Namespaces
namespace	dask

namespace	data

namespace	data_iter

namespace	metrics

namespace	params

namespace	ranking

namespace	shared

namespace	updater

Data Structures
class	DirectoryExcursion

class	IteratorForTest

class	TestDataset

Functions
bool	has_ipv6 ()

PytestSkip	no_mod (str name)

PytestSkip	no_ipv6 ()

PytestSkip	no_ubjson ()

PytestSkip	no_sklearn ()

PytestSkip	no_dask ()

PytestSkip	no_dask_ml ()

PytestSkip	no_spark ()

PytestSkip	no_pandas ()

PytestSkip	no_arrow ()

PytestSkip	no_modin ()

PytestSkip	no_dt ()

PytestSkip	no_matplotlib ()

PytestSkip	no_dask_cuda ()

PytestSkip	no_cudf ()

PytestSkip	no_cupy ()

PytestSkip	no_dask_cudf ()

PytestSkip	no_json_schema ()

PytestSkip	no_graphviz ()

PytestSkip	no_rmm ()

PytestSkip	no_multiple (*Any args)

PytestSkip	skip_s390x ()

Tuple[List[np.ndarray], List[np.ndarray], List[np.ndarray]]	make_batches (int n_samples_per_batch, int n_features, int n_batches, bool use_cupy=False, *bool vary_size=False)

Tuple[ArrayLike, ArrayLike, ArrayLike]	make_regression (int n_samples, int n_features, bool use_cupy)

Tuple[List[sparse.csr_matrix], List[np.ndarray], List[np.ndarray]]	make_batches_sparse (int n_samples_per_batch, int n_features, int n_batches, float sparsity)

Tuple[ArrayLike, np.ndarray]	make_categorical (int n_samples, int n_features, int n_categories, bool onehot, float sparsity=0.0, float cat_ratio=1.0, bool shuffle=False)

Tuple[np.ndarray, np.ndarray, np.ndarray, np.ndarray]	make_ltr (int n_samples, int n_features, int n_query_groups, int max_rel)

strategies.SearchStrategy	_cat_sampled_from ()

Tuple[Union[sparse.csr_matrix], np.ndarray]	make_sparse_regression (int n_samples, int n_features, float sparsity, bool as_dense)

Callable	make_datasets_with_margin (strategies.SearchStrategy unweighted_strategy)

Callable	make_dataset_strategy ()

bool	non_increasing (Sequence[float] L, float tolerance=1e-4)

bool	predictor_equal (xgb.DMatrix lhs, xgb.DMatrix rhs)

Tuple[str, np.float64]	eval_error_metric (np.ndarray predt, xgb.DMatrix dtrain)

np.float64	eval_error_metric_skl (np.ndarray y_true, np.ndarray y_score)

float	root_mean_square (np.ndarray y_true, np.ndarray y_score)

np.ndarray	softmax (np.ndarray x)

SklObjective	softprob_obj (int classes)

Generator[Tuple[StringIO, StringIO], None, None]	captured_output ()

Any	timeout (int sec, Any args, bool enable=True, *Any kwargs)

None	setup_rmm_pool (Any _, pytest.Config pytestconfig)

List[str]	get_client_workers (Any client)

str	demo_dir (str path)

str	normpath (str path)

str	data_dir (str path)

Tuple[xgb.DMatrix, xgb.DMatrix]	load_agaricus (str path)

str	project_root (str path)

Variables
	hypothesis = pytest.importorskip("hypothesis")

	datasets = pytest.importorskip("sklearn.datasets")

	PytestSkip = TypedDict("PytestSkip", {"condition": bool, "reason": str})

strategies	categorical_dataset_strategy = _cat_sampled_from()

	sparse_datasets_strategy

	_unweighted_multi_datasets_strategy

Callable	multi_dataset_strategy

	M = TypeVar("M", xgb.Booster, xgb.XGBModel)

Detailed Description

Utilities for defining Python tests. The module is private and subject to frequent
change without notice.

Function Documentation

◆ captured_output()

Generator[Tuple[StringIO, StringIO], None, None] xgboost.testing.captured_output ( )

Reassign stdout temporarily in order to test printed statements
Taken from:
https://stackoverflow.com/questions/4219717/how-to-assert-output-with-nosetest-unittest-in-python

Also works for pytest.

◆ demo_dir()

str xgboost.testing.demo_dir ( str path )

Look for the demo directory based on the test file name.

◆ eval_error_metric()

Tuple[str, np.float64] xgboost.testing.eval_error_metric	(	np.ndarray	predt,
		xgb.DMatrix	dtrain
	)

Evaluation metric for xgb.train

◆ eval_error_metric_skl()

np.float64 xgboost.testing.eval_error_metric_skl	(	np.ndarray	y_true,
		np.ndarray	y_score
	)

Evaluation metric that looks like metrics provided by sklearn.

◆ has_ipv6()

bool xgboost.testing.has_ipv6 ( )

Check whether IPv6 is enabled on this host.

◆ make_categorical()

Tuple[ArrayLike, np.ndarray] xgboost.testing.make_categorical	(	int	n_samples,
		int	n_features,
		int	n_categories,
		bool	onehot,
		float	sparsity = `0.0`,
		float	cat_ratio = `1.0`,
		bool	shuffle = `False`
	)

Generate categorical features for test.

Parameters
----------
n_categories:
    Number of categories for categorical features.
onehot:
    Should we apply one-hot encoding to the data?
sparsity:
    The ratio of the amount of missing values over the number of all entries.
cat_ratio:
    The ratio of features that are categorical.
shuffle:
    Whether we should shuffle the columns.

Returns
-------
X, y

◆ make_datasets_with_margin()

Callable xgboost.testing.make_datasets_with_margin ( strategies.SearchStrategy unweighted_strategy )

Factory function for creating strategies that generates datasets with weight and
base margin.

◆ make_ltr()

Tuple[np.ndarray, np.ndarray, np.ndarray, np.ndarray] xgboost.testing.make_ltr	(	int	n_samples,
		int	n_features,
		int	n_query_groups,
		int	max_rel
	)

Make a dataset for testing LTR.

◆ make_regression()

Tuple[ArrayLike, ArrayLike, ArrayLike] xgboost.testing.make_regression	(	int	n_samples,
		int	n_features,
		bool	use_cupy
	)

Make a simple regression dataset.

◆ make_sparse_regression()

Tuple[Union[sparse.csr_matrix], np.ndarray] xgboost.testing.make_sparse_regression	(	int	n_samples,
		int	n_features,
		float	sparsity,
		bool	as_dense
	)

Make sparse matrix.

Parameters
----------

as_dense:

  Return the matrix as np.ndarray with missing values filled by NaN

◆ no_ipv6()

PytestSkip xgboost.testing.no_ipv6 ( )

PyTest skip mark for IPv6.

◆ predictor_equal()

bool xgboost.testing.predictor_equal	(	xgb.DMatrix	lhs,
		xgb.DMatrix	rhs
	)

Assert whether two DMatrices contain the same predictors.

◆ timeout()

Any xgboost.testing.timeout	(	int	sec,
		*Any	args,
		bool	enable = `True`,
		**Any	kwargs
	)

Make a pytest mark for the `pytest-timeout` package.

Parameters
----------
sec :
    Timeout seconds.
enable :
    Control whether timeout should be applied, used for debugging.

Returns
-------
pytest.mark.timeout

Variable Documentation

◆ _unweighted_multi_datasets_strategy

xgboost.testing._unweighted_multi_datasets_strategy

protected

Initial value:

=  strategies.sampled_from(
    [
        TestDataset("digits", get_digits, "multi:softmax", "mlogloss"),
        TestDataset(
            "mtreg",
            lambda: datasets.make_regression(n_samples=128, n_features=2, n_targets=3),
            "reg:squarederror",
            "rmse",
        ),
        TestDataset(
            "mtreg-l1",
            lambda: datasets.make_regression(n_samples=128, n_features=2, n_targets=3),
            "reg:absoluteerror",
            "mae",
        ),
    ]
)

◆ multi_dataset_strategy

Callable xgboost.testing.multi_dataset_strategy