There may be some way to identify if an image pulled from the PDF already has alt text attached. Generally, we want to give preference to what's already there, since it is likely human curated and more accurate. If an image does have alt text, we should skip the alt text generation, so we don't overwrite what's already there.
There may be some way to identify if an image pulled from the PDF already has alt text attached. Generally, we want to give preference to what's already there, since it is likely human curated and more accurate. If an image does have alt text, we should skip the alt text generation, so we don't overwrite what's already there.