Missing font leads to missing characters at the Extraction API output

Report · Nov 22, 2021

When extracting text from Page 59 (zero based counting) at the following PDF, The word 2008 is extracted as unknown characters:

"Text": "WESTERN UNION  Annual Report "

PDF file

When watching the PDF with Acrobat Reader, it looks ok:

I tried with PDFBox, and got the following error for these characters:

No Unicode mapping for twoalt (2) in font HGLLLJ+BulmerMT-ItalicAlt

Any help with that would be highly appriciated !!!

Report · Nov 23, 2021

I haven't dug too far into the file but generally, if the font uses custom encoding and the ToUnicode map isn't in the font resource, we can't extract the text accurately.