This still look pretty strange. Could you send me your modified
makedata.sh file?

Just a few misc thoughts on makedata.sh - Whenever you run makedata.sh you
want to make sure that you don't have a LexSample directory present in
that directory
- you should rename any existing ones (if you want to keep them).  Your messages
below indicate that there is (perhaps?) already a LexSample directory present...

My hope is that you can go back to the original makedata.sh file - and
then after you
have created a LexSample directory with it, before you run it again,
delete or move
the LexSample directory before you try it again. I should probably
modify makedata.sh
to move an existing LexSample directory, to avoid any confusion on that point.

Anyway, it's good that the demo script is running, I hope you are able
to run the others
too.

Thanks!
Ted

On Tue, Apr 8, 2008 at 9:16 PM,  <wg2162@colu...> wrote:
> Hi Ted,
>
>  When I run makedata.sh, the error information is:
>  mv: overwrite `cool.a-test.xml'? y
>  mv: overwrite `day.n-test.xml'? y
>  mv: overwrite `facility.n-test.xml'? y
>  mv: overwrite `fine.a-test.xml'? y
>  mv: overwrite `free.a-test.xml'? y
>  mv: overwrite `grip.n-test.xml'? y
>  mv: overwrite `live.v-test.xml'? y
>  mv: overwrite `material.n-test.xml'?
>  mv: overwrite `mouth.n-test.xml'? y
>  mv: overwrite `natural.a-test.xml'? y
>  mv: overwrite `post.n-test.xml'? y
>  mv: overwrite `simple.a-test.xml'? y
>  ERROR(frequency.pl):
>         Source file <test-lexelts-test.xml> doesn't exist ...
>  ERROR(filter.pl):
>         Source file <test-lexelts-test.xml> doesn't exist...
>  mv: overwrite `train.v-test.xml'? y
>
>  But it doesn't influence the result, I guess, because I can use
> sc-toolkit.sh
>
>
>
>
>  Quoting Ted Pedersen <duluthted@gmai...>:
>
>
> > Hi, thanks very much for this feedback. I'm a little surprised that
> > you are having this particular error - could you run your script
> > without your fix and send me the error you get? I'm not able to
> > re-create it on my system.
> >
> > Thanks!
> > Ted
> >
> > On Fri, Apr 4, 2008 at 8:45 PM,  <wg2162@colu...> wrote:
> >
> > >
> > >
> > >  Thank you very much for the quick response!
> > >  So it does not matter if the results are different.
> > >
> > >  PS: When I run demo, it seems makedata.sh should be modified a little
> bit.
> > >  I inserted the sentences in line 49:
> > >    if($lexelt == "test-lexelts") then
> > >        echo "For directory test-lexelts, do nothing..";
> > >    else
> > >        ......
> > >    endif
> > >  Because there is no $lexelt-test.xml in directory test-lexelts and it
> will
> > > cause errors...
> > >
> > >
> > >
> > >
> > >
> > >  Quoting Ted Pedersen <tpederse@d.um...>:
> > >
> > >
> > > > Hi Weiwei,
> > > >
> > > > See responses inline...
> > > >
> > > >
> > > > >
> > > > >  (1) error information
> > > > >  The scripts in Testing/svd/svdpackout keep saying "Test errors".
> For
> > > > > exmaple, when running testA5.sh, the output is:
> > > > >  Test A5 for svdpackout.pl
> > > > >  Running las2
> > > > >  Running svdpackout.pl --rowonly --format f8.5 lav2 lao2 >
> > > test-A5.output
> > > > >  Test Error
> > > > >  When tested against test-A5.reqd
> > > > >  1,12c1,12
> > > > >  <  0.02330 0.78145
> > > > >  <  0.06605 0.99289
> > > > >  <  0.36126 0.00000
> > > > >  <  0.37638 0.43627
> > > > >  <  0.40459 0.00000
> > > > >  <  0.43953 0.06880
> > > > >  <  0.48444 0.17084
> > > > >  <  0.48444 0.17084
> > > > >  <  0.54986 0.00000
> > > > >  <  0.73770 0.09099
> > > > >  <  0.82344 0.00000
> > > > >  <  1.17799 0.00000
> > > > >  ---
> > > > >
> > > >
> > > > SVDPACKC seems to provide somewhat different results as we move from
> > > > system to system, so these test failures aren't too surprising. I
> > > > think right now the test cases don't really account for system
> > > > variation, so as long as las2 is producing output, things should be
> > > > ok. I'm in the process of modifying svdpackout.pl, and as a part of
> > > > that hope to resolve some of these issues with the test cases, but for
> > > > now this kind of variation is fairly normal (unfortunately).
> > > >
> > > >
> > > > >
> > > > > >  0.02330 0.00000
> > > > > >  0.05804 0.00000
> > > > > >  0.06605 0.00000
> > > > > >  0.36126 0.11493
> > > > > >  0.37638 0.00000
> > > > > >  0.40459 0.18044
> > > > > >  0.43953 0.00000
> > > > > >  0.48444 0.00000
> > > > > >  0.48444 0.00000
> > > > > >  0.54986 0.22522
> > > > > >  0.73770 0.00000
> > > > > >
> > > > >
> > > > >
> > > > >  (2) installation
> > > > >  I used CPAN to install all the perl package. Then I installed the
> > > CLUTO,
> > > > > SVD, Senseclusters. SVD is in /nlp/tools/SVDPACKC.
> > > > >  The environmental variables I set were:
> > > > >  export PERL5LIB=/usr/lib/perl5/site_perl/5.8.8:/usr/lib/perl5/5.8.8
> > > > >  # sense cluster
> > > > >  export HOMEDIR=/nlp/tools
> > > > >  export SVDPACK=$HOMEDIR/SVDPACKC
> > > > >  export CLUTO=$HOMEDIR/cluto-2.1.1
> > > > >  export MYCLUTO=$CLUTO/Linux
> > > > >  export SENSECLUSTERS=/usr
> > > > >  export NSP=/usr/lib/perl5/site_perl/5.8.8
> > > > >
> > > > >  PATH=$PATH:$HOME/bin:$CLUTO:$MYCLUTO:$SVDPACK:$SENSECLUSTERS
> > > > >
> > > >
> > > > How did you do the install, that is what sequence of commands did you
> > > > run to install? And then what was the message that SenseClusters
> > > > displayed after 'make install'?
> > > >
> > > > Also, what problem are you having? The fact that the svdpackout.pl
> > > > test is running suggests things are working at least in part...
> > > >
> > > > Thanks,
> > > > Ted
> > > >
> > > >
> > > >
> > > > >
> > > > >  Did I miss something? Thank you very much!
> > > > >
> > > > >  Regards,
> > > > >  Weiwei
> > > > >
> > > > >
> > > >
> > > >
> > > >
> > > > --
> > > > Ted Pedersen
> > > > http://www.d.umn.edu/~tpederse
> > > >
> > > >
> > >
> > >
> > >
> > >
> >
> >
> >
> > --
> > Ted Pedersen
> > http://www.d.umn.edu/~tpederse
> >
> >
>
>
>



--
Ted Pedersen
http://www.d.umn.edu/~tpederse

-------------------------------------------------------------------------
This SF.net email is sponsored by the 2008 JavaOne(SM) Conference
Don't miss this year's exciting event. There's still time to save $100.
Use priority code J8TL2D2.
http://ad.doubleclick.net/clk;198757673;13503038;p?http://java.sun.com/javaone
_______________________________________________
senseclusters-developers mailing list
senseclusters-developers@list...
https://lists.sourceforge.net/lists/listinfo/senseclusters-developers