luuik / foma

Automatically exported from code.google.com/p/foma
0 stars 0 forks source link

verb alone works fine; when combined woth other word types, some cases are missing #35

Closed GoogleCodeExporter closed 9 years ago

GoogleCodeExporter commented 9 years ago
What steps will reproduce the problem?
verb fst is called hun42.fst. It works fine alone:
en@en-desktop ~/program/foma/tktest $ kill_flookup.sh[1]+  Kilőve              

en@en-desktop ~/program/foma/tktest $ flookup -i -S -A 127.0.0.1 
/home/en/program/foma/tktest/hun42.fst &
[1] 4770
en@en-desktop ~/program/foma/tktest $ Started flookup server on 127.0.0.1 port 
6062

en@en-desktop ~/program/foma/tktest $ echo "abnormalitásoz+Verb+CondDefPl3" | 
nc -w 1 -u 127.0.0.1 6062
abnormalitásoz+Verb+CondDefPl3 Abnormalitásoznák
abnormalitásoz+Verb+CondDefPl3 abnormalitásoznák
-------------------------------------
Here I combine verbs with the other word types:
cat hfnnum.foma
read regex @"hun41.fst" | @"hun42.fst" | @"hunnum.fst" | @"hunadj.fst" | 
@"hunfxpp.fst" | @"hunmisc.fst";
en@en-desktop ~/program/foma/tktest $ cat crfnnum.sh
foma -l hfnnum.foma <savestackfnnum.sh
cat savestackfnnum.sh
save stack hunfnnum.fst
----------------------------------------------
Here I try the combined fst file:
kill_flookup.sh
[1]+  Kilőve                 
flookup -i -S -A 127.0.0.1 /home/en/program/foma/tktest/hun42.fst

$ cat do_test.sh
kill_flookup.sh
flookup -i -S -A 127.0.0.1 /home/en/program/foma/tktest/hunfnnum.fst &
en@en-desktop ~/program/foma/tktest/tools/fomaallchk $ sh do_test.sh
en@en-desktop ~/program/foma/tktest/tools/fomaallchk $ Started flookup server 
on 127.0.0.1 port 6062

en@en-desktop ~/program/foma/tktest $ echo "abnormalitásoz+Verb+CondDefPl3" | 
nc -w 1 -u 127.0.0.1 6062
abnormalitásoz+Verb+CondDefPl3 ?+

It works for lots of cases, but not for ConjIndef... , ConjDef..., 
CondIndef..., CondDef...

Also just using foma -l ....foma shows the same results; the problem is not 
flookup related.

What is the expected output? What do you see instead?
expected:
abnormalitásoz+Verb+CondDefPl3 Abnormalitásoznák
I see instead:
abnormalitásoz+Verb+CondDefPl3 ?+

What version of the product are you using? On what operating system?
foma 0.9.16alpha (from svn)
linux debian

Please provide any additional information below.

Original issue reported on code.google.com by eleonor...@gmx.net on 16 Aug 2012 at 2:58

GoogleCodeExporter commented 9 years ago
I have found the reason. I had a .lexc file containing:
----------------------------------
Multichar_Symbols  +Con

LEXICON Root
Con;

LEXICON Con
akár AddCon;
akárcsak AddCon;
...

LEXICON AddCon
+Con:0     #;
------------------------------------
I changed in this file Con to Koet, and that fixed the problem.

This file is part of hunmisc.fst
hunmisc is created by:
foma -l enhumisc.foma <savestackmisc.sh
where savestackmisc.sh is:
save stack hunmisc.fst

and enhumisc.foma is a .foma file containing:
--------------------------------------------------
definitions...
echo ">>> read in enhuadv <<<"                    
read lexc /home/en/program/foma/tktest/lexc/misc/enhuadv1.lexc
define Lexiconadv

define Grammaradv Lexiconadv           .o. 
               CleanupEndings    .o.
               Cleanup           .o.
               ToUpCase ;

...
regex Grammaradv | Grammarsentint | Grammarpost | Grammarvbpfx | Grammarcon | 
Grammarneg | Grammardet | Grammartwin ;
------------------------------------------------------------------------------

I still believe, this name clash is a foma problem, that needs to be fixed.
Verbs have nothing to do with misc type words, they just need to coexist.

Original comment by eleonor...@gmx.net on 16 Aug 2012 at 9:41

GoogleCodeExporter commented 9 years ago
Continued to find the reason, and found it.
I forgot to define in the verb lexc files
Multichar_Symbols
                  +CondIndefSg1   +CondIndefSg2   +CondIndefSg3   
                  +CondIndefPl1 +CondIndefPl2 +CondIndefPl3
                  +CondDefSg1   +CondDefSg2   +CondDefSg3  
                  +CondDefPl1 +CondDefPl2 +CondDefPl3
                  +ConjIndefSg1   +ConjIndefSg2   +ConjIndefSg3   
                  +ConjIndefPl1 +ConjIndefPl2 +ConjIndefPl3
                  +ConjDefSg1   +ConjDefSg2   +ConjDefSg3  
                  +ConjDefPl1 +ConjDefPl2 +ConjDefPl3
                  +PastIndefSg1   +PastIndefSg2   +PastIndefSg3   
                  +PastIndefPl1 +PastIndefPl2 +PastIndefPl3
                  +PastDefSg1   +PastDefSg2   +PastDefSg3  
                  +PastDefPl1 +PastDefPl2 +PastDefPl3
and this caused the mysterious error. Foma has no fault, my logic was not 
complete.
We can close this issue.

Original comment by eleonor...@gmx.net on 17 Aug 2012 at 3:57

GoogleCodeExporter commented 9 years ago

Original comment by mans.hul...@gmail.com on 21 Aug 2012 at 7:39