Comment on “Recognizing uric acid type of urinary stones by deep learning”: strengthening the reference standard and the image-to-label pipeline for reproducibility