my $class = shift @_;
my %Conf = @_;
+ my $index_path = $Conf{KinoPath} || die "no KinoPath";
+
+ my $self = bless {
+ index => $index_path,
+ first_time_indexing => ! -d $index_path,
+ }, $class;
+ warn "# ",dump($self);
+ return $self;
+}
+
+sub indexer {
+ my $self = shift;
+ return $self->{_indexer} if defined $self->{_indexer};
+
my $schema = KinoSearch::Plan::Schema->new;
);
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
- $schema->spec_field( name => $_, type => $string_type ) foreach ( qw/
- backup_date
+=for numeric-no-padding
+ fid
+ shareid
+ type
backupnum
+=cut
+ $self->{numeric_padding} = [ qw/
+ backup_date
date
+ size
+ / ];
+
+ # numeric
+ $schema->spec_field( name => $_, type => $string_type ) foreach ( qw/
+ backup_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
/);
-# $schema->spec_field( name => '_doc', type => $blob_type );
+ # sortable
+ $schema->spec_field( name => $_, type => $sort_type ) foreach (qw/
+ sname
+ filepath
+ /);
- my $index_path = $Conf{KinoPath} || die "no KinoPath";
+ # sortable numeric
+ $schema->spec_field( name => $_, type => $sort_type ) foreach (qw/
+ backupnum
+ 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(
schema => $schema,
- index => $index_path,
+ index => $self->{index},
create => 1,
);
- warn "# using $index_path";
+ warn "# created indexer";
- $indexer->commit; # make sure that index exists
+ return $self->{_indexer} = $indexer;
- my $self = bless {
- indexer => $indexer,
- searcher => KinoSearch::Search::IndexSearcher->new(
- index => $index_path,
- ),
+};
- }, $class;
- return $self;
+our $searcher;
+sub searcher {
+ my $self = shift;
+ return $self->{_searcher} if $self->{_searcher};
+ $self->{_searcher} =
+ KinoSearch::Search::IndexSearcher->new( index => $self->{index} )
}
sub exists {
my ($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" );
+ my $hits = $self->searcher->hits( query => "_uri:$uri" );
+
+
+ $self->{stat}->{exists}->{ $hits->total_hits }++;
+
return $hits->total_hits;
}
$path =~ s/(.)/$1 /g; # XXX our tokenize
$row->{_file_path_split} = $path;
+ $self->{stats}->{add_doc}++;
+
+ foreach my $col ( @{ $self->{numeric_padding} } ) {
+ $row->{$col} = sprintf "%011d", $row->{$col};
+ }
+
warn "XXX ",dump($row) if $ENV{DEBUG};
- $self->{indexer}->add_doc( $row );
+ $self->indexer->add_doc( $row );
}
sub commit {
my $self = shift;
- $self->{indexer}->commit;
- warn "# commit index";
+ $self->indexer->commit;
+ warn "# commit index ", dump($self->{stats});
+}
+
+sub _field_lower_upper_term {
+ my ( $field, $l, $u ) = @_;
+ my $range;
+ if ( $l ) {
+ $range->{lower_term} = $l;
+ $range->{include_lower} = 1;
+ }
+ if ( $u ) {
+ $range->{upper_term} = $u;
+ $range->{include_upper} = 1;
+ }
+ $range->{field} = $field if $range;
+ warn "# $field $l - $u ",dump($range);
+ return $range;
+}
+
+sub search {
+ my ( $self, $offset, $on_page, $sort, $q, $shareid, $backup_from, $backup_to, $files_from, $files_to ) = @_;
+
+ warn "# search $offset/$on_page [$q] shareid: $shareid backup: $backup_from - $backup_to files: $files_from - $files_to";
+
+ my $sort_field = (split(/_/,$sort,2))[0];
+
+ my $rules = [ KinoSearch::Search::SortRule->new( type => 'score' ) ];
+ $rules->[0] = KinoSearch::Search::SortRule->new( field => $sort_field, reverse => $sort =~ m/_a$/ ? 0 : 1 ) if $sort_field;
+
+ my $sort_spec = KinoSearch::Search::SortSpec->new( rules => $rules );
+
+ my $split = $q;
+ $split =~ s/(.)/$1 /g; # _file_path_split
+ my $split_query = KinoSearch::Search::TermQuery->new( field => '_file_path_split', term => $split );
+#warn "XXX ",dump($split_query);
+
+
+ my $query_parser = KinoSearch::Search::QueryParser->new(
+ schema => $self->searcher->get_schema,
+ fields => ['_file_path_split'],
+ );
+ my $query = $query_parser->parse( '"' . $split . '"' );
+
+ my @and_query;
+
+ if ( $shareid ) {
+ push @and_query, KinoSearch::Search::TermQuery->new( field => 'shareid', term => $shareid );
+ }
+
+ if ( my $range = _field_lower_upper_term( 'backup_date', $backup_from, $backup_to ) ) {
+ push @and_query, KinoSearch::Search::RangeQuery->new( %$range );
+ }
+ if ( my $range = _field_lower_upper_term( 'date', $files_from, $files_to ) ) {
+ push @and_query, KinoSearch::Search::RangeQuery->new( %$range );
+ }
+
+ if ( @and_query ) {
+ push @and_query, $query;
+ $query = KinoSearch::Search::ANDQuery->new( children => [ @and_query ] );
+ }
+
+ my $hits = $self->searcher->hits(
+ query => m/:/ ? $q : $query,
+ offset => $offset,
+ num_wanted => $on_page,
+ sort_spec => $sort_spec,
+ );
+
+
+ warn "# ", $hits->total_hits, " hits for $q\n";
+
+ return (0,[]) if $hits->total_hits == 0;
+
+ my $results;
+ while ( my $hit = $hits->next ) {
+ warn "## hit = ",dump($hit) if $ENV{DEBUG};
+ push @$results, $hit;
+ }
+
+ return ( $hits->total_hits, $results );
}
1;
projects / BackupPC.git / blobdiff