def test_calculate_average_semivariance_id_as_str(self):
# Data prepration
my_dir = os.path.dirname(__file__)
areal_dataset = os.path.join(
my_dir, '../sample_data/test_areas_pyinterpolate.shp')
subset = os.path.join(my_dir,
'../sample_data/test_points_pyinterpolate.shp')
a_id = 'idx'
areal_val = 'value'
points_val = 'value'
# Get maximum range and set step size
gdf = gpd.read_file(areal_dataset)
total_bounds = gdf.geometry.total_bounds
total_bounds_x = np.abs(total_bounds[2] - total_bounds[0])
total_bounds_y = np.abs(total_bounds[3] - total_bounds[1])
max_range = min(total_bounds_x, total_bounds_y)
step_size = max_range / 4
areal_data_prepared = prepare_areal_shapefile(areal_dataset, a_id,
areal_val)
points_in_area = get_points_within_area(areal_dataset,
subset,
areal_id_col_name=a_id,
points_val_col_name=points_val)
# Set areal semivariance class
areal_semivariance = ArealSemivariance(areal_data_prepared, step_size,
max_range, points_in_area)
areal_semivariance.regularize_semivariogram()
# Change ints to str in inblock semivariance for calculations
inblock = [[str(x[0]), x[1]]
for x in areal_semivariance.inblock_semivariance]
inblock = np.asarray(inblock, dtype=object)
ar_dists = [
areal_semivariance.distances_between_blocks[0],
[str(x) for x in areal_semivariance.distances_between_blocks[1]]
]
avg_semi = calculate_average_semivariance(
ar_dists, inblock, areal_semivariance.areal_lags,
areal_semivariance.areal_ss)
data_test_vals = avg_semi.astype(np.int)
desired_output = np.array([[0, 0], [1, 0], [1, 24]])
test_vals = np.equal(data_test_vals, desired_output).all()
self.assertTrue(
test_vals.all(),
f"Output is {data_test_vals} ut should be {desired_output}")
def test_calculate_semivariance_within_blocks(self):
# Data prepration
my_dir = os.path.dirname(__file__)
areal_dataset = os.path.join(
my_dir, '../sample_data/test_areas_pyinterpolate.shp')
subset = os.path.join(my_dir,
'../sample_data/test_points_pyinterpolate.shp')
a_id = 'id'
areal_val = 'value'
points_val = 'value'
# Get maximum range and set step size
gdf = gpd.read_file(areal_dataset)
total_bounds = gdf.geometry.total_bounds
total_bounds_x = np.abs(total_bounds[2] - total_bounds[0])
total_bounds_y = np.abs(total_bounds[3] - total_bounds[1])
max_range = min(total_bounds_x, total_bounds_y)
step_size = max_range / 4
areal_data_prepared = prepare_areal_shapefile(areal_dataset, a_id,
areal_val)
points_in_area = get_points_within_area(areal_dataset,
subset,
areal_id_col_name=a_id,
points_val_col_name=points_val)
# Get areal centroids with data values
areal_centroids = areal_data_prepared[:, 2:]
areal_centroids = np.array([[x[0], x[1], x[2]]
for x in areal_centroids])
gamma = calculate_semivariance(areal_centroids, step_size, max_range)
# Get theoretical semivariogram model
ts = TheoreticalSemivariogram(areal_centroids, gamma)
ts.find_optimal_model(number_of_ranges=8)
# Get centroids to calculate experimental semivariance
inblock_semivariance = calculate_semivariance_within_blocks(
points_in_area, ts)
inblock_semivariance = np.array(inblock_semivariance)
data_point = inblock_semivariance[inblock_semivariance[:, 0] == 1][0]
EXPECTED_OUTPUT = 24
output = int(data_point[1])
self.assertEqual(
output, EXPECTED_OUTPUT,
f"First data point's integer part should be equal to "
f"{EXPECTED_OUTPUT} but it is {output}")
def test_ata_pk(self):
my_dir = os.path.dirname(__file__)
areal_dataset = os.path.join(my_dir, '../sample_data/test_areas_pyinterpolate.shp')
subset = os.path.join(my_dir, '../sample_data/test_points_pyinterpolate.shp')
a_id = 'id'
areal_val = 'value'
points_val = 'value'
# Get maximum range and set step size
gdf = gpd.read_file(areal_dataset)
total_bounds = gdf.geometry.total_bounds
total_bounds_x = np.abs(total_bounds[2] - total_bounds[0])
total_bounds_y = np.abs(total_bounds[3] - total_bounds[1])
max_range = min(total_bounds_x, total_bounds_y)
step_size = max_range / 10
areal_data_prepared = prepare_areal_shapefile(areal_dataset, a_id, areal_val)
points_in_area = get_points_within_area(areal_dataset, subset, areal_id_col_name=a_id,
points_val_col_name=points_val)
# Get one area as unknown
unknown_area_id = [1]
u_points = points_in_area[points_in_area[:, 0] == unknown_area_id][0]
k_areas = areal_data_prepared[areal_data_prepared[:, 0] != unknown_area_id]
k_points = points_in_area[points_in_area[:, 0] != unknown_area_id]
# Semivariance deconvolution
semivar_modeling_data = set_areal_weights(k_areas, k_points)
smv_model = calculate_weighted_semivariance(semivar_modeling_data, step_size, max_range)
semivariogram = TheoreticalSemivariogram(k_areas[:, 2:], smv_model)
semivariogram.find_optimal_model()
# Poisson Kriging
search_radius = max_range / 2
number_of_observations = 3
pkc = AtAPoissonKriging(regularized_model=semivariogram,
known_areas=k_areas,
known_areas_points=k_points)
d = pkc.predict(u_points, number_of_observations, search_radius)
self.assertEqual(int(d[0]), 126, "Int of first value should be equal to 126")
def test_regularize_semivariogram_id_as_str(self):
reg_mod = RegularizedSemivariogram()
# Data prepration
my_dir = os.path.dirname(__file__)
areal_dataset = os.path.join(
my_dir, '../sample_data/test_areas_pyinterpolate.shp')
subset = os.path.join(my_dir,
'../sample_data/test_points_pyinterpolate.shp')
areal_id = 'idx'
areal_val = 'value'
points_val = 'value'
# Get maximum range and set step size
gdf = gpd.read_file(areal_dataset)
total_bounds = gdf.geometry.total_bounds
total_bounds_x = np.abs(total_bounds[2] - total_bounds[0])
total_bounds_y = np.abs(total_bounds[3] - total_bounds[1])
max_range = min(total_bounds_x, total_bounds_y)
step_size = max_range / 4
areal_data_prepared = prepare_areal_shapefile(areal_dataset, areal_id,
areal_val)
points_in_area = get_points_within_area(areal_dataset,
subset,
areal_id_col_name=areal_id,
points_val_col_name=points_val)
# Fit
reg_mod.fit(areal_data_prepared, step_size, max_range, points_in_area)
# Transform
reg_mod.transform()
regularized_smv = np.array([0, 75, 39])
test_output = reg_mod.final_optimal_model.astype(int)
check = np.equal(test_output, regularized_smv).all()
self.assertTrue(
check,
"Expected output is {} and an estimated output is {}".format(
regularized_smv, test_output))
def test_calculate_block_to_block_semivariance(self):
# Data prepration
my_dir = os.path.dirname(__file__)
areal_dataset = os.path.join(
my_dir, '../sample_data/test_areas_pyinterpolate.shp')
subset = os.path.join(my_dir,
'../sample_data/test_points_pyinterpolate.shp')
a_id = 'id'
areal_val = 'value'
points_val = 'value'
# Get maximum range and set step size
gdf = gpd.read_file(areal_dataset)
total_bounds = gdf.geometry.total_bounds
total_bounds_x = np.abs(total_bounds[2] - total_bounds[0])
total_bounds_y = np.abs(total_bounds[3] - total_bounds[1])
max_range = min(total_bounds_x, total_bounds_y)
step_size = max_range / 4
areal_data_prepared = prepare_areal_shapefile(areal_dataset, a_id,
areal_val)
points_in_area = get_points_within_area(areal_dataset,
subset,
areal_id_col_name=a_id,
points_val_col_name=points_val)
# Set areal semivariance class
areal_semivariance = ArealSemivariance(areal_data_prepared, step_size,
max_range, points_in_area)
areal_semivariance.regularize_semivariogram()
blocks = calculate_block_to_block_semivariance(
areal_semivariance.within_area_points,
areal_semivariance.distances_between_blocks,
areal_semivariance.theoretical_semivariance_model)
test_block = np.array(blocks[0][0])
mean_semivariance = int(np.mean(test_block[:, 1]))
EXPECTED_OUTPUT = 60
self.assertEqual(
mean_semivariance, EXPECTED_OUTPUT,
f"Average semivariance should be {EXPECTED_OUTPUT} "
f"but it is {mean_semivariance}")
def test_areal_semivariance(self):
# Data prepration
my_dir = os.path.dirname(__file__)
areal_dataset = os.path.join(
my_dir, '../sample_data/test_areas_pyinterpolate.shp')
subset = os.path.join(my_dir,
'../sample_data/test_points_pyinterpolate.shp')
a_id = 'id'
areal_val = 'value'
points_val = 'value'
# Get maximum range and set step size
gdf = gpd.read_file(areal_dataset)
total_bounds = gdf.geometry.total_bounds
total_bounds_x = np.abs(total_bounds[2] - total_bounds[0])
total_bounds_y = np.abs(total_bounds[3] - total_bounds[1])
max_range = min(total_bounds_x, total_bounds_y)
step_size = max_range / 4
areal_data_prepared = prepare_areal_shapefile(areal_dataset, a_id,
areal_val)
points_in_area = get_points_within_area(areal_dataset,
subset,
areal_id_col_name=a_id,
points_val_col_name=points_val)
# Set areal semivariance class
areal_semivariance = ArealSemivariance(areal_data_prepared, step_size,
max_range, points_in_area)
output = areal_semivariance.regularize_semivariogram()
test_d = output[:, 1].astype(np.int)
correct_output = np.array([0, 63, 46])
tescase = np.equal(test_d, correct_output).all()
self.assertTrue(
tescase,
f"Correct output is {correct_output} but calculated output is {test_d}"
)