R/formatFields.R

In track2KBA: Identifying Important Areas from Animal Tracking Data

## formatFields ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

#' Format tracking data
#'
#' \code{formatFields} formats the column names of a data frame so that they are
#'  accepted by track2KBA functions.
#'
#' If data are already in format of BirdLife Seabird tracking database
#' (\url{http://www.seabirdtracking.org/}), use \code{formatBL = TRUE} and formatting
#' conversion will occur automatically. I.e., data have following columns:
#' "latitude", "latitude", "date_gmt", "time_gmt". You must still specify the
#' ID column as either the track or animal identifier.
#'
#' By matching up the names of your existing columns with those recognized by
#' track2KBA functions, \code{formatFields} re-formats the data frame, and
#' converts the date/date-time fields into a singe date-time field of class
#' POSIXct.
#'
#'
#' If date-time is combined in a single column, please use \emph{fieldDateTime}
#' instead of \emph{fieldDate} and \emph{fieldTime}.
#'
#' @param dataGroup data.frame or data.table.
#' @param formatBL logical. Is data set already in format of BirdLife Seabird
#' tracking database? If so, indicate \code{TRUE}. \code{fieldID} must still be
#' specified and other fields may be ignored.
#' arguments.
#' @param fieldID character. Unique identifier; e.g. for individuals or
#' dataGroup.
#' @param fieldLat numeric. Name of column corresponding to the LATITUDINAL
#' positions.
#' @param fieldLon numeric. Name of column corresponding to the LONGITUDINAL
#' positions.
#' @param fieldDateTime character. If existing, this is the name of the column
#' corresponding to the combined DATE & TIME.
#' @param fieldDate character. Name of column corresponding to the DATE only.
#' @param fieldTime character. Name of column corresponding to the TIME only.
#' @param formatDT character. What is the format of the data in your DateTime,
#' Date, and Time columns (e.g. "ymd_HMS")? Specify the format following the
#' standard in \code{\link[lubridate]{parse_date_time}}.
#' @param cleanDF logical scalar (T/F). Should columns which are non-essential
#' for track2KBA analysis be removed from dataframe, or not? Removal will speed
#' analysis up a bit.
#'
#' @return Returns a data.frame, with 'ID', 'Latitude', 'Longitude', and
#' 'DateTime' (class POSIXct) columns.
#'
#' @examples
#' ## Load example dataset
#' tracks_raw <- track2KBA::boobies
#' ## using data with user-custom format i.e. with separate Date and Time fields
#' tracks_formatted <- formatFields(
#'   dataGroup = tracks_raw,
#'   fieldID   = "track_id",
#'   fieldLat  ="latitude",
#'   fieldLon  ="longitude",
#'   fieldDate ="date_gmt",
#'   fieldTime ="time_gmt"
#'   )
#' ## using data with only single Date field
#' tracks_formatted <- formatFields(
#'   dataGroup = tracks_raw,
#'   fieldID   = "track_id",
#'   fieldLat  ="latitude",
#'   fieldLon  ="longitude",
#'   fieldDate ="date_gmt",
#'   formatDT = "ymd"
#'   )
#' \dontrun{
#' ## if data were downloaded from Seabird Tracking Database
#' tracks_formatted <- formatFields(
#'   dataGroup=tracks_raw,
#'   formatBL,
#'   fieldID = "bird_id")
#'   }
#'   
#' @export
#' @importFrom rlang .data
#'
formatFields <- function(
  dataGroup, formatBL=FALSE, fieldID, fieldLat, fieldLon, fieldDateTime=NULL,
  fieldDate=NULL, fieldTime=NULL, formatDT=NULL, cleanDF=FALSE
  ) {

  #### INPUT CHECKS

  ## check that df is either a data.frame or a data.table
  if (! "data.frame" %in% class(dataGroup)) {
    if (! "data.table" %in% class(dataGroup)) {
    stop("Object is not a data.table or data.frame. Please try again.")
    }
  }
  #### convert df to data.frame instead of data.table
  dataGroup <- as.data.frame(dataGroup)

  #### if data are already in Seabird Database format use that

  if (formatBL == TRUE) {
    fieldLat  <- "latitude"
    fieldLon  <- "longitude"
    fieldDate <- "date_gmt"
    fieldTime <- "time_gmt"
  }

  # If no fieldID specified and there is an 'ID' field in dataframe, use that -
  if (missing(fieldID)) { # if fieldID missing
    if ("ID" %in% colnames(dataGroup)) { # AND no 'ID' field present in the DF
      message(
      "No fieldID specified, so the pre-existing column named 'ID' used.
        If another field desired as IDentifier, specify in fieldID argument.")
      fieldID <- "ID"
    } else { stop(
    "No fieldID specified. Specify desired IDentifier in the fieldID argument.")
      } # if no fieldID AND no pre-existing 'ID'
  }

  ### check that fieldID, fieldLat, fieldLon supplied are actually colnames ---
  if (! fieldID %in% colnames(dataGroup)) {
    stop("The fieldID supplied does not exist in the data frame.")
  } else if (! fieldLat %in% colnames(dataGroup)) {
    stop("The fieldLat supplied does not exist in the data frame.")
  } else if (! fieldLon %in% colnames(dataGroup)) {
    stop("The fieldLon supplied does not exist in the data frame.")
  }

  ### Date and Time, or DateTime?
  ## check: both Date and Time AND DateTime present? If so... -----------------
  if ((((!is.null(fieldDate)) &
      (!is.null(fieldTime))) &
      (!is.null(fieldDateTime)))) {
    message(
    "Both a DateTime, and separate Date and Time fields supplied. If separate
    Date and Time fields desired, remove the DateTime field from the function
    input argument 'fieldDateTime'.")
    fieldDate <- NULL
    fieldTime <- NULL
  }

  ## --------------------OPTION 1 - DateTime supplied -------------------------
  if (!is.null(fieldDateTime)) {
    if (!lubridate::is.POSIXct(dataGroup[, fieldDateTime])) {
      message(
      "Column supplied to 'fieldDateTime' is not of class POSIXct, the function
      will attempt to convert it.")
      if (is.null(formatDT)) {
        message(
        "No format was supplied for the DateTime field, default format
        ('ymd_HMS') was attempted. If error results, see help page
        ('?lubridate::parse_date_time') for information on formats.")
        formatDT <- "ymd_HMS"
        dataGroup$DateTime <- lubridate::parse_date_time(
          dataGroup[, fieldDateTime], formatDT, tz = "UTC"
          )
        }
      dataGroup$DateTime <- lubridate::parse_date_time(
        dataGroup[, fieldDateTime], formatDT, tz = "UTC"
        )
    } else {
    dataGroup <- dataGroup %>% dplyr::rename(DateTime = fieldDateTime)
    }
  }

  ## ------------- OPTION 2 - Date and Time, or only Date supplied ------------
  # conditions handle input format of Date, or Date and Time columns
  if (is.null(fieldDateTime) & ! is.null(fieldDate) & ! is.null(fieldTime)) {
    # if both Date and Time supplied ------------------------------------------

    if (is.null(formatDT)) {
      # if format of DateTime/(Date + Time) field(s) not supplied, use "ymd_HMS"
      if (formatBL == FALSE) {message(
      "No format supplied for Date and Time fields, a default format ('ymd_HMS')
      attempted when combining the fields. If error produced, see help page
      ('?lubridate::parse_date_time') for information on date formats."
      )
        }
      formatDT <- "ymd_HMS"
      dataGroup$DateTime <- lubridate::parse_date_time(
        paste(dataGroup[, fieldDate], dataGroup[, fieldTime]),
        formatDT, tz = "UTC"
        )
    } else {
    dataGroup$DateTime <- lubridate::parse_date_time(
      paste(dataGroup[, fieldDate], dataGroup[, fieldTime]),
      formatDT, tz = "UTC"
      )
    }
  } else {
    # if only Date supplied (and Date column not missing) ---------------------
    if (! is.null(fieldDate)) {
      message(
      "Only a Date column (fieldDate) supplied, this will be used to create the
      DateTime column. If you have a Time column, indicate it in the 'fieldTime'
      argument.")
      if (is.null(formatDT)) {
        # if format of Date field not supplied, set to default "ymd" ----------
        message(
        "No format supplied for the Date field, default ('ymd') attempted.
        If warning that 'no formats are found' produced, see help page
        ('?lubridate::parse_date_time') for information on Date formats.")
        formatDT <- "ymd"
        dataGroup$DateTime <- lubridate::parse_date_time(
          dataGroup[, fieldDate], formatDT, tz = "UTC"
          )
      } else {
     dataGroup$DateTime <- lubridate::parse_date_time(
       dataGroup[, fieldDate], formatDT, tz = "UTC"
       )
      }
    }
  }

  ## --------------------------- checks complete ------------------------------
  #### FORMAT COLUMNS
  if ("ID" %in% colnames(dataGroup) & fieldID != "ID") {
    message("NOTE: as fieldID !=  'ID' and a column of this name is present,
            this column has been renamed to 'origID' and ID corresponds to the
            fieldID specified.")
    dataGroup <- dataGroup %>% dplyr::rename(
      origID = .data$ID, ID = fieldID, Latitude = fieldLat, Longitude = fieldLon
    ) %>%
      dplyr::mutate(ID = as.character(.data$ID))
  } else {
    dataGroup <- dataGroup %>% dplyr::rename(
      ID = fieldID, Latitude = fieldLat, Longitude = fieldLon
    ) %>%
      dplyr::mutate(ID = as.character(.data$ID))
  }

  if (cleanDF == TRUE) {
    dataGroup <- dataGroup %>%
      dplyr::select(
        .data$ID, .data$Latitude, .data$Longitude, .data$DateTime
        ) %>%
      arrange(.data$ID, .data$DateTime)
  }
  return(dataGroup)
}