def test_fit(self, tmp_path):
data_path = _create_dummy_precip_data(tmp_path,
start_date="2000-01-01",
end_date="2010-01-01")
assert data_path.exists(), f"file not created correctly"
spi = SPI(data_path / "data_kenya.nc")
variable = "precip"
spi.fit(variable=variable)
coords = [c for c in spi.index.coords]
vars_ = [v for v in spi.index.variables if v not in coords]
assert "SPI3" in vars_, f"Expecting `v` variable in" "`self.index`"
assert spi.index.SPI3.isel(time=slice(0, 2)).isnull().all(), (
""
"The first two timesteps should be nan because using SPI3")
assert not spi.index.SPI3.isel(time=slice(0, 3)).isnull().all(), (
""
"The first two timesteps but not 3rd should be nan SPI3")
assert np.isclose(
spi.index.SPI3.mean(), 0,
atol=0.01), (""
"Expect the mean SPI value to be close to 0 because"
"converted to a standard normal distribution")
assert np.isclose(spi.index.SPI3.std(), 1, atol=0.01), (
""
"Expect the `std()` SPI value to be close to 1 because"
"converted to a standard normal distribution")
def test_fit(self, tmp_path):
data_path = _create_dummy_precip_data(tmp_path,
start_date='2000-01-01',
end_date='2010-01-01')
spi = SPI(data_path / 'chirps_kenya.nc')
variable = 'precip'
spi.fit(variable=variable)
coords = [c for c in spi.index.coords]
vars_ = [v for v in spi.index.variables if v not in coords]
assert "SPI3" in vars_, f"Expecting `v` variable in"\
"`self.index`"
assert spi.index.SPI3.isel(time=slice(0, 2)).isnull().all(), "" \
"The first two timesteps should be nan because using SPI3"
assert not spi.index.SPI3.isel(time=slice(0, 3)).isnull().all(), "" \
"The first two timesteps but not 3rd should be nan SPI3"
assert np.isclose(spi.index.SPI3.mean(), 0, atol=0.01), ""\
"Expect the mean SPI value to be close to 0 because" \
"converted to a standard normal distribution"
assert np.isclose(spi.index.SPI3.std(), 1, atol=0.01), ""\
"Expect the `std()` SPI value to be close to 0 because" \
"converted to a standard normal distribution"
def test_initialisation(self, tmp_path):
data_path = _create_dummy_precip_data(tmp_path)
spi = SPI(data_path / "data_kenya.nc")
assert spi.name == "spi", f"Expected name" f"to be `spi` got: {spi.name}"
with pytest.raises(AttributeError):
# assert error raised because haven't fit
spi.index
def test_different_scales(self, tmp_path):
data_path = _create_dummy_precip_data(tmp_path,
start_date='2000-01-01',
end_date='2010-01-01')
spi = SPI(data_path / 'chirps_kenya.nc')
variable = 'precip'
# SPI 6
spi.fit(variable=variable, scale=6)
assert spi.index.SPI6.isel(time=slice(0, 5)).isnull().all(), "" \
"The first two timesteps should be nan because using SPI6"
assert np.isclose(spi.index.SPI6.mean(), 0, atol=0.01), ""\
"Expect the mean SPI6 value to be close to 0 because" \
"converted to a standard normal distribution"
assert np.isclose(spi.index.SPI6.std(), 1, atol=0.01), ""\
"Expect the `std()` SPI6 value to be close to 0 because" \
"converted to a standard normal distribution"
ds.to_netcdf(out_dir / 'all_indices.nc')
# --------------------------------------------------------
# View correlation between all indices
# --------------------------------------------------------
# 1. as time series
ts = ds.drop('quintile').mean(dim=['lat', 'lon'])
# drop inf
ts = ts.where(~(ts.RAI == np.inf)).dropna(dim='time')
norm_df = ((ts - ts.mean()) / ts.std()).to_dataframe()
fig, ax = plt.subplots()
norm_df.loc[np.isin(norm_df.index.year,np.arange(2008, 2015))].plot(ax=ax)
ax.set_title('Normalised Index Values 2008 - 2014')
ax.axhline(0, 0, 1, color='k', ls='--', alpha=0.4)
# 2. over the whole spatial domain
spi = SPI(data_path)
spi.fit(variable='precip')
# k_2010 = s.index.sel(time=np.isin(s.index['time.year'], [2008, 2009, 2010, 2011, 2012]))
k_2010 = spi.index.sel(time=slice('2008','2013'))
fig, ax = plt.subplots()
k_2010.mean(dim=['lat','lon']).SPI.plot(ax=ax)
ax.set_title('SPI3 over time')
ax.axhline(y=0, xmin=0, xmax=1, color='k', ls='--', alpha=0.5)
plt.show()