Hendrik/xgboost
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

[CI] Refactor tests to reduce CI time. (#8312)

Browse Source
This commit is contained in:
Rory Mitchell
2022-10-12 11:32:06 +02:00
committed by GitHub
parent 39afdac3be
commit ce0382dcb0
19 changed files with 95 additions and 122 deletions
Download Patch File Download Diff File Expand all files Collapse all files

3
tests/python/test_data_iterator.py
View File

@@ -6,6 +6,7 @@ import pytest
from hypothesis import given, strategies, settings
from scipy.sparse import csr_matrix
pytestmark = pytest.mark.timeout(30)
def test_single_batch(tree_method: str = "approx") -> None:
from sklearn.datasets import load_breast_cancer
@@ -134,7 +135,7 @@ def run_data_iterator(
strategies.integers(0, 13),
strategies.booleans(),
)
@settings(deadline=None, print_blob=True)
@settings(deadline=None, max_examples=10, print_blob=True)
def test_data_iterator(
n_samples_per_batch: int,
n_features: int,

1
tests/python/test_demos.py
View File

@@ -4,6 +4,7 @@ import pytest
import testing as tm
import sys
pytestmark = pytest.mark.timeout(30)
ROOT_DIR = tm.PROJECT_ROOT
DEMO_DIR = os.path.join(ROOT_DIR, 'demo')

11
tests/python/test_linear.py
View File

@@ -1,7 +1,10 @@
import testing as tm
import pytest
from hypothesis import strategies, given, settings, note
import xgboost as xgb
pytestmark = pytest.mark.timeout(10)
parameter_strategy = strategies.fixed_dictionaries({
'booster': strategies.just('gblinear'),
'eta': strategies.floats(0.01, 0.25),
@@ -26,7 +29,7 @@ def train_result(param, dmat, num_rounds):
class TestLinear:
@given(parameter_strategy, strategies.integers(10, 50),
tm.dataset_strategy, coord_strategy)
@settings(deadline=None, print_blob=True)
@settings(deadline=None, max_examples=20, print_blob=True)
def test_coordinate(self, param, num_rounds, dataset, coord_param):
param['updater'] = 'coord_descent'
param.update(coord_param)
@@ -46,7 +49,7 @@ class TestLinear:
strategies.floats(1e-5, 0.8),
strategies.floats(1e-5, 0.8)
)
@settings(deadline=None, print_blob=True)
@settings(deadline=None, max_examples=20, print_blob=True)
def test_coordinate_regularised(self, param, num_rounds, dataset, coord_param, alpha, lambd):
param['updater'] = 'coord_descent'
param['alpha'] = alpha
@@ -59,7 +62,7 @@ class TestLinear:
@given(parameter_strategy, strategies.integers(10, 50),
tm.dataset_strategy)
@settings(deadline=None, print_blob=True)
@settings(deadline=None, max_examples=20, print_blob=True)
def test_shotgun(self, param, num_rounds, dataset):
param['updater'] = 'shotgun'
param = dataset.set_params(param)
@@ -76,7 +79,7 @@ class TestLinear:
@given(parameter_strategy, strategies.integers(10, 50),
tm.dataset_strategy, strategies.floats(1e-5, 1.0),
strategies.floats(1e-5, 1.0))
@settings(deadline=None, print_blob=True)
@settings(deadline=None, max_examples=20, print_blob=True)
def test_shotgun_regularised(self, param, num_rounds, dataset, alpha, lambd):
param['updater'] = 'shotgun'
param['alpha'] = alpha

14
tests/python/test_openmp.py
View File

@@ -8,6 +8,7 @@ import pytest
import testing as tm
pytestmark = pytest.mark.timeout(10)
class TestOMP:
def test_omp(self):
@@ -49,14 +50,15 @@ class TestOMP:
print('test approx ...')
param['tree_method'] = 'approx'
n_trials = 10
param['nthread'] = 1
auc_1, pred_1 = consist_test('approx_thread_1', 100)
auc_1, pred_1 = consist_test('approx_thread_1', n_trials)
param['nthread'] = 2
auc_2, pred_2 = consist_test('approx_thread_2', 100)
auc_2, pred_2 = consist_test('approx_thread_2', n_trials)
param['nthread'] = 3
auc_3, pred_3 = consist_test('approx_thread_3', 100)
auc_3, pred_3 = consist_test('approx_thread_3', n_trials)
assert auc_1 == auc_2 == auc_3
assert np.array_equal(auc_1, auc_2)
@@ -66,13 +68,13 @@ class TestOMP:
param['tree_method'] = 'hist'
param['nthread'] = 1
auc_1, pred_1 = consist_test('hist_thread_1', 100)
auc_1, pred_1 = consist_test('hist_thread_1', n_trials)
param['nthread'] = 2
auc_2, pred_2 = consist_test('hist_thread_2', 100)
auc_2, pred_2 = consist_test('hist_thread_2', n_trials)
param['nthread'] = 3
auc_3, pred_3 = consist_test('hist_thread_3', 100)
auc_3, pred_3 = consist_test('hist_thread_3', n_trials)
assert auc_1 == auc_2 == auc_3
assert np.array_equal(auc_1, auc_2)

45
tests/python/test_spark/test_spark_local.py
View File

@@ -16,10 +16,7 @@ if sys.platform.startswith("win") or sys.platform.startswith("darwin"):
pytest.skip("Skipping PySpark tests on Windows", allow_module_level=True)
from pyspark.ml import Pipeline, PipelineModel
from pyspark.ml.evaluation import (
BinaryClassificationEvaluator,
MulticlassClassificationEvaluator,
)
from pyspark.ml.evaluation import BinaryClassificationEvaluator
from pyspark.ml.feature import VectorAssembler
from pyspark.ml.functions import vector_to_array
from pyspark.ml.linalg import Vectors
@@ -40,6 +37,8 @@ from .utils import SparkTestCase
logging.getLogger("py4j").setLevel(logging.INFO)
pytestmark = pytest.mark.timeout(60)
class XgboostLocalTest(SparkTestCase):
def setUp(self):
@@ -711,17 +710,10 @@ class XgboostLocalTest(SparkTestCase):
estimatorParamMaps=paramMaps,
evaluator=BinaryClassificationEvaluator(),
seed=1,
numFolds=2,
)
cvBinModel = cvBin.fit(self.cls_df_train_large)
cvBinModel.transform(self.cls_df_test)
cvMulti = CrossValidator(
estimator=xgb_classifer,
estimatorParamMaps=paramMaps,
evaluator=MulticlassClassificationEvaluator(),
seed=1,
)
cvMultiModel = cvMulti.fit(self.multi_cls_df_train_large)
cvMultiModel.transform(self.multi_cls_df_test)
def test_callbacks(self):
from xgboost.callback import LearningRateScheduler
@@ -889,35 +881,6 @@ class XgboostLocalTest(SparkTestCase):
)
def test_classifier_with_weight_eval(self):
# with weight
classifier_with_weight = SparkXGBClassifier(weight_col="weight")
model_with_weight = classifier_with_weight.fit(
self.cls_df_train_with_eval_weight
)
pred_result_with_weight = model_with_weight.transform(
self.cls_df_test_with_eval_weight
).collect()
for row in pred_result_with_weight:
self.assertTrue(
np.allclose(row.probability, row.expected_prob_with_weight, atol=1e-3)
)
# with eval
classifier_with_eval = SparkXGBClassifier(**self.cls_params_with_eval)
model_with_eval = classifier_with_eval.fit(self.cls_df_train_with_eval_weight)
self.assertTrue(
np.isclose(
model_with_eval._xgb_sklearn_model.best_score,
self.cls_with_eval_best_score,
atol=1e-3,
)
)
pred_result_with_eval = model_with_eval.transform(
self.cls_df_test_with_eval_weight
).collect()
for row in pred_result_with_eval:
self.assertTrue(
np.allclose(row.probability, row.expected_prob_with_eval, atol=1e-3)
)
# with weight and eval
# Added scale_pos_weight because in 1.4.2, the original answer returns 0.5 which
# doesn't really indicate this working correctly.

38
tests/python/test_with_dask.py
View File

@@ -44,6 +44,8 @@ from xgboost.dask import DaskDMatrix
dask.config.set({"distributed.scheduler.allowed-failures": False})
pytestmark = pytest.mark.timeout(30)
if hasattr(HealthCheck, 'function_scoped_fixture'):
suppress = [HealthCheck.function_scoped_fixture]
else:
@@ -381,7 +383,7 @@ def test_categorical(client: "Client") -> None:
def test_dask_predict_shape_infer(client: "Client") -> None:
X, y = make_classification(n_samples=1000, n_informative=5, n_classes=3)
X, y = make_classification(n_samples=kRows, n_informative=5, n_classes=3)
X_ = dd.from_array(X, chunksize=100)
y_ = dd.from_array(y, chunksize=100)
dtrain = xgb.dask.DaskDMatrix(client, data=X_, label=y_)
@@ -522,8 +524,8 @@ def test_boost_from_prediction(tree_method: str, client: "Client") -> None:
def test_inplace_predict(client: "Client") -> None:
from sklearn.datasets import fetch_california_housing
X_, y_ = fetch_california_housing(return_X_y=True)
from sklearn.datasets import load_diabetes
X_, y_ = load_diabetes(return_X_y=True)
X, y = dd.from_array(X_, chunksize=32), dd.from_array(y_, chunksize=32)
reg = xgb.dask.DaskXGBRegressor(n_estimators=4).fit(X, y)
booster = reg.get_booster()
@@ -841,7 +843,7 @@ def run_empty_dmatrix_cls(client: "Client", parameters: dict) -> None:
def run_empty_dmatrix_auc(client: "Client", tree_method: str, n_workers: int) -> None:
from sklearn import datasets
n_samples = 100
n_features = 97
n_features = 7
rng = np.random.RandomState(1994)
make_classification = partial(
@@ -894,9 +896,9 @@ def run_empty_dmatrix_auc(client: "Client", tree_method: str, n_workers: int) ->
def test_empty_dmatrix_auc() -> None:
with LocalCluster(n_workers=8, dashboard_address=":0") as cluster:
with LocalCluster(n_workers=4, dashboard_address=":0") as cluster:
with Client(cluster) as client:
run_empty_dmatrix_auc(client, "hist", 8)
run_empty_dmatrix_auc(client, "hist", 4)
def run_auc(client: "Client", tree_method: str) -> None:
@@ -1033,7 +1035,7 @@ async def run_dask_classifier_asyncio(scheduler_address: str) -> None:
def test_with_asyncio() -> None:
with LocalCluster(dashboard_address=":0") as cluster:
with LocalCluster(n_workers=2, dashboard_address=":0") as cluster:
with Client(cluster) as client:
address = client.scheduler.address
output = asyncio.run(run_from_dask_array_asyncio(address))
@@ -1420,11 +1422,11 @@ class TestWithDask:
@given(params=hist_parameter_strategy,
dataset=tm.dataset_strategy)
@settings(deadline=None, suppress_health_check=suppress, print_blob=True)
@settings(deadline=None, max_examples=10, suppress_health_check=suppress, print_blob=True)
def test_hist(
self, params: Dict, dataset: tm.TestDataset, client: "Client"
) -> None:
num_rounds = 30
num_rounds = 10
self.run_updater_test(client, params, num_rounds, dataset, 'hist')
def test_quantile_dmatrix(self, client: Client) -> None:
@@ -1465,11 +1467,11 @@ class TestWithDask:
@given(params=exact_parameter_strategy,
dataset=tm.dataset_strategy)
@settings(deadline=None, suppress_health_check=suppress, print_blob=True)
@settings(deadline=None, max_examples=10, suppress_health_check=suppress, print_blob=True)
def test_approx(
self, client: "Client", params: Dict, dataset: tm.TestDataset
) -> None:
num_rounds = 30
num_rounds = 10
self.run_updater_test(client, params, num_rounds, dataset, 'approx')
def run_quantile(self, name: str) -> None:
@@ -1773,16 +1775,16 @@ class TestWithDask:
assert np.allclose(np.sum(shap, axis=len(shap.shape) - 1), margin, 1e-5, 1e-5)
def test_shap(self, client: "Client") -> None:
from sklearn.datasets import fetch_california_housing, load_digits
X, y = fetch_california_housing(return_X_y=True)
from sklearn.datasets import load_diabetes, load_iris
X, y = load_diabetes(return_X_y=True)
params: Dict[str, Any] = {'objective': 'reg:squarederror'}
self.run_shap(X, y, params, client)
X, y = load_digits(return_X_y=True)
params = {'objective': 'multi:softmax', 'num_class': 10}
X, y = load_iris(return_X_y=True)
params = {'objective': 'multi:softmax', 'num_class': 3}
self.run_shap(X, y, params, client)
params = {'objective': 'multi:softprob', 'num_class': 10}
params = {'objective': 'multi:softprob', 'num_class': 3}
self.run_shap(X, y, params, client)
self.run_shap_cls_sklearn(X, y, client)
@@ -1818,8 +1820,8 @@ class TestWithDask:
1e-5, 1e-5)
def test_shap_interactions(self, client: "Client") -> None:
from sklearn.datasets import fetch_california_housing
X, y = fetch_california_housing(return_X_y=True)
from sklearn.datasets import load_diabetes
X, y = load_diabetes(return_X_y=True)
params = {'objective': 'reg:squarederror'}
self.run_shap_interactions(X, y, params, client)

4
tests/python/test_with_modin.py
View File

@@ -14,10 +14,6 @@ except ImportError:
pytestmark = pytest.mark.skipif(**tm.no_modin())
dpath = 'demo/data/'
rng = np.random.RandomState(1994)
class TestModin:
@pytest.mark.xfail
def test_modin(self):

34
tests/python/test_with_sklearn.py
View File

@@ -12,7 +12,7 @@ import json
rng = np.random.RandomState(1994)
pytestmark = pytest.mark.skipif(**tm.no_sklearn())
pytestmark = [pytest.mark.skipif(**tm.no_sklearn()), pytest.mark.timeout(30)]
from sklearn.utils.estimator_checks import parametrize_with_checks
@@ -328,10 +328,10 @@ def test_select_feature():
def test_num_parallel_tree():
from sklearn.datasets import fetch_california_housing
from sklearn.datasets import load_diabetes
reg = xgb.XGBRegressor(n_estimators=4, num_parallel_tree=4, tree_method="hist")
X, y = fetch_california_housing(return_X_y=True)
X, y = load_diabetes(return_X_y=True)
bst = reg.fit(X=X, y=y)
dump = bst.get_booster().get_dump(dump_format="json")
assert len(dump) == 16
@@ -352,7 +352,7 @@ def test_num_parallel_tree():
)
def test_calif_housing_regression():
def test_regression():
from sklearn.metrics import mean_squared_error
from sklearn.datasets import fetch_california_housing
from sklearn.model_selection import KFold
@@ -381,7 +381,7 @@ def test_calif_housing_regression():
xgb_model.feature_names_in_
def run_calif_housing_rf_regression(tree_method):
def run_housing_rf_regression(tree_method):
from sklearn.metrics import mean_squared_error
from sklearn.datasets import fetch_california_housing
from sklearn.model_selection import KFold
@@ -401,8 +401,8 @@ def run_calif_housing_rf_regression(tree_method):
rfreg.fit(X, y, early_stopping_rounds=10)
def test_calif_housing_rf_regression():
run_calif_housing_rf_regression("hist")
def test_rf_regression():
run_housing_rf_regression("hist")
def test_parameter_tuning():
@@ -411,9 +411,9 @@ def test_parameter_tuning():
X, y = fetch_california_housing(return_X_y=True)
xgb_model = xgb.XGBRegressor(learning_rate=0.1)
clf = GridSearchCV(xgb_model, {'max_depth': [2, 4, 6],
'n_estimators': [50, 100, 200]},
cv=3, verbose=1)
clf = GridSearchCV(xgb_model, {'max_depth': [2, 4],
'n_estimators': [50, 200]},
cv=2, verbose=1)
clf.fit(X, y)
assert clf.best_score_ < 0.7
assert clf.best_params_ == {'n_estimators': 200, 'max_depth': 4}
@@ -840,13 +840,13 @@ def test_save_load_model():
def test_RFECV():
from sklearn.datasets import fetch_california_housing
from sklearn.datasets import load_diabetes
from sklearn.datasets import load_breast_cancer
from sklearn.datasets import load_iris
from sklearn.feature_selection import RFECV
# Regression
X, y = fetch_california_housing(return_X_y=True)
X, y = load_diabetes(return_X_y=True)
bst = xgb.XGBRegressor(booster='gblinear', learning_rate=0.1,
n_estimators=10,
objective='reg:squarederror',
@@ -861,7 +861,7 @@ def test_RFECV():
n_estimators=10,
objective='binary:logistic',
random_state=0, verbosity=0)
rfecv = RFECV(estimator=bst, step=1, cv=3, scoring='roc_auc')
rfecv = RFECV(estimator=bst, step=0.5, cv=3, scoring='roc_auc')
rfecv.fit(X, y)
# Multi-class classification
@@ -872,7 +872,7 @@ def test_RFECV():
objective='multi:softprob',
random_state=0, reg_alpha=0.001, reg_lambda=0.01,
scale_pos_weight=0.5, verbosity=0)
rfecv = RFECV(estimator=bst, step=1, cv=3, scoring='neg_log_loss')
rfecv = RFECV(estimator=bst, step=0.5, cv=3, scoring='neg_log_loss')
rfecv.fit(X, y)
X[0:4, :] = np.nan # verify scikit_learn doesn't throw with nan
@@ -881,7 +881,7 @@ def test_RFECV():
rfecv.fit(X, y)
cls = xgb.XGBClassifier()
rfecv = RFECV(estimator=cls, step=1, cv=3,
rfecv = RFECV(estimator=cls, step=0.5, cv=3,
scoring='neg_mean_squared_error')
rfecv.fit(X, y)
@@ -1155,7 +1155,7 @@ def run_boost_from_prediction_multi_clasas(
@pytest.mark.parametrize("tree_method", ["hist", "approx", "exact"])
def test_boost_from_prediction(tree_method):
from sklearn.datasets import load_breast_cancer, load_digits, make_regression
from sklearn.datasets import load_breast_cancer, load_iris, make_regression
import pandas as pd
X, y = load_breast_cancer(return_X_y=True)
@@ -1163,7 +1163,7 @@ def test_boost_from_prediction(tree_method):
run_boost_from_prediction_binary(tree_method, X, y, None)
run_boost_from_prediction_binary(tree_method, X, y, pd.DataFrame)
X, y = load_digits(return_X_y=True)
X, y = load_iris(return_X_y=True)
run_boost_from_prediction_multi_clasas(xgb.XGBClassifier, tree_method, X, y, None)
run_boost_from_prediction_multi_clasas(