improve doc of computeForecast for realtime usage
[talweg.git] / pkg / R / getData.R
1 #' @title Acquire data in a clean format
2 #'
3 #' @description Take in input data frames and/or files containing raw data, and
4 #' timezones, and output a Data object, roughly corresponding to a list where each cell
5 #' contains all value for one day (see \code{?Data}).
6 #'
7 #' @param ts_data Time-series, as a data frame (DB style: 2 columns, first is date/time,
8 #' second is value) or a CSV file
9 #' @param exo_data Exogenous variables, as a data frame or a CSV file; first comlumn is
10 #' dates, next block are measurements for the day, and final block are exogenous
11 #' forecasts
12 #' @param input_tz Timezone in the input files ("GMT" or e.g. "Europe/Paris")
13 #' @param date_format How date/time are stored (e.g. year/month/day hour:minutes;
14 #' see \code{strptime})
15 #' @param working_tz Timezone to work with ("GMT" or e.g. "Europe/Paris")
16 #' @param predict_at When does the prediction take place ? Integer, in hours. Default: 0
17 #' @param limit Number of days to extract (default: Inf, for "all")
18 #'
19 #' @return An object of class Data
20 #'
21 #' @examples
22 #' ts_data = read.csv(system.file("extdata","pm10_mesures_H_loc.csv",package="talweg"))
23 #' exo_data = read.csv(system.file("extdata","meteo_extra_noNAs.csv",package="talweg"))
24 #' data = getData(ts_data, exo_data, limit=120)
25 #' @export
26 getData = function(ts_data, exo_data, input_tz="GMT", date_format="%d/%m/%Y %H:%M",
27 working_tz="GMT", predict_at=0, limit=Inf)
28 {
29 # Sanity checks (not full, but sufficient at this stage)
30 if (!is.character(input_tz) || !is.character(working_tz))
31 stop("Bad timezone (see ?timezone)")
32 input_tz = input_tz[1]
33 working_tz = working_tz[1]
34 if ( (!is.data.frame(ts_data) && !is.character(ts_data)) ||
35 (!is.data.frame(exo_data) && !is.character(exo_data)) )
36 stop("Bad time-series / exogenous input (data frame or CSV file)")
37 if (is.character(ts_data))
38 ts_data = ts_data[1]
39 if (is.character(exo_data))
40 exo_data = exo_data[1]
41 predict_at = as.integer(predict_at)[1]
42 if (predict_at<0 || predict_at>23)
43 stop("Bad predict_at (0-23)")
44 if (!is.character(date_format))
45 stop("Bad date_format (character)")
46 date_format = date_format[1]
47
48 ts_df =
49 if (is.character(ts_data))
50 read.csv(ts_data)
51 else
52 ts_data
53 # Convert to the desired timezone (usually "GMT" or "Europe/Paris")
54 formatted_dates_POSIXlt = strptime(as.character(ts_df[,1]), date_format, tz=input_tz)
55 ts_df[,1] = format(
56 as.POSIXct(formatted_dates_POSIXlt, tz=input_tz), tz=working_tz, usetz=TRUE)
57
58 exo_df =
59 if (is.character(exo_data))
60 read.csv(exo_data)
61 else
62 exo_data
63 # Times in exogenous variables file are ignored: no conversions required
64
65 line = 1 #index in PM10 file (24 lines for 1 cell)
66 nb_lines = nrow(ts_df)
67 nb_exos = ( ncol(exo_df) - 1 ) / 2
68 data = Data$new()
69 i = 1 #index of a cell in data
70 while (line <= nb_lines)
71 {
72 time = c()
73 serie = c()
74 repeat
75 {
76 {
77 time = c(time, ts_df[line,1])
78 serie = c(serie, ts_df[line,2])
79 line = line + 1
80 };
81 if (line >= nb_lines + 1
82 || as.POSIXlt(ts_df[line-1,1],tz=working_tz)$hour == predict_at)
83 {
84 break
85 }
86 }
87
88 hat_exo = as.data.frame( exo_df[i,(1+nb_exos+1):(1+2*nb_exos)] )
89 exo = as.data.frame( exo_df[i,2:(1+nb_exos)] )
90 data$appendHat(time, hat_exo)
91 data$append(serie, exo) #in realtime, this call comes hours later
92 if (i >= limit)
93 break
94 i = i + 1
95 }
96 if (length(data$getCenteredSerie(1)) < length(data$getCenteredSerie(2)))
97 data$removeFirst()
98 if (length(data$getCenteredSerie(data$getSize()))
99 < length(data$getCenteredSerie(data$getSize()-1)))
100 {
101 data$removeLast()
102 }
103 data
104 }