Re: Is there some information that can be used for hyphenation using hunspell or aspell?

Hi Chen,

On Mon, Jun 20, 2011 at 3:49 PM, chenxiajian1985
<chenxiajian1985@gmail.com> wrote:
> Hi there :-)
>
> Dom, it's specially for you :-) I have talked with Pradeeban

I want to discuss with you the progress (a high-level update on how
far we have gone and what needs to be taken care of), as I asked in
this mail.

Dom is quite busy. So pls make sure to send the mail the the dev list,
instead of sending him (or anyone) personal mail, as we will miss the
mails.

We have to implement the hyphenation. So we will need some own
research by ourselves. :) Let's discuss further this week itself, to
address/clarify the issues, over irc/chat.

Regards,
Pradeeban.

>
> I have some problem while reading the source code of enchant
>
> 1. Some dict provide the function we need
> ispell_dict_suggest: provided by spell
> uspell_dict_suggest: we implemented it
> voikko_dict_suggest: we implemented it
> zemberek_dict_suggest : we implemented it
>
> but none of them provide the information about hyphenation, so I have some problem to implement hyphenation using the dict.
>
> there is the related hyphen project at
> http://sourceforge.net/projects/hunspell/files/Hyphen/ which includes an
> example.c for sample hyphenation. This "hyphen" is used by LibreOffice,
> OpenOffice.org and Scribus for their hyphenation.
>
> but since it is based on a rather limited algortihm (and there are
> issues in OOo/LO) its quality of hyphenation is poor for languages with
> long coumpounded words. Especially in case of ambiguous words.
>
>
> Is there some information that can be used for hyphenation using hunspell or aspell?
>
>
> Best Regard~
>
> chenxiajian
>
>
>
> 发件人： Kathiravelu Pradeeban
> 发送时间： 2011-06-18 22:37:39
> 收件人： chenxiajian1985
> 抄送： abiword-dev
> 主题： Re: chenxiajian_enchant_dict_hyphenationSuggest ?
>
> Hi Chen,
> As you discussed with me during our last chat, have you implemented
> this? Can you pls send the enchant specific code that you have
> implemented as a patch so that I can have a look?
> Can you also pls send a quick update of your progress to the mailing
> list upto now? We have been discussing over gtalk and personal mail.
> So a higher level update to the group would be great to see how far we
> have gone against the timeline that we have initially come up.
> We will also need a quick chat to ensure that we are on the track.
> Thank you.
> Regards,
> Pradeeban.
> On Tue, Jun 14, 2011 at 7:32 PM, Urmas <davian818@gmail.com> wrote:
>> From: "chenxiajian1985" <chenxiajian1985@gmail.com>
>> Subject: chenxiajian_enchant_dict_hyphenationSuggest ?
>>
>>> I try to implement hyphenation using enchant. so I add one method in enchant.
>>> ENCHANT_MODULE_EXPORT (char **)
>>> enchant_dict_hyphenationSuggest (EnchantDict * dict, const char *const word,
>>> ssize_t len, size_t * out_n_suggs);
>>>
>>
>> No, you'll need a separate object for doing hyphenation, something like:
>>
>> EnchantHyph* enchant_broker_request_hyph(EnchantBroker*, const char *const lang);
>>
>> to create an provider-dependant hyphenation object, and
>>
>> char *enchant_hyph_hyphenate(EnchantHyph *hyphenator, const char *const word);
>>
>> which would call a virtual method of "hyphenator" to obtain results;
>>
>>>
>>> 2. what is the best way to encode hyphenation results from enchant
>>> now I used (char **) to save the hyphenation results:
>>> such as : the word "hyphenation" can be save as:
>>> "hy-phen-ation" "hyphen-ation"....
>>
>> That's looks like terribly inefficient way, better just to return a string with embedded meta-characters (like '~' or control ones) in most desired and less desired breakpoints.
> --
> Kathiravelu Pradeeban.
> Software Engineer.
> WSO2 Inc.
> Blog: [Llovizna] http://kkpradeeban.blogspot.com/
>

-- 
Kathiravelu Pradeeban.
Software Engineer.
WSO2 Inc.
Blog: [Llovizna] http://kkpradeeban.blogspot.com/