[Dict-common-dev] MySpell hyphenation patterns
Agustin Martin Domingo
agustin.martin@hispalinux.es
Wed, 07 Jan 2004 19:22:53 +0100
Rene Engelhard wrote:
> Ih have patches for ifrench / ifrench-gut ready to build the
> dictionaries from the ispell ones directly which was how the zip was
> created anyhow. This was possible only now that we have alibmyspell-dev=
> in unstable which incidentially contains a script which was made for
> converting ifrench-gut :-)
Just for info, some time ago I was playing to write an aff converter and =
put it at
http://corbu.aq.upm.es/~agmartin/linux/store/ispell2myspell
It is a bit naive and somewhat unfinished perl script, but seems to work =
and do what is expected from it, including hardcoded support for latin1, =
latin2 and latin3 lc/uc conversion (it can also rely on locales for=20
this, but this might result in problems if the right locale is not=20
installed), as well as optional flag split into chunks for flags having=20
too many entries. It currently allows only alphabetic flags and also=20
requires an external header file with the myspell header, something like
SET ISO8859-1
# TRY
TRY aersoinltcdmubp=E1g=EDzfvh=E9jq=F3=F1xy=FA=FCkCMAIESPGJBRFTDVHUOwLKNZ=
=C1YX=DC=D3=DA=C9=D1QW=CD
It only deals with aff conversion, for the dict I hope something like
( echo `cat galician.words | wc -l`; cat galician.words ) > galician.dict=
to do the right thing, with .words being the munched wordlist
I wrote it some time ago and was waiting to take a look at other similar =
scripts flying around to try improving it, and eventually put it or=20
other similar script at dictionariers-common-dev package, but never had=20
the time for that. At some time I would like to add the possibility of=20
some non alphabetic flags to work with norwegian dicts, but cannot promis=
e.
Feel free to play with it
Cheers,
--=20
Agustin