added suggester to fix autocomplete

metacpan · Dec 5, 2016 · f4c7e13 · f4c7e13
1 parent 399132e
commit f4c7e13
Show file tree

Hide file tree

Showing 6 changed files with 87 additions and 34 deletions.
diff --git a/lib/MetaCPAN/Document/File.pm b/lib/MetaCPAN/Document/File.pm
@@ -363,6 +363,33 @@ sub _build_documentation {
     return undef;
 }
 
+=head2 suggest
+
+Autocomplete suggester data.
+
+=cut
+
+has suggest => (
+    is      => 'ro',
+    lazy    => 1,
+    builder => '_build_suggest',
+);
+
+sub _build_suggest {
+    my $self = shift;
+    my $doc  = $self->documentation;
+    return unless $doc;
+
+    my $weight = 1000 - length($doc);
+    $weight = 0 if $weight < 0;
+
+    return {
+        input   => [$doc],
+        payload => { doc_name => $doc },
+        weight  => $weight
+    };
+}
+
 =head2 indexed
 
 B<Default 0>

diff --git a/lib/MetaCPAN/Document/File/Set.pm b/lib/MetaCPAN/Document/File/Set.pm
@@ -2,6 +2,8 @@ package MetaCPAN::Document::File::Set;
 
 use Moose;
 
+use MetaCPAN::Util qw( single_valued_arrayref_to_scalar );
+
 extends 'ElasticSearchX::Model::Document::Set';
 
 my @ROGUE_DISTRIBUTIONS = qw(
@@ -442,37 +444,54 @@ sub autocomplete {
     my $query = join( q{ }, @terms );
     return $self unless $query;
 
-    return $self->search_type('dfs_query_then_fetch')->query(
+    my $suggestions
+        = $self->search_type('dfs_query_then_fetch')->es->suggest(
         {
-            filtered => {
-                query => {
-                    multi_match => {
-                        query    => $query,
-                        type     => 'most_fields',
-                        fields   => [ 'documentation', 'documentation.*' ],
-                        analyzer => 'camelcase',
-                        minimum_should_match => '80%'
-                    },
-                },
+            index => $self->index->name,
+            body  => {
+                documentation => {
+                    text       => $query,
+                    completion => {
+                        field => "suggest",
+                        size  => 50,
+                    }
+                }
+            },
+        }
+        );
+
+    my @docs
+        = map { $_->{text} } @{ $suggestions->{documentation}[0]{options} };
+
+    my $data = $self->es->search(
+        {
+            index => $self->index->name,
+            type  => 'file',
+            body  => {
+                query  => { match_all => {} },
                 filter => {
                     bool => {
                         must => [
-                            { exists => { field      => 'documentation' } },
-                            { term   => { status     => 'latest' } },
-                            { term   => { indexed    => 1 } },
-                            { term   => { authorized => 1 } }
-                        ],
-                        must_not => [
-                            {
-                                terms =>
-                                    { distribution => \@ROGUE_DISTRIBUTIONS }
-                            },
+                            { term  => { indexed             => 1 } },
+                            { term  => { authorized          => 1 } },
+                            { term  => { status              => 'latest' } },
+                            { terms => { 'documentation.raw' => \@docs } },
                         ],
                     }
                 }
-            }
+            },
+            fields => ['documentation'],
+            size   => 10
         }
-    )->sort( [ '_score', 'documentation' ] );
+    );
+
+    return +{
+        suggestions => [
+            sort { length($a) <=> length($b) || $a cmp $b }
+                map { $_->{fields}{documentation}[0] }
+                @{ $data->{hits}{hits} }
+        ]
+    };
 }
 
 __PACKAGE__->meta->make_immutable;

diff --git a/lib/MetaCPAN/Script/Mapping/Cpan/File.pm b/lib/MetaCPAN/Script/Mapping/Cpan/File.pm
@@ -7,6 +7,12 @@ sub mapping {
     '{
         "dynamic" : false,
         "properties" : {
+           "suggest": {
+              "type" : "completion",
+              "analyzer" : "simple",
+              "search_analyzer" : "simple",
+              "payloads" : true
+           },
            "pod" : {
               "index" : "no",
               "fields" : {
@@ -82,6 +88,10 @@ sub mapping {
                     "store" : true,
                     "type" : "string",
                     "analyzer" : "camelcase"
+                 },
+                 "raw" : {
+                    "type" : "string",
+                    "index" : "not_analyzed"
                  }
               },
               "type" : "string"
@@ -131,6 +141,10 @@ sub mapping {
                           "store" : true,
                           "type" : "string",
                           "analyzer" : "camelcase"
+                       },
+                       "raw" : {
+                          "type" : "string",
+                          "index" : "not_analyzed"
                        }
                     },
                     "type" : "string"

diff --git a/lib/MetaCPAN/Script/Release.pm b/lib/MetaCPAN/Script/Release.pm
@@ -255,7 +255,8 @@ sub import_archive {
             push( @provides, $_->name ) if $_->indexed && $_->authorized;
         }
         $file->clear_module if ( $file->is_pod_file );
-        $file->documentation;
+        $file->documentation;    # force build
+        $file->suggest;          # force build
         log_trace {"reindexing file $file->{path}"};
         $bulk->put($file);
         if ( !$document->has_abstract && $file->abstract ) {

diff --git a/lib/MetaCPAN/Server/Controller/Search/Autocomplete.pm b/lib/MetaCPAN/Server/Controller/Search/Autocomplete.pm
@@ -4,7 +4,6 @@ use strict;
 use warnings;
 
 use Moose;
-use MetaCPAN::Util qw( single_valued_arrayref_to_scalar );
 
 BEGIN { extends 'MetaCPAN::Server::Controller' }
 
@@ -17,12 +16,7 @@ sub get : Local : Path('') : Args(0) {
     my $model = $self->model($c);
     $model = $model->fields( [qw(documentation release author distribution)] )
         unless $model->fields;
-    my $data
-        = $model->autocomplete( $c->req->param("q") )->source(0)->raw->all;
-
-    single_valued_arrayref_to_scalar( $_->{fields} )
-        for @{ $data->{hits}{hits} };
-
+    my $data = $model->autocomplete( $c->req->param("q") );
     $c->stash($data);
 }
 

diff --git a/t/server/controller/search/autocomplete.t b/t/server/controller/search/autocomplete.t
@@ -13,9 +13,7 @@ test_psgi app, sub {
         ok( my $res = $cb->( GET '/search/autocomplete?q=Multiple::Modu' ),
             'GET' );
         my $json = decode_json_ok($res);
-
-        my $got = [ map { $_->{fields}{documentation} }
-                @{ $json->{hits}{hits} } ];
+        my $got  = $json->{suggestions};
 
         is_deeply $got, [
             qw(