def test_missing_mandatory_kwarg(self):
aggregator = WeightedPercentileAggregator()
emsg = "weighted_percentile aggregator requires " \
".* keyword argument 'percent'"
with self.assertRaisesRegexp(ValueError, emsg):
aggregator.aggregate('dummy', axis=0, weights=None)
emsg = "weighted_percentile aggregator requires " \
".* keyword argument 'weights'"
with self.assertRaisesRegexp(ValueError, emsg):
aggregator.aggregate('dummy', axis=0, percent=50)
def test_init(self):
name = 'weighted_percentile'
call_func = _weighted_percentile
units_func = sentinel.units_func
lazy_func = sentinel.lazy_func
aggregator = WeightedPercentileAggregator(units_func=units_func,
lazy_func=lazy_func)
self.assertEqual(aggregator.name(), name)
self.assertIs(aggregator.call_func, call_func)
self.assertIs(aggregator.units_func, units_func)
self.assertIs(aggregator.lazy_func, lazy_func)
self.assertIsNone(aggregator.cell_method)
def test_init(self):
name = 'weighted_percentile'
call_func = _weighted_percentile
units_func = sentinel.units_func
lazy_func = sentinel.lazy_func
aggregator = WeightedPercentileAggregator(units_func=units_func,
lazy_func=lazy_func)
self.assertEqual(aggregator.name(), name)
self.assertIs(aggregator.call_func, call_func)
self.assertIs(aggregator.units_func, units_func)
self.assertIs(aggregator.lazy_func, lazy_func)
self.assertIsNone(aggregator.cell_method)
def test_simple_single_point(self):
aggregator = WeightedPercentileAggregator()
percent = 50
kwargs = dict(percent=percent, weights=self.weights_simple)
data = np.empty(self.cube_simple.shape)
coords = [self.coord_simple]
actual = aggregator.post_process(self.cube_simple, data, coords,
**kwargs)
self.assertEqual(actual.shape, self.cube_simple.shape)
self.assertIs(actual.data, data)
name = 'weighted_percentile_over_time'
coord = actual.coord(name)
expected = AuxCoord(percent, long_name=name)
self.assertEqual(coord, expected)
def test_simple_single_point(self):
aggregator = WeightedPercentileAggregator()
percent = 50
kwargs = dict(percent=percent, weights=self.weights_simple)
data = np.empty(self.cube_simple.shape)
coords = [self.coord_simple]
actual = aggregator.post_process(self.cube_simple, data, coords,
**kwargs)
self.assertEqual(actual.shape, self.cube_simple.shape)
self.assertIs(actual.data, data)
name = 'weighted_percentile_over_time'
coord = actual.coord(name)
expected = AuxCoord(percent, long_name=name)
self.assertEqual(coord, expected)
def test_multi_single_point(self):
aggregator = WeightedPercentileAggregator()
percent = 70
kwargs = dict(percent=percent, weights=self.weights_multi)
data = np.empty(self.cube_multi.shape)
coords = [self.coord_multi_0]
actual = aggregator.post_process(self.cube_multi, data, coords,
**kwargs)
self.assertEqual(actual.shape, self.cube_multi.shape)
self.assertIs(actual.data, data)
name = "weighted_percentile_over_time"
coord = actual.coord(name)
expected = AuxCoord(percent, long_name=name, units="percent")
self.assertEqual(coord, expected)
def test_multi_multiple_points(self):
aggregator = WeightedPercentileAggregator()
percent = np.array([17, 29, 81])
kwargs = dict(percent=percent, weights=self.weights_multi)
shape = self.cube_multi.shape + percent.shape
data = np.empty(shape)
coords = [self.coord_multi_0]
actual = aggregator.post_process(self.cube_multi, data, coords,
**kwargs)
self.assertEqual(actual.shape, percent.shape + self.cube_multi.shape)
expected = np.rollaxis(data, -1)
self.assertArrayEqual(actual.data, expected)
name = 'weighted_percentile_over_time'
coord = actual.coord(name)
expected = AuxCoord(percent, long_name=name)
self.assertEqual(coord, expected)
def test_multi_multiple_points(self):
aggregator = WeightedPercentileAggregator()
percent = np.array([17, 29, 81])
kwargs = dict(percent=percent, weights=self.weights_multi)
shape = self.cube_multi.shape + percent.shape
data = np.empty(shape)
coords = [self.coord_multi_0]
actual = aggregator.post_process(self.cube_multi, data, coords,
**kwargs)
self.assertEqual(actual.shape, percent.shape + self.cube_multi.shape)
expected = np.rollaxis(data, -1)
self.assertArrayEqual(actual.data, expected)
name = 'weighted_percentile_over_time'
coord = actual.coord(name)
expected = AuxCoord(percent, long_name=name)
self.assertEqual(coord, expected)
def test_simple_multiple_points(self):
aggregator = WeightedPercentileAggregator()
percent = np.array([10, 20, 50, 90])
kwargs = dict(percent=percent, weights=self.weights_simple,
returned=True)
shape = self.cube_simple.shape + percent.shape
data = np.empty(shape)
total_weights = 1.
coords = [self.coord_simple]
actual = aggregator.post_process(
self.cube_simple, (data, total_weights), coords, **kwargs)
self.assertEqual(len(actual), 2)
self.assertEqual(actual[0].shape,
percent.shape + self.cube_simple.shape)
expected = np.rollaxis(data, -1)
self.assertArrayEqual(actual[0].data, expected)
self.assertIs(actual[1], total_weights)
name = 'weighted_percentile_over_time'
coord = actual[0].coord(name)
expected = AuxCoord(percent, long_name=name)
self.assertEqual(coord, expected)
def test_simple_multiple_points(self):
aggregator = WeightedPercentileAggregator()
percent = np.array([10, 20, 50, 90])
kwargs = dict(percent=percent, weights=self.weights_simple,
returned=True)
shape = self.cube_simple.shape + percent.shape
data = np.empty(shape)
total_weights = 1.
coords = [self.coord_simple]
actual = aggregator.post_process(
self.cube_simple, (data, total_weights), coords, **kwargs)
self.assertEqual(len(actual), 2)
self.assertEqual(actual[0].shape,
percent.shape + self.cube_simple.shape)
expected = np.rollaxis(data, -1)
self.assertArrayEqual(actual[0].data, expected)
self.assertIs(actual[1], total_weights)
name = 'weighted_percentile_over_time'
coord = actual[0].coord(name)
expected = AuxCoord(percent, long_name=name)
self.assertEqual(coord, expected)
def test_missing_mandatory_kwarg(self):
aggregator = WeightedPercentileAggregator()
emsg = "weighted_percentile aggregator requires " \
".* keyword argument 'percent'"
with self.assertRaisesRegexp(ValueError, emsg):
aggregator.aggregate('dummy', axis=0, weights=None)
emsg = "weighted_percentile aggregator requires " \
".* keyword argument 'weights'"
with self.assertRaisesRegexp(ValueError, emsg):
aggregator.aggregate('dummy', axis=0, percent=50)
