Re: wildcard search / Norwegian challenge
Huh, I see that you thought a lot about this problem. I have same problems
with slovenian language. I tried with some dictionary parsing but i don't
like tsearch parser because when i convert words with to_query It puts
wrong operator. Let me explain this.
let say I search word "psi", this is dogs in english. So possible words if I
use dictionary parser is
pes, psa, psu, psov.... but when this is converted in
psi & pes & psa $ psu. So nothing is found, because this is options. So it
would be better if generat or in between possible variants of original word
from dictionary. Something like this
psi | (pes | psa | psu)
and also results with psi in it has to be ranked higher then those with
pes, psa, psu ....
I hope my explanation is clear enough. Maybe this is planed in TODO as
Teodor said "In short we're thinking about more flexible parser and
I don't know if it's possible to have more wights (I'm speaking of A,B,C
and D) so document can be weighted more accurately. Let say I know title,
description, url, body, keywords so there's five of them. But this has
nothing to do with what was primary the problem here. I just mentioned.
I can put some ideas in some other mail and we can discus about this.