tests/testthat/test-format_weather.R

In ascotraceR: Simulate the Spread of Ascochyta Blight in Chickpea

# identify lon lat from file ---------------------------------------------------
test_that("`format_weather()` is able to identify the correct lat and lon values
          from file",
          {
            set.seed(27)
            # create data.frame of station coordinates
            write.csv(data.frame(
              station = c("69061", "16095"),
              lon = c(114.8627, 114.2627),
              lat = c(-28.5990, -28.1)
            ),
            file = file.path(tempdir(), "stat_coord.csv"))

            dat_minutes <- 10080 # equal to, 7 * 24 * 60

            weather_station_data <- data.table(
              Local.Time.YYYY = rep(2020, dat_minutes),
              Local.Time.MM = rep(6, dat_minutes),
              Local.Time.DD = rep(10:16, each = 24 * 60),
              Local.Time.HH24 = rep(1:24, each = 60, times = 7),
              Local.Time.MI = rep(0:59, times = 7 * 24),
              Precipitation.since.last.observation.in.mm = round(abs(rnorm(
                dat_minutes, mean = 0, sd = 0.2
              )), 1),
              Temperature.C = (sin(seq(
                0, (6.285) * 7,
                length.out = dat_minutes
              ) + 4) + 1) * 20,
              Wind.speed.in.km.h = abs(rnorm(
                dat_minutes, mean = 5, sd = 10
              )),
              Wind.direction.in.degrees.true = runif(n = dat_minutes,
                                                     min = 0, max = 359),
              Station.Number = "69061"
            )

            weather_dt <- format_weather(
              x = weather_station_data,
              YYYY = "Local.Time.YYYY",
              MM = "Local.Time.MM",
              DD = "Local.Time.DD",
              hh = "Local.Time.HH24",
              mm = "Local.Time.MI",
              temp = "Temperature.C",
              rain = "Precipitation.since.last.observation.in.mm",
              ws = "Wind.speed.in.km.h",
              wd = "Wind.direction.in.degrees.true",
              station = "Station.Number",
              lonlat_file = file.path(tempdir(), "stat_coord.csv"),
              time_zone = "Australia/Perth"

            )

            expect_s3_class(weather_dt, "asco.weather")
            expect_equal(
              names(weather_dt),
              c(
                "times",
                "temp",
                "rain",
                "ws",
                "wd",
                "wd_sd",
                "lon",
                "lat",
                "station",
                "YYYY",
                "MM",
                "DD",
                "hh",
                "mm"
              )
            )
            expect_equal(dim(weather_dt), c(dat_minutes / 60, 14))
            expect_true(anyNA(weather_dt$lon) == FALSE)
            expect_true(anyNA(weather_dt$lat) == FALSE)
            expect_true(weather_dt[, unique(lon)] == 114.8627)
            expect_equal(unique(weather_dt$lat), -28.5990)
            expect_is(weather_dt$times, "POSIXct")
            expect_equal(as.character(min(weather_dt$times)),
                         "2020-06-10 01:00:00")
            expect_equal(as.character(max(weather_dt$times)), "2020-06-17")
            expect_equal(round(min(weather_dt$rain), 0), 7)
            expect_equal(round(max(weather_dt$rain), 1), 12.4)
            expect_equal(round(min(weather_dt$ws), 1), 6.5)
            expect_equal(round(max(weather_dt$ws), 2), 10.99)
            expect_equal(round(min(weather_dt$wd), 0), 1)
            expect_equal(round(max(weather_dt$wd), 1), 358.3)
            expect_equal(round(min(weather_dt$wd_sd), 0), 82)
            expect_equal(round(max(weather_dt$wd_sd), 0), 195)
            expect_equal(mean(weather_dt$YYYY), 2020)
            expect_equal(mean(weather_dt$MM), 6)
            expect_equal(min(weather_dt$DD), 10)
            expect_equal(max(weather_dt$DD), 16)
            expect_equal(min(weather_dt$hh), 1)
            expect_equal(max(weather_dt$hh), 24)
            expect_equal(mean(weather_dt$mm), 0)

            unlink(file.path(tempdir(), "stat_coord.csv"))
          })

# Test that it is able to import supplied data suitable for `ascotraceR` model

# read in data associated with the ascotraceR model
newmarra_raw <-
  fread(
    system.file("extdata", "1998_Newmarracarra_weather_table.csv",
                package = "ascotraceR")
  )

test_that("format_weather() can format Newmarracarra weather data", {
  expect_silent(
    newM <- format_weather(
      x = newmarra_raw,
      POSIXct_time = "Local.Time",
      time_zone = "Australia/Perth",
      temp = "mean_daily_temp",
      rain = "rain_mm",
      ws = "ws",
      wd = "wd",
      wd_sd = "wd_sd",
      station = "Location",
      lat = NA,
      lon = NA
    )
  )
})

dat_no_location <- format_weather(
  x = newmarra_raw,
  POSIXct_time = "Local.Time",
  time_zone = "Australia/Perth",
  temp = "mean_daily_temp",
  rain = "rain_mm",
  ws = "ws",
  wd = "wd",
  wd_sd = "wd_sd",
  station = "Location",
  lat = NA,
  lon = NA
)

test_that("dat_no_location contains the correct output", {
  expect_false("lat" %in% colnames(dat_no_location))
  expect_false("lon" %in% colnames(dat_no_location))
  expect_true(all(
    c(
      "times",
      "temp",
      "rain",
      "ws",
      "wd",
      "wd_sd",
      "station",
      "YYYY",
      "MM",
      "DD",
      "hh",
      "mm"
    ) %in% colnames(dat_no_location)
  ))
})

# identify lon lat from cols ---------------------------------------------------
test_that("`format_weather()` works when lat lon are in data", {
  dat_minutes <- 10080 # equal to, 7 * 24 * 60

  weather_station_data <- data.table(
    Local.Time.YYYY = rep(2020, dat_minutes),
    Local.Time.MM = rep(6, dat_minutes),
    Local.Time.DD = rep(10:16, each = 24 * 60),
    Local.Time.HH24 = rep(1:24, each = 60, times = 7),
    Local.Time.MI = rep(0:59, times = 7 * 24),
    Precipitation.since.last.observation.in.mm = round(abs(rnorm(
      dat_minutes, mean = 0, sd = 0.2
    )), 1),
    Wind.speed.in.km.h = abs(rnorm(
      dat_minutes, mean = 5, sd = 10
    )),
    Wind.direction.in.degrees.true =
      runif(n = dat_minutes, min = 0, max = 359),
    Temperature.in.degrees.celcius =
      rep(c(11:22, 23:12) + rnorm(24, sd = 2), 420),
    Station.Number = "16096",
    lon = 135.7243,
    lat = -33.26625
  )

  weather_dt <- format_weather(
    x = weather_station_data,
    YYYY = "Local.Time.YYYY",
    MM = "Local.Time.MM",
    DD = "Local.Time.DD",
    hh = "Local.Time.HH24",
    mm = "Local.Time.MI",
    rain = "Precipitation.since.last.observation.in.mm",
    temp = "Temperature.in.degrees.celcius",
    ws = "Wind.speed.in.km.h",
    wd = "Wind.direction.in.degrees.true",
    station = "Station.Number",
    lon = "lon",
    lat = "lat",
    time_zone = "UTC"
  )


  expect_s3_class(weather_dt, "asco.weather")
  expect_s3_class(weather_dt, "data.table")

  expect_named(
    weather_dt,
    c(
      "times",
      "temp",
      "rain",
      "ws",
      "wd",
      "wd_sd",
      "lon",
      "lat",
      "station",
      "YYYY",
      "MM",
      "DD",
      "hh",
      "mm"
    )
  )
  expect_equal(dim(weather_dt), c(168, 14))
  expect_is(weather_dt$times, "POSIXct")
  expect_true(anyNA(weather_dt$times) == FALSE)
  expect_true(max(weather_dt$wd, na.rm = TRUE) < 360)
  expect_true(min(weather_dt$wd, na.rm = TRUE) > 0)
  expect_true(lubridate::tz(weather_dt$times) == "UTC")
})

# stop if `x` is not a data.frame object ---------------------------------------
test_that("`format_weather()` stops if `x` is not a data.frame object", {
  expect_error(
    weather_dt <- format_weather(
      x = list(),
      rain = "Precipitation.since.last.observation.in.mm",
      ws = "Wind.speed.in.km.h",
      wd = "Wind.direction.in.degrees.true",
      station = "Station.Number",
      lon = "lon",
      lat = "lat",
      r = "eyre",
      POSIXct_time = "times"
    ),
    regexp = "`x` must be provided as a `data.frame` object*"
  )
})

# stop if time isn't given in any col ------------------------------------------
test_that("`format_weather()` stops if time cols are not provided", {
  dat_minutes <- 10080 # equal to, 7 * 24 * 60

  weather_station_data <- data.table(
    Local.Time.YYYY = rep(2020, dat_minutes),
    Local.Time.MM = rep(6, dat_minutes),
    Local.Time.DD = rep(10:16, each = 24 * 60),
    Local.Time.HH24 = rep(1:24, each = 60, times = 7),
    Local.Time.MI = rep(0:59, times = 7 * 24),
    Precipitation.since.last.observation.in.mm = round(abs(rnorm(
      dat_minutes, mean = 0, sd = 0.2
    )), 1),
    Wind.speed.in.km.h = abs(rnorm(
      dat_minutes, mean = 5, sd = 10
    )),
    Wind.direction.in.degrees.true =
      runif(n = dat_minutes, min = 0, max = 359),
    Temperature.in.degrees.celcius =
      rep(c(11:22, 23:12) + rnorm(24, sd = 2), 420),
    Station.Number = "16096",
    lon = 135.7243,
    lat = -33.26625
  )

  weather_dt <- format_weather(
    x = weather_station_data,
    YYYY = "Local.Time.YYYY",
    MM = "Local.Time.MM",
    DD = "Local.Time.DD",
    hh = "Local.Time.HH24",
    mm = "Local.Time.MI",
    rain = "Precipitation.since.last.observation.in.mm",
    temp = "Temperature.in.degrees.celcius",
    ws = "Wind.speed.in.km.h",
    wd = "Wind.direction.in.degrees.true",
    station = "Station.Number",
    lon = "lon",
    lat = "lat",
    time_zone = "Australia/Adelaide"
  )

  expect_error(
    weather_dt <- format_weather(
      x = weather_station_data,
      rain = "Precipitation.since.last.observation.in.mm",
      ws = "Wind.speed.in.km.h",
      wd = "Wind.direction.in.degrees.true",
      station = "Station.Number",
      lon = "lon",
      lat = "lat",
      r = "eyre"
    ),
    regexp = "You must provide time values either as a*"
  )
})

# # stop if lonlat file lacks proper field names -------------------------------
test_that("`format_weather() stops if lonlat input lacks proper names", {
  # create a dummy .csv with misnamed cols
  write.csv(data.frame(
    stats = c("69061", "16096"),
    long = c(134.2734, 135.7243),
    lat = c(-33.52662, -33.26625)
  ),
  file = file.path(tempdir(), "stat_coord.csv"))

  dat_minutes <- 10080 # equal to, 7 * 24 * 60

  weather_station_data <- data.table(
    Local.Time.YYYY = rep(2020, dat_minutes),
    Local.Time.MM = rep(6, dat_minutes),
    Local.Time.DD = rep(10:16, each = 24 * 60),
    Local.Time.HH24 = rep(1:24, each = 60, times = 7),
    Local.Time.MI = rep(0:59, times = 7 * 24),
    Precipitation.since.last.observation.in.mm = round(abs(rnorm(
      dat_minutes, mean = 0, sd = 0.2
    )), 1),
    Wind.speed.in.km.h = abs(rnorm(
      dat_minutes, mean = 5, sd = 10
    )),
    Wind.direction.in.degrees.true =
      runif(n = dat_minutes, min = 0, max = 359),
    Station.Number = "16096",
    lon = 135.7243,
    lat = -33.26625
  )

  # while ascotraceR does not use rasters lets create one for testing
  eyre_temp <- terra::rast(
    nrows = 180,
    ncols = 360,
    nlyrs = 1,
    xmin = -135.7,
    xmax = 135.9,
    ymin = -33.4,
    ymax = -33.2
  )

  expect_error(
    weather_dt <- format_weather(
      x = weather_station_data,
      YYYY = "Local.Time.YYYY",
      MM = "Local.Time.MM",
      DD = "Local.Time.DD",
      hh = "Local.Time.HH24",
      mm = "Local.Time.MI",
      rain = "Precipitation.since.last.observation.in.mm",
      ws = "Wind.speed.in.km.h",
      wd = "Wind.direction.in.degrees.true",
      station = "Station.Number",
      lonlat_file = file.path(tempdir(), "stat_coord.csv"),
      r = eyre_temp
    ),
    regexp = "The CSV file of weather station coordinates *"
  )
  unlink(file.path(tempdir(), "stat_coord.csv"))
})

# stop if no lonlat info provided ----------------------------------------------
test_that("`format_weather() stops if lonlat input lacks proper names", {
  dat_minutes <- 10080 # equal to, 7 * 24 * 60

  weather_station_data <- data.table(
    Local.Time.YYYY = rep(2020, dat_minutes),
    Local.Time.MM = rep(6, dat_minutes),
    Local.Time.DD = rep(10:16, each = 24 * 60),
    Local.Time.HH24 = rep(1:24, each = 60, times = 7),
    Local.Time.MI = rep(0:59, times = 7 * 24),
    Precipitation.since.last.observation.in.mm = round(abs(rnorm(
      dat_minutes, mean = 0, sd = 0.2
    )), 1),
    Wind.speed.in.km.h = abs(rnorm(
      dat_minutes, mean = 5, sd = 10
    )),
    Wind.direction.in.degrees.true =
      runif(n = dat_minutes, min = 0, max = 359),
    Station.Number = "16096",
    lon = 135.7243,
    lat = -33.26625
  )

  expect_error(
    weather_dt <- format_weather(
      x = weather_station_data,
      YYYY = "Local.Time.YYYY",
      MM = "Local.Time.MM",
      DD = "Local.Time.DD",
      hh = "Local.Time.HH24",
      mm = "Local.Time.MI",
      rain = "Precipitation.since.last.observation.in.mm",
      ws = "Wind.speed.in.km.h",
      wd = "Wind.direction.in.degrees.true",
      station = "Station.Number",
      r = eyre
    ),
    regexp = "You must provide lonlat values for the weather *"
  )
  unlink(file.path(tempdir(), "stat_coord.csv"))
})


# fill missing mm --------------------------------------------------------------
test_that("`format_weather() creates a `mm` column if not provided", {
  dat_minutes <- 10080 # equal to, 7 * 24 * 60

  weather_station_data <- data.table(
    Local.Time.YYYY = rep(2020, dat_minutes),
    Local.Time.MM = rep(6, dat_minutes),
    Local.Time.DD = rep(10:16, each = 24 * 60),
    Local.Time.HH24 = rep(1:24, each = 60, times = 7),
    Precipitation.since.last.observation.in.mm = round(abs(rnorm(
      dat_minutes, mean = 0, sd = 0.2
    )), 1),
    Wind.speed.in.km.h = abs(rnorm(
      dat_minutes, mean = 5, sd = 10
    )),
    Wind.direction.in.degrees.true =
      runif(n = dat_minutes, min = 0, max = 359),
    Temperature.in.degrees.celcius =
      rep(c(11:22, 23:12) + rnorm(24, sd = 2), 420),
    Station.Number = "16096",
    lon = 135.7243,
    lat = -33.26625
  )

  expect_named(
    weather_dt <- format_weather(
      x = weather_station_data,
      YYYY = "Local.Time.YYYY",
      MM = "Local.Time.MM",
      DD = "Local.Time.DD",
      hh = "Local.Time.HH24",
      rain = "Precipitation.since.last.observation.in.mm",
      temp = "Temperature.in.degrees.celcius",
      ws = "Wind.speed.in.km.h",
      wd = "Wind.direction.in.degrees.true",
      station = "Station.Number",
      lat = "lat",
      lon = "lon",
      time_zone = "Australia/Adelaide"
    ),
    c(
      "times",
      "temp",
      "rain",
      "ws",
      "wd",
      "wd_sd",
      "lon",
      "lat",
      "station",
      "YYYY",
      "MM",
      "DD",
      "hh",
      "mm"
    )
  )
})

# fill create YYYY, MM, DD hhmm cols from POSIXct ------------------------------
test_that("`format_weather() creates a YYYY MM DD... cols", {
  dat_minutes <- 10080 # equal to, 7 * 24 * 60

  weather_station_data <- data.table(
    Precipitation.since.last.observation.in.mm = round(abs(rnorm(
      dat_minutes, mean = 0, sd = 0.2
    )), 1),
    Wind.speed.in.km.h = abs(rnorm(
      dat_minutes, mean = 5, sd = 10
    )),
    Wind.direction.in.degrees.true =
      runif(n = dat_minutes, min = 0, max = 359),
    Temperature.in.degrees.celcius =
      rep(c(11:22, 23:12) + rnorm(24, sd = 2), 420),
    Station.Number = "16096",
    Ptime = seq(ISOdate(2000, 1, 1), by = "1 min", length.out = dat_minutes),
    lon = 135.7243,
    lat = -33.26625,
    time_zone = c("Australia/Adelaide", "Australia/Brisbane")
  )

  # while ascotraceR does not use rasters lets create one for testing
  eyre_temp <- terra::rast(
    nrows = 180,
    ncols = 360,
    nlyrs = 1,
    xmin = -135.7,
    xmax = 135.9,
    ymin = -33.4,
    ymax = -33.2
  )

  expect_named(
    weather_dt <- format_weather(
      x = weather_station_data,
      YYYY = "Local.Time.YYYY",
      MM = "Local.Time.MM",
      DD = "Local.Time.DD",
      hh = "Local.Time.HH24",
      rain = "Precipitation.since.last.observation.in.mm",
      temp = "Temperature.in.degrees.celcius",
      ws = "Wind.speed.in.km.h",
      wd = "Wind.direction.in.degrees.true",
      station = "Station.Number",
      lat = "lat",
      lon = "lon",
      POSIXct_time = "Ptime",
      r = eyre_temp
    ),
    c(
      "times",
      "temp",
      "rain",
      "ws",
      "wd",
      "wd_sd",
      "lon",
      "lat",
      "station",
      "YYYY",
      "MM",
      "DD",
      "hh",
      "mm"
    )
  )
})

# stop if `wd_sd` is missing or cannot be calculated ---------------------------
test_that("`format_weather() stops if `wd_sd` is not available", {
  weather_station_data <- data.table(
    Precipitation.since.last.observation.in.mm = round(abs(rnorm(
      24, mean = 0, sd = 0.2
    )), 1),
    Wind.speed.in.km.h = abs(rnorm(24, mean = 5, sd = 10)),
    Wind.direction.in.degrees.true =
      runif(n = 24, min = 0, max = 359),
    Station.Number = "16096",
    Ptime = seq(ISOdate(2000, 1, 1), by = "1 hour", length.out = 24),
    lon = 135.7243,
    lat = -33.26625
  )

  expect_error(
    weather_dt <- format_weather(
      x = weather_station_data,
      YYYY = "Local.Time.YYYY",
      MM = "Local.Time.MM",
      DD = "Local.Time.DD",
      hh = "Local.Time.HH24",
      rain = "Precipitation.since.last.observation.in.mm",
      ws = "Wind.speed.in.km.h",
      wd = "Wind.direction.in.degrees.true",
      station = "Station.Number",
      lat = "lat",
      lon = "lon",
      POSIXct_time = "Ptime",
      time_zone = "UTC"
    ),
    regexp = "`format_weather*"
  )
})

# stop if no raster, `r` or `time_zone` provided -------------------------------
test_that("`format_weather() stops if `time_zone` cannot be determined", {
  weather_station_data <- data.table(
    Precipitation.since.last.observation.in.mm = round(abs(rnorm(
      24, mean = 0, sd = 0.2
    )), 1),
    Wind.speed.in.km.h = abs(rnorm(24, mean = 5, sd = 10)),
    Wind.direction.in.degrees.true =
      runif(n = 24, min = 0, max = 359),
    Station.Number = "16096",
    Ptime = seq(ISOdate(2000, 1, 1), by = "1 hour", length.out = 24),
    lon = 135.7243,
    lat = -33.26625
  )

  expect_error(
    weather_dt <- format_weather(
      x = weather_station_data,
      YYYY = "Local.Time.YYYY",
      MM = "Local.Time.MM",
      DD = "Local.Time.DD",
      hh = "Local.Time.HH24",
      rain = "Precipitation.since.last.observation.in.mm",
      ws = "Wind.speed.in.km.h",
      wd = "Wind.direction.in.degrees.true",
      station = "Station.Number",
      lat = "lat",
      lon = "lon",
      POSIXct_time = "Ptime"
    ),
    regexp =  "Please ensure that either a raster object for the area of * "
  )
})


test_that("format_weather detects impossible times", {
  raw_weather <- data.table(
    Year = rep(2020, 14 * 24 * 60),
    Month = rep(6, 14 * 24 * 60),
    Day = rep(rep(1:7, each = 24 * 60), 2),
    Hour = rep(rep(0:23, each = 60), 14),
    Minute = rep(1:60, 14 * 24),
    WindSpeed = abs(rnorm(14 * 24 * 60, 1, 3)),
    WindDirectionDegrees = round(runif(14 * 24 * 60, 0, 359)),
    Rainfall = floor(abs(rnorm(14 * 24 * 60, 0, 1))),
    stationID = rep(c("12345", "54321"), each = 7 * 24 * 60),
    StationLongitude = rep(c(134.123, 136.312), each = 7 * 24 * 60),
    StationLatitude = rep(c(-32.321, -33.123), each = 7 * 24 * 60)
  )



  expect_warning(
    expect_error(
      format_weather(
        x = raw_weather,
        YYYY = "Year",
        MM = "Month",
        DD = "Day",
        hh = "Hour",
        mm = "Minute",
        ws = "WindSpeed",
        rain = "Rainfall",
        wd = "WindDirectionDegrees",
        lon = "StationLongitude",
        lat = "StationLatitude",
        station = "stationID",
        time_zone = "UTC"
      ),
      regexp = "Time records contain NA values or impossible time*"
    )
  )

})

test_that("Incorrect column names are picked up and error is given", {
  Newmarracarra <-
    system.file("extdata",
                "1998_Newmarracarra_weather_table.csv",
                package = "ascotraceR")
  station_data <-
    system.file("extdata", "stat_dat.csv", package = "ascotraceR")
  expect_error(
    weather <- format_weather(
      x = read.csv(Newmarracarra),
      POSIXct_time = "Local.Time",
      ws = "ws",
      temp = "mean_daily_temp",
      wd_sd = "wd_sd",
      rain = "rain_mm",
      wd = "wdd",
      station = "Location",
      time_zone = "Australia/Perth",
      lonlat_file = station_data
    )
  )
})

test_that("function can reformat weather data previously saved as csv and
          read back in",
          {
            fileName <- paste0(tempfile(), ".csv")
            write.csv(x = dat_no_location,
                      file = fileName,
                      row.names = FALSE)
            w_dat <- read.csv(fileName)

            expect_equal(class(w_dat), "data.frame")
            expect_equal(
              colnames(w_dat),
              c(
                "times",
                "temp",
                "rain",
                "ws",
                "wd",
                "wd_sd",
                "station",
                "YYYY",
                "MM",
                "DD",
                "hh",
                "mm",
                "day",
                "hours_in_day",
                "wet_hours",
                "ws_sd"
              )
            )
            w_dat <- format_weather(w_dat, time_zone = "UTC")
            expect_equal(class(w_dat), c("asco.weather", "data.table", "data.frame"))
            expect_equal(
              colnames(w_dat),
              c(
                "times",
                "temp",
                "rain",
                "ws",
                "wd",
                "wd_sd",
                "station",
                "YYYY",
                "MM",
                "DD",
                "hh",
                "mm",
                "day",
                "hours_in_day",
                "wet_hours",
                "ws_sd"
              )
            )
          })