Creating a new dictionary
Christian Lohmaier
lohmaier at googlemail.com
Mon Mar 17 08:08:55 PDT 2014
Hi Jørgen, *,
On Mon, Mar 17, 2014 at 3:46 PM, Jørgen G. Bosoni <jgb at netcom.no> wrote:
>
> we would like to create an experimental spell checker (hunspell) for a
> language not yet implemented in LibreOffice.
> Where can we find information about how to proceed?
hunspell's project page at sf.net has some in-depth documentation on
how to create dictionaries.
Basically dictionaries consist of a list of stems with flags, and
corresponding affix rules identified by those flags.
The affix rules alter the stems to match multiple words, the simplest
example would be
"word/s" in the dictionary (i.e. word is the stem, and "s" names an
affix rules),
and the affix rule would say: "append an s to the stem", so both
"word" and "words" would be valid words.
http://sourceforge.net/projects/hunspell/files/Hunspell/Documentation/
> At a later stage we would also like to add hyphenation,
Hyphenation dictionaries use the TeX format, i.e. again stems/patterns
with weights where to break a word (or where it is forbidden)
The files are postprocessed though - see
https://wiki.openoffice.org/wiki/Documentation/SL/Using_TeX_hyphenation_patterns_in_OpenOffice.org
for details.
> thesaurus and
Default is to use the openthesaurus format -
http://www.openthesaurus.de/ /
https://github.com/danielnaber/openthesaurus/ probably is easiest
solution to create and maintain a thesaurus.
> grammar.
There are different options available - one is langaugetool, the other
is lightproof (the default that is used for the english-grammar
checking in LO)
http://cgit.freedesktop.org/libreoffice/lightproof/tree/doc/manual.txt
https://www.languagetool.org/
ciao
Christian
More information about the LibreOffice
mailing list