jcb_perl has asked for the wisdom of the Perl Monks concerning the following question:

Hi, I am having trouble using Text::Levenshtein. Here's an example: 
use Text::Levenshtein qw(fastdistance distance);

print fastdistance("history","hustory"), "\n";
 # prints "2"

[download]
IMHO, this should print "1" because there is only one substitution necessary to get from "history" to "hustory", but instead it returns "2". Interestingly enough, this works as expected: 
print fastdistance("gistory","history"), "\n";
 # prints "1"

[download]
Any idea what I am doing wrong? Or do I have the wrong notion about how Levenshtein distance actually works?? I also tried using distance, but then something seriously goes wrong: 
print distance("gistory","history"), "\n";
 # prints "1234561234561"

[download]
I am using ActivePerl on Windows 7. The input file is ANSI. I am thankful for all hints!

Cheers and thanks, Christian

Replies are listed 'Best First'.
Re: Text::Levenshtein - trouble
by Corion (Patriarch) on Jul 11, 2012 at 14:51 UTC
[reply]

      Indeed, the module is buggy. The comment here helped unbreak a script I wrote using the module. The lines that need to be commented are:

      # in distance()
        foreach my $i (1 .. $n) {
            #if ($i != $n && substr($s,$i) eq substr($t,$i)) {
            #    push @result,$i;next;
            #}
            $d[$i][0]=$i;
        }
        foreach my $j (1 .. $m) {
            #if ($j != $m && substr($s,$j) eq substr($t,$j)) {
            #    push @result,$j;next;
            #}
            $d[0][$j]=$j;
        }

# and the two early returns in fastdistance()

      [download]
[reply]
[d/l]
Re: Text::Levenshtein - trouble
by Khen1950fx (Canon) on Jul 12, 2012 at 13:25 UTC
    Text::Levenshtein is returning in list context, but scalar context is what should be happening. You could force scalar context by using scalar. For example: 
    #!/usr/bin/perl -l

use autodie;
use strictures 1;
use Devel::SimpleTrace;

use strict qw/refs/;
use Text::LevenshteinXS qw/distance/;
    
my $x = scalar distance("history", "history");
my $y = scalar distance("gistory", "history");
my $z = scalar distance("history", "hustory");
print $x, "\n", $y, "\n", $z;

use strict qw/refs/;
use Text::Levenshtein::Damerau qw/edistance/;
    
$x = scalar edistance("history", "history");
$y = scalar edistance("gistory", "history");
$z = scalar edistance("history", "hustory");
print $x, "\n", $y, "\n", $z;

use strict qw/refs/;
use Text::Levenshtein;

$x = scalar distance("history", "history");
$y = scalar distance("gistory", "history");
$z = scalar distance("history", "hustory");
print $x, "\n", $y, "\n", $z;

    [download]
[reply]
[d/l]

Back to Seekers of Perl Wisdom