Re: processing urls with tsearch2

On Thu, 13 Sep 2007, Laimonas Simutis wrote:

> Hey guys,
>
> maybe anyone using tsearch2 could advise on this. With the default
> installation, url, host and some other tokens are processed with the simple
> dictionary. Thus term like mywebsite.com gets stored as 'mywebsite.com'. The
> parser correctly assigns token id of type host to the term, but then the
> dictionary the terms gets routed through is simple and what gets stored is
> mywebsite.com
>
> The questions are:
>
> 1) is there a dictionary available that I could utilize that will remove
> .com, .net, .org, etc? I could write one myself, but after seeing some
> sample dictionary implementations and C code I try to avoid, I got scared a
> bit.

Yes, we have dict_regex, which was developed by Sergey Karpov, see details
http://lynx.sao.ru/~karpov/software/postgres_dict_regex.html
It uses pcre library and you need to know perl regexps.

>
> 2) has anyone else dealt with this maybe in a different way?

sure, preprocess text using prefered language before passing to ro_tsvector

>
>
> Thanks for any suggestions and help,
>
> Laimis
>

Regards,
Oleg
_____________________________________________________________
Oleg Bartunov, Research Scientist, Head of AstroNet (www.astronet.ru)
Sternberg Astronomical Institute, Moscow University, Russia
Internet: oleg(at)sai(dot)msu(dot)su, http://www.sai.msu.su/~megera/
phone: +007(495)939-16-83, +007(495)939-23-83