weft qda users - Apr 2011 - [Weft QDA users] Shifty Markings - round 2

Good Afternoon,
 
I''m new to the mailing list and was wondering if anyone could help me
with my current headache.  I saw in the previous posts that someone else has had
the problem of shifty markings. I too am finding the text immediately above what
I wrote appearing in the coding reports, even though the coding in the documents
themselves remains as it should.  The project involves multiple files that have
become what I would call "massively huge."  I had reviewed the coding
part way through some of the files and the error was not there - is there any
way to get my work back or undo the shiftiness?  I''m quite desperate at
this point as I need to deliver coding reports to my advisor and
there''s a lot of information in the files that I need.  I''m
running v. 1.0.1 on a PC, but have access to a linux-based system if
there''s any way to undo the error there....
 
Any help would be greatly appreciated.
 
Laura 		 	   		  
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://rubyforge.org/pipermail/weft-qda-users/attachments/20110408/08262f9a/attachment-0001.html>

In my experience (also cross-platform) the shiftiness always seemed
to increase the extent of the marked text, which is not ideal but
can be worked around...

If you really want to get perfect text, you could try hitting
the database directly (the Weft info is just stored in a
SQLite2 database), and trimming it programmatically e.g; to
punctuation. But I found doing the extraction myself to be
sufficient, and allowed for other powerful post-proessing too.

Here''s the main magic of the script I used to extract the marked
text from the database for my thesis:

    Select code.catid, substr(doc.doctext, code.offset, code.length),
    code.offset, doc.docid from code, document doc where code.catid in
    (select catid from category where parent > -2) AND
    doc.docid=code.docid ORDER BY code.catid;

There are are a variety of free SQLite browsers available,
including sqlitebrowser2 on Linux. You should be able to paste
the above into the query tool for one and see your text, but
will have to do catid to catname separately if you have nested
categories, otherwise you could try a join.

For my nested categories I used the Perl code below (with EZDBI):

$r = Select(''catid, parent, catname from category where parent >
-2'');
my(%catsP, %catsN, %catsZ);
while( $_ = $r->([]) ){ $catsP{$_->[0]} = $_->[1]; $catsN{$_->[0]} =
$_->[2] }
if( $full ){
    %catsZ = %catsN;
}
else{
    foreach my $cat ( keys %catsP ){
        $catsZ{$cat} = join(''/'', map {$catsN{$_}} reverse
cat($cat)) }
}
sub cat{
    return if $_[0] == 1;
    return ($_[0] , cat($catsP{$_[0]}));
}

Thanks so much for the ideas!  I''ll try it out asap.

Laura
> To: weft-qda-users at rubyforge.org; glowin_glowworm at hotmail.com
> Subject: Re: [Weft QDA users] Shifty Markings - round 2
> Date: Fri, 8 Apr 2011 13:52:15 -0400
> From: belg4mit at pthbb.org
> 
> In my experience (also cross-platform) the shiftiness always seemed
> to increase the extent of the marked text, which is not ideal but
> can be worked around...
> 
> If you really want to get perfect text, you could try hitting
> the database directly (the Weft info is just stored in a
> SQLite2 database), and trimming it programmatically e.g; to
> punctuation. But I found doing the extraction myself to be
> sufficient, and allowed for other powerful post-proessing too.
> 
> Here''s the main magic of the script I used to extract the marked
> text from the database for my thesis:
> 
>     Select code.catid, substr(doc.doctext, code.offset, code.length),
>     code.offset, doc.docid from code, document doc where code.catid in
>     (select catid from category where parent > -2) AND
>     doc.docid=code.docid ORDER BY code.catid;
> 
> There are are a variety of free SQLite browsers available,
> including sqlitebrowser2 on Linux. You should be able to paste
> the above into the query tool for one and see your text, but
> will have to do catid to catname separately if you have nested
> categories, otherwise you could try a join.
> 
> For my nested categories I used the Perl code below (with EZDBI):
> 
> $r = Select(''catid, parent, catname from category where parent
> -2'');
> my(%catsP, %catsN, %catsZ);
> while( $_ = $r->([]) ){ $catsP{$_->[0]} = $_->[1];
$catsN{$_->[0]} = $_->[2] }
> if( $full ){
>     %catsZ = %catsN;
> }
> else{
>     foreach my $cat ( keys %catsP ){
>         $catsZ{$cat} = join(''/'', map {$catsN{$_}} reverse
cat($cat)) }
> }
> sub cat{
>     return if $_[0] == 1;
>     return ($_[0] , cat($catsP{$_[0]}));
> }
 		 	   		  
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://rubyforge.org/pipermail/weft-qda-users/attachments/20110409/d668e2ff/attachment.html>