UTF-8 handles most cases but I still have to deal with emails in ISO2022-JP, Shift-JIS and EUC-JP. After some research it seems Xapian has no support for Asian languages. I will try to make some tests and open an issue if I cannot make it work.
I can see in the sup configuration file that the stem language can be configured but there are no CJK stemmers for Xapian that I can find. On Thu, May 2, 2013 at 5:17 PM, Gaute Hope <e...@gaute.vetsj.com> wrote: > > > On 30. april 2013 11:44, Horacio Sanson wrote: > > Great to see Sup getting back on track again.. > > > > I submitted some patches for the Gmail dumper of Heliotrope some time ago > > but the lack of non alphabet languages (Japanese, Chinese) made it > > impossible for me to keep using heliotrope/turnesole. > > > > The main issue to support Japanese/Chinese with heliotrope was that > > whistlepig (indexer) lacked the ability to tokenize these languages. Also > > the half baked UTF-8 support caused several issues with these languages. > > > > I would like to help in testing/implementing support for these languages, > > starting with Japanese, but I would require some guidance. First I would > > like to know is there is a way to configure the Xapian tokenizer > > (segmenter) within sup? Please consider that I am new to both sup and to > > Xapian. > > Hi Horacio, > > consider opening an issue at > https://github.com/sup-heliotrope/sup/issues to make sure this doesn't > disappear. Some changes will probably be made to the indexer when going > to Mail (from RMail), but I hope to be able to migrate the existing > index. Perhaps its time to get it right for arbitrary languages as well. > I am unfamiliar with Japanes/Chinese - does UTF-8 cover the needs? > > Mail is better at handling UTF-8 and I think there was some fork that > had some extra support for Japanese. > > Regards, Gaute >
_______________________________________________ Sup-devel mailing list Sup-devel@rubyforge.org http://rubyforge.org/mailman/listinfo/sup-devel
