OCR extraction. Input is either a PDF (pages render at dpi and the engine recognizes them, imported lazily) or an already-recognized word dump (the offline replay input — see --words-json).
dpi
--words-json
{ pdfPath } to recognize, or { words } to replay a dump
{ pdfPath }
{ words }
see ExtractOcrOptions
OCR extraction. Input is either a PDF (pages render at
dpiand the engine recognizes them, imported lazily) or an already-recognized word dump (the offline replay input — see--words-json).