1 package MojoFacets::Data;
6 use base 'Mojolicious::Controller';
8 #use Data::Dump qw(dump); # broken with Mojo::JSON, see https://rt.cpan.org/Public/Bug/Display.html?id=86592
11 sub dump { Dumper(@_) };
18 use Time::HiRes qw(time);
19 use File::Path qw(mkpath);
20 use Text::Unaccent::PurePerl;
22 use Statistics::Descriptive;
25 foreach my $module ( glob('lib/MojoFacets/Import/*.pm') ) {
26 $module =~ s{lib/(\w+)/(\w+)/(.*)\.pm}{$1::$2::$3};
28 die "$module: $!" if $!;
29 my ( $ext, $priority ) = $module->ext;
30 $imports->{$priority || 'file'}->{$ext} = $module;
31 warn "# import $ext $module\n";
34 warn "# import loaded ",dump( $imports );
37 my $full_path = shift;
39 # warn "# import_module $full_path\n";
41 return if $full_path =~ m/\.columns$/;
43 foreach my $ext ( keys %{ $imports->{file} } ) {
44 if ( -f $full_path && $full_path =~ m/$ext/i ) {
45 return $imports->{file}->{$ext};
50 foreach my $ext ( keys %{ $imports->{directory} } ) {
51 if ( -f $full_path && $full_path =~ m/$ext/i ) {
52 return $imports->{directory}->{$ext};
64 my $data_dir = $self->app->home->rel_file('data');
65 die "no data dir $data_dir" unless -d $data_dir;
71 my $file = $File::Find::name;
73 next if $file =~ m/.timefmt$/;
75 if ( -f $file && $file =~ m/([^\/]+)\.changes\/(\d+[\.,]\d+.+)/ ) {
76 push @{ $changes->{$1} }, $2
77 } elsif ( import_module( $file ) ) {
78 my $mtime = (stat($file))[9]; # mtime
79 $file =~ s/$data_dir\/*//;
81 $loaded->{$file}->{mtime} ||= $mtime;
83 #warn "IGNORE: $file\n";
87 no warnings qw(uninitialized); # mtime
88 @files = sort { $loaded->{$b}->{mtime} <=> $loaded->{$a}->{mtime} || lc $a cmp lc $b } @files,
89 grep { defined $loaded->{$_}->{generated} } keys %$loaded;
92 $size->{$_} = -s "$data_dir/$_" foreach @files;
99 dump_path => { map { $_ => $self->_dump_path($_) } @files },
105 my ( $self, $name ) = @_;
106 my $dir = $self->app->home->rel_file('data');
109 return '/tmp/mojo_facets.' . $name . '.storable';
113 my ( $self, $path ) = @_;
115 my $dump_path = $self->_dump_path( $path );
116 my $first_load = ! -e $dump_path;
117 warn "save loaded to $dump_path";
118 my $info = $loaded->{$path};
119 store $info, $dump_path;
122 my $mtime = $loaded->{$path}->{mtime};
123 utime $mtime, $mtime, $dump_path;
124 warn "sync time to $path at $mtime\n";
127 warn $dump_path, ' ', -s $dump_path, " bytes\n";
136 my $nr_items = $#{ $_[0] } + 1;
138 warn "__stats $nr_items\n";
140 # foreach my $e ( @{ $_[0] } ) {
141 foreach my $i ( 0 .. $#{$_[0]} ) {
142 print STDERR " $i" if $i % 5000;
144 foreach my $n ( keys %$e ) {
145 $stats->{$n}->{count}++;
147 if ( ref $e->{$n} eq 'ARRAY' ) {
148 $stats->{$n}->{array} += $#{ $e->{$n} } + 1;
154 foreach my $x ( @v ) {
155 if ( ! defined $x ) { # FIXME really null
156 $stats->{$n}->{empty}++;
160 $stats->{$n}->{numeric}++
161 if $x =~ m/^[-+]?([0-9]*\.[0-9]+|[0-9]+)$/;
162 $stats->{$n}->{empty}++
163 if length $x == 0; # faster than $x =~ m/^\s*$/;
169 foreach my $n ( keys %$stats ) {
170 my $s = $stats->{$n};
171 next unless defined $s->{array};
172 if ( $s->{array} == $s->{count} ) {
174 if ( $s->{count} == $nr_items ) {
175 warn "check $n for uniqeness\n";
177 foreach my $e ( @{ $_[0] } ) {
178 if ( ++$unique->{ $e->{$n}->[0] } == 2 ) {
184 $stats->{$n}->{unique} = 1;
185 #warn "# $n unique ",dump( $unique );
191 warn "# __stats ",dump($stats);
196 sub _param_or_session {
197 $_[0]->param( $_[1] ) || $_[0]->session( $_[1] )
202 my $path = $self->_param_or_session('path');
203 warn "stats $path\n";
204 delete $loaded->{$path}->{stats};
205 return $self->redirect_to( '/data/columns' );
210 my ( $self, $path ) = @_;
212 return if defined $loaded->{$path}->{generated};
214 my $full_path = $self->app->home->rel_file( 'data/' . $path );
215 return $self->redirect_to('/data/index') unless -r $full_path;
217 my $dump_path = $self->_dump_path( $path );
219 if ( defined $loaded->{$path}->{data} ) {
220 my $mtime = (stat($full_path))[9];
221 return if $loaded->{$path}->{mtime} == $mtime;
222 warn "reload $full_path, modified ", time() - $mtime, " seconds ago\n";
223 } elsif ( -e $dump_path ) {
224 warn "dump_path $dump_path ", -s $dump_path, " bytes loading...\n";
225 my $info = retrieve $dump_path;
226 $loaded->{ $path } = $info;
231 if ( my $module = import_module( $full_path ) ) {
232 $data = $module->new( full_path => $full_path )->data;
234 die "can't load $full_path";
239 if ( defined $data->{header} ) {
240 if ( ref $data->{header} eq 'ARRAY' ) {
241 @header = @{ $data->{header} };
243 warn "header not array ", dump( $data->{header} );
247 my $stats = __stats( $data->{items} );
250 sort { $stats->{$b}->{count} <=> $stats->{$a}->{count} }
251 grep { defined $stats->{$_}->{count} } keys %$stats
255 header => [ @header ],
257 full_path => $full_path,
258 size => -s $full_path,
259 mtime => (stat($full_path))[9],
261 defined $data->{generated} ? ( generated => 1 ) : (),
264 $loaded->{ $path } = $info;
265 $self->_save( $path ) unless $info->{generated};
274 my @paths = @{ $self->every_param('paths') };
275 warn "# paths ", dump @paths;
277 foreach my $p ( keys %$loaded ) {
278 next if grep { /^\Q$p\E$/ } @paths;
279 warn "remove $p from memory\n";
280 delete $loaded->{$p};
283 $self->_load_path( $_ ) foreach @paths;
285 my $path = $self->param('path') || $self->session('path') || $paths[0] || $self->redirect_to('/data/index');
287 warn "# path $path\n";
288 $self->_load_path( $path );
290 $self->session( 'path' => $path );
292 my $timefmt_path = $self->_permanent_path( 'timefmt' );
293 if ( -e $timefmt_path ) {
294 my $timefmt = read_file $timefmt_path;
295 $self->session( 'timefmt', $timefmt );
296 warn "timefmt = ", $timefmt;
299 my $redirect_to = '/data/items';
301 $self->session( 'header' => $loaded->{$path}->{header} );
302 if ( ! defined $loaded->{$path}->{columns} ) {
303 my $columns_path = $self->_permanent_path( 'columns' );
304 if ( -e $columns_path ) {
305 my @columns = map { s/[\r\n]+$//; $_ } read_file $columns_path, binmode => ':utf8';
306 $loaded->{$path}->{columns} = [ @columns ];
307 warn "# columns_path $columns_path ",dump(@columns);
309 $loaded->{$path}->{columns} = $loaded->{$path}->{header}
312 $redirect_to = '/data/columns';
314 $self->session( 'columns' => $loaded->{$path}->{columns} );
315 $self->session( 'order' => $loaded->{$path}->{columns}->[0] );
316 return $self->redirect_to( $redirect_to );
321 my ( $self, $name ) = @_;
322 my $path = $self->session('path') || $self->param('path');
323 return $self->redirect_to('/data/index') unless $path;
325 if ( defined $loaded->{$path}->{modified} && $loaded->{$path}->{modified} > 1 ) {
326 my $caller = (caller(1))[3];
327 if ( $caller =~ m/::edit/ ) {
328 warn "rebuild stats for $path ignored caller $caller\n";
330 warn "rebuild stats for $path FORCED by modified caller $caller\n";
331 # $loaded->{$path}->{stats} = __stats( $loaded->{$path}->{data}->{items} );
332 $loaded->{$path}->{rebuild_stats} = 1;
333 $loaded->{$path}->{modified} = 1;
337 if ( defined $loaded->{$path}->{rebuild_stats} ) {
338 warn "rebuild_stats $path";
339 $loaded->{$path}->{stats} = __stats( $loaded->{$path}->{data}->{items} );
340 delete $loaded->{$path}->{rebuild_stats};
343 if ( ! defined $loaded->{$path}->{$name} ) {
344 warn "$path $name isn't loaded\n";
345 $self->_load_path( $path );
346 if ( ! defined $loaded->{$path}->{stats} ) {
347 warn "rebuild stats for $path\n";
348 $loaded->{$path}->{stats} = __stats( $loaded->{$path}->{data}->{items} );
350 if ( ! defined $loaded->{$path}->{$name} ) {
351 warn "MISSING $name for $path\n";
352 return $self->redirect_to('/data/index')
356 $self->session( 'modified' => $loaded->{$path}->{modified} );
358 return $loaded->{$path}->{$name};
365 $checked->{$_}++ foreach @_;
366 # warn "# _checked ",dump($checked);
370 sub _permanent_path {
372 my $path = $self->_param_or_session('path');
373 $self->app->home->rel_file('data') . '/' . join('.', $path, @_);
378 $n = unac_string('utf-8',$n);
383 sub _column_from_unac {
384 my ($self,$name) = @_;
386 my $stats = $self->_loaded('stats');
387 my $cols_norm = { map { __unac( $_ ) => $_ } keys %$stats };
389 $cols_norm->{$name} || die "can't find column $name in ", dump($cols_norm);
394 my $path = $self->_param_or_session('path');
396 warn "no path in param or session";
399 my $dir = $self->app->home->rel_file('public') . "/export/$path";
400 mkpath $dir unless -e $dir;
401 my $name = __export_path_name( $path, @_ );
402 my $full = $dir . '/' . $name;
403 $full =~ s/\/+$// if -d $full; # strip trailing slash for dirs
407 sub __export_path_name {
411 my $name = join('.', map { __unac($_) } @_ );
412 if ( length($name) > $max_length ) {
413 $name = substr($name,0,$max_length) . Digest::MD5::md5_hex substr($name,$max_length);
421 my $view_path = $self->_permanent_path( 'views' );
423 if ( $self->param('columns') ) {
424 my @columns = $self->_param_array('columns');
425 write_file( $self->_permanent_path( 'columns' ), { binmode => ':utf8' }, map { "$_\n" } @columns );
426 if ( my $view = $self->param('view') ) {
427 mkdir $view_path unless -e $view_path;
428 write_file( "$view_path/$view", { binmode => ':utf8' }, map { "$_\n" } @columns );
429 warn "view $view_path/$view ", -s "$view_path/$view", " bytes\n";
432 return $self->redirect_to('/data/items');
434 } elsif ( ! $self->session('header') ) {
435 return $self->redirect_to('/');
436 return $self->redirect_to('/data/load');
439 if ( my $id = $self->param('id') ) {
440 my $view_full = "$view_path/$id";
441 if ( -f $view_full ) {
442 my @columns = map { chomp; $_ } read_file $view_full, binmode => ':utf8';
443 warn "view $view_full loaded ", dump @columns;
444 $self->session( 'columns' => [ @columns ] );
445 return $self->redirect_to('/data/items');
449 my $stats = $self->_loaded( 'stats' );
452 @columns = grep { defined $stats->{$_}->{count} } @{ $self->session('columns') } if $self->session('columns');
454 foreach my $c ( @{ $self->session( 'header' ) } ) {
455 push @columns, $c unless grep { /^\Q$c\E$/ } @columns;
459 if ( -d $view_path ) {
460 @views = map { s{^\Q$view_path\E/*}{}; $_ } glob "$view_path/*";
461 warn "# views ",dump @views;
465 message => 'Select columns to display',
467 columns => \@columns,
468 checked => $self->_checked( $self->_param_array('columns') ),
474 my ($self,$name) = @_;
476 my @array = @{ $self->every_param($name) };
477 my $path = $self->session('path');
480 $self->session($name => [ @array ]);
481 $loaded->{$path}->{$name} = [ @array ];
482 } elsif ( my $session = $self->session($name) ) {
483 if ( ref $session eq 'ARRAY' ) {
486 die "$name not array ",dump($session);
489 #warn "# $name ",dump @array;
494 my ($self,$name,$default) = @_;
496 my $scalar = $self->param($name);
498 if ( defined $scalar ) {
499 $self->session($name => $scalar);
501 $scalar = $self->session($name);
504 if ( ! defined $scalar ) {
506 if ( defined $scalar ) {
507 $self->session($name => $scalar);
509 warn "no default for $name";
513 warn "# _param_scalar $name ",dump $scalar;
520 my $name = $self->param('filter_name') || die "name?";
521 my @vals = @{ $self->every_param('filter_vals') };
523 $self->_remove_filter( $name );
525 $self->_filter_on_data( $name, @vals );
526 if ( my $permanent = $self->param('_permanent') ) {
527 my $permanent_path = $self->_export_path( 'filter', $name, $permanent );
528 write_file $permanent_path, { binmode => ':utf8' }, map { "$_\n" } @vals;
529 warn "permanent filter $permanent_path ", -s $permanent_path;
533 $self->session( 'offset' => 0 );
535 return $self->redirect_to('/data/items');
538 sub _filter_on_data {
539 my ( $self, $name, @vals ) = @_;
541 my $path = $self->session('path');
543 if ( ! defined $loaded->{$path}->{stats}->{ $name } ) {
544 warn "filter $name not found in data set";
548 $filters->{$name} = [ @vals ];
549 warn "_filter_on_data $name ", $#vals + 1, " values on $path\n";
552 $filter_hash->{$_}++ foreach @vals;
554 #warn "# filter_hash ",dump( $filter_hash );
556 my $items = $self->_loaded('data')->{items};
558 my $include_missing = defined $filter_hash->{_missing};
561 foreach my $i ( 0 .. $#$items ) {
563 if ( defined $items->[$i]->{$name} ) {
564 my $row = $items->[$i]->{$name};
565 $row = [ $row ] unless ref $row eq 'ARRAY'; # FIXME probably wrong place
566 foreach my $v ( @$row ) {
567 if ( defined $filter_hash->{ $v } ) {
568 $filtered_items->{$i}++;
571 } elsif ( $include_missing ) {
572 $filtered_items->{$i}++;
576 #warn "# filter $name ",dump($filtered_items);
578 # invalidate filters on other datasets
579 foreach my $dataset ( grep { exists $loaded->{$_}->{filters}->{$name} } keys %$loaded ) {
580 delete $loaded->{$dataset}->{filters}->{$name};
581 delete $loaded->{$dataset}->{filtered};
584 $loaded->{$path}->{filters}->{$name} = $filtered_items;
585 warn "filter $name with ", scalar keys %$filtered_items, " items created\n";
589 sub _current_filters {
592 my $stats = $self->_loaded('stats');
594 $current_filters->{ $_ } = $filters->{ $_ }
595 foreach ( grep { defined $filters->{ $_ } } keys %$stats )
597 warn "# _current_filters ",dump( keys %$current_filters );
598 return $current_filters;
601 sub _data_sorted_by {
602 my ( $self, $order ) = @_;
604 my $path = $self->session('path');
606 warn "_data_sorted_by $order from $path";
608 if ( defined $loaded->{$path}->{sorted}->{$order} ) {
609 return $loaded->{$path}->{sorted}->{$order};
612 my $data = $self->_loaded( 'data' );
613 my $numeric = $self->_is_numeric($order);
614 my $missing = $numeric ? 0 : '';
615 no warnings qw(numeric);
627 if ( ! exists $_->{$order} ) {
629 } elsif ( ref $_->{$order} eq 'ARRAY' ) {
630 $v = join('', @{$_->{$order}});
635 } @{ $data->{items} }
638 warn "sorted: $order numeric: $numeric items: ", $#sorted + 1, "\n";
639 #warn "# sorted ",dump( @sorted );
641 $loaded->{$path}->{sorted}->{$order} = [ @sorted ];
647 join(',', sort(@_), 'order', $order);
650 our ($out, $key,$value);
652 our $lookup_path_col;
656 warn "__commit_begin";
657 $lookup_path_col = undef;
663 $lookup_path_col = undef; # cleanup memory
668 warn "# lookup ",dump @_;
669 my ( $vals, $on_path, $on_col, $code, $stat_code ) = @_;
670 die "code is not sub{ ... } but ", dump $code unless ref $code eq 'CODE';
672 if ( ! exists $loaded->{$on_path} ) {
673 my @possible_paths = grep { /\Q$on_path\E/ } keys %$loaded;
674 die "more than one dataset available for '$on_path' ",dump @possible_paths if $#possible_paths > 0;
675 $on_path = shift @possible_paths;
676 warn "## fuzzy selected path $on_path";
679 my $items = $loaded->{$on_path}->{data}->{items} || die "no items for $on_path";
681 if ( ! exists $lookup_path_col->{$on_path}->{$on_col} ) {
682 warn "create lookup_path_col $on_path $on_col";
683 foreach my $i ( 0 .. $#$items ) {
684 my $item = $items->[$i];
685 if ( exists $item->{$on_col} ) {
686 if ( ref $item->{$on_col} eq 'ARRAY' ) {
687 foreach my $v ( @{ $item->{$on_col} } ) {
688 push @{ $lookup_path_col->{$on_path}->{$on_col}->{$v} }, $i;
690 } elsif ( ! ref $item->{$on_col} ) { # scalar
691 my $v = $item->{$on_col};
692 push @{ $lookup_path_col->{$on_path}->{$on_col}->{$v} }, $i;
694 die "unknown type of ",dump $item->{$on_col};
698 warn "XXX ",dump $lookup_path_col->{$on_path}->{$on_col} if $ENV{DEBUG};
702 $stat = Statistics::Descriptive::Full->new() if $stat_code;
704 foreach my $v ( ref $vals eq 'ARRAY' ? @$vals : ( $vals ) ) {
705 foreach my $i ( @{ $lookup_path_col->{$on_path}->{$on_col}->{$v} } ) {
707 warn "XXX lookup code $v $i ",dump $on if $ENV{DEBUG};
710 $stat_code->( $stat ) if $stat_code;
714 sub __commit_path_code {
715 my ( $path, $i, $code, $commit_changed ) = @_;
717 my $items = $loaded->{$path}->{data}->{items} || die "no items for $path";
718 my $row = $items->[$i];
721 foreach ( keys %$update ) {
722 $$commit_changed->{$_}++;
723 $loaded->{$path}->{data}->{items}->[$i]->{$_} = $update->{$_};
725 #warn "__commit_path_code $path $i ",dump( $update );
728 # uses templates/admin.html.ep
729 sub _switch_dataset {
734 foreach my $path ( keys %$loaded ) {
735 next unless exists $loaded->{$path}->{data};
736 push @$datasets, $path;
739 warn "# datasets ",dump($datasets);
741 $self->stash( 'datasets' => $datasets );
747 $self->_switch_dataset;
749 if ( my $timefmt = $self->param('timefmt') ) {
750 $self->session('timefmt', $timefmt);
751 warn "session store timefmt $timefmt\n";
752 my $timefmt_path = $self->_permanent_path( 'timefmt' );
753 write_file $timefmt_path, $timefmt;
754 warn "## $timefmt_path $timefmt"
757 if ( my $show = $self->param('id') ) {
758 $self->param('show', $show);
762 my $path = $self->_param_scalar('path');
764 my @columns = $self->_param_array('columns');
765 return $self->redirect_to('/data/columns') unless @columns;
766 my $order = $self->_param_scalar('order', $columns[0]);
767 my $sort = $self->_param_scalar('sort', 'a');
768 my $offset = $self->_param_scalar('offset', 0);
769 my $limit = $self->_param_scalar('limit', 20);
770 $self->_param_scalar('show', 'table');
772 # fix offset when changing limit
773 $offset = int( $offset / $limit ) * $limit;
775 if ( ! grep { /^\Q$order\E$/ } @columns ) {
776 $order = $columns[0];
777 $self->session( order => $order );
779 my $sorted = $self->_data_sorted_by( $order );
783 foreach my $name ( keys %$filters ) {
784 if ( ! defined $loaded->{$path}->{stats}->{ $name } ) {
785 warn "skip filter $name not found in $path\n";
788 push @filter_names, $name;
790 warn "filter_names ",dump( @filter_names );
791 foreach my $name ( @filter_names ) {
792 next if ref $loaded->{$path}->{filters}->{$name} eq 'ARRAY';
793 $self->_filter_on_data( $name, @{ $filters->{$name} } );
797 my $all_filters = __all_filters( @filter_names,$order );
799 # warn "# all_filters $all_filters ", dump( $loaded->{$path}->{filtered}->{$all_filters} );
801 if ( ! defined $loaded->{$path}->{filtered}->{$all_filters} ) {
803 my $path_filters = $loaded->{$path}->{filters};
805 warn "create combined filter for $all_filters from ", $#$sorted + 1, " items\n";
808 foreach my $i ( 0 .. $#$sorted ) {
809 my $pos = $sorted->[$i];
811 if ( $#filter_names == -1 ) {
812 push @filtered, $pos;
817 foreach ( @filter_names ) {
818 $skip ||= 1 if ! defined $path_filters->{$_}->{$pos};
822 push @filtered, $pos;
825 $loaded->{$path}->{filtered}->{$all_filters} = [ @filtered ];
828 my $filtered = $loaded->{$path}->{filtered}->{$all_filters}
829 if defined $loaded->{$path}->{filtered}->{$all_filters};
831 warn "all_filters $all_filters produced ", $#$filtered + 1, " items\n" if $filtered;
833 my $data = $self->_loaded('data');
835 my $code = $self->_param_scalar('code','');
837 $code =~ s{\n+$}{\n}s;
839 # XXX convert @row->{foo} into @{$row->{foo}}
840 $code =~ s|\@(row->\{[^}]+\})|\@{\$$1}|gs;
842 my $commit = $self->param('commit');
843 my $test = $self->param('test');
848 if ( $code && ( $test || $commit ) ) {
849 # XXX find columns used in code snippet and show them to user
851 foreach my $column ( $code =~ m/\$row->\{([^}]+)\}/g ) {
852 if ( $column =~ s/^(['"])// ) {
855 next if $column =~ m/\$/; # hide columns with vars in them
856 $commit_changed->{$column} = 0;
860 my $code_path = $self->app->home->rel_file('public') . "/code";
863 __path_modified( $path, 'commit' );
865 warn "# commit on ", $#$filtered + 1, " items:\n$code\n";
866 ( $key, $value, $out ) = ( 'key', 'value' );
867 foreach ( 0 .. $#$filtered ) {
868 my $i = $filtered->[$_];
869 __commit_path_code( $path, $i, $code, \$commit_changed );
872 # this might move before $out to recalculate stats on source dataset?
873 __path_rebuild_stats( $path );
874 my $c = { map { $_ => 1 } @columns };
875 my @added_columns = sort grep { ! $c->{$_} } keys %$commit_changed;
876 warn "# added_columns ",dump( @added_columns );
877 unshift @columns, @added_columns;
879 $loaded->{$path}->{columns} = [ @columns ];
880 warn "# new columns ",dump( @columns );
882 __invalidate_path_column( $path, $_ ) foreach keys %$commit_changed;
884 $self->_save_change({
886 time => $self->param('time') || time(),
887 user => $self->param('user') || $ENV{'LOGNAME'},
889 commit_changed => $commit_changed,
892 if ( my $description = $self->param('code_description') ) {
893 my $depends = $self->param('code_depends') || die "no code_depends?";
894 my $path = "$code_path/$depends.$description.pl";
895 if ( -e $path && ! $self->param('overwrite') ) {
896 warn "# code $path not saved\n";
898 write_file( $path, { binmode => ':utf8' }, "$code\n" );
899 warn "code $path ", -s $path, " bytes saved\n";
906 my $commit_dataset = join('.'
907 , $self->param('code_depends')
908 , $self->param('code_description')
913 warn "key $key value $value";
915 foreach my $n ( keys %$out ) {
916 my $i = { $key => [ $n ] };
917 my $ref = ref $out->{$n};
918 if ( $ref eq 'HASH' ) {
919 $i->{$_} = [ $out->{$n}->{$_} ] foreach keys %{ $out->{$n} };
920 } elsif ( $ref eq 'ARRAY' ) {
921 $i->{$_} = $out->{$n};
923 $i->{$value} = [ $out->{$n} ];
925 $i->{_error} = [ dump($out->{$n}) ];
930 my $stats = __stats( $items );
931 my @columns = grep { ! m/^\Q$key\E$/ } sort keys %$stats;
932 unshift @columns, $key;
934 $loaded->{$commit_dataset} = {
935 header => [ @columns ],
936 columns => [ @columns ],
938 data => { items => $items },
942 warn "# loaded out ", dump( $loaded->{$commit_dataset} );
943 $self->session('path', $commit_dataset);
944 $self->session('columns', [ @columns ]);
945 $self->session('order', $key);
946 return $self->redirect_to('/data/items');
949 $self->session('columns', [ @columns ]);
953 my $from_end = $sort eq 'd' ? $#$filtered : 0;
955 my ( $key, $value, $out ) = ( 'key', 'value' ); # XXX make local
956 foreach ( 0 .. $limit ) {
957 my $i = $_ + $offset;
958 last unless defined $filtered->[$i];
959 $i = $from_end - $i if $from_end;
960 my $id = $filtered->[$i];
961 my $row = Storable::dclone $data->{items}->[ $id ];
962 if ( $code && $test ) {
966 warn "ERROR evaling $@", dump($code);
967 $self->stash('eval_error', $@) if $@;
969 warn "EVAL ",dump($update);
970 foreach ( keys %$update ) {
971 $test_changed->{$_}++;
972 $row->{$_} = $update->{$_};
976 $row->{_row_id} ||= $id;
977 push @$sorted_items, $row;
980 if ( $self->param('export') ) {
981 my $export_path = $self->_export_path( 'items', @columns);
982 open(my $fh, '>', $export_path) || die "ERROR: can't open $export_path: $!";
983 print $fh "#",join("\t",@columns),"\n";
984 foreach my $f ( 0 .. $#$filtered ) {
985 print $fh join("\t", map {
986 my $i = $data->{items}->[ $filtered->[$f] ];
988 if ( ! defined $i->{$_} ) {
990 } elsif ( ref $i->{$_} eq 'ARRAY' ) {
991 $v =join(',', @{ $i->{$_} });
992 $v = '\N' if length($v) == 0;
993 } elsif ( ! ref $i->{$_} ) {
1002 warn "export $export_path ", -s $export_path, " bytes\n";
1005 my ( $code_depends, $code_description );
1009 warn "# test_changed ",dump( $test_changed );
1010 my $c = { map { $_ => 1 } @columns };
1011 my @added_columns = sort grep { ! $c->{$_} } keys %$test_changed;
1012 unshift @columns, @added_columns;
1014 warn "# sorted_items ", $#$sorted_items + 1, " offset $offset limit $limit order $sort";
1017 my $tmp = $code; $tmp =~ s/\$row->\{(['"]?)([\w\s]+)\1/$depends_on->{$2}++/gse;
1018 warn "# depends_on ",dump $depends_on;
1020 my $test_added = Storable::dclone $test_changed;
1021 delete $test_added->{$_} foreach keys %$depends_on;
1023 $code_depends = $self->param('code_depends')
1024 || join(',', keys %$depends_on);
1026 $code_description = $self->param('code_description') ||
1027 join(',', keys %$test_added);
1029 $code_depends ||= $code_description; # self-modifing
1030 if ( ! $code_depends && $out ) {
1031 $code_depends = $key;
1032 $code_description = $value;
1035 warn "# test_changed ",dump( $test_changed, $code_depends, $code_description );
1045 sorted => $sorted_items,
1046 columns => [ @columns ],
1047 rows => $#$filtered + 1,
1048 numeric => { map { $_, $self->_is_numeric($_) } @columns },
1049 unique => { map { $_, $self->_is_unique( $_) } @columns },
1050 filters => $self->_current_filters,
1052 cols_changed => $commit ? $commit_changed : $test_changed,
1053 code_depends => $code_depends,
1054 code_description => $code_description,
1055 code_path => $code_path,
1064 $self->session('order', $self->param('order'));
1065 $self->session('sort', $self->param('sort'));
1066 return $self->redirect_to('/data/items');
1070 my ( $self, $name ) = @_;
1072 my $stats = $self->_loaded( 'stats' );
1074 # sort facet numerically if more >50% elements are numeric
1075 my $count = $stats->{$name}->{count};
1076 $count -= $stats->{$name}->{empty} if defined $stats->{$name}->{empty};
1077 defined $stats->{$name}->{numeric} &&
1078 $stats->{$name}->{numeric} > $count / 2;
1082 my ( $self, $name ) = @_;
1083 my $stats = $self->_loaded( 'stats' );
1084 defined $stats->{$name}->{unique};
1087 sub _remove_filter {
1088 my ($self,$name) = @_;
1089 warn "_remove_filter $name\n";
1091 my $path = $self->session('path');
1093 delete $filters->{$name};
1094 delete $loaded->{$path}->{filters}->{$name};
1095 warn "filters left: ", keys %{ $loaded->{$path}->{filters} };
1098 grep { /\Q$name\E/ }
1099 keys %{ $loaded->{$path}->{filtered} }
1101 delete $loaded->{$path}->{filtered}->{$_};
1102 warn "remove filtered cache $_";
1109 my $path = $self->session('path') || return $self->redirect_to( '/data/index' );
1111 if ( my $name = $self->param('remove') ) {
1112 $self->_remove_filter( $name );
1113 return $self->redirect_to( '/data/items' );
1117 my $name = $self->param('name') || die "no name";
1119 my $all = $self->_param_scalar('all', 1);
1120 my $data = $self->_loaded('data');
1122 my $filters = $self->_current_filters;
1123 my $all_filters = __all_filters( keys %$filters,$self->session('order') );
1124 my $filtered = $loaded->{$path}->{filtered}->{$all_filters}
1125 if defined $loaded->{$path}->{filtered}->{$all_filters};
1127 if ( ! $filtered || $all ) {
1128 $filtered = [ 0 .. $#{ $data->{items} } ];
1129 warn "filter all values\n";
1131 warn "filter using $all_filters\n";
1134 foreach my $i ( @$filtered ) {
1135 my $item = $data->{items}->[$i];
1136 if ( ! exists $item->{$name} || ! defined $item->{$name} ) {
1137 $facet->{ _missing }++;
1138 } elsif ( ref $item->{$name} eq 'ARRAY' ) {
1139 $facet->{$_}++ foreach @{ $item->{$name} };
1141 $facet->{ $item->{$name} }++;
1145 my $checked_values = $self->_checked( @{ $filters->{$name} } ) if defined $filters->{$name};
1147 if ( my $code = $self->param('code') ) {
1149 foreach my $value ( keys %$facet ) {
1150 my $count = $facet->{$value};
1151 my $checked = $checked_values->{$value};
1155 warn "ERROR: $@\n$code\n";
1157 } elsif ( $checked != $checked_values->{$value} ) {
1158 warn "checked $value $count -> $checked\n";
1159 $checked_values->{$value} = $checked;
1162 warn "out ",dump( $out );
1163 $self->stash( out => $out );
1166 # warn "# facet $name ",dump $facet;
1170 : defined $filters->{$name} ? @{ $filters->{$name} }
1173 my $numeric = $self->_is_numeric($name);
1175 my $sort = $self->param('sort');
1176 # sort numeric facets with more than 5 values ascending
1177 $sort ||= $numeric && $#facet_names > 4 ? 'a' : 'c';
1179 @facet_names = sort {
1181 if ( $sort eq 'a' ) {
1182 $result = $numeric ? $a <=> $b : lc $a cmp lc $b;
1183 } elsif ( $sort eq 'd' ) {
1184 $result = $numeric ? $b <=> $a : lc $b cmp lc $a;
1185 } elsif ( $sort eq 'c' ) {
1186 $result = ( $facet->{$b} || -1 ) <=> ( $facet->{$a} || -1 )
1188 warn "unknown sort: $sort";
1189 $result = $a cmp $b;
1191 $result = $a cmp $b unless defined $result; # FIXME cludge for numeric facets with invalid data
1195 $self->render( name => $name, facet => $facet, checked => $checked_values,
1196 facet_names => \@facet_names, sort => $sort, numeric => $numeric,
1201 sub __invalidate_path_column {
1202 my ( $path, $name ) = @_;
1204 if ( defined $loaded->{$path}->{sorted}->{$name} ) {
1205 delete $loaded->{$path}->{sorted}->{$name};
1206 warn "# invalidate $path sorted $name\n";
1209 foreach ( grep { m/\Q$name\E/ } keys %{ $loaded->{$path}->{filtered} } ) {
1210 delete $loaded->{$path}->{filtered}->{$_};
1211 warn "# invalidate $path filtered $_\n";
1215 sub __path_modified {
1216 my ( $path, $value ) = @_;
1217 $value = 1 unless defined $value;
1219 $loaded->{$path}->{modified} = $value;
1221 warn "# __path_modified $path $value\n";
1224 sub __path_rebuild_stats { $loaded->{ $_[0] }->{rebuild_stats} = 1 };
1227 my ($self,$change) = @_;
1229 my $change_path = $self->_permanent_path( 'changes' );
1230 mkdir $change_path unless -d $change_path;
1231 $change_path .= '/' . $change->{time};
1232 store $change, $change_path;
1233 utime $change->{time}, $change->{time}, $change_path;
1234 warn "_save_change $change_path ", dump($change);
1239 my $new_content = $self->param('new_content');
1240 $new_content ||= $self->param('content'); # backward compatibility with old actions
1242 my $i = $self->param('_row_id');
1243 die "invalid _row_id ",dump($i) unless $i =~ m/^\d+$/;
1244 my $path = $self->param('path') || die "no path";
1245 my $name = $self->param('name') || die "no name";
1246 my $status = 200; # 200 = OK, 201 = Created
1248 my $data = $self->_loaded('data');
1250 if ( defined $loaded->{$path}->{data}->{items}->[$i] ) {
1251 $new_content =~ s/^\s+//s;
1252 $new_content =~ s/\s+$//s;
1254 if ( $new_content =~ /\xB6/ ) { # para
1255 $v = [ split(/\s*\xB6\s*/, $new_content) ];
1257 $v = [ $new_content ];
1260 my $old = dump $loaded->{$path}->{data}->{items}->[$i]->{$name};
1263 && ! ( $old eq 'undef' && length($new_content) == 0 ) # new value empty, previous undef
1265 $self->_save_change({
1269 old => $loaded->{$path}->{data}->{items}->[$i]->{$name},
1271 time => $self->param('time') || time(),
1272 user => $self->param('user') || $ENV{'LOGNAME'},
1274 map { $_ => $loaded->{$path}->{data}->{items}->[$i]->{$_}->[0] }
1275 grep { defined $loaded->{$path}->{stats}->{$_}->{unique} }
1276 keys %{ $loaded->{$path}->{stats} }
1280 warn "# change $path $i $old -> $new\n";
1281 $loaded->{$path}->{data}->{items}->[$i]->{$name} = $v;
1283 __invalidate_path_column( $path, $name );
1285 $status = 201; # created
1286 __path_rebuild_stats( $path );
1288 $new_content = join("\xB6",@$v);
1291 warn "# unchanged $path $i $old\n";
1295 $new_content = "$path $i $name doesn't exist\n";
1299 warn "# edit $status ", dump $new_content;
1303 new_content => scalar $new_content,
1310 my $path = $self->_param_or_session('path');
1311 my $dump_path = $self->_save( $path );
1312 __path_modified( $path, 0 );
1314 return $self->redirect_to( '/data/items' );
1320 my $dir = $self->app->home->rel_file('public');
1322 if ( my $import = $self->param('import') ) {
1324 if ( $import =~ m{/filter\.(.+?)\..+} ) {
1325 my $name = $self->_column_from_unac( $1 );
1327 my @vals = map { chomp; $_ }
1328 read_file "$dir/export/$import", binmode => ':utf8';
1330 $self->_remove_filter( $name );
1331 $self->_filter_on_data( $name, @vals );
1332 $self->session( 'offset' => 0 );
1333 return $self->redirect_to('/data/items');
1335 warn "UNKNOWN IMPORT $import";
1339 if ( my $remove = $self->param('remove') ) {
1340 my $path = "$dir/export/$remove";
1341 unlink $path if -e $path;
1343 unlink $path if -e $path;
1346 my $path = $self->_export_path || return $self->redirect_to('/data/index');
1348 my @files = grep { ! /\.png$/ } glob "$path/*";
1349 my $mtime = { map { $_ => (stat($_))[9] } @files };
1350 @files = sort { $mtime->{$b} <=> $mtime->{$a} } @files;
1351 $self->render( export => [ @files ] );
1354 sub __loaded_paths {
1356 grep { defined $loaded->{$_}->{data} }
1362 $self->stash( reload => 1 );
1364 # $self->_load_path( $self->param('path') );
1365 $self->redirect_to('/data/load?path=' . $self->param('path') );
1370 my $path = $self->param('path');
1371 if ( $path =~ m{^/tmp/mojo_facets\.} ) {
1373 warn "# unlink $path";
1374 if ( my $name = $self->param('name') ) {
1375 delete $loaded->{$name};
1376 warn "# remove $name from memory";
1379 warn "WARNING: $path unlink ignored";
1381 return if $self->stash('reload');
1382 return $self->redirect_to( '/data/load' );