Re[2]: [infomap-nlp-users] document x document, add

[Beate D. <do...@IM...>]

Hi Mich,

To answer your second question first:

Do you get the same error message when you list the absolute pathnames of 
your corpus files (relative to the root directory) in the file of 
filenames? If you also use the absolute pathname of the reference file in 
the "-m" option, infomap-build should be able to locate the files. The 
format of your reference file is just fine, it's one file name per line.

Let me know if you are still having problems building the model.
Good luck,
Beate


On Thu, 28 Oct 2004, Mich wrote:

> While I'm - actually again - at it, could i ask a very basic question?
> What exactly is the format for a multiple file list of corpora? Right
> now, i have a small amount of .txt files in a single directory, and
> according to the manual, another file should point towards the other
> files. I have done this by numbering the txt filenames as 1.txt 2.txt
> 3.txt etc, and have made a 'reference file' with all these filenames
> underneath another:
> 1.txt
> 2.txt
> 3.txt
> [..]
> 19.txt
> etc.
>
> However, infomap-build doesn't seem to recognize this, stopping with
> 'can't open current corpus file'
> and
> 'make *** [/home/jrandom/infomap-models/ned/wordlist] Error 1'
>
> so, i figured the text file probably has a different format from what
> i thought it would be. Actually, i had to guess, since the manual
> states that
>
> "In a multiple-file corpus, each disk file that is part of the corpus
> must contain exactly one document. No tags are used; the entire
> contents of the file are considered to make up the text of the
> document and are processed by the Infomap software."
>
> which really leaves me puzzled as to the exact specifications of the
> reference file. I have tried several alterations in my reference file,
> but infomap-build seems either to stop with mentioned error message,
> or continue and treat the reference file as a single corpus anyway. If
> someone would send an example of a multifile reference-file, i would
> be most pleased (as the one in the documentation seems lacking).
>
> Thank you kindly
>
> Mich
>
>
>
>
> -------------------------------------------------------
> This SF.Net email is sponsored by:
> Sybase ASE Linux Express Edition - download now for FREE
> LinuxWorld Reader's Choice Award Winner for best database on Linux.
> http://ads.osdn.com/?ad_id=5588&alloc_id=12065&op=click
> _______________________________________________
> infomap-nlp-users mailing list
> inf...@li...
> https://lists.sourceforge.net/lists/listinfo/infomap-nlp-users
>