Merge pull request #999 from openvinotoolkit/vsaltykovx/add_deepobjec…

…treid_input_parameters_validation

[OTE] Added deep-object-reid input parameters validation and tests

external/anomaly/ote_tests_pytest.ini

@@ -1,2 +1,2 @@
 [pytest]
-python_files = test_ote_api.py test_ote_inference.py test_ote_training.py
+python_files = test_ote_api.py test_ote_inference.py test_ote_training.py

external/deep-object-reid/tests/sc_input_params_validation/helpers.py

@@ -0,0 +1,72 @@
+"""
+OTE parameters validation tests helpers
+"""
+
+import numpy as np
+
+from ote_sdk.entities.annotation import (
+    Annotation,
+    AnnotationSceneEntity,
+    AnnotationSceneKind,
+)
+from ote_sdk.entities.dataset_item import DatasetItemEntity
+from ote_sdk.entities.datasets import DatasetEntity
+from ote_sdk.entities.id import ID
+from ote_sdk.entities.image import Image
+from ote_sdk.entities.label import Domain, LabelEntity
+from ote_sdk.entities.scored_label import ScoredLabel
+from ote_sdk.entities.shapes.rectangle import Rectangle
+from ote_sdk.entities.subset import Subset
+
+
+def load_test_dataset():
+    """Helper to create test dataset"""
+
+    def gen_image(resolution, x1, y1, x2, y2):
+        w, h = resolution
+        image = np.full([h, w, 3], fill_value=255, dtype=np.uint8)
+        image[int(y1 * h): int(y2 * h), int(x1 * w): int(x2 * w), :] = np.array(
+            [0, 128, 128], dtype=np.uint8
+        )[None, None, :]
+        return image, Rectangle(x1=x1, y1=y1, x2=x2, y2=y2)
+
+    images = [
+        gen_image((640, 480), 0.0, 0.0, 0.5, 0.5),
+        gen_image((640, 480), 0.5, 0.0, 1.0, 0.5),
+        gen_image((640, 480), 0.0, 0.5, 0.5, 1.0),
+        gen_image((640, 480), 0.5, 0.5, 1.0, 1.0),
+    ]
+    labels = [LabelEntity(name="rect", domain=Domain.DETECTION, id=ID("0"))]
+
+    def get_image(i, subset):
+        image, bbox = images[i]
+        return DatasetItemEntity(
+            media=Image(data=image),
+            annotation_scene=AnnotationSceneEntity(
+                annotations=[Annotation(bbox, labels=[ScoredLabel(label=labels[0])])],
+                kind=AnnotationSceneKind.ANNOTATION,
+            ),
+            subset=subset,
+        )
+
+    items = [
+        get_image(0, Subset.TRAINING),
+        get_image(1, Subset.TRAINING),
+        get_image(2, Subset.TRAINING),
+        get_image(3, Subset.TRAINING),
+        get_image(0, Subset.TRAINING),
+        get_image(1, Subset.TRAINING),
+        get_image(2, Subset.TRAINING),
+        get_image(3, Subset.TRAINING),
+        get_image(0, Subset.TRAINING),
+        get_image(1, Subset.TRAINING),
+        get_image(0, Subset.VALIDATION),
+        get_image(1, Subset.VALIDATION),
+        get_image(2, Subset.VALIDATION),
+        get_image(3, Subset.VALIDATION),
+        get_image(0, Subset.TESTING),
+        get_image(1, Subset.TESTING),
+        get_image(2, Subset.TESTING),
+        get_image(3, Subset.TESTING),
+    ]
+    return DatasetEntity(items), labels

external/deep-object-reid/tests/sc_input_params_validation/test_ote_classification_input_params_validation.py

@@ -0,0 +1,294 @@
+import numpy as np
+import pytest
+from ote_sdk.test_suite.e2e_test_system import e2e_pytest_unit
+from ote_sdk.tests.parameters_validation.validation_helper import (
+    check_value_error_exception_raised,
+)
+
+from torchreid_tasks.model_wrappers.classification import (
+    OteClassification,
+    sigmoid_numpy,
+    softmax_numpy,
+    get_hierarchical_predictions,
+    get_multiclass_predictions,
+    get_multilabel_predictions,
+    preprocess_features_for_actmap,
+    get_actmap,
+)
+
+
+class MockClassification(OteClassification):
+    def __init__(self):
+        pass
+
+
+class TestClassificationFunctionsParamsValidation:
+    @e2e_pytest_unit
+    def test_preprocess_features_for_actmap_parameters_params_validation(self):
+        """
+        <b>Description:</b>
+        Check "preprocess_features_for_actmap" function input parameters validation
+
+        <b>Input data:</b>
+        "features" non-expected type object
+
+        <b>Expected results:</b>
+        Test passes if ValueError exception is raised when unexpected type object is specified as
+        input parameter for "preprocess_features_for_actmap" function
+        """
+        with pytest.raises(ValueError):
+            preprocess_features_for_actmap(features=None)  # type: ignore
+
+    @e2e_pytest_unit
+    def test_get_actmap_params_validation(self):
+        """
+        <b>Description:</b>
+        Check "get_actmap" function input parameters validation
+
+        <b>Input data:</b>
+        "get_actmap" unexpected type parameters
+
+        <b>Expected results:</b>
+        Test passes if ValueError exception is raised when unexpected type object is specified as
+        input parameter for "get_actmap" function
+        """
+        correct_values_dict = {
+            "features": ["some", "features"],
+            "output_res": ("iterable", "object")
+        }
+        unexpected_values = [
+            # Unexpected dictionary is specified as "features" parameter
+            ("features", None),
+            # Unexpected dictionary is specified as "output_res" parameter
+            ("output_res", None),
+        ]
+
+        check_value_error_exception_raised(
+            correct_parameters=correct_values_dict,
+            unexpected_values=unexpected_values,
+            class_or_function=get_actmap,
+        )
+
+    @e2e_pytest_unit
+    def test_sigmoid_numpy_params_validation(self):
+        """
+        <b>Description:</b>
+        Check "sigmoid_numpy" function input parameters validation
+
+        <b>Input data:</b>
+        "x" non-nd.array parameter
+
+        <b>Expected results:</b>
+        Test passes if ValueError exception is raised when unexpected type object is specified as
+        input parameter for "sigmoid_numpy" function
+        """
+        with pytest.raises(ValueError):
+            sigmoid_numpy(x="unexpected string")  # type: ignore
+
+    @e2e_pytest_unit
+    def test_softmax_numpy_params_validation(self):
+        """
+        <b>Description:</b>
+        Check "softmax_numpy" function input parameters validation
+
+        <b>Input data:</b>
+        "x" non-nd.array parameter
+
+        <b>Expected results:</b>
+        Test passes if ValueError exception is raised when unexpected type object is specified as
+        input parameter for "softmax_numpy" function
+        """
+        with pytest.raises(ValueError):
+            softmax_numpy(x="unexpected string")  # type: ignore
+
+    @e2e_pytest_unit
+    def test_get_hierarchical_predictions_params_validation(self):
+        """
+        <b>Description:</b>
+        Check "get_hierarchical_predictions" function input parameters validation
+
+        <b>Input data:</b>
+        "get_hierarchical_predictions" unexpected type parameters
+
+        <b>Expected results:</b>
+        Test passes if ValueError exception is raised when unexpected type object is specified as
+        input parameter for "get_hierarchical_predictions" function
+        """
+        correct_values_dict = {
+            "logits": np.random.randint(low=0, high=255, size=(10, 16, 3)),
+            "multihead_class_info": {"multihead": "dictionary"},
+        }
+        unexpected_str = "unexpected string"
+        unexpected_values = [
+            # Unexpected string is specified as "logits" parameter
+            ("logits", unexpected_str),
+            # Unexpected string is specified as "multihead_class_info" parameter
+            ("multihead_class_info", unexpected_str),
+            # Unexpected string is specified as "pos_thr" parameter
+            ("pos_thr", unexpected_str),
+            # Unexpected string is specified as "activate" parameter
+            ("activate", unexpected_str),
+        ]
+
+        check_value_error_exception_raised(
+            correct_parameters=correct_values_dict,
+            unexpected_values=unexpected_values,
+            class_or_function=get_hierarchical_predictions,
+        )
+
+    @e2e_pytest_unit
+    def test_get_multiclass_predictions_params_validation(self):
+        """
+        <b>Description:</b>
+        Check "get_multiclass_predictions" function input parameters validation
+
+        <b>Input data:</b>
+        "get_multiclass_predictions" unexpected type parameters
+
+        <b>Expected results:</b>
+        Test passes if ValueError exception is raised when unexpected type object is specified as
+        input parameter for "get_multiclass_predictions" function
+        """
+        correct_values_dict = {
+            "logits": np.random.randint(low=0, high=255, size=(10, 16, 3)),
+            "activate": True,
+        }
+        unexpected_str = "unexpected string"
+        unexpected_values = [
+            # Unexpected string is specified as "logits" parameter
+            ("logits", unexpected_str),
+            # Unexpected string is specified as "activate" parameter
+            ("activate", unexpected_str),
+        ]
+
+        check_value_error_exception_raised(
+            correct_parameters=correct_values_dict,
+            unexpected_values=unexpected_values,
+            class_or_function=get_multiclass_predictions,
+        )
+
+    @e2e_pytest_unit
+    def test_get_multilabel_predictions_params_validation(self):
+        """
+        <b>Description:</b>
+        Check "get_multilabel_predictions" function input parameters validation
+
+        <b>Input data:</b>
+        "get_multilabel_predictions" unexpected type parameters
+
+        <b>Expected results:</b>
+        Test passes if ValueError exception is raised when unexpected type object is specified as
+        input parameter for "get_multilabel_predictions" function
+        """
+        correct_values_dict = {
+            "logits": np.random.randint(low=0, high=255, size=(10, 16, 3)),
+            "activate": True,
+        }
+        unexpected_str = "unexpected string"
+        unexpected_values = [
+            # Unexpected string is specified as "logits" parameter
+            ("logits", unexpected_str),
+            # Unexpected string is specified as "pos_thr" parameter
+            ("pos_thr", unexpected_str),
+            # Unexpected string is specified as "activate" parameter
+            ("activate", unexpected_str),
+        ]
+
+        check_value_error_exception_raised(
+            correct_parameters=correct_values_dict,
+            unexpected_values=unexpected_values,
+            class_or_function=get_multilabel_predictions,
+        )
+
+
+class TestOteClassificationParamsValidation:
+    @e2e_pytest_unit
+    def test_ote_classification_preprocess_params_validation(self):
+        """
+        <b>Description:</b>
+        Check OteClassification object "preprocess" method input parameters validation
+
+        <b>Input data:</b>
+        OteClassification object. "image" non-ndarray object
+
+        <b>Expected results:</b>
+        Test passes if ValueError exception is raised when unexpected type object is specified as
+        input parameter for "preprocess" method
+        """
+        classification = MockClassification()
+        with pytest.raises(ValueError):
+            classification.preprocess(image="unexpected string")  # type: ignore
+
+    @e2e_pytest_unit
+    def test_ote_classification_postprocess_params_validation(self):
+        """
+        <b>Description:</b>
+        Check OteClassification object "postprocess" method input parameters validation
+
+        <b>Input data:</b>
+        OteClassification object. "postprocess" method unexpected type parameters
+
+        <b>Expected results:</b>
+        Test passes if ValueError exception is raised when unexpected type object is specified as
+        input parameter for "postprocess" method
+        """
+        classification = MockClassification()
+        correct_values_dict = {
+            "outputs": {"output_1": np.random.rand(2, 2)},
+            "metadata": {"metadata_1": "some_data"},
+        }
+        unexpected_int = 1
+        unexpected_values = [
+            # Unexpected integer is specified as "outputs" parameter
+            ("outputs", unexpected_int),
+            # Unexpected integer is specified as "outputs" dictionary key
+            ("outputs", {unexpected_int: np.random.rand(2, 2)}),
+            # Unexpected integer is specified as "outputs" dictionary value
+            ("outputs", {"output_1": unexpected_int}),
+            # Unexpected integer is specified as "metadata" parameter
+            ("metadata", unexpected_int),
+            # Unexpected integer is specified as "metadata" dictionary key
+            ("metadata", {unexpected_int: "some_data"}),
+        ]
+        check_value_error_exception_raised(
+            correct_parameters=correct_values_dict,
+            unexpected_values=unexpected_values,
+            class_or_function=classification.postprocess,
+        )
+
+    @e2e_pytest_unit
+    def test_ote_classification_postprocess_aux_outputs_params_validation(self):
+        """
+        <b>Description:</b>
+        Check OteClassification object "postprocess_aux_outputs" method input parameters validation
+
+        <b>Input data:</b>
+        OteClassification object. "postprocess_aux_outputs" method unexpected type parameters
+
+        <b>Expected results:</b>
+        Test passes if ValueError exception is raised when unexpected type object is specified as
+        input parameter for "postprocess_aux_outputs" method
+        """
+        classification = MockClassification()
+        correct_values_dict = {
+            "outputs": {"output_1": np.random.rand(2, 2)},
+            "metadata": {"metadata_1": "some_data"},
+        }
+        unexpected_int = 1
+        unexpected_values = [
+            # Unexpected integer is specified as "outputs" parameter
+            ("outputs", unexpected_int),
+            # Unexpected integer is specified as "outputs" dictionary key
+            ("outputs", {unexpected_int: np.random.rand(2, 2)}),
+            # Unexpected integer is specified as "outputs" dictionary value
+            ("outputs", {"output_1": unexpected_int}),
+            # Unexpected integer is specified as "metadata" parameter
+            ("metadata", unexpected_int),
+            # Unexpected integer is specified as "metadata" dictionary key
+            ("metadata", {unexpected_int: "some_data"}),
+        ]
+        check_value_error_exception_raised(
+            correct_parameters=correct_values_dict,
+            unexpected_values=unexpected_values,
+            class_or_function=classification.postprocess_aux_outputs,
+        )