Hi all, sorry to step in suddenly... but recently I've been using
Compass framework with excellent results. Compass has a framework called
XSEM (Xml Search Engine Mapping) which makes a really nice abstraction
for Lucene. Maybe this is not a convenient solution for Xindice, but, I
believe it is worth a try!
Best regards,
Nacho G. Mac Dowell
Vadim Gritsenko escribió:
Natalia Shilenkova wrote:
All,
I am interested in adding full text indexing capability to Xindice,
which is, by the way, one of the items on TODO list. The items refers
to the scratchpad code, but there is not much, just couple of classes
that were moved out of main codebase at some (rather distant) point.
However, mail list archive has several threads about that very same
subject, with most interesting being 'Xindice full text indexing' [1]
and 'Full text indexing patch' [2].
Does anyone remember why that effort was abandoned?
I think you'd have to ask Andy about that :)
Is that still a
preferred way of indexing implementation?
Yes, I think so, it has all necessary pieces for initial
implementation: text search service, text indexer, text query
resolver... More thought should be put into how it indexes (I do not
remember what/how exactly this patch indexes...), what sort of fields
it creates in Lucene index and such, but this can come later as
further enhancements as soon as basics are taken care of.
And even next step after that would be to add full text XPath
functions - most probably as extension functions.
Vadim
[1] http://marc.info/?t=107721013100003&r=1&w=2
[2] http://marc.info/?t=107900573100002&r=1&w=2
Regards,
Natalia
