ispellaff2myspell

Langue: en

Version: 2009-06-04 (debian - 07/07/09)

Section: 1 (Commandes utilisateur)

NAME

ispellaff2myspell - A program to convert ispell affix tables to myspell format.

SYNOPSIS

  ispellaff2myspell [options] <affixfile> --myheader your_header
 
    Options:
 
     --affixfile=s      Affix file
     --bylocale         Use current locale setup for upper/lowercase 
                        conversion
     --charset=s        Use specified charset for upper/lowercase 
                        conversion (defaults to latin1)
     --debug            Print debugging info
     --extraflags=s     Allow some non alphabetic flags
     --lowercase=s      Lowercase string
     --myheader=s       Header file 
     --printcomments    Print commented lines in output 
     --replacements=s   Replacements file 
     --split=i          Split flags with more that i entries
     --uppercase=s      Uppercase string
 
 

DESCRIPTION

ispellaff2myspell is a script that will convert ispell affix tables to myspell format in a more or less successful way.

This script does not create the dict file. Something like

( echo `cat mydict.words+ | wc -l`; cat mydict.words+ ) > mydict.dict

should do the work, with mydict.words+ being the munched wordlist

OPTIONS

--affixfile=s
Affix file. You can put it directly in the command line.
--bylocale
Use current locale setup for upper/lowercase conversion. Make sure that the selected locale match the dictionary one, or you might get into trouble.
--charset=s
Use specified charset for upper/lowercase conversion (defaults to latin1). Currently allowed values for charset are: latin0, latin1, latin2, latin3.
--debug
Print some debugging info.
--extraflags:s
Allows some non alphabetic flags.

When invoked with no value the supported flags are currently those corresponding to chars represented with the escape char \ as first char. \ will be stripped.

When given with the flag prefix will allow that flag and strip the given prefix. Be careful when giving the prefix to properly escape chars, e.g. you will need -e ``\\\\'' or -e '\\' for flags like \[ to be stripped to [. Otherwise you might even get errors. Use -e ``^'' to allow all flags and pass them unmodified.

You will need a call to -e for each flag type, e.g., -e ``\\\\'' -e ``~\\\\'' (or -e '\\' -e '~\\').

When a prefix is explicitely set, the default value (anything starting by \) is disabled and you need to enable it explicitely as in previous example.

--lowercase=s
Lowercase string. Manually set the string of lowercase chars. This requires --uppercase having exactly that string but uppercase.
--myheader=s
Header file. The myspell aff header. You need to write it manually. This can contain everything you want to be before the affix table
--printcomments
Print commented lines in output.
--replacements=file
Add a pre-defined replacements table taken from 'file' to the .aff file. Will skip lines not beginning with REP, and set the replacements number appropriately.
--split=i
Split flags with more that i entries. This can be of interest for flags having a lot of entries. Will split the flag in chunks containing i entries.
--uppercase=s
Uppercase string. Manually set the sring of uppercase chars. This requires --lowercase having exactly that string but lowercase.

If your encoding is currently unsupported you can send me a file with the two strings of lower and uppercase chars. Note that they must match exactly but case changed. It will look something like

   $lowercase='a-zaaaa~a.a°aeceeee.iiii.d-n~oooo~o.ouuuu.yIo';
   $uppercase='A-ZAAAA~A.A°CEEEE.IIII.D-N~OOOO~O.OUUUU.YIo';
 
 

SEE ALSO

The OpenOffice.org Lingucomponent Project home page

<http://lingucomponent.openoffice.org/index.html>

and the document

<http://lingucomponent.openoffice.org/affix.readme>

that provides information about the basics of the myspell affix file format.

You can also take a look at

  /usr/share/doc/libmyspell-dev/affix.readme.gz
  /usr/share/doc/libmyspell-dev/README.compoundwords
  /usr/share/doc/libmyspell-dev/README.replacetable
 
 

in your Debian system.

AUTHORS

Agustin Martin <agustin.martin@hispalinux.es>