Paste your own OCR text and extraction JSON to test the correction rules.
Raw text from OCR (may contain errors like O instead of 0)
Extracted values (may have inherited OCR errors)