Make a start at a Lucy backend.

Build.PL

@@ -85,7 +85,7 @@ my %yn_vars = map { $_ => 1 }
    qw(use_plucene enable_page_deletion navbar_on_home_page backlinks_in_title
       moderation_requires_password enable_node_image enable_common_categories
       enable_common_locales recent_changes_on_home_page use_leaflet
-      random_page_omits_locales random_page_omits_categories
+      random_page_omits_locales random_page_omits_categories use_lucy
       content_above_navbar_in_html show_gmap_in_node_display force_wgs84
       send_moderation_notifications read_only);
 
@@ -135,14 +135,15 @@ push @answers, { question => $dbtype_qu,
 
 my $install_directory; # used to suggest template paths
 my $use_plucene = 1; # keep track of this so we know what to put in prereqs
+my $use_lucy = 0;
 my $use_leaflet; # if true, we skip the GMaps questions
 my %gmaps_vars = map { $_ => 1 } qw( gmaps_api_key centre_long centre_lat
                                 default_gmaps_zoom default_gmaps_search_zoom );
 my $centre_lat = ''; # contains centre lat derived from Google Maps URL
 foreach my $var ( qw(
    dbname dbuser dbpass dbhost dbport script_name
    install_directory template_path custom_template_path script_url
-   custom_lib_path use_plucene indexing_directory enable_page_deletion
+   custom_lib_path use_lucy use_plucene indexing_directory enable_page_deletion
    admin_pass stylesheet_url site_name navbar_on_home_page
    recent_changes_on_home_page random_page_omits_locales
    random_page_omits_categories content_above_navbar_in_html home_name
@@ -178,15 +179,26 @@ foreach my $var ( qw(
         next;
     }
 
-    # We don't ask this for new installs as Search::InvertedIndex is
-    # deprecated
-    if ( $var eq "use_plucene" and $existing_config->$var == 1) {
-        print "Skipping question about plucene\n"
-            unless $skip_config;
-        push @answers, { question => $qu,
-                         variable => $var,
-                         value => 1 };
-        next;
+    # Don't ask about Plucene if we've said to use Lucy.  If we haven't said
+    # to use Lucy, ask about Plucene only for existing installs, since the old
+    # Search::InvertedIndex method is deprecated.
+    if ( $var eq "use_plucene" ) {
+        if ( $use_lucy ) {
+            print "Skipping question about Plucene - we're using Lucy.\n"
+                unless $skip_config;
+            push @answers, { question => $qu,
+                             variable => $var,
+                             value => 0 };
+            next;
+        } elsif ( $existing_config->$var == 1 ) {
+            print "Skipping question about Plucene - we'll use it by "
+                  . "default.\n"
+                unless $skip_config;
+            push @answers, { question => $qu,
+                             variable => $var,
+                             value => 1 };
+            next;
+        }
     }
 
     # If we're using Leaflet, we can skip the GMaps stuff.  Don't erase any
@@ -245,6 +257,7 @@ foreach my $var ( qw(
 
     # Keep track of chosen search method so we know what to put in prereqs.
     # From Module::Build docs: ->y_n returns a Perl boolean true or false.
+    $use_lucy = 1 if $var eq "use_lucy" and $val;
     $use_plucene = 1 if $var eq "use_plucene" and $val;
 
     # If they've just chosen to use Leaflet, we won't need to ask any of the

Changes

@@ -3,6 +3,10 @@
 More detailed changelogs can be found at
 <http://github.com/OpenGuides/OpenGuides>
 
+0.75
+        Added a Lucy backend for searching.  This is still experimental,
+          so we recommend most people stick with Plucene for now.
+
 0.74    19 July 2013
         Skip the correct number tests in t/210_json.t
 

INSTALL

@@ -103,6 +103,19 @@ to search here just as you would enter them in a 'use lib qw( ... );'
 in a Perl script.
 
 
+  "Do you want to use Lucy for searching?"
+
+This is currently experimental - don't say "yes" to this question unless
+you know what you're doing.
+
+If you are changing to Lucy from another search method (Plucene or
+Search::InvertedIndex), you will need to do two things:
+    - either delete your old indexes (they're just files in the index
+      directory) or use a different index directory
+    - reindex your entire wiki (see reindex.pl in the examples/
+      directory of this distribution)
+
+
   "Do you want to use Plucene for searching?"
 
 This question is no longer asked, but documentation here is retained for

lib/OpenGuides/Config.pm

@@ -12,7 +12,7 @@ use base qw( Class::Accessor );
 my @variables = qw(
    dbtype dbname dbuser dbpass dbport dbhost dbencoding
    script_name install_directory script_url
-   custom_lib_path use_plucene indexing_directory enable_page_deletion
+   custom_lib_path use_plucene use_lucy indexing_directory enable_page_deletion
    admin_pass stylesheet_url site_name navbar_on_home_page
    recent_changes_on_home_page random_page_omits_locales
    random_page_omits_categories content_above_navbar_in_html home_name
@@ -78,6 +78,7 @@ sub _init {
                      script_name => "wiki.cgi",
                      install_directory => "/usr/lib/cgi-bin/openguides/",
                      use_plucene => 1,
+                     use_lucy => 0,
                      indexing_directory => "/usr/lib/cgi-bin/openguides/indexes/",
                      enable_page_deletion => 0,
                      moderation_requires_password => 1,
@@ -163,6 +164,7 @@ sub _init {
         script_url => "What URL does the install directory map to?",
         custom_lib_path => "Do you want me to munge a custom lib path into the scripts?  If so, enter it here.  Separate path entries with whitespace.",
         use_plucene => "Do you want to use Plucene for searching? (recommended, but see Changes file before saying yes to this if you are upgrading)",
+        use_lucy => "Do you want to use Lucy for searching? (experimental)",
         indexing_directory => "What directory can I use to store indexes in for searching? ***NOTE*** This directory must exist and be writeable by the user that your script will run as.  See README for more on this.",
         enable_page_deletion => "Do you want to enable page deletion?",
         moderation_requires_password => "Is the admin password required for moderating pages?",
@@ -267,6 +269,8 @@ sub script_url {
 
 =item * use_plucene (default: true)
 
+=item * use_lucy (default: false)
+
 =item * indexing_directory (default: C</usr/lib/cgi-bin/openguides/indexes>)
 
 =item * enable_page_deletion (default: false)

lib/OpenGuides/Search.pm

@@ -340,6 +340,15 @@ sub run_text_search {
     my $self = shift;
     my $searchstr = $self->{search_string};
     my $wiki = $self->wiki;
+    my $config = $self->config;
+
+    if ( $config->use_lucy ) {
+        require OpenGuides::Search::Lucy;
+        my $lucy = OpenGuides::Search::Lucy->new( config => $config );
+        my %results = $lucy->run_text_search( search_string => $searchstr );
+        $self->{results} = \%results;
+        return $self;
+    }
 
     # Create parser to parse the search string.
     my $parser = Parse::RecDescent->new( q{

lib/OpenGuides/Search/Lucy.pm

@@ -0,0 +1,94 @@
+package OpenGuides::Search::Lucy;
+use strict;
+
+use Wiki::Toolkit::Search::Lucy;
+
+=head1 NAME
+
+OpenGuides::Search::Lucy - Run Lucy-backed text searches for OpenGuides.
+
+=head1 DESCRIPTION
+
+Does search stuff for OpenGuides.  Distributed and installed as part of
+the OpenGuides project, not intended for independent installation.
+This documentation is probably only useful to OpenGuides developers.
+
+=head1 SYNOPSIS
+
+  use OpenGuides::Config;
+  use OpenGuides::Search::Lucy;
+
+  my $config = OpenGuides::Config->new( file => "wiki.conf" );
+  my $search = OpenGuides::Search::Lucy->new( config => $config );
+  $search->run_text_search( search_string => "wombat defenestration" );
+
+=head1 METHODS
+
+=over 4
+
+=item B<new>
+
+  my $config = OpenGuides::Config->new( file => "wiki.conf" );
+  my $search = OpenGuides::Search::Lucy->new( config => $config );
+
+=cut
+
+sub new {
+    my ($class, %args) = @_;
+    my $config = $args{config};
+    my $searcher = Wiki::Toolkit::Search::Lucy->new(
+      path => $config->indexing_directory );
+
+    my $self = {
+      config => $config,
+      searcher => $searcher,
+    };
+
+    bless $self, $class;
+}
+
+=item B<run_text_search>
+
+  my $config = OpenGuides::Config->new( file => "wiki.conf" );
+  my $search = OpenGuides::Search::Lucy->new( config => $config );
+  $search->run_text_search( search_string => "wombat defenestration" );
+
+=cut
+
+sub run_text_search {
+    my ( $self, %args ) = @_;
+
+    # If there are commas in the search string, we're looking at an OR search.
+    my $str = $args{search_string};
+    my $and_or = ( $str =~ /,/ ) ? "OR" : "AND";
+
+    my %finds = $self->{searcher}->search_nodes( $str, $and_or );
+
+    # Package the finds in a way that OpenGuides::Search expects.
+    my %results = map { $_ => { name => $_, score => $finds{$_} } }
+                      keys %finds;
+    return %results;
+}
+
+=back
+
+=cut
+
+=head1 AUTHOR
+
+The OpenGuides Project (openguides-dev@lists.openguides.org)
+
+=head1 COPYRIGHT
+
+     Copyright (C) 2013 The OpenGuides Project.  All Rights Reserved.
+
+The OpenGuides distribution is free software; you can redistribute it
+and/or modify it under the same terms as Perl itself.
+
+=head1 SEE ALSO
+
+L<OpenGuides>
+
+=cut
+
+1;

lib/OpenGuides/Search/Lucy.pm~

@@ -0,0 +1,63 @@
+package OpenGuides::Search::Lucy;
+use strict;
+
+=head1 NAME
+
+OpenGuides::Search::Lucy - Run Lucy-backed text searches for OpenGuides.
+
+=head1 DESCRIPTION
+
+Does search stuff for OpenGuides.  Distributed and installed as part of
+the OpenGuides project, not intended for independent installation.
+This documentation is probably only useful to OpenGuides developers.
+
+=head1 SYNOPSIS
+
+  use OpenGuides::Config;
+  use OpenGuides::Search::Lucy;
+
+  my $config = OpenGuides::Config->new( file => "wiki.conf" );
+  OpenGuides::Search::Lucy->run_text_search(
+            config => $config,
+            search_string => "wombat defenestration",
+  );
+
+=head1 METHODS
+
+=over 4
+
+=item B<run_text_search>
+
+  my $config = OpenGuides::Config->new( file => "wiki.conf" );
+  OpenGuides::Search::Lucy->run_text_search(
+            config => $config,
+            search_string => "wombat defenestration",
+  );
+
+=cut
+
+sub run_text_search {
+}
+
+=back
+
+=cut
+
+=head1 AUTHOR
+
+The OpenGuides Project (openguides-dev@lists.openguides.org)
+
+=head1 COPYRIGHT
+
+     Copyright (C) 2013 The OpenGuides Project.  All Rights Reserved.
+
+The OpenGuides distribution is free software; you can redistribute it
+and/or modify it under the same terms as Perl itself.
+
+=head1 SEE ALSO
+
+L<OpenGuides>
+
+=cut
+
+1;

lib/OpenGuides/Test.pm

@@ -208,7 +208,8 @@ sub make_cgi_object {
 
   Openguides::Test::refresh_db();
 
-Unlink the existing SQLite database t/node.db and plucene indexes. Then create a new SQLite database t/node.db
+Unlink the existing SQLite database t/node.db and Plucene/Lucy indexes.
+Then create a new SQLite database t/node.db
 
 =cut
 sub refresh_db {

lib/OpenGuides/Utils.pm

@@ -60,7 +60,8 @@ dbname
 
 =item *
 
-indexing_directory - for the L<Search::InvertedIndex> or L<Plucene> files to go
+indexing_directory - for the L<Search::InvertedIndex>, L<Plucene>,
+or C<Lucy> files to go in
 
 =back
 
@@ -97,7 +98,11 @@ sub make_wiki_object {
 
     # Make search.
     my $search;
-    if ( $config->use_plucene
+    if ( $config->use_lucy ) {
+        require Wiki::Toolkit::Search::Lucy;
+        $search = Wiki::Toolkit::Search::Lucy->new(
+                      path => $config->indexing_directory );
+    } elsif ( $config->use_plucene
          && ( lc($config->use_plucene) eq "y"
               || $config->use_plucene == 1 )
        ) {

t/206_geo_data_search_form.t

@@ -10,9 +10,10 @@ if ( $@ ) {
     plan skip_all => "DBD::SQLite could not be used - no database to test with. ($error)";
 }
 
-eval { require Plucene; };
-if ( $@ ) {
-    plan skip_all => "Plucene not installed";
+my $have_lucy = eval { require Lucy; } ? 1 : 0;
+my $have_plucene = eval { require Plucene; } ? 1 : 0;
+unless ( $have_lucy || $have_plucene ) {
+    plan skip_all => "Neither Lucy nor Plucene is installed";
 }
 
 eval { require Test::HTML::Content; };
@@ -26,7 +27,11 @@ plan tests => 12;
     OpenGuides::Test::refresh_db();
 
 my $config = OpenGuides::Test->make_basic_config;
-$config->use_plucene( 1 );
+if ( $have_lucy ) {
+  $config->use_lucy ( 1 );
+} else {
+  $config->use_plucene( 1 );
+}
 
 # British National Grid guides should have os_x/os_y/os_dist search fields.
 my $guide = OpenGuides->new( config => $config );