pkg/R/utils.R

   1 #' dateIndexToInteger
   2 #'
   3 #' Transform a (potential) date index into an integer (relative to data)
   4 #'
   5 #' @param index Date (or integer) index
   6 #' @param data Object of class \code{Data}
   7 #'
   8 #' @export
   9 dateIndexToInteger = function(index, data)
  10 {
  11     #works on integers too: trust input
  12     if (is.numeric(index))
  13         index = as.integer(index)
  14     if (is.integer(index))
  15         return (index)
  16
  17     if (inherits(index, "Date") || is.character(index))
  18     {
  19         tryCatch(indexAsDate <- as.Date(index), error=function(e) stop("Unrecognized index format"))
  20         #TODO: tz arg to difftime ?
  21         integerIndex <- round( as.numeric(
  22             difftime(indexAsDate, as.Date(data$getTime(1)[1])) ) ) + 1
  23         if (integerIndex >= 1 && integerIndex <= data$getSize())
  24             return (integerIndex)
  25         stop("Date outside data range")
  26     }
  27     stop("Unrecognized index format")
  28 }
  29
  30 #' integerIndexToDate
  31 #'
  32 #' Transform an integer index to date index (relative to data)
  33 #'
  34 #' @param index Date (or integer) index
  35 #' @param data Object of class \code{Data}
  36 #'
  37 #' @export
  38 integerIndexToDate = function(index, data)
  39 {
  40     #works on dates too: trust input
  41     if (is.character(index))
  42         index = as.Date(index)
  43     if (is(index,"Date"))
  44         return (index)
  45
  46     index = index[1]
  47     if (is.numeric(index))
  48         index = as.integer(index)
  49     if (!is.integer(index))
  50         stop("'index' should be a date or integer")
  51     as.Date( data$getTime(index)[1] )
  52 }
  53
  54 #' getSimilarDaysIndices
  55 #'
  56 #' Find similar days indices in the past.
  57 #'
  58 #' @param index Day index (numeric or date)
  59 #' @param data Reference dataset, object output of \code{getData}
  60 #' @param limit Maximum number of indices to return
  61 #' @param same_season Should the indices correspond to day in same season?
  62 #' @param days_in Optional set to intersect with results (NULL to discard)
  63 #'
  64 #' @export
  65 getSimilarDaysIndices = function(index, data, limit, same_season, days_in=NULL)
  66 {
  67     index = dateIndexToInteger(index, data)
  68
  69     # Look for similar days (optionally in same season)
  70     i = index - 1
  71     days = c()
  72     dt_ref = as.POSIXlt(data$getTime(index)[1]) #first date-time of current day
  73     day_ref = dt_ref$wday #1=monday, ..., 6=saturday, 0=sunday
  74     month_ref = as.POSIXlt(data$getTime(index)[1])$mon+1 #month in 1...12
  75     while (i >= 1 && length(days) < limit)
  76     {
  77         dt = as.POSIXlt(data$getTime(i)[1])
  78         if ((is.null(days_in) || i %in% days_in) && .isSameDay(dt$wday, day_ref))
  79         {
  80             if (!same_season || .isSameSeason(dt$mon+1, month_ref))
  81                 days = c(days, i)
  82         }
  83         i = i - 1
  84     }
  85     return ( days )
  86 }
  87
  88 # isSameSeason
  89 #
  90 # Check if two months fall in the same "season" (defined by estimated pollution rate)
  91 #
  92 # @param month month index to test
  93 # @param month_ref month to compare to
  94 #
  95 .isSameSeason = function(month, month_ref)
  96 {
  97     if (month_ref %in% c(11,12,1,2)) #~= mid-polluted
  98         return (month %in% c(11,12,1,2))
  99     if (month_ref %in% c(3,4,9,10)) #~= high-polluted
 100         return (month %in% c(3,4,9,10))
 101     return (month %in% c(5,6,7,8)) #~= non polluted
 102 }
 103
 104 # isSameDay
 105 #
 106 # Monday=Tuesday=Wednesday=Thursday ; Friday, Saturday, Sunday: specials
 107 #
 108 # @param day day index to test
 109 # @param day_ref day index to compare to
 110 #
 111 .isSameDay = function(day, day_ref)
 112 {
 113     if (day_ref == 0)
 114         return (day==0)
 115     if (day_ref <= 4)
 116         return (day <= 4)
 117     return (day == day_ref)
 118 }
 119
 120 #' getNoNA2
 121 #'
 122 #' Get indices in data of no-NA series followed by no-NA, within [first,last] range.
 123 #'
 124 #' @param data Object of class Data
 125 #' @param first First index (included)
 126 #' @param last Last index (included)
 127 #'
 128 #' @export
 129 getNoNA2 = function(data, first, last)
 130 {
 131     (first:last)[ sapply(first:last, function(i)
 132         !any( is.na(data$getCenteredSerie(i)) | is.na(data$getCenteredSerie(i+1)) )
 133     ) ]
 134 }