PDF-native extraction vs. vision models for document processingpymupdf.io6 points·by akaybk·há 8 meses·1 comments