Hi Francis,
now it works in the other direction too. Basically, what you have done
is a general splitting of unknown words, haven't you? Without changing
the paradigms for every noun (in the pair nn-nb almost every noun calls
a paradigm with splitting options). But how is the splitting of unknown
words done in the pair nn-nb? I cannot find any equivalent to what you
have done.

But what now? How to produce a translation? When I call Apertium,
instead of  lt-proc, I don't get any splitting. How do I achieve that?

echo "Polisen hade hittat deras tjuvgömma. Tjuvgodset var redan sålt av
hälaren." | apertium -d /home/per/Repository/apertium-sv-da sv-da
gives me:
Politiet havde fundet deres *tjuvgömma. *Tjuvgodset var allerede *sålt
af *hälaren.

But calling lt-proc works:

echo "Polisen hade hittat deras tjuvgömma. Tjuvgodset var redan sålt av
hälaren." | lt-proc -e sv-da.automorf.bin

gives:

^Polisen/Polis<n><ut><sg><def><nom>/Polis<n><ut><sg><def><nom>$
^hade/ha<vbhaver><past><actv>$
^hittat/hitta<vblex><supn><actv>/hitta<vblex><pp><nt><sg><ind>$
^deras/de<prn><pers><p3><un><pl><gen>/deras<det><pos><ut><sg>/deras<det><pos><nt><sg>/deras<det><pos><un><pl>$
^tjuvgömma/tjuv<n><ut><sg><ind><cmp>+gömma<n><ut><sg><ind><nom>$^./.<sent>$
^Tjuvgodset/Tjuv<n><ut><sg><ind><cmp>+gods<n><nt><sg><def><nom>$
^var/var<adv>/vara<vbser><past><actv>$
^redan/redan<adv>/reda<n><ut><sg><def><nom>/reda<n><ut><sg><def><nom>$
^sålt/*sålt$ ^av/av<pr>$
^hälaren/häl<n><ut><sg><ind><cmp>+ar<n><nt><pl><def><nom>$^./.<sent>$

Yours,
Per Tunedal

On Sat, Nov 10, 2012, at 17:11, Francis Tyers wrote:
> El ds 10 de 11 de 2012 a les 17:09 +0100, en/na Per Tunedal va escriure:
> > Hi,
> > yes, your examples work. Now the first part works: splitting unknown
> > words into parts. What exactly have you changed in the Danish monodix?
> 
> You can use svn diff -r HEAD or look at the viewvc log. 
> http://apertium.svn.sourceforge.net/viewvc/apertium/trunk/apertium-sv-da/apertium-sv-da.da.dix?view=log
> 
> But basically, I called the compound pardefs from the case pardefs.
> 
> > I tried some changes in the Swedish monodix: modified the pardefs for
> > some words to open for compounding. That wasn't successful.
> 
> Just copy what I did with Danish.
> 
> > I guess that is the next step, after what ever you did:
> > 
> > Adding compounding to the pardefs for nouns in the source language.
> > And then doing the same in the target language.
> 
> Fran
> 

------------------------------------------------------------------------------
Everyone hates slow websites. So do we.
Make your web apps faster with AppDynamics
Download AppDynamics Lite for free today:
http://p.sf.net/sfu/appdyn_d2d_nov
_______________________________________________
Apertium-stuff mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/apertium-stuff

Reply via email to