def test_obj_3D_rectangular(use_weights):
x = da.random.standard_normal(size=(2000, 3), chunks=(400, 3))
if use_weights:
weights = da.random.uniform(0.5,
0.75,
size=x.shape[0],
chunks=x.chunksize[0])
storage = dhb.storage.Weight()
else:
weights = None
storage = dhb.storage.Double()
h = dhb.Histogram(
dhb.axis.Regular(8, -3.5, 3.5),
dhb.axis.Regular(7, -3.3, 3.3),
dhb.axis.Regular(9, -3.2, 3.2),
storage=storage,
)
h.fill(x, weight=weights)
h.compute()
control = bh.Histogram(*h.axes, storage=h._storage_type())
if use_weights:
control.fill(*(x.compute().T), weight=weights.compute())
else:
control.fill(*(x.compute().T))
assert np.allclose(h.counts(), control.counts())
if use_weights:
assert np.allclose(h.variances(), control.variances())
def test_clear_fills():
x = da.random.standard_normal(size=(8, 2), chunks=(4, 2))
h = dhb.Histogram(
dhb.axis.Regular(8, -3.5, 3.5),
dhb.axis.Regular(9, -3.2, 3.2),
)
h.fill(x)
assert h.staged_fills()
h.clear_fills()
assert not h.staged_fills()
def test_to_da(flow):
x = da.random.standard_normal(size=(3_000, ), chunks=500)
h1 = dhb.Histogram(dhb.axis.Regular(10, -3, 3))
h2 = bh.Histogram(bh.axis.Regular(10, -3, 3))
h1.fill(x)
h2.fill(x.compute())
h1c, h1e = h1.to_dask_array(dd=True, flow=flow)
h2c, h2e = h2.to_numpy(dd=True, flow=flow)
np.testing.assert_array_almost_equal(h1c.compute(), h2c)
np.testing.assert_array_almost_equal(h1e[0].compute(), h2e[0])
def test_to_delayed():
x = da.random.standard_normal(size=(2_000, 3), chunks=(500, 3))
bins = [
[-3, -2.2, 0, 1.1, 2.2, 3.3],
[-4, -1.1, 0, 2.2, 3.3, 4.4],
[-2, -0.9, 0, 0.3, 2.2],
]
dh = dhb.Histogram(
dhb.axis.Variable(bins[0]),
dhb.axis.Variable(bins[1]),
dhb.axis.Variable(bins[2]),
)
dh.fill(x)
dh.compute()
dh.fill(x)
ch = dhc.clone(dh)
ch.fill(*(x.compute().T))
ch.fill(*(x.compute().T))
np.testing.assert_array_almost_equal(
dh.to_delayed().compute().to_numpy()[0],
ch.to_numpy()[0])
def test_obj_1D(use_weights):
x = da.random.standard_normal(size=(2000, ), chunks=(400, ))
if use_weights:
weights = da.random.uniform(0.5, 0.75, size=x.shape, chunks=x.chunks)
storage = dhb.storage.Weight()
else:
weights = None
storage = dhb.storage.Double()
h = dhb.Histogram(dhb.axis.Regular(12, -3, 3), storage=storage)
h.fill(x, weight=weights)
h.compute()
control = bh.Histogram(*h.axes, storage=h._storage_type())
if use_weights:
control.fill(x.compute(), weight=weights.compute())
else:
control.fill(x.compute())
assert np.allclose(h.counts(), control.counts())
if use_weights:
assert np.allclose(h.variances(), control.variances())
def test_concrete_fill():
x = da.random.standard_normal(size=(500, 2), chunks=(4, 2))
h = dhb.Histogram(
dhb.axis.Regular(8, -3.5, 3.5),
dhb.axis.Regular(9, -3.2, 3.2),
)
h.fill(x)
material_x = x.compute()
h.fill(*material_x.T)
h.compute()
h2 = bh.Histogram(*h.axes)
h2.fill(*material_x.T)
h2.fill(*material_x.T)
np.testing.assert_array_almost_equal(h2.values(), h.values())
with pytest.raises(
TypeError,
match="concrete_fill does not support Dask collections, "
"only materialized data; use the Histogram.fill method.",
):
h.concrete_fill(x)