Mike Schilli

NAME

Text::Language::Guess - Trained module to guess a document's language

SYNOPSIS

    use Text::Language::Guess;

    my $guesser = Text::Language::Guess->new();
    my $lang = $guesser->language_guess("bill.txt");

        # prints 'en'
    print "Best fit: $lang\n";

DESCRIPTION

Text::Language::Guess guesses a document's language. Its implementation is simple: Using Text::ExtractWords and Lingua::StopWords from CPAN, it determines how many of the known stopwords the document contains for each language supported by Lingua::StopWords.

Each word in the document recognized as stopword of a particular language scores one point for this language.

The language_guess() function takes a document as a parameter and returns the abbreviation of the language that it is most likely written in.

Supported Languages:

  • English (en)

  • French (fr)

  • Spanish (es)

  • Portugese (pt)

  • Italian (it)

  • German (de)

  • Dutch (nl)

  • Swedish (sv)

  • Norwegian (no)

  • Danish (da)

Methods

new()

Initializes the guesser with all stopwords available for all supported languges. If new has been called before, subsequent calls will return the same precomputed stoplist map, avoiding collecting all stopwords again (as long as the number of languages stays the same, see next paragraph).

You can limit the number of searched languages by specifying the language parameter and passing it an array ref of wanted languages:

        # Only guess between English and German
    $guesser = Text::Language::Guess->new(languages => ['en', 'de']);
language_guess($textfile)

Reads in a text file, extracts all words, scores them using the stopword maps and returns a single two-letter string indicating the language the document is most likely written in.

language_guess_string($string)

Just like language_guess, but takes a string instead of a file name.

scores($textfile)

Like language_guess($textfile), just returning a ref to a hash mapping language strings (e.g. 'en') to a score number. The entry with the highest score is the most likely one.

scores_string($string)

Like scores, but takes a string instead of a file name.

EXAMPLES

    use Text::Language::Guess;

        # Guess language in a string instead of a file
    my $guesser = Text::Language::Guess->new();
    my $lang = $guesser->language_guess_string("Make love not war");
        # 'en'


        # Limit number of languages to choose from
    my $guesser = Text::Language::Guess->new(languages => ['da', 'nl']);
    my $lang = $guesser->language_guess_string(
                   "Which is closer to English, danish or dutch?");
        # 'nl'


        # Show different scores
    my $guesser = Text::Language::Guess->new();
    my $scores = $guesser->scores_string(
        "This text is English, but other languages are scoring as well");
    use Data::Dumper;
    print Dumper($scores);

        # $VAR1 = {
        #   'pt' => 1,
        #   'en' => 6,
        #   'fr' => 1,
        #   'nl' => 1
        # };

LEGALESE

Copyright 2005 by Mike Schilli, all rights reserved. This program is free software, you can redistribute it and/or modify it under the same terms as Perl itself.

AUTHOR

2005, Mike Schilli <cpan@perlmeister.com>




Hosting generously
sponsored by Bytemark