_row_id now include offset in dataset array
[MojoFacets.git] / lib / MojoFacets / Data.pm
index 1cc4507..7eccd2c 100644 (file)
@@ -11,6 +11,7 @@ use JSON;
 use Encode;
 use locale;
 use File::Find;
+use Storable;
 
 our $loaded;
 our $filters;
@@ -44,14 +45,38 @@ sub index {
        );
 }
 
-sub _load_path {
+sub _save {
        my ( $self, $path ) = @_;
 
-       return if defined $loaded->{$path}->{data};
+       my $name = $path;
+       my $dir = $self->app->home->rel_dir('data');
+       $name =~ s/^$dir//;
+       $name =~ s/\/+/_/g;
+       my $dump_path = '/tmp/mojo_facets.' . $name . '.storable';
+
+       warn "save loaded to $dump_path";
+       my $info = $loaded->{$path};
+       store $info, $dump_path;
+
+       # sync timestamp
+       my $mtime = $loaded->{$path}->{mtime};
+       utime $mtime, $mtime, $dump_path;
+
+       warn $dump_path, ' ', -s $dump_path, " bytes\n";
+}
+
+sub _load_path {
+       my ( $self, $path ) = @_;
 
        my $full_path = $self->app->home->rel_file( 'data/' . $path );
        die "$full_path $!" unless -r $full_path;
 
+       if ( defined $loaded->{$path}->{data} ) {
+               my $mtime = (stat($full_path))[9];
+               return if $loaded->{$path}->{mtime} == $mtime;
+               warn "reload $full_path, modified ", time() - $mtime, " seconds ago\n";
+       }
+
        # we could use Mojo::JSON here, but it's too slow
 #      $data = from_json read_file $path;
        my $data = read_file $full_path;
@@ -91,7 +116,9 @@ sub _load_path {
                                }
                        }
                        my $item;
-                       $item->{ $header[$_] || "f_$_" } = [ $v[$_] ] foreach ( 0 .. $#v );
+                       foreach my $i ( 0 .. $#v ) {
+                               $item->{ $header[$i] || "f_$i" } = [ $v[$i] ];
+                       }
                        push @{ $data->{items} }, $item;
                }
        } else {
@@ -115,7 +142,7 @@ sub _load_path {
                                $stats->{$n}->{numeric}++
                                        if $x =~ m/^[-+]?([0-9]*\.[0-9]+|[0-9]+)$/;
                                $stats->{$n}->{empty}++
-                                       if $x =~ m/^\s*$/;
+                                       if length $x == 0; # faster than $x =~ m/^\s*$/;
                        }
 
                }
@@ -127,6 +154,16 @@ sub _load_path {
                        if $stats->{$n}->{array} == $stats->{$n}->{count};
        }
 
+       if ( ! @header ) {
+               if ( defined $data->{header} ) {
+                       if ( ref $data->{header} eq 'ARRAY' ) {
+                               @header = @{ $data->{header} };
+                       } else {
+                               warn "header not array ", dump( $data->{header} );
+                       }
+               }
+       }
+
        @header =
                sort { $stats->{$b}->{count} <=> $stats->{$a}->{count} }
                grep { defined $stats->{$_}->{count} } keys %$stats
@@ -134,14 +171,18 @@ sub _load_path {
 
        warn dump($stats);
 
-       $loaded->{ $path } = {
+       my $info = {
                header => [ @header ],
                stats  => $stats,
                full_path => $full_path,
                size => -s $full_path,
+               mtime => (stat($full_path))[9],
                data => $data,
        };
 
+       $loaded->{ $path } = $info;
+       $self->_save( $path );
+
 }
 
 
@@ -160,20 +201,23 @@ sub load {
        $self->session( 'header' => $loaded->{$path}->{header} );
        if ( ! defined $loaded->{$path}->{columns} ) {
                $self->session( 'columns' => $loaded->{$path}->{header} );
+               $self->session( 'order'   => $loaded->{$path}->{header}->[0] );
                $self->redirect_to( '/data/columns' );
        } else {
                $self->session( 'columns' => $loaded->{$path}->{columns} );
+               $self->session( 'order'   => $loaded->{$path}->{columns}->[0] );
                $self->redirect_to( '/data/items' );
        }
-               
 }
 
 
 sub _loaded {
        my ( $self, $name ) = @_;
        my $path = $self->session('path');
-       die "$path $name doesn't exist in loaded ",dump( $loaded )
-               unless defined $loaded->{$path}->{$name};
+       if ( ! defined $loaded->{$path}->{$name} ) {
+               warn "$path $name doesn't exist in loaded ",dump( $loaded );
+               $self->redirect_to('/data/index');
+       }
        return $loaded->{$path}->{$name};
 }
 
@@ -228,7 +272,7 @@ sub _perm_array {
                        die "$name not array ",dump($session);
                }
        }
-       warn "# $name ",dump @array;
+       #warn "# $name ",dump @array;
        return @array;
 }
 
@@ -259,55 +303,56 @@ sub filter {
        my $name = $self->param('filter_name') || die "name?";
        my @vals = $self->param('filter_vals');
 
-#      warn "# filter $name vals ",dump(@vals);
+       $self->_remove_filter( $name );
+       $self->_filter_on_data( $name, @vals ) if @vals;
+
+       $self->session( 'offset' => 0 );
+
+       $self->redirect_to('/data/items');
+}
+
+sub _filter_on_data {
+       my ( $self, $name, @vals ) = @_;
 
        my $path = $self->session('path');
 
-       if ( @vals ) {
-               $filters->{$name} = [ @vals ];
-               warn "# filter + $name $#vals\n";
+       if ( ! defined $loaded->{$path}->{stats}->{ $name } ) {
+               warn "filter $name not found in data set";
+               return;
+       }
+
+       $filters->{$name} = [ @vals ];
+       warn "_filter_on_data $name ", $#vals + 1, " values on $path\n";
 
-               my $filter_hash;
-               $filter_hash->{$_}++ foreach @vals;
+       my $filter_hash;
+       $filter_hash->{$_}++ foreach @vals;
 
-               warn "# filter_hash ",dump( $filter_hash );
+       #warn "# filter_hash ",dump( $filter_hash );
 
-               my $items = $self->_loaded('data')->{items};
+       my $items = $self->_loaded('data')->{items};
 
-               my $include_missing = defined $filter_hash->{_missing};
-               my $filtered_items;
+       my $include_missing = defined $filter_hash->{_missing};
+       my $filtered_items;
 
-               foreach my $i ( 0 .. $#$items ) {
+       foreach my $i ( 0 .. $#$items ) {
 
-                       if ( defined $items->[$i]->{$name} ) {
-                               foreach my $v ( @{ $items->[$i]->{$name} } ) {
-                                       if ( defined $filter_hash->{ $v } ) {
-                                               $filtered_items->{$i}++;
-                                       }
+               if ( defined $items->[$i]->{$name} ) {
+                       foreach my $v ( @{ $items->[$i]->{$name} } ) {
+                               if ( defined $filter_hash->{ $v } ) {
+                                       $filtered_items->{$i}++;
                                }
-                       } elsif ( $include_missing ) {
-                               $filtered_items->{$i}++;
                        }
+               } elsif ( $include_missing ) {
+                       $filtered_items->{$i}++;
                }
-
-               warn "# filter $name ",dump($filtered_items);
-
-               $loaded->{$path}->{filters}->{$name} = $filtered_items;
-
-       } else {
-               warn "# filter - $name\n";
-               delete $filters->{$name};
-               delete $loaded->{$path}->{filters}->{$name};
        }
 
-       warn "# filters ",dump($filters);
+       #warn "# filter $name ",dump($filtered_items);
 
-       $self->session( 'offset' => 0 );
-
-       $self->redirect_to('/data/items');
+       $loaded->{$path}->{filters}->{$name} = $filtered_items;
+       warn "filter $name with ", scalar keys %$filtered_items, " items created\n";
 }
 
-
 sub _data_items {
        my ( $self, $all ) = @_;
        my $data = $self->_loaded( 'data' );
@@ -359,7 +404,7 @@ sub _current_filters {
                        grep { defined $filters->{ $_ } }
                        @{ $self->_loaded('header') }
                );
-       warn "# current_filters ",dump($current_filters);
+       #warn "# current_filters ",dump($current_filters);
        return $current_filters;
 }
 
@@ -368,6 +413,8 @@ sub _data_sorted_by {
 
        my $path = $self->session('path');
 
+       warn "_data_sorted_by $order from $path";
+
        if ( defined $loaded->{$path}->{sorted}->{$order} ) {
                return $loaded->{$path}->{sorted}->{$order};
        }
@@ -390,7 +437,8 @@ sub _data_sorted_by {
        } @{ $data->{items} }
        ;
 
-       warn "sorted $order"; # ,dump( @sorted );
+       warn "sorted: $order numeric: $numeric items: ", $#sorted + 1, "\n";
+       #warn "# sorted ",dump( @sorted );
 
        $loaded->{$path}->{sorted}->{$order} = [ @sorted ];
 }
@@ -415,31 +463,72 @@ sub items {
 
        my $sorted = $self->_data_sorted_by( $order );
 
-       my $path_filters = $loaded->{$path}->{filters};
-       my @filter_names = keys %$path_filters;
+       my @filter_names;
+       if ( $filters ) {
+               foreach my $name ( keys %$filters ) {
+                       if ( ! defined $loaded->{$path}->{stats}->{ $name } ) {
+                               warn "skip filter $name not found in $path\n";
+                               next;
+                       }
+                       push @filter_names, $name;
+               }
+               warn "filter_names ",dump( @filter_names );
+               foreach my $name ( @filter_names ) {
+                       next if ref $loaded->{$path}->{filters}->{$name} eq 'ARRAY';
+                       $self->_filter_on_data( $name, @{ $filters->{$name} } );
+               }
+       }
+
+       my $all_filters = join(' ',sort @filter_names,'order:',$order);
+
+#      warn "# all_filters $all_filters ", dump( $loaded->{$path}->{filtered}->{$all_filters} );
+
+       if ( ! defined $loaded->{$path}->{filtered}->{$all_filters} ) {
+
+               my $path_filters = $loaded->{$path}->{filters};
 
-       my @filtered;
-       foreach my $i ( 0 .. $#$sorted ) {
-               my $pos = $sort eq 'd' ? $sorted->[$i] : $sorted->[ $#$sorted - $i ];
+               warn "create combined filter for $all_filters\n";
 
-               my $skip = 0;
-               foreach ( @filter_names ) {
-                       $skip ||= 1 if ! defined $path_filters->{$_}->{$pos};
+               my @filtered;
+               foreach my $i ( 0 .. $#$sorted ) {
+                       my $pos = $sorted->[$i];
+
+                       if ( $#filter_names == -1 ) {
+                               push @filtered, $pos;
+                               next;
+                       }
+
+                       my $skip = 0;
+                       foreach ( @filter_names ) {
+                               $skip ||= 1 if ! defined $path_filters->{$_}->{$pos};
+                       }
+                       next if $skip;
+
+                       push @filtered, $pos;
                }
-               next if $skip;
 
-               push @filtered, $pos;
+               $loaded->{$path}->{filtered}->{$all_filters} = [ @filtered ];
        }
 
+       my $filtered = $loaded->{$path}->{filtered}->{$all_filters}
+               if defined $loaded->{$path}->{filtered}->{$all_filters};
+
+       warn "all_filters $all_filters produced ", $#$filtered + 1, " items\n" if $filtered;
+
        my $sorted_items;
        my $data = $self->_loaded('data');
-       foreach ( $offset .. $offset + $limit ) {
-               last unless defined $filtered[$_];
-               push @$sorted_items, $data->{items}->[ $filtered[$_] ];
+       my $from_end = $sort eq 'd' ? $#$filtered : 0;
+       foreach ( 0 .. $limit ) {
+               my $i = $_ + $offset;
+               last unless defined $filtered->[$i];
+               $i = $from_end - $i if $from_end;
+               my $id = $filtered->[$i];
+               push @$sorted_items,
+               my $item = $data->{items}->[ $id ];
+               $item->{_row_id} ||= $id;
        }
 
-
-#      warn "# sorted ", dump $sorted;
+       warn "# sorted_items ", $#$sorted_items + 1, " offset $offset limit $limit order $sort";
 
        $self->render(
                order => $order,
@@ -447,7 +536,7 @@ sub items {
                limit => $limit,
                sorted => $sorted_items,
                columns => [ @columns ],
-               rows => $#filtered + 1,
+               rows => $#$filtered + 1,
                numeric => { map { $_, $self->_is_numeric($_) } @columns },
                filters => $self->_current_filters,
        );
@@ -474,14 +563,32 @@ sub _is_numeric {
                $stats->{$name}->{numeric} > $count / 2;
 }
 
+sub _remove_filter {
+       my ($self,$name) = @_;
+       warn "_remove_filter $name\n";
+
+       my $path = $self->session('path');
+
+       delete $filters->{$name};
+       delete $loaded->{$path}->{filters}->{$name};
+       warn "filters left: ", keys %{ $loaded->{$path}->{filters} };
+
+       foreach (
+                       grep { /\b$name\b/ }
+                       keys %{ $loaded->{$path}->{filtered} }
+       ) {
+               delete $loaded->{$path}->{filtered}->{$_};
+               warn "remove filtered cache $_";
+       }
+}
+
 sub facet {
        my $self = shift;
 
        my $path = $self->session('path') || $self->redirect_to( '/data/index' );
 
-       if ( my $remove = $self->param('remove') ) {
-               delete $filters->{$remove};
-               delete $loaded->{$path}->{filters}->{$remove};
+       if ( my $name = $self->param('remove') ) {
+               $self->_remove_filter( $name );
                $self->redirect_to( '/data/items' );
        }
 
@@ -489,14 +596,28 @@ sub facet {
        my $name = $self->param('name') || die "no name";
 
        my $all = $self->_perm_scalar('all', 1);
+       my $data = $self->_loaded('data');
 
-       foreach my $i ( $self->_data_items($all) ) {
-               if ( ! exists $i->{$name} ) {
+       my $filters = $self->_current_filters;
+       my $all_filters = join(' ',sort keys %$filters,'order:',$self->session('order'));
+       my $filtered = $loaded->{$path}->{filtered}->{$all_filters}
+               if defined $loaded->{$path}->{filtered}->{$all_filters};
+
+       if ( ! $filtered || $all ) {
+               $filtered = [ 0 .. $#{ $data->{items} } ];
+               warn "filter all values\n";
+       } else {
+               warn "filter using $all_filters\n";
+       }
+
+       foreach my $i ( @$filtered ) {
+               my $item = $data->{items}->[$i];
+               if ( ! exists $item->{$name} ) {
                        $facet->{ _missing }++;
-               } elsif ( ref $i->{$name} eq 'ARRAY' ) {
-                       $facet->{$_}++ foreach @{ $i->{$name} };
+               } elsif ( ref $item->{$name} eq 'ARRAY' ) {
+                       $facet->{$_}++ foreach @{ $item->{$name} };
                } else {
-                       $facet->{ $i->{$name} }++;
+                       $facet->{ $item->{$name} }++;
                }
        }
 
@@ -510,22 +631,26 @@ sub facet {
 
        $checked = $self->_checked( @{ $filters->{$name} } ) if defined $filters->{$name};
 
-       my $sort = $self->param('sort') || 'c';
-
-       # sort facet numerically if more >50% elements are numeric
        my $numeric = $self->_is_numeric($name);
 
+       my $sort = $self->param('sort');
+       # sort numeric facets with more than 5 values ascending
+       $sort ||= $numeric && $#facet_names > 4 ? 'a' : 'c';
+
        @facet_names = sort {
-               if ( $sort =~ m/a/i ) {
-                       $numeric ? $a <=> $b : lc $a cmp lc $b;
-               } elsif ( $sort =~ m/d/i ) {
-                       $numeric ? $b <=> $a : lc $b cmp lc $a;
-               } elsif ( $sort =~ m/c/i ) {
-                       $facet->{$b} <=> $facet->{$a};
+               my $result;
+               if ( $sort eq 'a' ) {
+                       $result = $numeric ? $a <=> $b : lc $a cmp lc $b;
+               } elsif ( $sort eq 'd' ) {
+                       $result = $numeric ? $b <=> $a : lc $b cmp lc $a;
+               } elsif ( $sort eq 'c' ) {
+                       $result = ( $facet->{$b} || -1 ) <=> ( $facet->{$a} || -1 )
                } else {
                        warn "unknown sort: $sort";
-                       $a cmp $b;
+                       $result = $a cmp $b;
                }
+               $result = $a cmp $b unless defined $result; # FIXME cludge for numeric facets with invalid data
+               $result;
        } @facet_names;
 
        $self->render( name => $name, facet => $facet, checked => $checked,