more specific schema desing for sorting

[BackupPC.git] / lib / BackupPC / Search / KinoSearch.pm

diff --git a/lib/BackupPC/Search/KinoSearch.pm b/lib/BackupPC/Search/KinoSearch.pm
index 3efcf8c..0301fa3 100644 (file)
--- a/lib/BackupPC/Search/KinoSearch.pm
+++ b/lib/BackupPC/Search/KinoSearch.pm
@@ -17,7 +17,10 @@ sub new {
 
        my $index_path = $Conf{KinoPath} || die "no KinoPath";
 
-       my $self = bless { index => $index_path }, $class;
+       my $self = bless {
+               index => $index_path,
+               first_time_indexing => ! -d $index_path,
+       }, $class;
        warn "# ",dump($self);
        return $self;
 }
@@ -40,22 +43,39 @@ sub indexer {
        );
        my $blob_type = KinoSearch::Plan::BlobType->new( stored => 1 );
        my $string_type = KinoSearch::Plan::StringType->new; # non-tokenized
-       my $num_type = KinoSearch::Plan::Int64Type->new;
+       my $num_type = KinoSearch::Plan::Int64Type->new( sortable => 1 );
+       my $sort_type = KinoSearch::Plan::StringType->new( sortable => 1 ); # non-tokenized
 
+       # numeric
        $schema->spec_field( name => $_, type => $string_type ) foreach ( qw/
                backup_date
                backupnum
-               date
                fid
                shareid
-               size
                type
        / );
 
+       # non-tokenized strings
        $schema->spec_field( name => $_, type => $string_type ) foreach ( qw/
-               _uri _file_path_split filepath hname sname
+               _uri
+               hname
+       /);
+
+       # sortable
+       $schema->spec_field( name => $_, type => $sort_type ) foreach (qw/
+               sname
+               filepath
+       /);
+
+       # sortable numeric
+       $schema->spec_field( name => $_, type => $sort_type ) foreach (qw/
+               date
+               size
        /);
 
+       # tokenized magic columns for infix search
+       $schema->spec_field( name => '_file_path_split', type => $ft_type );
+
 #      $schema->spec_field( name => '_doc', type => $blob_type );
 
        my $indexer = KinoSearch::Index::Indexer->new(
@@ -64,8 +84,6 @@ sub indexer {
                create => 1,
        );
 
-       $indexer->commit; # make sure that index exists
-
        warn "# created indexer";
 
        return $self->{_indexer} = $indexer;
@@ -83,8 +101,14 @@ sub searcher {
 sub exists {
        my ($self,$row) = @_;
 
+       return 0 if $self->{first_time_indexing};
+
        my $uri = $row->{hname} . ':' . $row->{sname} . '#' . $row->{backupnum} . ' ' . $row->{filepath};
        my $hits = $self->searcher->hits( query => "_uri:$uri" );
+
+
+       $self->{stat}->{exists}->{ $hits->total_hits }++;
+
        return $hits->total_hits;
 }
 
@@ -98,6 +122,8 @@ sub add_doc {
 
        warn "XXX ",dump($row) if $ENV{DEBUG};
 
+       $self->{stats}->{add_doc}++;
+
        $self->indexer->add_doc( $row );
 
 }
@@ -105,7 +131,7 @@ sub add_doc {
 sub commit {
        my $self = shift;
        $self->indexer->commit;
-       warn "# commit index";
+       warn "# commit index ", dump($self->{stats});
 }
 
 sub search {
@@ -120,7 +146,7 @@ sub search {
 
        my $sort_spec = KinoSearch::Search::SortSpec->new( rules => $rules );
 
-#      $q =~ s/(.)/$1 /g;
+       $q =~ s/(.)/$1 /g; # _file_path_split
        my $hits = $self->searcher->hits(
                query => $q,
                sort_spec => $sort_spec,