# Copyright 2024 Google LLC # # 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 # # https://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. # pylint: disable=missing-module-docstring import numpy as np from numpy.testing import assert_array_equal import PIL.Image import pytest from speciesnet.constants import Detection from speciesnet.detector import SpeciesNetDetector from speciesnet.utils import load_rgb_image from speciesnet.utils import PreprocessedImage class TestDetector: """Tests for the detector component.""" @pytest.fixture(scope="class") def detector(self, model_name: str) -> SpeciesNetDetector: return SpeciesNetDetector(model_name) @pytest.fixture def img_green_w500_h700(self) -> PIL.Image.Image: return PIL.Image.new("RGB", (500, 700), color=(0, 255, 0)) @pytest.fixture def img_gray_green_gray_w23_w914_w23_h1280(self) -> PIL.Image.Image: img_gray = PIL.Image.new("RGB", (23, 1280), color=(114, 114, 114)) img_green = PIL.Image.new("RGB", (914, 1280), color=(0, 255, 0)) img = PIL.Image.new("RGB", (960, 1280)) img.paste(img_gray, (0, 0)) img.paste(img_green, (23, 0)) img.paste(img_gray, (937, 0)) return img @pytest.fixture def img_green_w2000_h1111(self) -> PIL.Image.Image: return PIL.Image.new("RGB", (2000, 1111), color=(0, 255, 0)) @pytest.fixture def img_gray_green_gray_w1280_h28_h711_h29(self) -> PIL.Image.Image: img_gray = PIL.Image.new("RGB", (1280, 29), color=(114, 114, 114)) img_green = PIL.Image.new("RGB", (1280, 711), color=(0, 255, 0)) img = PIL.Image.new("RGB", (1280, 768)) img.paste(img_gray, (0, 0)) img.paste(img_green, (0, 28)) img.paste(img_gray, (0, 739)) return img @pytest.fixture def img_green_w1280_h1280(self) -> PIL.Image.Image: return PIL.Image.new("RGB", (1280, 1280), color=(0, 255, 0)) def test_preprocess( # pylint: disable=too-many-positional-arguments self, detector, img_green_w500_h700, # input img_gray_green_gray_w23_w914_w23_h1280, # output img_green_w2000_h1111, # input img_gray_green_gray_w1280_h28_h711_h29, # output ) -> None: assert detector.preprocess(None) is None preprocessed = detector.preprocess(img_green_w500_h700) assert preprocessed assert preprocessed.orig_width == 500 assert preprocessed.orig_height == 700 assert preprocessed.arr.shape == (1280, 960, 3) assert_array_equal(preprocessed.arr, img_gray_green_gray_w23_w914_w23_h1280) preprocessed = detector.preprocess(img_green_w2000_h1111) assert preprocessed assert preprocessed.orig_width == 2000 assert preprocessed.orig_height == 1111 assert preprocessed.arr.shape == (768, 1280, 3) assert_array_equal(preprocessed.arr, img_gray_green_gray_w1280_h28_h711_h29) def test_predict(self, detector, img_green_w1280_h1280) -> None: filepath = "missing.jpg" prediction = detector.predict(filepath, None) assert prediction["filepath"] == filepath assert "failures" in prediction filepath = "green.png" prediction = detector.predict( filepath, PreprocessedImage(np.asarray(img_green_w1280_h1280), 1280, 1280) ) assert prediction["filepath"] == filepath assert "failures" not in prediction assert "detections" in prediction scores = [det["conf"] for det in prediction["detections"]] assert scores == sorted(scores, reverse=True) @pytest.fixture( # Expected MDv5 detections. params=[ ("test_data/blank.jpg", []), ("test_data/blank2.jpg", []), ( "test_data/blank3.jpg", [ { "category": "1", "conf": 0.0101, "bbox": [0.0009191, 0, 0.996, 0.9571], } ], ), ( "test_data/african_elephants.jpg", [ { "category": "1", "conf": 0.0105, "bbox": [0.4863, 0.4466, 0.04394, 0.0996], }, { "category": "1", "conf": 0.0126, "bbox": [0.1416, 0.4082, 0.1049, 0.1217], }, { "category": "1", "conf": 0.0135, "bbox": [0.6386, 0.4231, 0.02685, 0.03385], }, { "category": "1", "conf": 0.0175, "bbox": [0.3198, 0.4127, 0.05126, 0.05598], }, { "category": "1", "conf": 0.0419, "bbox": [0.6406, 0.429, 0.05712, 0.05403], }, { "category": "1", "conf": 0.0517, "bbox": [0.145, 0.4088, 0.07177, 0.07617], }, { "category": "1", "conf": 0.0605, "bbox": [0.3769, 0.444, 0.04687, 0.07096], }, { "category": "1", "conf": 0.0798, "bbox": [0.3208, 0.414, 0.07226, 0.1158], }, { "category": "1", "conf": 0.142, "bbox": [0.102, 0.4055, 0.1118, 0.11], }, { "category": "1", "conf": 0.636, "bbox": [0.3403, 0.4433, 0.08349, 0.09114], }, { "category": "1", "conf": 0.703, "bbox": [0.854, 0.4785, 0.08593, 0.08268], }, { "category": "1", "conf": 0.816, "bbox": [0.9418, 0.4674, 0.0581, 0.1673], }, { "category": "1", "conf": 0.903, "bbox": [0.8291, 0.511, 0.1044, 0.1106], }, { "category": "1", "conf": 0.912, "bbox": [0.5644, 0.4218, 0.08496, 0.1673], }, { "category": "1", "conf": 0.935, "bbox": [0.7041, 0.4765, 0.1108, 0.125], }, ], ), ( "test_data/african_elephants_bw.jpg", [ { "category": "1", "conf": 0.0101, "bbox": [0.8535, 0.4791, 0.03808, 0.03515], }, { "category": "1", "conf": 0.0105, "bbox": [0.3637, 0.4453, 0.08251, 0.1002], }, { "category": "1", "conf": 0.0148, "bbox": [0.4531, 0.4427, 0.07275, 0.1035], }, { "category": "1", "conf": 0.0151, "bbox": [0.4965, 0.4518, 0.03466, 0.08854], }, { "category": "1", "conf": 0.0284, "bbox": [0.03417, 0.3561, 0.1689, 0.1699], }, { "category": "1", "conf": 0.0348, "bbox": [0.3305, 0.4088, 0.05908, 0.1152], }, { "category": "1", "conf": 0.0434, "bbox": [0.499, 0.4505, 0.06787, 0.1087], }, { "category": "1", "conf": 0.0841, "bbox": [0.1123, 0.3541, 0.1, 0.1412], }, { "category": "1", "conf": 0.139, "bbox": [0.3666, 0.442, 0.05615, 0.06966], }, { "category": "1", "conf": 0.484, "bbox": [0.3378, 0.4316, 0.08447, 0.1035], }, { "category": "1", "conf": 0.665, "bbox": [0.8525, 0.4791, 0.08691, 0.09635], }, { "category": "1", "conf": 0.818, "bbox": [0.9414, 0.468, 0.05859, 0.1692], }, { "category": "1", "conf": 0.874, "bbox": [0.5644, 0.4173, 0.09423, 0.1705], }, { "category": "1", "conf": 0.883, "bbox": [0.8295, 0.511, 0.1044, 0.1113], }, { "category": "1", "conf": 0.931, "bbox": [0.7031, 0.4765, 0.1118, 0.1256], }, ], ), ( "test_data/african_elephants_cmyk.jpg", [ { "category": "1", "conf": 0.0101, "bbox": [0.4858, 0.4466, 0.04443, 0.0996], }, { "category": "1", "conf": 0.0104, "bbox": [0.6386, 0.4231, 0.02783, 0.0345], }, { "category": "1", "conf": 0.012, "bbox": [0.142, 0.4082, 0.1035, 0.1217], }, { "category": "1", "conf": 0.0221, "bbox": [0.3198, 0.4127, 0.05126, 0.05598], }, { "category": "1", "conf": 0.0415, "bbox": [0.6411, 0.4296, 0.05664, 0.05338], }, { "category": "1", "conf": 0.0544, "bbox": [0.145, 0.4082, 0.07128, 0.07682], }, { "category": "1", "conf": 0.0667, "bbox": [0.3769, 0.444, 0.04687, 0.07161], }, { "category": "1", "conf": 0.0841, "bbox": [0.3208, 0.4147, 0.07177, 0.1152], }, { "category": "1", "conf": 0.156, "bbox": [0.101, 0.4055, 0.1123, 0.11], }, { "category": "1", "conf": 0.637, "bbox": [0.3403, 0.4433, 0.08349, 0.09114], }, { "category": "1", "conf": 0.706, "bbox": [0.854, 0.4785, 0.08593, 0.08268], }, { "category": "1", "conf": 0.815, "bbox": [0.9418, 0.4674, 0.0581, 0.1679], }, { "category": "1", "conf": 0.903, "bbox": [0.8295, 0.511, 0.104, 0.1106], }, { "category": "1", "conf": 0.914, "bbox": [0.5644, 0.4218, 0.08496, 0.1673], }, { "category": "1", "conf": 0.934, "bbox": [0.7041, 0.4765, 0.1108, 0.125], }, ], ), ( "test_data/african_elephants_truncated_file.jpg", [ { "category": "1", "conf": 0.0107, "bbox": [0.4199, 0.4733, 0.01318, 0.02734], }, { "category": "1", "conf": 0.0182, "bbox": [0.3198, 0.4127, 0.05078, 0.05468], }, { "category": "1", "conf": 0.0186, "bbox": [0.6381, 0.4238, 0.0288, 0.0332], }, { "category": "1", "conf": 0.0526, "bbox": [0.6406, 0.429, 0.05712, 0.05273], }, { "category": "1", "conf": 0.0565, "bbox": [0.1494, 0.4088, 0.07031, 0.07552], }, { "category": "1", "conf": 0.0597, "bbox": [0.3764, 0.4433, 0.04492, 0.05533], }, { "category": "1", "conf": 0.0759, "bbox": [0.3247, 0.4212, 0.06542, 0.1067], }, { "category": "1", "conf": 0.193, "bbox": [0.1103, 0.4062, 0.1044, 0.1035], }, { "category": "1", "conf": 0.602, "bbox": [0.3403, 0.444, 0.08398, 0.08789], }, { "category": "1", "conf": 0.691, "bbox": [0.854, 0.4785, 0.08593, 0.08333], }, { "category": "1", "conf": 0.822, "bbox": [0.9414, 0.4674, 0.05859, 0.1679], }, { "category": "1", "conf": 0.9, "bbox": [0.83, 0.5117, 0.1035, 0.11], }, { "category": "1", "conf": 0.902, "bbox": [0.5644, 0.4218, 0.08496, 0.1666], }, { "category": "1", "conf": 0.931, "bbox": [0.7045, 0.4759, 0.1103, 0.1256], }, ], ), ( "test_data/african_elephants_with_exif_orientation.jpg", [ { "category": "1", "conf": 0.0127, "bbox": [0.1435, 0.4075, 0.1147, 0.1256], }, { "category": "1", "conf": 0.0147, "bbox": [0.6386, 0.4231, 0.02685, 0.03385], }, { "category": "1", "conf": 0.02, "bbox": [0.3198, 0.4127, 0.05126, 0.05533], }, { "category": "1", "conf": 0.0514, "bbox": [0.6411, 0.4296, 0.05664, 0.05273], }, { "category": "1", "conf": 0.0556, "bbox": [0.3769, 0.444, 0.04687, 0.07096], }, { "category": "1", "conf": 0.0558, "bbox": [0.145, 0.4082, 0.07226, 0.07812], }, { "category": "1", "conf": 0.0714, "bbox": [0.3208, 0.4147, 0.07275, 0.1152], }, { "category": "1", "conf": 0.146, "bbox": [0.103, 0.4062, 0.1113, 0.108], }, { "category": "1", "conf": 0.65, "bbox": [0.3398, 0.4433, 0.08398, 0.09179], }, { "category": "1", "conf": 0.698, "bbox": [0.854, 0.4785, 0.08593, 0.08268], }, { "category": "1", "conf": 0.815, "bbox": [0.9418, 0.4674, 0.0581, 0.1673], }, { "category": "1", "conf": 0.903, "bbox": [0.8295, 0.511, 0.104, 0.1106], }, { "category": "1", "conf": 0.912, "bbox": [0.5644, 0.4218, 0.08496, 0.1673], }, { "category": "1", "conf": 0.934, "bbox": [0.7041, 0.4765, 0.1108, 0.125], }, ], ), ( "test_data/american_black_bear.jpg", [ { "category": "1", "conf": 0.959, "bbox": [0.4331, 0.4283, 0.3232, 0.3222], } ], ), ( "test_data/domestic_cattle.jpg", [ { "category": "1", "conf": 0.942, "bbox": [0.8886, 0.483, 0.1113, 0.1093], }, { "category": "1", "conf": 0.952, "bbox": [0.5791, 0.3723, 0.07617, 0.1295], }, { "category": "1", "conf": 0.954, "bbox": [0.3803, 0.4348, 0.1176, 0.1575], }, ], ), ( "test_data/domestic_dog.jpg", [ { "category": "1", "conf": 0.0137, "bbox": [0.2402, 0.3092, 0.2548, 0.2109], }, { "category": "1", "conf": 0.929, "bbox": [0.2377, 0.08398, 0.5161, 0.6497], }, ], ), ( "test_data/human.jpg", [ { "category": "1", "conf": 0.0118, "bbox": [0.07877, 0.64, 0.107, 0.173], }, { "category": "1", "conf": 0.0738, "bbox": [0.4104, 0.6707, 0.09472, 0.1238], }, { "category": "2", "conf": 0.933, "bbox": [0.7115, 0.4976, 0.0664, 0.2424], }, ], ), ( "test_data/ocelot.jpg", [ { "category": "1", "conf": 0.971, "bbox": [0.0478, 0.5548, 0.521, 0.2743], } ], ), ( "test_data/vehicle.jpg", [ { "category": "3", "conf": 0.977, "bbox": [0, 0.02083, 1.0, 0.6816], } ], ), ] ) def predicted_vs_expected(self, detector, request) -> tuple[list, list]: filepath, label = request.param img = detector.preprocess(load_rgb_image(filepath)) assert img is not None return detector.predict(filepath, img)["detections"], label def test_detections(self, predicted_vs_expected) -> None: predicted, expected = predicted_vs_expected assert predicted == sorted(predicted, key=lambda det: det["conf"], reverse=True) expected = sorted(expected, key=lambda det: det["conf"], reverse=True) for pred_det, exp_det in zip(predicted, expected): assert pred_det["category"] == exp_det["category"] assert pred_det["label"] == Detection.from_category(exp_det["category"]) assert pred_det["conf"] == pytest.approx(pred_det["conf"], abs=1.5e-3) assert pred_det["bbox"] == pytest.approx(exp_det["bbox"], abs=1.5e-3)