|
|
|
|
|
import numpy as np |
|
import albumentations as A |
|
|
|
from src.utils import get_images_list, load_image, load_augmentations_config |
|
|
|
|
|
def test_get_images_list(): |
|
images_list = get_images_list("images") |
|
assert isinstance(images_list, list) |
|
assert len(images_list) > 0 |
|
assert isinstance(images_list[0], str) |
|
|
|
|
|
def test_load_image(): |
|
images_list = get_images_list("images") |
|
for image_name in images_list: |
|
image = load_image(image_name, path_to_folder="images", bgr2rgb=True) |
|
assert len(image.shape) == 3, f"error in {image_name}" |
|
assert image.shape[2] == 3, f"error in {image_name}" |
|
assert image.max() <= 255, f"error in {image_name}" |
|
assert image.min() >= 0, f"error in {image_name}" |
|
|
|
|
|
def test_load_augmentations_config(): |
|
image = np.random.randint(0, 255, (100, 100, 3)).astype(np.uint8) |
|
placeholder_params = { |
|
"image_width": image.shape[1], |
|
"image_height": image.shape[0], |
|
"image_half_width": int(image.shape[1] / 2), |
|
"image_half_height": int(image.shape[0] / 2), |
|
} |
|
augmentations = load_augmentations_config( |
|
placeholder_params, path_to_config="configs/augmentations.json" |
|
) |
|
|
|
for transform_name in augmentations.keys(): |
|
if transform_name in [ |
|
"CenterCrop", |
|
"RandomCrop", |
|
"RandomResizedCrop", |
|
"Resize", |
|
]: |
|
param_values = {"p": 1.0, "height": 10, "width": 10} |
|
elif transform_name in ["RandomSizedCrop"]: |
|
param_values = { |
|
"p": 1.0, |
|
"height": 10, |
|
"width": 10, |
|
"min_max_height": (50, 50), |
|
} |
|
elif transform_name in ["Crop"]: |
|
param_values = {"p": 1.0, "x_max": 10, "y_max": 10} |
|
else: |
|
param_values = {"p": 1.0} |
|
transform = getattr(A, transform_name)(**param_values) |
|
transformed_image = transform(image=image)["image"] |
|
assert len(transformed_image.shape) == 3, f"error in {str(transform)}" |
|
assert transformed_image.shape[2] == 3, f"error in {str(transform)}" |
|
assert transformed_image.max() <= 255, f"error in {str(transform)}" |
|
assert transformed_image.min() >= 0, f"error in {str(transform)}" |
|
|
