pdf to txt library?

Is there a simple library on aos4 which allows to extract text from a pdf file in c++ programms?

Thanks,
nexus

@nexus

Are you trying to do something as a one off or coding something new?

@nexus

Quote:
AFAIK no. There is libharu on OS4depot but it can only put text to PDF, not the other way round.

EDIT: Perhaps you can coax someone into porting Poppler for you: http://en.wikipedia.org/wiki/Poppler_%28software%29

hm. ok, i see.

poppler .. yes, it's a fork of xpdf and as we already have amipdf, which is a xpdf clone, i hoped, that maybe there's also an appropriate library.

thanks,
nexus

@nexus

Not that I know of, but I had a similar problem converting pdf's for my kindle. After hunting through aminet and os4depot I found a couple of programs that let me do it:

pdftohtml: http://os4depot.net/download.php?file ... ext/convert/pdftohtml.lha

html2txt: http://uk.aminet.net/text/hyper/html2txt.lha

together with a little amiblitz program I wrote: http://www.am1ga.net/Left$

I then wrote the following dos script to convert pdf to html:

.bra {
.ket }
.key filename

set title `left$ "{filename}" -4` ; removes the .pdf extension
if exists "${title}.txt"
echo "File already converted."
quit
endif

echo "Converting to html."
pdftohtml -q "{filename}"
echo "converting to txt."
html2txt "${title}s.html" "${title}.txt" >nil:
;clean up the files created by pdftohtml and set
delete "${title}.html" "${title}s.html" "${title}_ind.html" quiet
unset title
echo "Done."

not very elegent and the conversion is a bit dodgy with some pdf's but it does the job.

@Severin

hm. that's somehow unsatisfiable. I thought that at least the xpdf tools like pdftotxt were available but I've just found out that amipdf doesn't have them included. With my question about the library, i wanted to avoid the inelegant way of using an external tools to convert a pdf on dis and using then a file reader.

Your suggestion showed me that it's even more inelegant because one have to use a complete chain of external tools and hope that in the end, the result is sufficient "useful" for further processing.

Anyway, thanks..
nexus