Commit 3cccdc35 authored by Birte Kristina Friesel's avatar Birte Kristina Friesel
Browse files

use Text::Levenshtein(XS) for fuzzy station name matching

parent 4caa67e1

Build.PL

+1 −0
Original line number Diff line number Diff line
@@ -29,6 +29,7 @@ Module::Build->new( 
		'List::MoreUtils' => 0,

		'List::Util' => 0,

		'LWP::UserAgent' => 0,

		'Text::LevenshteinXS' => 0,

		'XML::LibXML' => 0,

	},

	sign => 1,

Changelog

+4 −0
Original line number Diff line number Diff line 
git HEAD



    * Result: Add info key 900

    * Station: Improve get_station matching quality by using the Levenshtein

      edit distance instead of simple substring matching

    * new dependency: Text::LevenshteinXS (see README for notes about

      drop-in replacements)



Travel::Status::DE::IRIS 1.02 - Tue May 26 2015

README

+14 −0
Original line number Diff line number Diff line
@@ -14,8 +14,22 @@ Dependencies 
* List::Compare

* List::MoreUtils

* LWP::UserAgent

* Text::LevenshteinXS

* XML::LibXML



Note about Text::LevenshteinXS: This module is old and unmaintained, but

appears to be packaged for slightly more distros than its successor

Text::Levenshtein::XS. If it is not available for your distro (and you do

not wish to build it), the following drop-in replacements are available:



* Text::Levenshtein::XS

* Text::Levenshtein (about 10 times slower than the XS modules)



To use them, simply run:

> sed -i 's/Text::LevenshteinXS/Text::Levenshtein::XS/g' Build.PL lib/Travel/Status/DE/IRIS/Stations.pm

or

> sed -i 's/Text::LevenshteinXS/Text::Levenshtein/g' Build.PL lib/Travel/Status/DE/IRIS/Stations.pm



Installation

------------

bin/db-iris

+2 −1
Original line number Diff line number Diff line
@@ -153,7 +153,8 @@ sub get_station { 
	else {

		say STDERR "The input '$input_name' is ambiguous. Please choose one "

		  . 'of the following:';

		say STDERR join( "\n", map { $_->[1] } @stations );

		say STDERR

		  join( "\n", map { $_->[1] . ' (' . $_->[0] . ')' } @stations );

		exit(1);

	}

}

lib/Travel/Status/DE/IRIS/Stations.pm

+22 −2
Original line number Diff line number Diff line
@@ -5,7 +5,11 @@ use warnings; 
use 5.014;

use utf8;



use List::MoreUtils qw(firstval);

use List::Util qw(min);

use List::MoreUtils qw(firstval pairwise);

use Text::LevenshteinXS qw(distance);



# TODO switch to Text::Levenshtein::XS once AUR/Debian packages become available



our $VERSION = '1.02';
@@ -15265,7 +15269,19 @@ sub get_station_by_name { 
		return ($actual_match);

	}



	return ( grep { $_->[1] =~ m{$name}i } @stations );

	my @distances   = map { distance( $nname, $_->[1] ) } @stations;

	my $min_dist    = min(@distances);

	my $minp1_dist  = min( grep { $_ != $min_dist } @distances );

	my @station_map = pairwise { [ $a, $b ] } @stations, @distances;



	# arbitrary selection: edit distance < 5 is probably a typo, >= 5

	# probably means the station does not exist / has an odd name

	if ( $min_dist < 5 ) {

		return map { $_->[0] } grep { $_->[1] == $min_dist } @station_map;

	}



	# always return a list when the edit distance is large

	return map { $_->[0] } grep { $_->[1] <= $minp1_dist } @station_map;

}



1;
@@ -15355,6 +15371,10 @@ None. 


=item * List::MoreUtils(3pm)



=item * List::Util(3pm)



=item * Text::LevenshteinXS(3pm)



=back



=head1 BUGS AND LIMITATIONS