Extracting text from PDF : PDFLib vs PDF extract vs pdf2xml [closed]

Question

I am looking for a library (if possible available in Java or PHP) in order to extract text from a PDF. There is a lot of software available, including:

• 3-Heights™ PDF Extract http://www.pdf-tools.com/pdf/pdf-extract-content-metadata-text.aspx

• PDFlib TET – Text Extraction Toolkit http://www.pdflib.com/products/tet/

• PDF2XML http://sourceforge.net/projects/pdf2xml/

Which tools would you choose? What do you think of them?

Thank you very much for your kind help!

Answer 1

My favourite is iText (java) but extracting text from a PDF can be fraught with difficulties as the text in the PDF is not alway stored in the order in which it appears.

Answer 2

itext had trouble properly extracting text from the irs i1040.pdf as reported here:

<1> article.gmane.org/gmane.comp.java.lib.itext.general/65680

As near as I can figure out, when the fonts are embedded, it's not always possible to get the right text extracted. See: <2> www.verypdf.com/wordpress/201109/pdf-to-text-converter-cant-extract-text-which-render-by-embedded-fonts-2452.html <3> section 9.10.1 of: www.adobe.com/content/dam/Adobe/en/devnet/pdf/pdfs/PDF32000_2008.pdf

<3> says:

If a font is not defined in one of these ways ... the characters cannot be converted to Unicode values without additional information.

I assume that "conversion to unicode values" is essential for text extraction.