Hello!
You can easily implement almost all of described features
by using UdmSearch "external parsers". Take a look at Mime
indexer.conf command. You need to make a parser that takes
text/html as input, then does something with document and
then return text/html again to indexer.
You can not change the weight of the words though.
But at least I can't see any problems with filters or language
detection or content based indexing.
Regards
Martin Ebert wrote:
> I think, it is better to do first another thing:
>
> We need a slot (or more then one) for these things.
> This slot (I 've told Alex this a few days before)
> should be called by indexer :
>
> - after receiving the document
> - has documented inputs/outputs
> - can give back (output) a value index/not index/Slot[number]Weight
> [another_number]
> - can give back a specific value: detected language or so
> - can give back the *filtered* document
>
> This generally slot (interface) can give
>
> - you the chance to implement your searchlist
> - other people the chance to implement language detection
> - other people the chance to implement content based indexing (not Indexing)
> - other people the chance to filter bad things (whatever this means)
> and so on.
--
Alexander Barkov
IZHCOM, Izhevsk
email: [EMAIL PROTECTED] | http://www.izhcom.ru
Phone: +7 (3412) 51-55-45 | Fax: +7 (3412) 78-70-10
ICQ: 7748759
______________
If you want to unsubscribe send "unsubscribe udmsearch"
to [EMAIL PROTECTED]