Skip to content


Subversion checkout URL

You can clone with
Download ZIP
Branch: master
Fetching latest commit...
Cannot retrieve the latest commit at this time.



Text::Fingerprint - perform simple text clustering by key collision


version 0.006


    #!/usr/bin/env perl
    use common::sense;

    use Text::Fingerprint qw(:all);

    my $str = q(
        À noite, vovô Kowalsky vê o ímã cair no pé do pingüim
        queixoso e vovó põe açúcar no chá de tâmaras do jabuti feliz.

    say fingerprint($str);
    # a acucar cair cha de do e feliz ima jabuti kowalsky
    # no noite o pe pinguim poe queixoso tamaras ve vovo

    say fingerprint_ngram($str);
    # abacadaialamanarasbucachcudedoeaedeieleoetevfeg
    # uhaifiminiritixizjakokylilsmamqngnoocoeoiojokop
    # osovowpepipoqurarnsdsksotatetiucueuiutvevowaxoyv

    say fingerprint_ngram($str, 1);
    # abcdefghijklmnopqrstuvwxyz


Text clustering functions borrowed from the Google Refine. Can be useful for finding groups of different values that might be alternative representations of the same thing. For example, the two strings "New York" and "new york" are very likely to refer to the same concept and just have capitalization differences. Likewise, "Gödel" and "Godel" probably refer to the same person.



The process that generates the key from a $string value is the following (note that the order of these operations is significant):

  • normalize extended western characters to their ASCII representation (for example "gödel" → "godel")
  • change all characters to their lowercase representation
  • remove leading and trailing whitespace
  • split the string into punctuation, whitespace and control characters-separated tokens (using /[\W_]/ regexp)
  • sort the tokens and remove duplicates
  • join the tokens back together

fingerprint_ngram($string, $n)

The n-gram fingerprint method is similar to the fingerprint method described above but instead of using whitespace separated tokens, it uses n-grams, where the $n (or the size in chars of the token) can be specified by the user (default: 2). Algorithm steps:

  • normalize extended western characters to their ASCII representation
  • change all characters to their lowercase representation
  • remove all punctuation, whitespace, and control characters (using /[\W_]/ regexp)
  • obtain all the string n-grams
  • sort the n-grams and remove duplicates
  • join the sorted n-grams back together


Fingerprint functions are not exactly the same as those found in Google Refine! They were slightly changed to take advantage of the superb Perl handling of Unicode characters.



Stanislaw Pusep <>


This software is copyright (c) 2014 by Stanislaw Pusep.

This is free software; you can redistribute it and/or modify it under the same terms as the Perl 5 programming language system itself.

Something went wrong with that request. Please try again.