[developers] problems with rmrs output using tsdb

[Sergio Roa]

Hello,

I am using the "test" version of delph-in since some weeks ago and the
svn  repository version  of pet  on a  linux.x86.64  architecture.  As
described  in   the  thread  called  "Downloading   problems"  in  the
developers  list, I also  have to  use the  "test" version  instead of
"latest",  although in  my  case a  lack  of a  .tgz  package was  not
reported  by  the  install  script,  but after  the  installation  the
$DELPHINHOME/lkb/src/.l8c4  directory was  missing and  I need  it for
using tsdb++.

Well, my problem now is  different, using this "test" revision. When I
use cheap with fsr preprocessing from the command line I get a correct
character  span  and xml  output  of rmrs.  However,  when  I use  the
following code in lkb using tsdb:

(setf tsdb::*pvm-cpus* (list  (tsdb::make-cpu :host \"${HOST}\" :class
'(:cheap    :${CPU})   :spawn   \"${CHEAP}\"    :options   '(\"-tsdb\"
\"-tok=fsr\"  \"-packing=15\"\"${GRAMMAR}/english.grm\") :name \"pet\"
:threshold 2)))

and after using  the export script provided with lkb,  i get an output
without character spans like this:

<rmrs cfrom='-1' cto='-1'I show you my work'1 @ 0 @ '>
<label vid='1'/>
<ep cfrom='-1' cto='-1'><gpred>prop-or-ques_m_rel</gpred><label vid='1'/><var 
sort='e' vid='2'/></ep>
<ep cfrom='-1' cto='-1'><gpred>pron_rel</gpred><label vid='6'/><var sort='x' 
vid='7'/></ep>
<ep cfrom='-1' cto='-1'><gpred>pronoun_q_rel</gpred><label vid='8'/><var 
sort='x' vid='7'/></ep>
<ep cfrom='-1' cto='-1'><realpred lemma='show' pos='v' sense='1'/><label 
vid='11'/><var sort='e' vid='2'/></ep>
<ep cfrom='-1' cto='-1'><gpred>pron_rel</gpred><label vid='14'/><var sort='x' 
vid='12'/></ep>
<ep cfrom='-1' cto='-1'><gpred>pronoun_q_rel</gpred><label vid='15'/><var 
sort='x' vid='12'/></ep>
<ep cfrom='-1' cto='-1'><gpred>def_explicit_q_rel</gpred><label vid='18'/><var 
sort='x' vid='13'/></ep>
<ep cfrom='-1' cto='-1'><gpred>poss_rel</gpred><label vid='21'/><var sort='e' 
vid='23'/></ep>
<ep cfrom='-1' cto='-1'><gpred>pronoun_q_rel</gpred><label vid='24'/><var 
sort='x' vid='22'/></ep>
<ep cfrom='-1' cto='-1'><gpred>pron_rel</gpred><label vid='27'/><var sort='x' 
vid='22'/></ep>
<ep cfrom='-1' cto='-1'><realpred lemma='work' pos='n' sense='1'/><label 
vid='10001'/><var sort='x' vid='13'/></ep>
[...]
</rmrs>

I am getting  the same output using the 'tok=yy"  option for cheap and
the lkb  preprocessor. It seems to  be an unparseable  xml output (see
first line -  I am not sure).  As I mentioned  above, when using cheap
from the command line I get the first lines as follows:

[...]
69132 types in 7.6 s
(1) `i show you my work' [0] --- 1 (0.03|0.03s) <21:172> (2060.3K) [0.0s]
derivation[1] (0.3977):i show you my work 

<rmrs cfrom='-1' cto='-1'>
<label vid='1'/>
<ep cfrom='0' cto='18'><gpred>prop-or-ques_m_rel</gpred><label vid='1'/><var 
sort='e' vid='2'/></ep>
<ep cfrom='0' cto='1'><gpred>pron_rel</gpred><label vid='6'/><var sort='x' 
vid='7'/></ep>
<ep cfrom='0' cto='1'><gpred>pronoun_q_rel</gpred><label vid='8'/><var 
sort='x' vid='7'/></ep>
<ep cfrom='2' cto='6'><realpred lemma='show' pos='v' sense='1'/><label 
vid='11'/><var sort='e' vid='2'/></ep>
<ep cfrom='7' cto='10'><gpred>pron_rel</gpred><label vid='14'/><var sort='x' 
vid='12'/></ep>
<ep cfrom='7' cto='10'><gpred>pronoun_q_rel</gpred><label vid='15'/><var 
sort='x' vid='12'/></ep>
<ep cfrom='11' cto='13'><gpred>def_explicit_q_rel</gpred><label vid='18'/><var 
sort='x' vid='13'/></ep>
<ep cfrom='11' cto='13'><gpred>poss_rel</gpred><label vid='10001'/><var 
sort='e' vid='22'/></ep>
<ep cfrom='11' cto='13'><gpred>pronoun_q_rel</gpred><label vid='23'/><var 
sort='x' vid='21'/></ep>
<ep cfrom='11' cto='13'><gpred>pron_rel</gpred><label vid='10002'/><var 
sort='x' vid='21'/></ep>
<ep cfrom='14' cto='18'><realpred lemma='work' pos='n' sense='1'/><label 
vid='10003'/><var sort='x' vid='13'/></ep>
[...]

Any help or ideas would be appreciated.

Cheers,

Sergio.


-- 
Sergio Roa Ovalle
http://www.informatik.uni-freiburg.de/~roa