[developers] cheap MWE problem and solution.

[Francis Bond]

G'day,

Sorry for the extended series of pet issues.  Please let Stephan and I
offer something constructive as well.

Pet has at least two output routines for dependency trees: one used by
the debugger, and one used by itsdb.  Pet 0.99.7 fixed a problem with
MWEs in the output routine used by the debugger, but not for itsdb.

The problem is with things like edge 125:

Debugger output:
                
 (695 ru-lexeme-infl-rule 0.00 7 9
   (125 dearu-v-cop-stem/reg-cop-stem-lex 0.00 7 9 []
     ("で" 0.00 7 8)
     ("ある" 0.00 8 9))))))))

ITSDB output, reformatted slightly:

 (695 ru-lexeme-infl-rule  7 9 
   (125 dearu-v-cop-stem 0 7 9 ("ある" 8 9))))))))

Only the lexeme for the key daughter is given.

Stephan produced a fix (given below) which outputs:

 (695 ru-lexeme-infl-rule  7 9 
   (125 dearu-v-cop-stem 0 7 9 ("で ある" 7 9))))))))

For what it's worth, there was also a problem with this in Stephan's
branch.  It appears to be a bug with a long pedigree.

-- 
Francis Bond  <www.kecl.ntt.co.jp/icl/mtg/members/bond/>
NTT Communication Science Laboratories | Machine Translation Research Group

P.S. The only change was to <tLexItem::tsdb_derivation>

New:
string
tLexItem::tsdb_derivation(int protocolversion)
{
    ostringstream res;

    res << "(" << _id << " " << _stem->printname()
        << " " << score() << " " << _start <<  " " << _end
        << " " << "(\"" << orth() << "\" " 
	<< _start << " " << _end << ")";

    return res.str();
}
Old:

string
tLexItem::tsdb_derivation(int protocolversion)
{
    ostringstream res;
  
    res << "(" << _id << " " << _stem->printname()
        << " " << score() << " " << _start <<  " " << _end
        << " " << _keydaughter->tsdb_derivation(protocolversion);

    return res.str();
}

-- 
Francis Bond  <www.kecl.ntt.co.jp/icl/mtg/members/bond/>
NTT Communication Science Laboratories | Machine Translation Research Group