I've checked for copy and paste, there's so many character flaws, their OCR must have sucked really bad, I may try with deepseekOCR or something. I mean the database would probably more searchable if someone ran every file through a better OCR.