|
|
"""Tests for preprocessing module.""" |
|
|
|
|
|
import numpy as np |
|
|
import pytest |
|
|
from src.inference.preprocess import preprocess, preprocess_batch, crop |
|
|
|
|
|
|
|
|
class TestPreprocess: |
|
|
def test_output_shape(self, dummy_face_crop): |
|
|
"""Output shape is (C, H, W).""" |
|
|
result = preprocess(dummy_face_crop, 128) |
|
|
assert result.shape == (3, 128, 128) |
|
|
|
|
|
def test_output_range(self, dummy_face_crop): |
|
|
"""Output values in [0, 1].""" |
|
|
result = preprocess(dummy_face_crop, 128) |
|
|
assert result.min() >= 0.0 |
|
|
assert result.max() <= 1.0 |
|
|
|
|
|
def test_output_dtype(self, dummy_face_crop): |
|
|
"""Output is float32.""" |
|
|
result = preprocess(dummy_face_crop, 128) |
|
|
assert result.dtype == np.float32 |
|
|
|
|
|
def test_different_input_sizes(self): |
|
|
"""Handles various input sizes.""" |
|
|
for size in [64, 100, 200, 256]: |
|
|
img = np.random.randint(0, 255, (size, size, 3), dtype=np.uint8) |
|
|
result = preprocess(img, 128) |
|
|
assert result.shape == (3, 128, 128) |
|
|
|
|
|
|
|
|
class TestPreprocessBatch: |
|
|
def test_batch_shape(self, dummy_face_crop): |
|
|
"""Batch output shape is (N, C, H, W).""" |
|
|
batch = [dummy_face_crop, dummy_face_crop] |
|
|
result = preprocess_batch(batch, 128) |
|
|
assert result.shape == (2, 3, 128, 128) |
|
|
|
|
|
def test_empty_batch_raises(self): |
|
|
"""Empty batch raises ValueError.""" |
|
|
with pytest.raises(ValueError): |
|
|
preprocess_batch([], 128) |
|
|
|
|
|
|
|
|
class TestCrop: |
|
|
def test_crop_square_output(self): |
|
|
"""Crop returns square image.""" |
|
|
img = np.random.randint(0, 255, (480, 640, 3), dtype=np.uint8) |
|
|
bbox = (100, 100, 200, 200) |
|
|
result = crop(img, bbox, 1.5) |
|
|
|
|
|
assert result.shape[0] == result.shape[1] |
|
|
|
|
|
def test_crop_invalid_bbox_raises(self): |
|
|
"""Invalid bbox raises ValueError.""" |
|
|
img = np.random.randint(0, 255, (480, 640, 3), dtype=np.uint8) |
|
|
bbox = (100, 100, 100, 100) |
|
|
|
|
|
with pytest.raises(ValueError): |
|
|
crop(img, bbox, 1.5) |
|
|
|
