All, I just finished drafting a high level "lab report" comparing pdftotext and Tika/PDFBox on the PDFs in our refreshed regression corpus: https://wiki.apache.org/tika/ComparisonTikaAndPDFToText201811. The more interesting bits are in the actual reports from tika-eval and/or the comparison database available here: http://162.242.228.174/pdf_parsing/pdftotextVPDFBox_201811/
Let me know what you think.
Cheers,
Tim
