scikit-learn-intelex 2024.2.0__py38-none-manylinux1_x86_64.whl → 2024.4.0__py38-none-manylinux1_x86_64.whl

sklearnex/basic_statistics/tests/test_incremental_basic_statistics.py

@@ -0,0 +1,386 @@
+# ===============================================================================
+# Copyright 2024 Intel Corporation
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+# ===============================================================================
+
+import numpy as np
+import pytest
+from numpy.testing import assert_allclose
+
+from onedal.basic_statistics.tests.test_incremental_basic_statistics import (
+    expected_max,
+    expected_mean,
+    expected_sum,
+    options_and_tests,
+)
+from onedal.tests.utils._dataframes_support import (
+    _convert_to_dataframe,
+    get_dataframes_and_queues,
+)
+from sklearnex.basic_statistics import IncrementalBasicStatistics
+
+
+@pytest.mark.parametrize("dataframe,queue", get_dataframes_and_queues())
+@pytest.mark.parametrize("weighted", [True, False])
+@pytest.mark.parametrize("dtype", [np.float32, np.float64])
+def test_partial_fit_multiple_options_on_gold_data(dataframe, queue, weighted, dtype):
+    X = np.array([[0, 0], [1, 1]])
+    X = X.astype(dtype=dtype)
+    X_split = np.array_split(X, 2)
+    if weighted:
+        weights = np.array([1, 0.5])
+        weights = weights.astype(dtype=dtype)
+        weights_split = np.array_split(weights, 2)
+
+    incbs = IncrementalBasicStatistics()
+    for i in range(2):
+        X_split_df = _convert_to_dataframe(
+            X_split[i], sycl_queue=queue, target_df=dataframe
+        )
+        if weighted:
+            weights_split_df = _convert_to_dataframe(
+                weights_split[i], sycl_queue=queue, target_df=dataframe
+            )
+            result = incbs.partial_fit(X_split_df, sample_weight=weights_split_df)
+        else:
+            result = incbs.partial_fit(X_split_df)
+
+    if weighted:
+        expected_weighted_mean = np.array([0.25, 0.25])
+        expected_weighted_min = np.array([0, 0])
+        expected_weighted_max = np.array([0.5, 0.5])
+        assert_allclose(expected_weighted_mean, result.mean)
+        assert_allclose(expected_weighted_max, result.max)
+        assert_allclose(expected_weighted_min, result.min)
+    else:
+        expected_mean = np.array([0.5, 0.5])
+        expected_min = np.array([0, 0])
+        expected_max = np.array([1, 1])
+        assert_allclose(expected_mean, result.mean)
+        assert_allclose(expected_max, result.max)
+        assert_allclose(expected_min, result.min)
+
+
+@pytest.mark.parametrize("dataframe,queue", get_dataframes_and_queues())
+@pytest.mark.parametrize("num_batches", [2, 10])
+@pytest.mark.parametrize("option", options_and_tests)
+@pytest.mark.parametrize("row_count", [100, 1000])
+@pytest.mark.parametrize("column_count", [10, 100])
+@pytest.mark.parametrize("weighted", [True, False])
+@pytest.mark.parametrize("dtype", [np.float32, np.float64])
+def test_partial_fit_single_option_on_random_data(
+    dataframe, queue, num_batches, option, row_count, column_count, weighted, dtype
+):
+    result_option, function, tols = option
+    fp32tol, fp64tol = tols
+    seed = 77
+    gen = np.random.default_rng(seed)
+    X = gen.uniform(low=-0.3, high=+0.7, size=(row_count, column_count))
+    X = X.astype(dtype=dtype)
+    X_split = np.array_split(X, num_batches)
+    if weighted:
+        weights = gen.uniform(low=-0.5, high=+1.0, size=row_count)
+        weights = weights.astype(dtype=dtype)
+        weights_split = np.array_split(weights, num_batches)
+    incbs = IncrementalBasicStatistics(result_options=result_option)
+
+    for i in range(num_batches):
+        X_split_df = _convert_to_dataframe(
+            X_split[i], sycl_queue=queue, target_df=dataframe
+        )
+        if weighted:
+            weights_split_df = _convert_to_dataframe(
+                weights_split[i], sycl_queue=queue, target_df=dataframe
+            )
+            result = incbs.partial_fit(X_split_df, sample_weight=weights_split_df)
+        else:
+            result = incbs.partial_fit(X_split_df)
+
+    res = getattr(result, result_option)
+    if weighted:
+        weighted_data = np.diag(weights) @ X
+        gtr = function(weighted_data)
+    else:
+        gtr = function(X)
+
+    tol = fp32tol if res.dtype == np.float32 else fp64tol
+    assert_allclose(gtr, res, atol=tol)
+
+
+@pytest.mark.parametrize("dataframe,queue", get_dataframes_and_queues())
+@pytest.mark.parametrize("num_batches", [2, 10])
+@pytest.mark.parametrize("row_count", [100, 1000])
+@pytest.mark.parametrize("column_count", [10, 100])
+@pytest.mark.parametrize("weighted", [True, False])
+@pytest.mark.parametrize("dtype", [np.float32, np.float64])
+def test_partial_fit_multiple_options_on_random_data(
+    dataframe, queue, num_batches, row_count, column_count, weighted, dtype
+):
+    seed = 42
+    gen = np.random.default_rng(seed)
+    X = gen.uniform(low=-0.3, high=+0.7, size=(row_count, column_count))
+    X = X.astype(dtype=dtype)
+    X_split = np.array_split(X, num_batches)
+    if weighted:
+        weights = gen.uniform(low=-0.5, high=+1.0, size=row_count)
+        weights = weights.astype(dtype=dtype)
+        weights_split = np.array_split(weights, num_batches)
+    incbs = IncrementalBasicStatistics(result_options=["mean", "max", "sum"])
+
+    for i in range(num_batches):
+        X_split_df = _convert_to_dataframe(
+            X_split[i], sycl_queue=queue, target_df=dataframe
+        )
+        if weighted:
+            weights_split_df = _convert_to_dataframe(
+                weights_split[i], sycl_queue=queue, target_df=dataframe
+            )
+            result = incbs.partial_fit(X_split_df, sample_weight=weights_split_df)
+        else:
+            result = incbs.partial_fit(X_split_df)
+
+    res_mean, res_max, res_sum = result.mean, result.max, result.sum
+    if weighted:
+        weighted_data = np.diag(weights) @ X
+        gtr_mean, gtr_max, gtr_sum = (
+            expected_mean(weighted_data),
+            expected_max(weighted_data),
+            expected_sum(weighted_data),
+        )
+    else:
+        gtr_mean, gtr_max, gtr_sum = (
+            expected_mean(X),
+            expected_max(X),
+            expected_sum(X),
+        )
+
+    tol = 1e-5 if res_mean.dtype == np.float32 else 1e-7
+    assert_allclose(gtr_mean, res_mean, atol=tol)
+    assert_allclose(gtr_max, res_max, atol=tol)
+    assert_allclose(gtr_sum, res_sum, atol=tol)
+
+
+@pytest.mark.parametrize("dataframe,queue", get_dataframes_and_queues())
+@pytest.mark.parametrize("num_batches", [2, 10])
+@pytest.mark.parametrize("row_count", [100, 1000])
+@pytest.mark.parametrize("column_count", [10, 100])
+@pytest.mark.parametrize("weighted", [True, False])
+@pytest.mark.parametrize("dtype", [np.float32, np.float64])
+def test_partial_fit_all_option_on_random_data(
+    dataframe, queue, num_batches, row_count, column_count, weighted, dtype
+):
+    seed = 77
+    gen = np.random.default_rng(seed)
+    X = gen.uniform(low=-0.3, high=+0.7, size=(row_count, column_count))
+    X = X.astype(dtype=dtype)
+    X_split = np.array_split(X, num_batches)
+    if weighted:
+        weights = gen.uniform(low=-0.5, high=+1.0, size=row_count)
+        weights = weights.astype(dtype=dtype)
+        weights_split = np.array_split(weights, num_batches)
+    incbs = IncrementalBasicStatistics(result_options="all")
+
+    for i in range(num_batches):
+        X_split_df = _convert_to_dataframe(
+            X_split[i], sycl_queue=queue, target_df=dataframe
+        )
+        if weighted:
+            weights_split_df = _convert_to_dataframe(
+                weights_split[i], sycl_queue=queue, target_df=dataframe
+            )
+            result = incbs.partial_fit(X_split_df, sample_weight=weights_split_df)
+        else:
+            result = incbs.partial_fit(X_split_df)
+
+    if weighted:
+        weighted_data = np.diag(weights) @ X
+
+    for option in options_and_tests:
+        result_option, function, tols = option
+        print(result_option)
+        fp32tol, fp64tol = tols
+        res = getattr(result, result_option)
+        if weighted:
+            gtr = function(weighted_data)
+        else:
+            gtr = function(X)
+        tol = fp32tol if res.dtype == np.float32 else fp64tol
+        assert_allclose(gtr, res, atol=tol)
+
+
+@pytest.mark.parametrize("dataframe,queue", get_dataframes_and_queues())
+@pytest.mark.parametrize("weighted", [True, False])
+@pytest.mark.parametrize("dtype", [np.float32, np.float64])
+def test_fit_multiple_options_on_gold_data(dataframe, queue, weighted, dtype):
+    X = np.array([[0, 0], [1, 1]])
+    X = X.astype(dtype=dtype)
+    X_df = _convert_to_dataframe(X, sycl_queue=queue, target_df=dataframe)
+    if weighted:
+        weights = np.array([1, 0.5])
+        weights = weights.astype(dtype=dtype)
+        weights_df = _convert_to_dataframe(weights, sycl_queue=queue, target_df=dataframe)
+    incbs = IncrementalBasicStatistics(batch_size=1)
+
+    if weighted:
+        result = incbs.fit(X_df, sample_weight=weights_df)
+    else:
+        result = incbs.fit(X_df)
+
+    if weighted:
+        expected_weighted_mean = np.array([0.25, 0.25])
+        expected_weighted_min = np.array([0, 0])
+        expected_weighted_max = np.array([0.5, 0.5])
+        assert_allclose(expected_weighted_mean, result.mean)
+        assert_allclose(expected_weighted_max, result.max)
+        assert_allclose(expected_weighted_min, result.min)
+    else:
+        expected_mean = np.array([0.5, 0.5])
+        expected_min = np.array([0, 0])
+        expected_max = np.array([1, 1])
+        assert_allclose(expected_mean, result.mean)
+        assert_allclose(expected_max, result.max)
+        assert_allclose(expected_min, result.min)
+
+
+@pytest.mark.parametrize("dataframe,queue", get_dataframes_and_queues())
+@pytest.mark.parametrize("num_batches", [2, 10])
+@pytest.mark.parametrize("option", options_and_tests)
+@pytest.mark.parametrize("row_count", [100, 1000])
+@pytest.mark.parametrize("column_count", [10, 100])
+@pytest.mark.parametrize("weighted", [True, False])
+@pytest.mark.parametrize("dtype", [np.float32, np.float64])
+def test_fit_single_option_on_random_data(
+    dataframe, queue, num_batches, option, row_count, column_count, weighted, dtype
+):
+    result_option, function, tols = option
+    fp32tol, fp64tol = tols
+    seed = 77
+    gen = np.random.default_rng(seed)
+    batch_size = row_count // num_batches
+    X = gen.uniform(low=-0.3, high=+0.7, size=(row_count, column_count))
+    X = X.astype(dtype=dtype)
+    X_df = _convert_to_dataframe(X, sycl_queue=queue, target_df=dataframe)
+    if weighted:
+        weights = gen.uniform(low=-0.5, high=1.0, size=row_count)
+        weights = weights.astype(dtype=dtype)
+        weights_df = _convert_to_dataframe(weights, sycl_queue=queue, target_df=dataframe)
+    incbs = IncrementalBasicStatistics(
+        result_options=result_option, batch_size=batch_size
+    )
+
+    if weighted:
+        result = incbs.fit(X_df, sample_weight=weights_df)
+    else:
+        result = incbs.fit(X_df)
+
+    res = getattr(result, result_option)
+    if weighted:
+        weighted_data = np.diag(weights) @ X
+        gtr = function(weighted_data)
+    else:
+        gtr = function(X)
+
+    tol = fp32tol if res.dtype == np.float32 else fp64tol
+    assert_allclose(gtr, res, atol=tol)
+
+
+@pytest.mark.parametrize("dataframe,queue", get_dataframes_and_queues())
+@pytest.mark.parametrize("num_batches", [2, 10])
+@pytest.mark.parametrize("row_count", [100, 1000])
+@pytest.mark.parametrize("column_count", [10, 100])
+@pytest.mark.parametrize("weighted", [True, False])
+@pytest.mark.parametrize("dtype", [np.float32, np.float64])
+def test_partial_fit_multiple_options_on_random_data(
+    dataframe, queue, num_batches, row_count, column_count, weighted, dtype
+):
+    seed = 77
+    gen = np.random.default_rng(seed)
+    batch_size = row_count // num_batches
+    X = gen.uniform(low=-0.3, high=+0.7, size=(row_count, column_count))
+    X = X.astype(dtype=dtype)
+    X_df = _convert_to_dataframe(X, sycl_queue=queue, target_df=dataframe)
+    if weighted:
+        weights = gen.uniform(low=-0.5, high=1.0, size=row_count)
+        weights = weights.astype(dtype=dtype)
+        weights_df = _convert_to_dataframe(weights, sycl_queue=queue, target_df=dataframe)
+    incbs = IncrementalBasicStatistics(
+        result_options=["mean", "max", "sum"], batch_size=batch_size
+    )
+
+    if weighted:
+        result = incbs.fit(X_df, sample_weight=weights_df)
+    else:
+        result = incbs.fit(X_df)
+
+    res_mean, res_max, res_sum = result.mean, result.max, result.sum
+    if weighted:
+        weighted_data = np.diag(weights) @ X
+        gtr_mean, gtr_max, gtr_sum = (
+            expected_mean(weighted_data),
+            expected_max(weighted_data),
+            expected_sum(weighted_data),
+        )
+    else:
+        gtr_mean, gtr_max, gtr_sum = (
+            expected_mean(X),
+            expected_max(X),
+            expected_sum(X),
+        )
+
+    tol = 3e-4 if res_mean.dtype == np.float32 else 1e-7
+    assert_allclose(gtr_mean, res_mean, atol=tol)
+    assert_allclose(gtr_max, res_max, atol=tol)
+    assert_allclose(gtr_sum, res_sum, atol=tol)
+
+
+@pytest.mark.parametrize("dataframe,queue", get_dataframes_and_queues())
+@pytest.mark.parametrize("num_batches", [2, 10])
+@pytest.mark.parametrize("row_count", [100, 1000])
+@pytest.mark.parametrize("column_count", [10, 100])
+@pytest.mark.parametrize("weighted", [True, False])
+@pytest.mark.parametrize("dtype", [np.float32, np.float64])
+def test_fit_all_option_on_random_data(
+    dataframe, queue, num_batches, row_count, column_count, weighted, dtype
+):
+    seed = 77
+    gen = np.random.default_rng(seed)
+    batch_size = row_count // num_batches
+    X = gen.uniform(low=-0.3, high=+0.7, size=(row_count, column_count))
+    X = X.astype(dtype=dtype)
+    X_df = _convert_to_dataframe(X, sycl_queue=queue, target_df=dataframe)
+    if weighted:
+        weights = gen.uniform(low=-0.5, high=+1.0, size=row_count)
+        weights = weights.astype(dtype=dtype)
+        weights_df = _convert_to_dataframe(weights, sycl_queue=queue, target_df=dataframe)
+    incbs = IncrementalBasicStatistics(result_options="all", batch_size=batch_size)
+
+    if weighted:
+        result = incbs.fit(X_df, sample_weight=weights_df)
+    else:
+        result = incbs.fit(X_df)
+
+    if weighted:
+        weighted_data = np.diag(weights) @ X
+
+    for option in options_and_tests:
+        result_option, function, tols = option
+        print(result_option)
+        fp32tol, fp64tol = tols
+        res = getattr(result, result_option)
+        if weighted:
+            gtr = function(weighted_data)
+        else:
+            gtr = function(X)
+        tol = fp32tol if res.dtype == np.float32 else fp64tol
+        assert_allclose(gtr, res, atol=tol)

sklearnex/cluster/dbscan.py

@@ -26,7 +26,7 @@ from daal4py.sklearn._n_jobs_support import control_n_jobs
 from daal4py.sklearn._utils import sklearn_check_version
 from onedal.cluster import DBSCAN as onedal_DBSCAN
 
-from .._device_offload import dispatch, wrap_output_data
+from .._device_offload import dispatch
 from .._utils import PatchingConditionsChain
 
 if sklearn_check_version("1.1") and not sklearn_check_version("1.2"):
@@ -186,3 +186,5 @@ class DBSCAN(sklearn_DBSCAN, BaseDBSCAN):
         )
 
         return self
+
+    fit.__doc__ = sklearn_DBSCAN.fit.__doc__

sklearnex/conftest.py

@@ -0,0 +1,63 @@
+# ==============================================================================
+# Copyright 2024 Intel Corporation
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+# ==============================================================================
+
+import io
+import logging
+
+import pytest
+
+from sklearnex import patch_sklearn, unpatch_sklearn
+
+
+def pytest_configure(config):
+    config.addinivalue_line(
+        "markers", "allow_sklearn_fallback: mark test to not check for sklearnex usage"
+    )
+
+
+@pytest.hookimpl(hookwrapper=True)
+def pytest_runtest_call(item):
+    # setup logger to check for sklearn fallback
+    if not item.get_closest_marker("allow_sklearn_fallback"):
+        log_stream = io.StringIO()
+        log_handler = logging.StreamHandler(log_stream)
+        sklearnex_logger = logging.getLogger("sklearnex")
+        level = sklearnex_logger.level
+        sklearnex_stderr_handler = sklearnex_logger.handlers
+        sklearnex_logger.handlers = []
+        sklearnex_logger.addHandler(log_handler)
+        sklearnex_logger.setLevel(logging.INFO)
+        log_handler.setLevel(logging.INFO)
+
+        yield
+
+        sklearnex_logger.handlers = sklearnex_stderr_handler
+        sklearnex_logger.setLevel(level)
+        sklearnex_logger.removeHandler(log_handler)
+        text = log_stream.getvalue()
+        if "fallback to original Scikit-learn" in text:
+            raise TypeError(
+                f"test did not properly evaluate sklearnex functionality and fell back to sklearn:\n{text}"
+            )
+    else:
+        yield
+
+
+@pytest.fixture
+def with_sklearnex():
+    patch_sklearn()
+    yield
+    unpatch_sklearn()