Unittests for creators

requirements-dev.txt

@@ -1,2 +1,7 @@
 pre-commit>=3.2.1
-pydoctor
+pydoctor
+pytest-dependency>=0.6.0
+pytest-cov>=4.1.0
+pytest-subtests>=0.12.1
+pytest-md>=0.2.0
+coverage-badge>=1.1.0

tests/unittests/test_creators/test_classification.py

@@ -0,0 +1,168 @@
+import pytest
+
+from depthai_nodes.ml.messages import Classifications
+from depthai_nodes.ml.messages.creators.classification import (
+    create_classification_message,
+)
+
+
+def test_none_classe():
+    with pytest.raises(ValueError):
+        create_classification_message(None, [0.5, 0.2, 0.3])
+
+
+def test_none_scores():
+    with pytest.raises(ValueError):
+        create_classification_message(["cat", "dog", "bird"], None)
+
+
+def test_none_both():
+    with pytest.raises(ValueError):
+        create_classification_message(None, None)
+
+
+def test_empty_classes():
+    with pytest.raises(ValueError):
+        create_classification_message([], [0.5, 0.2, 0.3])
+
+
+def test_empty_scores():
+    with pytest.raises(ValueError):
+        create_classification_message(["cat", "dog", "bird"], [])
+
+
+def test_non_list_classes():
+    with pytest.raises(ValueError):
+        create_classification_message("cat", [0.5, 0.2, 0.3])
+
+
+def test_non_np_scores():
+    with pytest.raises(ValueError):
+        create_classification_message(["cat", "dog", "bird"], [int(1), int(2), int(3)])
+
+
+def test_tuple_scores():
+    with pytest.raises(ValueError):
+        create_classification_message(["cat", "dog", "bird"], (0.5, 0.2, 0.3))
+
+
+def test_nd_scores():
+    with pytest.raises(ValueError):
+        create_classification_message(["cat", "dog", "bird"], [[0.5, 0.2, 0.3]])
+
+
+def test_mixed_scores():
+    with pytest.raises(ValueError):
+        create_classification_message(["cat", "dog", "bird"], [0.5, 0.2, "30"])
+
+
+def test_non_probability_scores():
+    with pytest.raises(ValueError):
+        create_classification_message(["cat", "dog", "bird"], [0.2, 0.3, 0.4])
+
+
+def test_non_probability_scores_2():
+    with pytest.raises(ValueError):
+        create_classification_message(["cat", "dog", "bird"], [0.5, 0.5, 0.5])
+
+
+def test_sum_above_upper_thr():  # upper thr is 1.01001001
+    with pytest.raises(ValueError):
+        create_classification_message(["cat", "dog", "bird"], [0.5, 0.11001001, 0.4])
+
+
+def test_sum_below_upper_thr():
+    create_classification_message(["cat", "dog", "bird"], [0.5, 0.11001000, 0.4])
+
+
+def test_sum_above_lower_thr():  # lower thr is 0.98999001
+    create_classification_message(["cat", "dog", "bird"], [0.5, 0.18999001, 0.3])
+
+
+def test_sum_below_bottom_thr():
+    with pytest.raises(ValueError):
+        create_classification_message(["cat", "dog", "bird"], [0.5, 0.18999, 0.3])
+
+
+def test_mismatch_lengths():
+    with pytest.raises(ValueError):
+        create_classification_message(["cat", "dog", "bird"], [0.5, 0.2, 0.3, 0.4])
+
+
+def test_correct_input():
+    classes = ["cat", "dog", "bird"]
+    scores = [0.2, 0.5, 0.3]
+
+    correct_classes = ["dog", "bird", "cat"]
+    correct_scores = [0.5, 0.3, 0.2]
+
+    message = create_classification_message(classes, scores)
+
+    assert isinstance(message, Classifications)
+    assert message.classes == correct_classes
+    assert message.scores == correct_scores
+    assert isinstance(message.classes, list)
+    assert isinstance(message.scores, list)
+
+
+def test_single_class_and_score():
+    classes = ["cat"]
+    scores = [1.0]
+
+    message = create_classification_message(classes, scores)
+    assert message.classes == ["cat"]
+    assert message.scores == [1.0]
+
+
+def test_correct_input_with_mixed_classes():
+    classes = ["cat", 1, None]
+    scores = [0.2, 0.5, 0.3]
+
+    correct_classes = [1, None, "cat"]
+    correct_scores = [0.5, 0.3, 0.2]
+
+    message = create_classification_message(classes, scores)
+
+    assert isinstance(message, Classifications)
+    assert message.classes == correct_classes
+    assert message.scores == correct_scores
+    assert isinstance(message.classes, list)
+    assert isinstance(message.scores, list)
+
+
+def test_very_small_scores():
+    classes = ["cat", "dog", "bird"]
+    scores = [1e-10, 1e-10, 1 - 2e-10]
+
+    message = create_classification_message(classes, scores)
+
+    assert isinstance(message, Classifications)
+    assert message.classes == ["bird", "cat", "dog"]
+    assert message.scores == [1 - 2e-10, 1e-10, 1e-10]
+
+
+def test_identical_scores():
+    classes = ["cat", "dog", "bird"]
+    scores = [1 / 3, 1 / 3, 1 / 3]
+
+    message = create_classification_message(classes, scores)
+
+    assert message.classes == classes
+    assert message.scores == scores
+
+
+def test_duplicate_scores():
+    classes = ["cat", "dog", "bird"]
+    scores = [0.4, 0.2, 0.4]
+
+    correct_classes = ["cat", "bird", "dog"]
+    correct_scores = [0.4, 0.4, 0.2]
+
+    message = create_classification_message(classes, scores)
+
+    assert message.classes == correct_classes
+    assert message.scores == correct_scores
+
+
+if __name__ == "__main__":
+    pytest.main()

tests/unittests/test_creators/test_depth.py

@@ -0,0 +1,81 @@
+import re
+
+import depthai as dai
+import numpy as np
+import pytest
+
+from depthai_nodes.ml.messages.creators.depth import create_depth_message
+
+UINT16_MAX_VALUE = 65535
+np.random.seed(0)
+
+
+def test_not_numpy_array():
+    with pytest.raises(ValueError, match="Expected numpy array, got <class 'list'>."):
+        create_depth_message([1, 2, 3], "relative")
+
+
+def test_wrong_literal_type():
+    with pytest.raises(ValueError):
+        create_depth_message(np.array([1, 2, 3]), "wrong")
+
+
+def test_not_3d_input():
+    with pytest.raises(ValueError, match="Expected 3D input, got 1D input."):
+        create_depth_message(np.array([1, 2, 3]), "relative")
+
+
+def test_wrong_input_shape():
+    with pytest.raises(
+        ValueError,
+        match=re.escape("Unexpected image shape. Expected CHW or HWC, got (3, 1, 3)."),
+    ):
+        create_depth_message(
+            np.array([[[1, 2, 3]], [[4, 5, 6]], [[7, 8, 9]]]), "relative"
+        )
+
+
+def test_metric_not_implemented():
+    with pytest.raises(
+        NotImplementedError,
+        match="The message for 'metric' depth type is not yet implemented.",
+    ):
+        create_depth_message(np.array([[[1, 2, 3]]]), "metric")
+
+
+def test_depth_map():
+    depth_map = np.random.rand(320, 640, 1)
+
+    message = create_depth_message(depth_map, "relative")
+    depth_map = depth_map[:, :, 0]
+
+    assert isinstance(message, dai.ImgFrame)
+    assert message.getType() == dai.ImgFrame.Type.RAW16
+    assert message.getWidth() == 640
+    assert message.getHeight() == 320
+
+    frame = message.getFrame()
+    assert frame.shape == depth_map.shape
+    scaled_depth_map = (
+        (depth_map - depth_map.min())
+        / (depth_map.max() - depth_map.min())
+        * UINT16_MAX_VALUE
+    )
+    scaled_depth_map = scaled_depth_map.astype(np.uint16)
+    assert np.all(np.isclose(frame, scaled_depth_map))
+
+
+def test_same_depth():
+    depth_map = np.ones((320, 640, 1))
+    message = create_depth_message(depth_map, "relative")
+
+    assert isinstance(message, dai.ImgFrame)
+    assert message.getType() == dai.ImgFrame.Type.RAW16
+    assert message.getWidth() == 640
+    assert message.getHeight() == 320
+    frame = message.getFrame()
+    assert np.all(np.isclose(frame, np.zeros((320, 640))))
+
+
+if __name__ == "__main__":
+    pytest.main()