Text this: Choosing preclinical study models of diabetic retinopathy: key problems for consideration