matching both accented and non-accented character for non-accented characters?

matching both accented and non-accented character for non-accented characters?

[Bence Ferdinandy]

[-- Attachment #1.1: Type: text/plain, Size: 699 bytes --]

Hi,

I'm in the process of trying to set up reading email in the terminal and
just installed notmuch, which looks like a pretty awesome tool. I currently
have one question nagging me:

I have a lot of mail in my native Hungarian, which properly written is full
of characters like éáűúü, but if someone's writing on a non-Hungarian
keyboard, or just quickly writing an email from a phone, they often drop
the accents as it's faster and we'll likely understand anyway. Is it
possible to set it up that if I search for "lanc" it would also match
"lánc" other than going `notmuch search lanc OR lánc`?

Thanks very much,
Bence Ferdinandy

-- 
+36305425054
bence.ferdinandy.com

[-- Attachment #1.2: Type: text/html, Size: 1022 bytes --]

[-- Attachment #2: Type: text/plain, Size: 0 bytes --]

Re: matching both accented and non-accented character for non-accented characters?

[David Bremner]

Bence Ferdinandy <bence@ferdinandy.com> writes:

> Hi,
>
> I'm in the process of trying to set up reading email in the terminal and
> just installed notmuch, which looks like a pretty awesome tool. I currently
> have one question nagging me:
>
> I have a lot of mail in my native Hungarian, which properly written is full
> of characters like éáűúü, but if someone's writing on a non-Hungarian
> keyboard, or just quickly writing an email from a phone, they often drop
> the accents as it's faster and we'll likely understand anyway. Is it
> possible to set it up that if I search for "lanc" it would also match
> "lánc" other than going `notmuch search lanc OR lánc`?

There is some previous discussion at

      https://nmbug.notmuchmail.org/nmweb/search/id%3A87efp2b9er.fsf%40tethera.net

I don't think anyone worked on this in the meantime, so I guess the
short answer is that there is currently no support, but people have
tossed around some ideas.

d\r

Re: matching both accented and non-accented character for non-accented characters?

[Bence Ferdinandy]

[-- Attachment #1.1: Type: text/plain, Size: 1650 bytes --]

Thanks! I didn't know unicode equivalence existed, but it seems to be the
feature I want, so at least now I have a name for it :) And yes, actually
setting the stemmer would also be cool, I saw that Xapian has a Hungarian
stemmer but I kind of assumed all stemmers are applied somehow (although it
makes sense they're not). Is stemming done during search or would it affect
the database as well? Just to have a notion of how complicated a settable
stemmer feature would be.

David Bremner <david@tethera.net> ezt írta (időpont: 2022. aug. 8., H,
16:58):

> Bence Ferdinandy <bence@ferdinandy.com> writes:
>
> > Hi,
> >
> > I'm in the process of trying to set up reading email in the terminal and
> > just installed notmuch, which looks like a pretty awesome tool. I
> currently
> > have one question nagging me:
> >
> > I have a lot of mail in my native Hungarian, which properly written is
> full
> > of characters like éáűúü, but if someone's writing on a non-Hungarian
> > keyboard, or just quickly writing an email from a phone, they often drop
> > the accents as it's faster and we'll likely understand anyway. Is it
> > possible to set it up that if I search for "lanc" it would also match
> > "lánc" other than going `notmuch search lanc OR lánc`?
>
> There is some previous discussion at
>
>
> https://nmbug.notmuchmail.org/nmweb/search/id%3A87efp2b9er.fsf%40tethera.net
>
> I don't think anyone worked on this in the meantime, so I guess the
> short answer is that there is currently no support, but people have
> tossed around some ideas.
>
> d
>


-- 
+36305425054
bence.ferdinandy.com

[-- Attachment #1.2: Type: text/html, Size: 2400 bytes --]

[-- Attachment #2: Type: text/plain, Size: 0 bytes --]

Re: matching both accented and non-accented character for non-accented characters?

[David Bremner]

Bence Ferdinandy <bence@ferdinandy.com> writes:

> Thanks! I didn't know unicode equivalence existed, but it seems to be the
> feature I want, so at least now I have a name for it :) And yes, actually
> setting the stemmer would also be cool, I saw that Xapian has a Hungarian
> stemmer but I kind of assumed all stemmers are applied somehow (although it
> makes sense they're not). Is stemming done during search or would it affect
> the database as well? Just to have a notion of how complicated a settable
> stemmer feature would be.
>

Stemming happens both during search (unless turned off for given term)
and at indexing time. So yeah, changing the stemming algorithm with
change the database (and require a re-index).