def test_serializable_groupby_agg(c, s, a, b):
pd = pytest.importorskip("pandas")
dd = pytest.importorskip("dask.dataframe")
df = pd.DataFrame({"x": [1, 2, 3, 4], "y": [1, 0, 1, 0]})
ddf = dd.from_pandas(df, npartitions=2)
result = ddf.groupby("y").agg("count")
yield c.compute(result)
def test_blockwise_concatenate(c):
"""Test a blockwise operation with concatenated axes"""
da = pytest.importorskip("dask.array")
np = pytest.importorskip("numpy")
def f(x, y):
da.assert_eq(y, [[0, 1, 2]])
return x
x = da.from_array(np.array([0, 1, 2]))
y = da.from_array(np.array([[0, 1, 2]]))
z = da.blockwise(
f,
("i"),
x,
("i"),
y,
("ij"),
dtype=x.dtype,
concatenate=True,
)
c.compute(z, optimize_graph=False)
da.assert_eq(z, x, scheduler=c)
def test_predict_with_meta(client):
X, y, w = generate_array(with_weights=True)
partition_size = 20
margin = da.random.random(kRows, partition_size) + 1e4
dtrain = DaskDMatrix(client, X, y, weight=w, base_margin=margin)
booster = xgb.dask.train(
client, {}, dtrain, num_boost_round=4)['booster']
prediction = xgb.dask.predict(client, model=booster, data=dtrain)
assert prediction.ndim == 1
assert prediction.shape[0] == kRows
prediction = client.compute(prediction).result()
assert np.all(prediction > 1e3)
m = xgb.DMatrix(X.compute())
m.set_info(label=y.compute(), weight=w.compute(), base_margin=margin.compute())
single = booster.predict(m) # Make sure the ordering is correct.
assert np.all(prediction == single)
