import numpy as np
import cv2
import os
import json
from photo_quality_analyzer_core.analyzer import evaluate_photo_quality
TEST_DIR = "test_data"
def generate_test_images():
os.makedirs(TEST_DIR, exist_ok=True)
size = (512, 512)
# 1. Pure Sharp Edges (Solid black background with white rectangle)
img_sharp = np.zeros((*size, 3), dtype=np.uint8)
cv2.rectangle(img_sharp, (100, 100), (412, 412), (255, 255, 255), -1)
cv2.imwrite(os.path.join(TEST_DIR, "sharp_test.png"), img_sharp)
# 2. Pure Blurry Edges (Gaussian Blur on Sharp)
img_blurry = cv2.GaussianBlur(img_sharp, (21, 21), 0)
cv2.imwrite(os.path.join(TEST_DIR, "blurry_test.png"), img_blurry)
# 3. Pure Noise (Gaussian noise)
img_noise = np.random.normal(128, 50, (*size, 3)).astype(np.uint8)
cv2.imwrite(os.path.join(TEST_DIR, "noise_test.png"), img_noise)
# 4. Blown Highlights (Pure white)
img_white = np.ones((*size, 3), dtype=np.uint8) * 255
cv2.imwrite(os.path.join(TEST_DIR, "white_test.png"), img_white)
def run_suite():
print(f"Running automated validation suite...")
results = {}
images = [
("Sharp", "sharp_test.png"),
("Blurry", "blurry_test.png"),
("Noise", "noise_test.png"),
("White", "white_test.png")
]
for label, filename in images:
path = os.path.join(TEST_DIR, filename)
print(f"Analyzing {label}...")
try:
res = evaluate_photo_quality(path)
results[label] = {
"Overall": res["overallConfidence"],
"Sharpness": res["metrics"]["sharpness"]["score"],
"Exposure": res["metrics"]["exposure"]["score"],
"Noise": res["metrics"]["noise"]["score"]
}
except Exception as e:
print(f"Error analyzing {label}: {e}")
print("\n--- Validation Results ---")
print(json.dumps(results, indent=2))
# Simple Assertions for tuning
if results.get("Blurry") and results["Blurry"]["Sharpness"] > 0.4:
print("WARNING: Blurry image rated too sharp!")
if results.get("Noise") and results["Noise"]["Sharpness"] > 0.4:
print("WARNING: Noise rated too sharp (Noise/Sharpness confusion)!")
if results.get("White") and results["White"]["Exposure"] > 0.3:
print("WARNING: Blown highlights rated with good exposure!")
if __name__ == "__main__":
generate_test_images()
run_suite()
