cropmodel still fails

tests/test_active_learning.py

@@ -0,0 +1,24 @@
+import pytest
+from src.active_learning import choose_train_images, choose_test_images
+
+@pytest.fixture
+def performance():
+    return {"detection": 0.9, "classification": 0.8}
+
+@pytest.fixture
+def detection_model():
+    from deepforest import main
+    return main.deepforest()
+
+@pytest.fixture
+def classification_model():
+    from deepforest.model import CropModel
+    return CropModel()
+
+def test_choose_train_images(performance, detection_model, config):
+    train_images_to_annotate = choose_train_images(performance, detection_model, **config.active_learning)
+    assert len(train_images_to_annotate) > 0
+
+def test_choose_test_images(performance, detection_model, config):
+    test_images_to_annotate = choose_test_images(performance, **config.active_testing)
+    assert len(test_images_to_annotate) > 0

tests/test_model.py

@@ -0,0 +1,16 @@
+import pytest
+import pandas as pd
+from deepforest import main
+from deepforest.model import CropModel
+from src.model import preprocess_and_train
+
+@pytest.mark.parametrize("model_type", ["detection", "classification"])
+def test_preprocess_and_train(config, model_type, crop_model):
+    if model_type == "detection":
+        model = main.deepforest()
+    else:
+        model = CropModel()
+    model.create_trainer(fast_dev_run=True)
+
+    validation_df = pd.read_csv(config.detection_model.validation_csv_path)
+    trained_model = preprocess_and_train(config, m=model, validation_df=validation_df, model_type=model_type)