Prescription parsing typically follows a two-stage process: an OCR engine converts the prescription image into raw text, then a large language model performs semantic extraction