The actual
situation is as follows:

I got a
wordlist with lemmas and Part-of-Speech tags and created an affix file by a
grammar. To connect the words and the rules I wrote a Java program which
replaces some PoS-tags by the rule name. Some of them were connected manually.
Yes, I can imagine, that the tool works well for maintenance, so I can use it
now. 

 

Exactly, now
I have to remove redundancy. And the compounding isn’t working yet. 

For example:

 

I have the words tschun/Zxyz, in/Zxyz and melli/xyz in the dictionary and
the rule Z in my affixfile:

SFX Z Y 26

SFX Z    0        onta/WXZxyz       [gt]        #
10

SFX Z    0        tschien/Xxyz      .
          #
100

SFX Z    0        melli/Xxyz        .           # 1000

SFX Z    0        s/V               [ai]        # dua, trei --> duas, treis

SFX Z    a        s/V               dua

SFX Z    in       endisch           in          #
11

SFX Z    a        disch/W           a           # dudisch (12)

SFX Z    i        disch/W           i           #
tredisch (13)

SFX Z    ater     itordisch/W       quater      # quitordisch (14)

SFX Z    t        disch/W           quent       #
quendisch (15)

SFX Z    is       edisch/W          sis         # sedisch (16)

SFX Z    dua      vegn/ZW   
       dua   
     # 20

SFX Z    0        conta/WXZxyz      un   
      # 50

SFX Z    0        sonta/WXZxyz   
  s       
   # 60

SFX Z    ov       avonta/WXZxyz     nov   
     # 90

SFX Z    a        in/ZXxyz          a       
   # trentin(31), curontin(41)

SFY Z    in       emprem/W   
      in   
      # 1.

SFY Z    in       prem              in          # 1.

SFX Z    dua      secund/W   
      dua   
     # 2.

SFX Z    rei      ierz/W   
        trei   
    # tierz (3.)

SFX Z    rei      erzio   
         trei   
    # terzio (3.)

SFX Z    ter      rt/W   
          quater
     # quart (4.)

SFX Z    tschun   quint/W   
       tschun   
  # quint (5.)

SFX Z    0        avel/W   
        [uiaton][nstgv]   
    # 6.,7. etc.

SFX Z    rei      iarza   
         trei   
    # tiarza (1/3)

SFX Z    ent   
  int               quent        



tschuncontamelli is correctly identified and
tschuncontainmelli (not correct) is suggested, but the right word
tschuncontinmelli isnt found. Why not? How is the compounding working? 

 

I am testing on the prompt, because I cant see
the extension in the OO Writer. I imported it via extension manager but it isn’t
available for the spell check yet. 

My extension consist of the dic, the aff, the description.xml,
the dictionary.xcu and a readme.txt with the license, so whats wrong?

 

Regards, 

Maria

 

 

> Date: Fri, 15 Aug 2014 17:44:04 +0200
> From: pesce...@apache.org
> To: l10n@openoffice.apache.org
> Subject: Re: Spellchecker for Rhaeto-Romanic
> 
> On 12/08/2014 Maria Hartmann wrote:
> > Good morning,
> > the proofing tool gui is no help, sadly.
> 
> I had understood that what you wanted to do is maintenance and checks on 
> a dictionary, so that you had already created you .dic file, your .aff 
> file and added (manually) the rules to the .dic file. This is how 
> maintenance is normally done, and it is the scenario in which the tool 
> is useful.
> 
> Are you at another stage instead? I mean, what does your .aff file look 
> like? Are you at the stage where you have a huge wordlist collected 
> randomly (so, not by hand) and you created some .aff rules and want to 
> remove redundancy (i.e., obtaining an equivalent dictionary with 
> annotated words in the .dic file)?
> 
> It will greatly help (if you are still stuck here) if you copy and paste 
> ten lines (no attachments please) from your .dic file and .aff file.
> 
> Regards,
>    Andrea.
> 
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: l10n-unsubscr...@openoffice.apache.org
> For additional commands, e-mail: l10n-h...@openoffice.apache.org
> 
                                          

Reply via email to