The problem with this seems to be that the pdftohtml program used by
Greenstone doesn't handle right-to-left text (which is perhaps not very
The solution to this is to apply the Unicode Bidirectional Algorithm to
the HTML file produced by pdftohtml. One program that does this is
"fribidi" (http://fribidi.org), which is installed by default on most
Linux computers. I'm not sure whether this is available for Windows, but
I imagine there will be an equivalent program available if not.
I've added an option to ConvertToPlug called "-apply_fribidi"; this will
run fribidi after pdftohtml is finished. You can download the new
version of ConvertToPlug.pm from
Replace your existing perllib/plugins/ConvertToPlug.pm file with this
new one, then edit your collection's collect.cfg file and add "-fribidi"
to the PDFPlug instance, eg:
plugin PDFPlug -fribidi
Then re-import and re-build the collection.
The "fribidi" program doesn't seem to be perfect (you'll be able to tell
much better than I can!) -- if you find any problems with it please
e-mail the fribidi people.
MOH Scorpion wrote:
>Thanks to Ms. Katherine don for fixing bug in
>pdfplugin but it has another problem it joins the
>letter in wrong direction :Left To Right .
>it must be right to left .
>please help me fix this .
>Yahoo! Mail - PC Magazine Editors' Choice 2005
>greenstone-users mailing list