troubleshooting xgb==2.0.0

setup.py

@@ -55,7 +55,7 @@
     include_package_data=True,
     install_requires=install_requires,
     extras_require={
-        "tests": ["flake8", "pytest", "coverage", "pre-commit"],
+        "tests": ["flake8", "pytest", "coverage", "pre-commit", "hypothesis", "distutils"],
         "sparkml": ["pyspark>=2.4.4", "pyarrow>1.0"],
         "onnx": onnx_requires,
         "extra": extra_requires,

tests/test_onnxml_decision_tree_converter.py

@@ -8,6 +8,7 @@
 from sklearn.tree import DecisionTreeClassifier, DecisionTreeRegressor
 from sklearn.ensemble import RandomForestClassifier, RandomForestRegressor
 from onnxconverter_common.data_types import FloatTensorType
+from distutils.version import LooseVersion
 
 from hummingbird.ml import convert
 from hummingbird.ml._utils import onnx_ml_tools_installed, onnx_runtime_installed, xgboost_installed
@@ -263,6 +264,7 @@ def test_random_forest_multi_random(self):
         not (onnx_ml_tools_installed() and onnx_runtime_installed()), reason="ONNXML test require ONNX, ORT and ONNXMLTOOLS"
     )
     @unittest.skipIf(not xgboost_installed(), reason="ONNXML BRANCH_LT test requires XGBoost")
+    @unittest.skipIf(LooseVersion(xgb.__version__) > LooseVersion("2.0"), reason="Unsupported XGBoost version")
     def test_xgboost_branch_lt(self):
         warnings.filterwarnings("ignore")
         n_features = 28

tests/test_xgboost_converter.py

@@ -15,6 +15,7 @@
 
 if xgboost_installed():
     import xgboost as xgb
+    from xgboost import testing as tm
 
 if pandas_installed():
     import pandas as pd
@@ -98,13 +99,11 @@ def test_xgb_perf_tree_trav_multi_classifier_converter(self):
     def _run_xgb_ranker_converter(self, num_classes, extra_config={}):
         warnings.filterwarnings("ignore")
         for max_depth in [1, 3, 8, 10, 12]:
-            model = xgb.XGBRanker(n_estimators=10, max_depth=max_depth)
-            np.random.seed(0)
-            X = np.random.rand(100, 200)
-            X = np.array(X, dtype=np.float32)
-            y = np.random.randint(num_classes, size=100)
+            model = xgb.XGBRanker(n_estimators=10, max_depth=max_depth, ndcg_exp_gain=True)  # ndcg is default
+            X, y, q, w = tm.make_ltr(n_samples=1024, n_features=num_classes, n_query_groups=3, max_rel=3)
 
-            model.fit(X, y, group=[X.shape[0]])
+            model.fit(X, y, qid=q, sample_weight=w, eval_set=[(X, y)],
+                      eval_qid=(q,), sample_weight_eval_set=(w,), verbose=True)
 
             torch_model = hummingbird.ml.convert(model, "torch", X, extra_config=extra_config)
             self.assertIsNotNone(torch_model)
@@ -256,7 +255,7 @@ def test_xgb_regressor_converter_torchscript(self):
         import torch
 
         for max_depth in [1, 3, 8, 10, 12]:
-            model = xgb.XGBRegressor(n_estimators=10, max_depth=max_depth)
+            model = xgb.XGBRegressor(n_estimators=10, max_depth=max_depth, objective="rank:pairwise")
             np.random.seed(0)
             X = np.random.rand(100, 200)
             X = np.array(X, dtype=np.float32)
@@ -275,7 +274,7 @@ def test_xgb_classifier_converter_torchscript(self):
         import torch
 
         for max_depth in [1, 3, 8, 10, 12]:
-            model = xgb.XGBClassifier(n_estimators=10, max_depth=max_depth)
+            model = xgb.XGBClassifier(n_estimators=10, max_depth=max_depth, objective="rank:pairwise")
             np.random.seed(0)
             X = np.random.rand(100, 200)
             X = np.array(X, dtype=np.float32)