infomap-nlp-users — General questions and discussion about how to use the software.

[infomap-nlp-users] Spam apology re. Generating new Releases

[Dominic W. <dwi...@cs...>]

My apologies, the previous message was meant to go to the developers list
not the users list. Sorry for cluttering everybody's inbox: however, if
you *do* have a better idea of how to get releases automatically generated
from cvs and successfully uploaded to SourceForge, please write to me!

(and cc. inf...@li...)

-Dominic

On Fri, 23 Apr 2004, Dominic Widdows wrote:

>
> Dear All,
>
> In reponse to a problem Beate had earlier with making new releases.
>
> The cvs tag, autoconf, automake, make dist process that's supposed to
> generate new distibutions automatically from the cvs information didn't
> work for me either.
>
> Instead I've started with the 2nd step of the process, chaning the
> configure file (in the base directory) so that the version tags are up to
> date. (I did this manually just editing configure in emacs and replacing
> 0.8.3 with 0.8.4, 0.8.4 with 0.8.5, etc.)
>
> Running "configure" and then "make dist" from here produces the correct
> tarball.
>
> You can then navigate through the tortuous sourceforge documentation from
> the project admin page, got to "File Releases", upload the new tarball via
> anaonymous ftp to "upload.sourceforge.net" (remember to set the type to
> "bin", I uploaded garbage the first time I tried yesterday!), fill in the
> release notes forms, hope that doing these steps in a different order
> doesn't erase the notes you've already typed in, and hope for the best.
>
> If anyone who is more SourceForge savvy can give me tips to make this
> process better(I'm still not sure I'm doing things right), pleae chime in.
>
> Best wishes,
> Dominic
>
>
>
> -------------------------------------------------------
> This SF.net email is sponsored by: The Robotic Monkeys at ThinkGeek
> For a limited time only, get FREE Ground shipping on all orders of $35
> or more. Hurry up and shop folks, this offer expires April 30th!
> http://www.thinkgeek.com/freeshipping/?cpg=12297
> _______________________________________________
> infomap-nlp-users mailing list
> inf...@li...
> https://lists.sourceforge.net/lists/listinfo/infomap-nlp-users
>

[infomap-nlp-users] Releasing new verions: how to get around

[Dominic W. <dwi...@cs...>]

Dear All,

In reponse to a problem Beate had earlier with making new releases.

The cvs tag, autoconf, automake, make dist process that's supposed to
generate new distibutions automatically from the cvs information didn't
work for me either.

Instead I've started with the 2nd step of the process, chaning the
configure file (in the base directory) so that the version tags are up to
date. (I did this manually just editing configure in emacs and replacing
0.8.3 with 0.8.4, 0.8.4 with 0.8.5, etc.)

Running "configure" and then "make dist" from here produces the correct
tarball.

You can then navigate through the tortuous sourceforge documentation from
the project admin page, got to "File Releases", upload the new tarball via
anaonymous ftp to "upload.sourceforge.net" (remember to set the type to
"bin", I uploaded garbage the first time I tried yesterday!), fill in the
release notes forms, hope that doing these steps in a different order
doesn't erase the notes you've already typed in, and hope for the best.

If anyone who is more SourceForge savvy can give me tips to make this
process better(I'm still not sure I'm doing things right), pleae chime in.

Best wishes,
Dominic

Re: [infomap-nlp-users] still print_doc problems ...

[Dominic W. <dwi...@cs...>]

Dear German,

This seems to have been a pathnames problem: adding the line

 model_data_dir = strcat( model_dir_parent, model_tag );

seems to have fixed it (and another ++argv; --argc; has prevented the
printing of trivial files at the beginning of the results.

I've released a version 0.8.5 with these fixes - please let me know if
this solves it for you.

Best wishes,
Dominic

On Fri, 23 Apr 2004, German Rigau wrote:

>
> Dear all,
>
> I'm using infomap-nlp-0.8.4 and I also have the print_doc problems when
> retrieving documents.
>
> [rigau@meaning infomap-nlp-0.8.4]$ associate -t -c kjbible god
> god:1.000000
> lord:0.670154
> disgrace:0.617541
> scoffers:0.567710
> deceivableness:0.553055
> ...
>
> [rigau@meaning infomap-nlp-0.8.4]$ associate -d -c kjbible god
> 3999484:0.800569
> 4300759:0.799777
> 4175212:0.796587
> 4179848:0.795709
> ...
>
> Everything seems to be in place. However, print_doc don't find the models:
>
> [rigau@meaning infomap-nlp-0.8.4]$ print_doc -c kjbible 3999484
> model_params.c: read_model_params: can't stat model params file: No such
> file or directory
> Can't read model params file "(null)/model_params.bin"
>
> [rigau@meaning infomap-nlp-0.8.4]$ print_doc -m
> /home/rigau/tools/infomap-nlp-0.8.4/models -c kjbible 3999484
> model_params.c: read_model_params: can't stat model params file: No such
> file or directory
> Can't read model params file "(null)/model_params.bin"
>
> Any suggestion? Thanks in advance,
>
> German
>
>
>
>
>
>
>
> -------------------------------------------------------
> This SF.net email is sponsored by: The Robotic Monkeys at ThinkGeek
> For a limited time only, get FREE Ground shipping on all orders of $35
> or more. Hurry up and shop folks, this offer expires April 30th!
> http://www.thinkgeek.com/freeshipping/?cpg=12297
> _______________________________________________
> infomap-nlp-users mailing list
> inf...@li...
> https://lists.sourceforge.net/lists/listinfo/infomap-nlp-users
>

[infomap-nlp-users] still print_doc problems ...

[German R. <ri...@ls...>]

Dear all,

I'm using infomap-nlp-0.8.4 and I also have the print_doc problems when
retrieving documents.

[rigau@meaning infomap-nlp-0.8.4]$ associate -t -c kjbible god
god:1.000000
lord:0.670154
disgrace:0.617541
scoffers:0.567710
deceivableness:0.553055
...

[rigau@meaning infomap-nlp-0.8.4]$ associate -d -c kjbible god
3999484:0.800569
4300759:0.799777
4175212:0.796587
4179848:0.795709
...

Everything seems to be in place. However, print_doc don't find the models:

[rigau@meaning infomap-nlp-0.8.4]$ print_doc -c kjbible 3999484
model_params.c: read_model_params: can't stat model params file: No such 
file or directory
Can't read model params file "(null)/model_params.bin"

[rigau@meaning infomap-nlp-0.8.4]$ print_doc -m 
/home/rigau/tools/infomap-nlp-0.8.4/models -c kjbible 3999484
model_params.c: read_model_params: can't stat model params file: No such 
file or directory
Can't read model params file "(null)/model_params.bin"

Any suggestion? Thanks in advance,

German

Re: [infomap-nlp-users] Collection size limit?

[Dominic W. <dwi...@cs...>]

Dear Massimo,

I recently built a model from New York Times newswire data which contains
some 170M words. Is that comparable to your 4M document collection?

I used 1000 column-labels and 75000 words - it took a while but it seems
to work pretty well. I've attached the params file I used in case it's
helpgul to you.

Best wishes and thanks for using our software.
Dominic

On Tue, 30 Mar 2004, Massimo Miccoli wrote:

> Dear Developpers.
>
> I'm have some question about your software (sorry for my english, I'm
> italian).
> Have Infomap-nlp a limit about document collection? If I want to try
> indexing about 4000.000 of docs on sigle file I nedd to modify defaul
> params?  What's the best setting for ROWS and COLUNMs for a large
> collection like above?
>
> Many thanks,
>
> Massimo
>
>
>
> -------------------------------------------------------
> This SF.Net email is sponsored by: IBM Linux Tutorials
> Free Linux tutorial presented by Daniel Robbins, President and CEO of
> GenToo technologies. Learn everything from fundamentals to system
> administration.http://ads.osdn.com/?ad_id=1470&alloc_id=3638&op=click
> _______________________________________________
> infomap-nlp-users mailing list
> inf...@li...
> https://lists.sourceforge.net/lists/listinfo/infomap-nlp-users
>

[infomap-nlp-users] Collection size limit?

[Massimo M. <mmi...@il...>]

Dear Developpers.

I'm have some question about your software (sorry for my english, I'm 
italian).
Have Infomap-nlp a limit about document collection? If I want to try 
indexing about 4000.000 of docs on sigle file I nedd to modify defaul 
params?  What's the best setting for ROWS and COLUNMs for a large 
collection like above?

Many thanks,

Massimo

[infomap-nlp-users] Question about retrieving documents

[Sophia K. <sop...@gm...>]

Dear Sirs,

I am interested in LSA, lexical semantics etc., so I found your software
very useful. Since I have some problems while using it, I would like to put
several questions.
I've tried this software using slightly more than 1,000 docs from Reuters
collection. Running associate with option -c (eg., associate -c my_model
year
), I got words which are the closest to a given one. Afterwards, I tried to
get an output as a list  of documents (using ?d, eg. associate -d -c
my_model
year). Unfortunately, I didn?t succeed. I went back to a step of building a
model and discovered that ?FEWER THAN EXPECTED SINGULAR VALUES ??I assume
some
problems have arisen when building a model...

Another thing I?d like to ask you about is whether you?re going to add
probabilistic LSA to this software.

Thank you very much in advance.
Sophia Katrenko
sop...@gm...

-- 
+++ NEU bei GMX und erstmalig in Deutschland: TÜV-geprüfter Virenschutz +++
100% Virenerkennung nach Wildlist. Infos: http://www.gmx.net/virenschutz

Re: [infomap-nlp-users] Infomap. Can I choose and feed "content-bearing words" to "count_wordvec"?

[Scott J. C. <ced...@cs...>]

Shuji,

   Thanks for taking up the use of infomap-nlp-users.  Hopefully the
   archives of this list will come to be a useful resource.

   Unfortunately, allowing specification of content-bearing words (or
   control over how they are selected) is not yet available.  I think
   adding a feature to allow a user-specified list of content-bearing
   words would be straightforward (see next message).  

   I'm backlogged at the moment and need to be looking for work, but
   I'm hoping to get to adding such a feature next week.  

                                                          Scott

On Fri, Mar 05, 2004 at 06:53:17PM -0800, Shuji Yamaguchi wrote:
> Hi InfoMap admin and users,
> 
> I wonder whether I could choose the "content-bearing words" myself and feed
> them into the pre-processing of InfoMap.
> The count_wordvec appears to be the program that does it. According to its
> man page, the content words are chosen from the ones in "ranking 50-1049".
> Are there any way to customize this by use of options and/or parameters?
> 
> Thank you for your support.
> Regards, Shuji
> 
> Shuji Yamaguchi, 
> Fellow, Reuters Digital Vision Program, CSLI, Stanford.
> 

[infomap-nlp-users] Infomap. Can I choose and feed "content-bearing words" to "count_wordvec"?

[Shuji Y. <yam...@ya...>]

Hi InfoMap admin and users,

I wonder whether I could choose the "content-bearing words" myself and feed
them into the pre-processing of InfoMap.
The count_wordvec appears to be the program that does it. According to its
man page, the content words are chosen from the ones in "ranking 50-1049".
Are there any way to customize this by use of options and/or parameters?

Thank you for your support.
Regards, Shuji

Shuji Yamaguchi, 
Fellow, Reuters Digital Vision Program, CSLI, Stanford.

[infomap-nlp-users] make data / Tutorial

[Scott J. C. <ced...@cs...>]

Dear Marco,

  The Tutorial is, unfortunately, outdated and in need of revision.
  Please use the User Manual instead.  I'm sorry for the confusion;
  I've just added prominent notices in the web documentation to the effect
  that the tutorial is outdated.

  Models were built using "make data" in older, project-internal
  versions of the software.  In the current version, the more
  conveninet "infomap-build" program is used instead.  (infomap-build
  is actually a wrapper around "make data" that knows where to find
  the appropriate Makefile and a default parameter file.)

  Please consult the User Manual and the infomap-build(1) manual page
  distributed with the software, and let us know if you have any
  further troubles.

  Please send future questions and problems to
  inf...@li..., so that they can be
  archived for other users' benefit.

  Thank you for using our software!

                                             Scott
-- 
Scott Cederberg
Researcher

Infomap Project 
Computational Semantics Lab
Center for the Study of Language and Information (CSLI)
Stanford University

http://infomap.stanford.edu/

Re: [infomap-nlp-users] print_doc problem

[Scott J. C. <ced...@cs...>]

Hello,

   print_doc has some bugs in the way it handles its command-line
   arguments.  I'm hoping to fix those in the next couple of days.

   In the mean time, you could try 
 
   print_doc -m /users1/gosse/Infomap/Models -c test2 863882

   You may want to subscribe to
   inf...@li... to receive announcements
   of future releases (e.g. bug fixes for print_doc).

   Thanks for using our software!

                                                       Scott


On Wed, Mar 03, 2004 at 03:50:37PM +0100, Bouma G. wrote:
> 
> Hi,
> 
> I tried to use the print_doc script which comes with the infomap
> software, but couldn't get it to work.
> 
> I made a small single file test corpus consisting of 10K documents
> (each doc consists of just a single sentence).
> 
> Associate works, and returns doc ids:
> 
> > associate -t -d -c test2 miljoen
> 863882:0.922600
> 793603:0.913477
> 751758:0.899828
> 252246:0.897580
> ....
> 
> However, if I do
> > print_doc -t -c test2 863882
> 
> nothing happens (a print_doc process is running forever it seems, but no output
> is produced).
> 
> If I try to use an `installed model', I get:
> 
> > print_doc  -c test 863882
> model_params.c: read_model_params: can't stat model params file: No such file or directory
> Can't read model params file "(null)/model_params.bin" 
> 
> I have set the relevant parameters:
> 
> INFOMAP_MODEL_DIR=/users1/gosse/Infomap/Models
> INFOMAP_MODEL_PATH=/users1/gosse/Infomap/Models
> INFOMAP_WORKING_DIR=/users1/gosse/Infomap/Working
> 
> I am using suse linux 9.
> 
> best,
> 
> Gosse.
> -- 
> Gosse Bouma, Informatiekunde, RUG, Postbus 716, 9700 AS Groningen
> go...@le...     tel. +31-50-3635937      fax  +31-50-3636855
> 
> 
> -------------------------------------------------------
> SF.Net is sponsored by: Speed Start Your Linux Apps Now.
> Build and deploy apps & Web services for Linux with
> a free DVD software kit from IBM. Click Now!
> http://ads.osdn.com/?ad_id=1356&alloc_id=3438&op=click
> _______________________________________________
> infomap-nlp-users mailing list
> inf...@li...
> https://lists.sourceforge.net/lists/listinfo/infomap-nlp-users

-- 
Scott Cederberg
Researcher

Infomap Project 
Computational Semantics Lab
Center for the Study of Language and Information (CSLI)
Stanford University

http://infomap.stanford.edu/

[infomap-nlp-users] print_doc problem

[Bouma G. <go...@le...>]

Hi,

I tried to use the print_doc script which comes with the infomap
software, but couldn't get it to work.

I made a small single file test corpus consisting of 10K documents
(each doc consists of just a single sentence).

Associate works, and returns doc ids:

> associate -t -d -c test2 miljoen
863882:0.922600
793603:0.913477
751758:0.899828
252246:0.897580
....

However, if I do
> print_doc -t -c test2 863882

nothing happens (a print_doc process is running forever it seems, but no output
is produced).

If I try to use an `installed model', I get:

> print_doc  -c test 863882
model_params.c: read_model_params: can't stat model params file: No such file or directory
Can't read model params file "(null)/model_params.bin" 

I have set the relevant parameters:

INFOMAP_MODEL_DIR=/users1/gosse/Infomap/Models
INFOMAP_MODEL_PATH=/users1/gosse/Infomap/Models
INFOMAP_WORKING_DIR=/users1/gosse/Infomap/Working

I am using suse linux 9.

best,

Gosse.
-- 
Gosse Bouma, Informatiekunde, RUG, Postbus 716, 9700 AS Groningen
go...@le...     tel. +31-50-3635937      fax  +31-50-3636855

[infomap-nlp-users] Dude, where's the software?

[Scott C. <ced...@st...>]

Um, where's the software?

(This note is to test the infomap-nlp-users list.  The software is coming=

soon.)

                                                        Scott


____________________________________________________________________
   =