> What do you mean with 'virtual XML file' ? Perhaps the word 'virtual' puts
> on the wrong track here. Do you mean: simply create a shellscript that
> generates an XML file as explained above and then index this xml file?
if you had the disk space, then you could create a real .xml file. That
effectively doubles the required disk space for your doc collection. I
personally wouldn't want to use that much space, unless I was re-indexing
often, so I would just write a script (similar to DirTree.pl or spider.pl)
that created a XML doc in memory and then printed to stdout. swish-e can
read from stdin, so it would look like:
myscript | swish-e -S prog -i stdin
If you do it in Perl, look at the SWISH::Filter modules for examples of
creating the virtual doc and filtering from PDF, etc.
Peter Karman . http://peknet.com/ . peter(at)not-real.peknet.com
Received on Thu Dec 1 08:41:49 2005