("Sunday", "Monday", "Tuesday", "Wednesday", "Thursday", 
                      "Friday", "Saturday")[as.POSIXlt(data$Date)$wday + 1]
  # TODO: check if time zones are correctly intepredted when calculating data$"Day"
  data$"Day" <- format(data$Date, format="%d.%m.%Y")
  return(data)
  
}

########## Below are functions for debugging #############
# 
# # set path to CSV file exported form AIRS
# filePath <- "C:/Users/Crt Ahlin/Documents/Dropbox/DataSets/AIRS/1385030889684_2.txt"
# 
# # set desired sensor
# sensor <- "HP"
# 
# # extract the timestamp of data export from CSV file name
# # assumptions: the origin is "1970-01-01 00:00.00 CET"
# string <- regexpr(pattern="/?[[:digit:]]+.txt", text=filePath)
# string <- string + 1
# attr(string, "match.length") <- attr(string,"match.length") - 5
# miliseconds <- regmatches(filePath, string)
# seconds <- as.double(miliseconds)/1000
# str(miliseconds)
# fileTimestamp <- as.POSIXlt(x=seconds, tz="CET", origin="1970-01-01 00:00.00 CET")
# fileTimestampMiliseconds <- miliseconds
# 
# # list valid entries for chosen sensor
# data[validEntries]
# 
# # name the columns of the HP measurements
# colnames(AIRS_HP) <- c("Miliseconds","Sensor","Value")
# 
# head(AIRS_HP)
# str(AIRS_HP)
# 
# plot(AIRS_HP$Value ~ AIRS_HP$Miliseconds,
#      type="l")
#