def test_multiple_cubes_same_validity_time():
"""Basic test using several input cubes with same validity times
but with different forecast_period."""
# Set up expected dataframe.
validity_date = dt.utcfromtimestamp(1487311200).date()
data = [[
validity_date, 600, 1000, "air_temperature", "IMPRO", 280., 281.,
282.
]]
columns = [
"validity_date", "validity_time", "station_id", "cf_name",
"exp_id", "fcr_tplus000", "fcr_tplus001", "fcr_tplus002"
]
expected_df = pd.DataFrame(data, columns=columns)
expected_df = expected_df.set_index([
"validity_date", "validity_time", "station_id", "cf_name", "exp_id"
])
expected_df.columns.name = "forecast_period"
# Call the plugin.
cubes = [
set_up_spot_cube(280 + i, forecast_period=i, number_of_sites=1)
for i in range(3)
]
cubes = iris.cube.CubeList(cubes)
plugin = VerificationTable("output", "csv", "improver", "IMPRO", 0)
plugin.to_dataframe(cubes)
result = plugin.df
assert_frame_equal(expected_df, result)
def test_single_cube_extra_data(self):
"""Basic test using one input cube with an extra point in the
percentile dimension."""
# Set up cubes
cube = set_up_spot_cube(280)
second_cube = cube.copy()
second_cube.coord("percentile").points = np.array([60.0])
cubelist = iris.cube.CubeList([cube, second_cube])
cubes = cubelist.concatenate()
plugin = VerificationTable("csv", "output", "improver", "IMPRO", 0)
message = "Dimensions that are not described by the pivot_dim or "\
"coord_to_slice_over must only have one point in. "\
"Dimension '1' has length '2' and is associated with the "\
"'percentile' coordinate."
with self.assertRaisesRegexp(ValueError, message):
plugin.to_dataframe(cubes)
def test_single_cube_single_site():
"""Basic test using one input cube with a single site in it."""
# Set up expected dataframe.
validity_date = dt.utcfromtimestamp(1487311200).date()
data = [[validity_date, 600, 1000, "air_temperature", "IMPRO", 280.]]
columns = [
"validity_date", "validity_time", "station_id", "cf_name",
"exp_id", "fcr_tplus000"
]
expected_df = pd.DataFrame(data, columns=columns)
expected_df = expected_df.set_index([
"validity_date", "validity_time", "station_id", "cf_name", "exp_id"
])
expected_df.columns.name = "forecast_period"
# Call the plugin.
cubes = iris.cube.CubeList([set_up_spot_cube(280, number_of_sites=1)])
plugin = VerificationTable("output", "csv", "improver", "IMPRO", 0)
plugin.to_dataframe(cubes)
result = plugin.df
assert_frame_equal(expected_df, result)