Initial commit

[talweg.git] / R / getData.R

#' @title Acquire data in a clean format
#'
#' @description Take in input data frames and/or files containing raw data, and timezones, and
#'   output a Data object, roughly corresponding to a list where each cell contains all value
#'   for one day (see \code{?Data}). Current limitation: series (in working_tz) must start at
#'   right after midnight (to keep in sync with exogenous vars)
#'
#' @param ts_data Time-series, as a data frame (DB style: 2 columns, first is date/time,
#'   second is value) or a CSV file
#' @param exo_data Exogenous variables, as a data frame or a CSV file; first comlumn is dates,
#'   next block are measurements for the day, and final block are exogenous forecasts
#' @param input_tz Timezone in the input files ("GMT" or e.g. "Europe/Paris")
#' @param date_format How date/time are stored (e.g. year/month/day hour:minutes;
#'   see \code{strptime})
#' @param working_tz Timezone to work with ("GMT" or e.g. "Europe/Paris")
#' @param predict_at When does the prediction take place ? ab[:cd][:ef] where a,b,c,d,e,f
#'   in (0,9) and define an hour[minute[second]]; time must be present in the file
#'
#' @return An object of class Data
#'
#' @export
getData = function(ts_data, exo_data,
	input_tz="GMT", date_format="%d/%m/%Y %H:%M", working_tz="GMT", predict_at="00")
{
	# Sanity checks (not full, but sufficient at this stage)
	if (!is.character(input_tz) || !is.character(working_tz))
		stop("Bad timezone (see ?timezone)")
	input_tz = input_tz[1]
	working_tz = working_tz[1]
	if (!is.data.frame(ts_data) && !is.character(ts_data))
		stop("Bad time-series input (data frame or CSV file)")
	if (is.character(ts_data))
		ts_data = ts_data[1]
	pattern_index_in_predict_at = grep("^[0-9]{2}(:[0-9]{2}){0,2}$", predict_at)
	if (!is.character(predict_at) || length(pattern_index_in_predict_at) == 0)
		stop("Bad predict_at ( ^[0-9]{2}(:[0-9]{2}){0,2}$ )")
	predict_at = predict_at[ pattern_index_in_predict_at[1] ]
	if (!is.character(date_format))
		stop("Bad date_format (character)")
	date_format = date_format[1]

	ts_df =
		if (is.character(ts_data)) {
			read.csv(ts_data)
		} else {
			ts_data
		}
	exo_df =
		if (is.character(exo_data)) {
			read.csv(exo_data)
		} else {
			exo_data
		}
	# Convert to the desired timezone (usually "GMT" or "Europe/Paris")
	formatted_dates_POSIXlt = strptime(as.character(ts_df[,1]), date_format, tz=input_tz)
	ts_df[,1] = format(as.POSIXct(formatted_dates_POSIXlt), tz=working_tz, usetz=TRUE)

	if (nchar(predict_at) == 2)
		predict_at = paste(predict_at,":00",sep="")
	if (nchar(predict_at) == 5)
		predict_at = paste(predict_at,":00",sep="")

	line = 1 #index in PM10 file (24 lines for 1 cell)
	nb_lines = nrow(ts_df)
	nb_exos = ( ncol(exo_df) - 1 ) / 2
	data = list() #new("Data")
	i = 1 #index of a cell in data
	while (line <= nb_lines)
	{
		time = c()
		serie = c()
		repeat {
		{
			time = c(time, ts_df[line,1])
			serie = c(serie, ts_df[line,2])
			line = line + 1
		};
		if (line >= nb_lines + 1
			# NOTE: always second part of date/time, because it has been formatted
			|| strsplit(as.character(ts_df[line-1,1])," ")[[1]][2] == predict_at)
		{
			break
		}}

		# NOTE: if predict_at does not cut days at midnight,
		# for the exogenous to be synchronized they need to be shifted
		if (predict_at != "00:00:00")
		{
			exo_hat = as.data.frame(exo_df[max(1,i-1),(1+nb_exos+1):(1+2*nb_exos)])
			exo_Dm1 = if (i>=3) as.data.frame(exo_df[i-1,2:(1+nb_exos)]) else NA
		} else
		{
			exo_hat = as.data.frame(exo_df[i,(1+nb_exos+1):(1+2*nb_exos)])
			exo_Dm1 = if (i>=2) as.data.frame(exo_df[i-1,2:(1+nb_exos)]) else NA
		}
		i = i + 1
		#center data
		level = mean(serie, na.rm=TRUE)
		centered_serie = serie - level
#		data$append(time, centered_serie, level, exo_hat, exo_Jm1) #TODO: slow: why ?
		data[[length(data)+1]] = list("time"=time, "serie"=centered_serie, "level"=level,
			"exo_hat"=exo_hat, "exo_Dm1"=exo_Dm1)
	}
	new("Data",data=data)
}