added remove profile link
[MojoFacets.git] / lib / MojoFacets / Data.pm
1 package MojoFacets::Data;
2
3 use strict;
4 use warnings;
5
6 use base 'Mojolicious::Controller';
7
8 use Data::Dump qw(dump);
9 use File::Slurp;
10 use Encode;
11 use locale;
12 use File::Find;
13 use Storable;
14 use Time::HiRes qw(time);
15 use File::Path qw(mkpath);
16 use Text::Unaccent::PurePerl;
17
18 use MojoFacets::Import::File;
19 use MojoFacets::Import::HTMLTable;
20 use MojoFacets::Import::CSV;
21
22 our $loaded;
23 our $filters;
24
25 sub index {
26         my $self = shift;
27
28         my $data_dir = $self->app->home->rel_dir('data');
29         die "no data dir $data_dir" unless -d $data_dir;
30
31         my @files;
32         my $changes;
33         find( sub {
34                 my $file = $File::Find::name;
35                 if ( -f $file && $file =~ m/\.(js(on)?|txt)$/ ) {
36                         $file =~ s/$data_dir\/*//;
37                         push @files, $file;
38                 } elsif ( -f $file && $file =~ m/([^\/]+)\.changes\/(\d+\.\d+.+)/ ) {
39                         push @{ $changes->{$1} }, $2
40                 } elsif ( -d $file && $file =~ m/\.html$/ ) {
41                         $file =~ s/$data_dir\/*//;
42                         push @files, $file;
43                 } elsif ( -f $file && $file =~ m/\.csv$/i ) {
44                         $file =~ s/$data_dir\/*//;
45                         push @files, $file;
46                 } elsif ( -f $file && $file =~ m/\.storable/i ) {
47                         $file =~ s/$data_dir\/*//;
48                         push @files, $file;
49                 } else {
50                         #warn "IGNORE: $file\n";
51                 }
52         }, $data_dir);
53
54         no warnings qw(uninitialized); # mtime
55         @files = sort { $loaded->{$b}->{mtime} <=> $loaded->{$a}->{mtime} || lc $a cmp lc $b } @files,
56                         grep { defined $loaded->{$_}->{generated} } keys %$loaded;
57         my $size;
58         $size->{$_} = -s "$data_dir/$_" foreach @files;
59
60         $self->render(
61                 files => [ @files ],
62                 size => $size,
63                 loaded => $loaded,
64                 filters => $filters,
65                 dump_path => { map { $_ => $self->_dump_path($_) } @files },
66                 changes => $changes,
67         );
68 }
69
70 sub _dump_path {
71         my ( $self, $name ) = @_;
72         my $dir = $self->app->home->rel_dir('data');
73         $name =~ s/^$dir//;
74         $name =~ s/\/+/_/g;
75         return '/tmp/mojo_facets.' . $name . '.storable';
76 }
77
78 sub _save {
79         my ( $self, $path ) = @_;
80
81         my $dump_path = $self->_dump_path( $path );
82         my $first_load = ! -e $dump_path;
83         warn "save loaded to $dump_path";
84         my $info = $loaded->{$path};
85         store $info, $dump_path;
86
87         if ( $first_load ) {
88                 my $mtime = $loaded->{$path}->{mtime};
89                 utime $mtime, $mtime, $dump_path;
90                 warn "sync time to $path at $mtime\n";
91         }
92
93         warn $dump_path, ' ', -s $dump_path, " bytes\n";
94         return $dump_path;
95 }
96
97
98 sub __stats {
99
100         my $stats;
101
102         my $nr_items = $#{ $_[0] } + 1;
103
104         warn "__stats $nr_items\n";
105
106         foreach my $e ( @{ $_[0] } ) {
107                 foreach my $n ( keys %$e ) {
108                         $stats->{$n}->{count}++;
109                         my @v;
110                         if ( ref $e->{$n} eq 'ARRAY' ) {
111                                 $stats->{$n}->{array} += $#{ $e->{$n} } + 1;
112                                 @v = @{ $e->{$n} };
113                         } else {
114                                 @v = ( $e->{$n} );
115                         }
116
117                         foreach my $x ( @v ) {
118                                 $stats->{$n}->{numeric}++
119                                         if $x =~ m/^[-+]?([0-9]*\.[0-9]+|[0-9]+)$/;
120                                 $stats->{$n}->{empty}++
121                                         if length $x == 0; # faster than $x =~ m/^\s*$/;
122                         }
123
124                 }
125         }
126
127         foreach my $n ( keys %$stats ) {
128                 my $s = $stats->{$n};
129                 next unless defined $s->{array};
130                 if ( $s->{array} == $s->{count} ) {
131                         delete $s->{array};
132                         if ( $s->{count} == $nr_items ) {
133                                 warn "check $n for uniqeness\n";
134                                 my $unique;
135                                 foreach my $e ( @{ $_[0] } ) {
136                                         if ( ++$unique->{ $e->{$n}->[0] } == 2 ) {
137                                                 $unique = 0;
138                                                 last;
139                                         }
140                                 }
141                                 if ( $unique ) {
142                                         $stats->{$n}->{unique} = 1;
143                                         #warn "# $n unique ",dump( $unique );
144                                 }
145                         }
146                 }
147         }
148
149         warn "# __stats ",dump($stats);
150
151         return $stats;
152 }
153
154 sub _param_or_session {
155         $_[0]->param( $_[1] ) || $_[0]->session( $_[1] )
156 }
157
158 sub stats {
159         my $self = shift;
160         my $path = $self->_param_or_session('path');
161         warn "stats $path\n";
162         delete $loaded->{$path}->{stats};
163         $self->redirect_to( '/data/columns' );
164 }
165
166
167 sub _load_path {
168         my ( $self, $path ) = @_;
169
170         return if defined $loaded->{$path}->{generated};
171
172         my $full_path = $self->app->home->rel_file( 'data/' . $path );
173         $self->redirect_to('/data/index') unless -r $full_path;
174
175         my $dump_path = $self->_dump_path( $path );
176
177         if ( defined $loaded->{$path}->{data} ) {
178                 my $mtime = (stat($full_path))[9];
179                 return if $loaded->{$path}->{mtime} == $mtime;
180                 warn "reload $full_path, modified ", time() - $mtime, " seconds ago\n";
181         } elsif ( -e $dump_path ) {
182                 warn "dump_path $dump_path ", -s $dump_path, " bytes loading...\n";
183                 my $info = retrieve $dump_path;
184                 $loaded->{ $path } = $info;
185                 return;
186         }
187
188         my $data;
189         if ( -f $full_path ) {
190                 if ( $full_path =~ m/.csv/i ) {
191                         $data = MojoFacets::Import::CSV->new( full_path => $full_path )->data;
192                 } elsif ( $full_path =~ m/.storable/ ) {
193                         warn "open $full_path ", -s $full_path, " bytes";
194                         open(my $pipe, "<", $full_path) || die $!;
195                         while ( my $o = eval { Storable::fd_retrieve $pipe } ) {
196                                 if ( defined $o->{item} ) {
197                                         push @{ $data->{items} }, $o->{item};
198                                 } else {
199                                         warn "SKIP ",dump($o);
200                                 }
201                         }
202                         close($pipe);
203                         warn "loaded ", $#{ $data->{items} } + 1, " items from $full_path\n";
204                         $data->{generated}++;
205                 } else {
206                         $data = MojoFacets::Import::File->new( full_path => $full_path, path => $path )->data;
207                 }
208         } elsif ( -d $full_path && $full_path =~ m/.html/ ) {
209                 $data = MojoFacets::Import::HTMLTable->new( dir => $full_path )->data;
210         } else {
211                 die "can't load $full_path";
212         }
213
214         my @header;
215
216         if ( defined $data->{header} ) {
217                 if ( ref $data->{header} eq 'ARRAY' ) {
218                         @header = @{ $data->{header} };
219                 } else {
220                         warn "header not array ", dump( $data->{header} );
221                 }
222         }
223
224         my $stats = __stats( $data->{items} );
225
226         @header =
227                 sort { $stats->{$b}->{count} <=> $stats->{$a}->{count} }
228                 grep { defined $stats->{$_}->{count} } keys %$stats
229                 unless @header;
230
231         my $info = {
232                 header => [ @header ],
233                 stats  => $stats,
234                 full_path => $full_path,
235                 size => -s $full_path,
236                 mtime => (stat($full_path))[9],
237                 data => $data,
238                 defined $data->{generated} ? ( generated => 1 ) : (),
239         };
240
241         $loaded->{ $path } = $info;
242         $self->_save( $path ) unless $info->{generated};
243
244 }
245
246
247 sub load {
248         my $self = shift;
249
250         my $path = $self->param('path') || $self->redirect_to( '/data/index' );
251
252         my @paths = $self->param('paths');
253         warn "# paths ", dump @paths;
254
255         foreach my $p ( keys %$loaded ) {
256                 next if grep { /^\Q$p\E$/ } @paths;
257                 warn "remove $p from memory\n";
258                 delete $loaded->{$p};
259         }
260
261         $self->_load_path( $_ ) foreach @paths;
262
263         warn "# path $path\n";
264         $self->_load_path( $path );
265
266         $self->session( 'path' => $path );
267
268         my $redirect_to = '/data/items';
269
270         $self->session( 'header' => $loaded->{$path}->{header} );
271         if ( ! defined $loaded->{$path}->{columns} ) {
272                 my $columns_path = $self->_permanent_path( 'columns' );
273                 if ( -e $columns_path ) {
274                         my @columns = map { s/[\r\n]+$//; $_ } read_file $columns_path, binmode => ':utf8';
275                         $loaded->{$path}->{columns} = [ @columns ];
276                         warn "# columns_path $columns_path ",dump(@columns);
277                 } else {
278                         $loaded->{$path}->{columns} = $loaded->{$path}->{header}
279                 }
280
281                 $redirect_to = '/data/columns';
282         }
283         $self->session( 'columns' => $loaded->{$path}->{columns} );
284         $self->session( 'order'   => $loaded->{$path}->{columns}->[0] );
285         $self->redirect_to( $redirect_to );
286 }
287
288
289 sub _loaded {
290         my ( $self, $name ) = @_;
291         my $path = $self->session('path') || $self->param('path');
292         $self->redirect_to('/data/index') unless $path;
293
294         if ( defined $loaded->{$path}->{modified} && $loaded->{$path}->{modified} > 1 ) {
295                 my $caller = (caller(1))[3];
296                 if ( $caller =~ m/::edit/ ) {
297                         warn "rebuild stats for $path ignored caller $caller\n";
298                 } else {
299                         warn "rebuild stats for $path FORCED by modified caller $caller\n";
300 #                       $loaded->{$path}->{stats} = __stats( $loaded->{$path}->{data}->{items} );
301                         $loaded->{$path}->{rebuild_stats} = 1;
302                         $loaded->{$path}->{modified} = 1;
303                 }
304         }
305
306         if ( defined $loaded->{$path}->{rebuild_stats} ) {
307                 warn "rebuild_stats $path";
308                 $loaded->{$path}->{stats} = __stats( $loaded->{$path}->{data}->{items} );
309                 delete $loaded->{$path}->{rebuild_stats};
310         }
311
312         if ( ! defined $loaded->{$path}->{$name} ) {
313                 warn "$path $name isn't loaded\n";
314                 $self->_load_path( $path );
315                 if ( ! defined $loaded->{$path}->{stats} ) {
316                         warn "rebuild stats for $path\n";
317                         $loaded->{$path}->{stats} = __stats( $loaded->{$path}->{data}->{items} );
318                 }
319                 if ( ! defined $loaded->{$path}->{$name} ) {
320                         warn "MISSING $name for $path\n";
321                         $self->redirect_to('/data/index')
322                 }
323         }
324
325         $self->session( 'modified' => $loaded->{$path}->{modified} );
326
327         return $loaded->{$path}->{$name};
328 }
329
330
331 sub _checked {
332         my $self = shift;
333         my $checked;
334         $checked->{$_}++ foreach @_;
335 #       warn "# _checked ",dump($checked);
336         return $checked;
337 }
338
339 sub _permanent_path {
340         my $self = shift;
341         my $path = $self->_param_or_session('path');
342         $self->app->home->rel_dir('data') . '/' . join('.', $path, @_);
343 }
344
345 sub __unac {
346         my $n = shift;
347         $n = unac_string($n);
348         $n =~ s/\W+/_/g;
349         return $n;
350 }
351
352 sub _column_from_unac {
353         my ($self,$name) = @_;
354
355         my $stats = $self->_loaded('stats');
356         my $cols_norm = { map { __unac( $_ ) => $_ } keys %$stats };
357
358         $cols_norm->{$name} || die "can't find column $name in ", dump($cols_norm);
359 }
360
361 sub _export_path {
362         my $self = shift;
363         my $path = $self->_param_or_session('path');
364         if ( ! $path ) {
365                 warn "no path in param or session";
366                 return;
367         }
368         my $dir = $self->app->home->rel_dir('public') . "/export/$path";
369         mkpath $dir unless -e $dir;
370         my $name = join('.', map { __unac($_) } @_ );
371         my $full = $dir . '/' . $name;
372         $full =~ s/\/+$// if -d $full; # strip trailing slash for dirs
373         return $full;
374 }
375
376 sub columns {
377     my $self = shift;
378
379         if ( $self->param('columns') ) {
380                 my @columns = $self->_param_array('columns');
381                 write_file( $self->_permanent_path( 'columns' ), { binmode => ':utf8' }, map { "$_\n" } @columns );
382                 $self->redirect_to('/data/items');
383         }
384
385         my $stats = $self->_loaded( 'stats' );
386
387         my @columns;
388         @columns = grep { defined $stats->{$_}->{count} } @{ $self->session('columns') } if $self->session('columns');
389
390         foreach my $c ( @{ $self->session( 'header' ) } ) {
391                 push @columns, $c unless grep { /^\Q$c\E$/ } @columns;
392         }
393
394     $self->render(
395                 message => 'Select columns to display',
396                 stats => $stats,
397                 columns => \@columns,
398                 checked => $self->_checked( $self->_param_array('columns') ),
399         );
400 }
401
402 sub _param_array {
403     my ($self,$name) = @_;
404
405         my @array = $self->param($name);
406         my $path  = $self->session('path');
407
408         if ( @array ) {
409                 $self->session($name => [ @array ]);
410                 $loaded->{$path}->{$name} = [ @array ];
411         } elsif ( my $session = $self->session($name) ) {
412                 if ( ref $session eq 'ARRAY' ) {
413                         @array = @$session;
414                 } else {
415                         die "$name not array ",dump($session);
416                 }
417         }
418         #warn "# $name ",dump @array;
419         return @array;
420 }
421
422 sub _param_scalar {
423     my ($self,$name,$default) = @_;
424
425         my $scalar = $self->param($name);
426
427         if ( defined $scalar ) {
428                 $self->session($name => $scalar);
429         } else {
430                 $scalar = $self->session($name);
431         }
432
433         if ( ! defined $scalar ) {
434                 $scalar = $default;
435                 die "no default for $name" unless defined $scalar;
436                 $self->session($name => $scalar);
437         }
438
439         warn "# _param_scalar $name ",dump $scalar;
440         return $scalar;
441 }
442
443 sub filter {
444         my $self = shift;
445
446         my $name = $self->param('filter_name') || die "name?";
447         my @vals = $self->param('filter_vals');
448
449         $self->_remove_filter( $name );
450         if ( @vals ) {
451                 $self->_filter_on_data( $name, @vals );
452                 if ( my $permanent = $self->param('_permanent') ) {
453                         my $permanent_path = $self->_export_path( 'filter', $name, $permanent );
454                         write_file $permanent_path, { binmode => ':utf8' }, map { "$_\n" } @vals;
455                         warn "permanent filter $permanent_path ", -s $permanent_path;
456                 }
457         }
458
459         $self->session( 'offset' => 0 );
460
461         $self->redirect_to('/data/items');
462 }
463
464 sub _filter_on_data {
465         my ( $self, $name, @vals ) = @_;
466
467         my $path = $self->session('path');
468
469         if ( ! defined $loaded->{$path}->{stats}->{ $name } ) {
470                 warn "filter $name not found in data set";
471                 return;
472         }
473
474         $filters->{$name} = [ @vals ];
475         warn "_filter_on_data $name ", $#vals + 1, " values on $path\n";
476
477         my $filter_hash;
478         $filter_hash->{$_}++ foreach @vals;
479
480         #warn "# filter_hash ",dump( $filter_hash );
481
482         my $items = $self->_loaded('data')->{items};
483
484         my $include_missing = defined $filter_hash->{_missing};
485         my $filtered_items;
486
487         foreach my $i ( 0 .. $#$items ) {
488
489                 if ( defined $items->[$i]->{$name} ) {
490                         my $row = $items->[$i]->{$name};
491                         $row = [ $row ] unless ref $row eq 'ARRAY'; # FIXME probably wrong place
492                         foreach my $v ( @$row ) {
493                                 if ( defined $filter_hash->{ $v } ) {
494                                         $filtered_items->{$i}++;
495                                 }
496                         }
497                 } elsif ( $include_missing ) {
498                         $filtered_items->{$i}++;
499                 }
500         }
501
502         #warn "# filter $name ",dump($filtered_items);
503
504         $loaded->{$path}->{filters}->{$name} = $filtered_items;
505         warn "filter $name with ", scalar keys %$filtered_items, " items created\n";
506 }
507
508
509 sub _current_filters {
510         my $self = shift;
511         my $current_filters;
512         my $stats = $self->_loaded('stats');
513
514         $current_filters->{ $_ } = $filters->{ $_ }
515                 foreach ( grep { defined $filters->{ $_ } } keys %$stats )
516         ;
517         warn "# _current_filters ",dump( keys %$current_filters );
518         return $current_filters;
519 }
520
521 sub _data_sorted_by {
522         my ( $self, $order ) = @_;
523
524         my $path = $self->session('path');
525
526         warn "_data_sorted_by $order from $path";
527
528         if ( defined $loaded->{$path}->{sorted}->{$order} ) {
529                 return $loaded->{$path}->{sorted}->{$order};
530         }
531
532         my $data = $self->_loaded( 'data' );
533         my $numeric = $self->_is_numeric($order);
534         my $missing = $numeric ? 0 : '';
535         no warnings qw(numeric);
536         my $nr = 0;
537         my @sorted = map {
538                 $_->[0]
539         } sort {
540                 if ( $numeric ) {
541                         $a->[1] <=> $b->[1]
542                 } else {
543                         $a->[1] cmp $b->[1]
544                 }
545         } map {
546                 my $v;
547                 if ( ! exists $_->{$order} ) {
548                         $v = $missing;
549                 } elsif ( ref $_->{$order} eq 'ARRAY' ) {
550                         $v = join('', @{$_->{$order}});
551                 } else {
552                         $v = $_->{$order};
553                 }
554                 [ $nr++, $v ]
555         } @{ $data->{items} }
556         ;
557
558         warn "sorted: $order numeric: $numeric items: ", $#sorted + 1, "\n";
559         #warn "# sorted ",dump( @sorted );
560
561         $loaded->{$path}->{sorted}->{$order} = [ @sorted ];
562 }
563
564
565 sub __all_filters {
566         my $order = pop @_;
567         join(',', sort(@_), 'order', $order);
568 }
569
570 our ($out, $key,$value);
571
572 sub __commit_path_code {
573         my ( $path, $i, $code, $commit_changed ) = @_;
574
575         my $items = $loaded->{$path}->{data}->{items} || die "no items for $path";
576         my $row = $items->[$i];
577         my $update;
578         eval $code;
579         foreach ( keys %$update ) {
580                 $$commit_changed->{$_}++;
581                 $loaded->{$path}->{data}->{items}->[$i]->{$_} = $update->{$_};
582         }
583         #warn "__commit_path_code $path $i ",dump( $update );
584 }
585
586 sub items {
587         my $self = shift;
588
589         if ( my $show = $self->param('id') ) {
590                 $self->param('show', $show);
591                 warn "show $show\n";
592         }
593
594         my $path = $self->session('path');
595
596         my @columns = $self->_param_array('columns');
597         $self->redirect_to('/data/columns') unless @columns;
598         my $order   = $self->_param_scalar('order', $columns[0]);
599         my $sort    = $self->_param_scalar('sort', 'a');
600         my $offset  = $self->_param_scalar('offset', 0);
601         my $limit   = $self->_param_scalar('limit', 20);
602         $self->_param_scalar('show', 'table');
603
604         # fix offset when changing limit
605         $offset = int( $offset / $limit ) * $limit;
606
607         if ( ! grep { /^\Q$order\E$/ } @columns ) {
608                 $order = $columns[0];
609                 $self->session( order => $order );
610         }
611         my $sorted = $self->_data_sorted_by( $order );
612
613         my @filter_names;
614         if ( $filters ) {
615                 foreach my $name ( keys %$filters ) {
616                         if ( ! defined $loaded->{$path}->{stats}->{ $name } ) {
617                                 warn "skip filter $name not found in $path\n";
618                                 next;
619                         }
620                         push @filter_names, $name;
621                 }
622                 warn "filter_names ",dump( @filter_names );
623                 foreach my $name ( @filter_names ) {
624                         next if ref $loaded->{$path}->{filters}->{$name} eq 'ARRAY';
625                         $self->_filter_on_data( $name, @{ $filters->{$name} } );
626                 }
627         }
628
629         my $all_filters = __all_filters( @filter_names,$order );
630
631 #       warn "# all_filters $all_filters ", dump( $loaded->{$path}->{filtered}->{$all_filters} );
632
633         if ( ! defined $loaded->{$path}->{filtered}->{$all_filters} ) {
634
635                 my $path_filters = $loaded->{$path}->{filters};
636
637                 warn "create combined filter for $all_filters from ", $#$sorted + 1, " items\n";
638
639                 my @filtered;
640                 foreach my $i ( 0 .. $#$sorted ) {
641                         my $pos = $sorted->[$i];
642
643                         if ( $#filter_names == -1 ) {
644                                 push @filtered, $pos;
645                                 next;
646                         }
647
648                         my $skip = 0;
649                         foreach ( @filter_names ) {
650                                 $skip ||= 1 if ! defined $path_filters->{$_}->{$pos};
651                         }
652                         next if $skip;
653
654                         push @filtered, $pos;
655                 }
656
657                 $loaded->{$path}->{filtered}->{$all_filters} = [ @filtered ];
658         }
659
660         my $filtered = $loaded->{$path}->{filtered}->{$all_filters}
661                 if defined $loaded->{$path}->{filtered}->{$all_filters};
662
663         warn "all_filters $all_filters produced ", $#$filtered + 1, " items\n" if $filtered;
664
665         my $data = $self->_loaded('data');
666
667         my $code = $self->_param_scalar('code','');
668         $code =~ s{\r}{}gs;
669         $code =~ s{\n+$}{\n}s;
670
671         my $commit = $self->param('commit');
672         my $test = $self->param('test');
673
674         my $commit_changed;
675
676         if ( $code && ( $test || $commit ) ) {
677                 # XXX find columns used in code snippet and show them to user
678                 my $order = 0;
679                 foreach my $column ( $code =~ m/\$row->{([^}]+)}/g ) {
680                         if ( $column =~ s/^(['"])// ) {
681                                 $column =~ s/$1$//;
682                         }
683                         next if $column =~ m/\$/; # hide columns with vars in them
684                         $commit_changed->{$column} = 0;
685                 }
686         }
687
688         my $code_path = $self->app->home->rel_dir('public') . "/code";
689         if ( $commit ) {
690
691                 warn "# commit on ", $#$filtered + 1, " items:\n$code\n";
692                 ( $key, $value, $out ) = ( 'key', 'value' );
693                 foreach ( 0 .. $#$filtered ) {
694                         my $i = $filtered->[$_];
695                         __commit_path_code( $path, $i, $code, \$commit_changed );
696                 }
697
698                 $self->_save_change({
699                         path => $path,
700                         time => $self->param('time') || time(),
701                         user => $self->param('user') || $ENV{'LOGNAME'},
702                         code => $code,
703                         commit_changed => $commit_changed,
704                 });
705
706                 if ( my $description = $self->param('code_description') ) {
707                         my $depends = $self->param('code_depends') || die "no code_depends?";
708                         my $path = "$code_path/$depends.$description.pl";
709                         if ( -e $path && ! $self->param('overwrite') ) {
710                                 warn "# code $path not saved\n";
711                         } else {
712                                 write_file(  $path, { binmode => ':utf8' }, "$code\n" );
713                                 warn "code $path ", -s $path, " bytes saved\n";
714                         }
715                 }
716
717                 # remove console
718                 $code = '';
719                 if ( $out ) {
720                         my $commit_dataset = join('.'
721                                 , $self->param('code_depends')
722                                 , $self->param('code_description')
723                                 , time()
724                         );
725                         $key ||= 'key';
726                         $value ||= 'value';
727                         warn "key $key value $value";
728                         my $items;
729                         foreach my $n ( keys %$out ) {
730                                 my $i = { $key => [ $n ] };
731                                 my $ref = ref $out->{$n};
732                                 if ( $ref eq 'HASH' ) {
733                                         $i->{$_} = [ $out->{$n}->{$_} ] foreach keys %{ $out->{$n} };
734                                 } elsif ( $ref eq 'ARRAY' ) {
735                                         $i->{$_} = $out->{$n};
736                                 } elsif ( ! $ref ) {
737                                         $i->{$value} = [ $out->{$n} ];
738                                 } else {
739                                         $i->{_error} = [ dump($out->{$n}) ];
740                                 }
741                                 push @$items, $i;
742                         };
743                         undef $out;
744                         my $stats = __stats( $items );
745                         my @columns = grep { ! m/^\Q$key\E$/ } sort keys %$stats;
746                         unshift @columns, $key;
747
748                         $loaded->{$commit_dataset} = {
749                                 header => [ @columns ],
750                                 columns => [ @columns ],
751                                 mtime => time(),
752                                 data => { items => $items },
753                                 stats => $stats,
754                                 generated => 1,
755                         };
756                         warn "# loaded out ", dump( $loaded->{$commit_dataset} );
757                         $self->session('path', $commit_dataset);
758                         $self->session('columns', [ @columns ]);
759                         $self->session('order', $key);
760                         $self->redirect_to('/data/items');
761                         return; # FIXME needed to correctly show columns
762                 }
763
764                 # this might move before $out to recalculate stats on source dataset?
765                 __path_rebuild_stats( $path );
766                 my $c = { map { $_ => 1 } @columns };
767                 my @added_columns = sort grep { ! $c->{$_} } keys %$commit_changed;
768                 warn "# added_columns ",dump( @added_columns );
769                 unshift @columns, @added_columns;
770
771                 $self->session('columns', [ @columns ]);
772                 $loaded->{$path}->{columns} = [ @columns ];
773                 warn "# new columns ",dump( @columns );
774
775                 __invalidate_path_column( $path, $_ ) foreach keys %$commit_changed;
776         }
777
778         my $sorted_items;
779         my $from_end = $sort eq 'd' ? $#$filtered : 0;
780         my $test_changed;
781         my ( $key, $value, $out ) = ( 'key', 'value' ); # XXX make local
782         foreach ( 0 .. $limit ) {
783                 my $i = $_ + $offset;
784                 last unless defined $filtered->[$i];
785                 $i = $from_end - $i if $from_end;
786                 my $id = $filtered->[$i];
787                 my $row = Storable::dclone $data->{items}->[ $id ];
788                 if ( $code && $test ) {
789                         my $update;
790                         eval $code;
791                         if ( $@ ) {
792                                 warn "ERROR evaling $@", dump($code);
793                                 $self->stash('eval_error', $@) if $@;
794                         } else {
795                                 warn "EVAL ",dump($update);
796                                 foreach ( keys %$update ) {
797                                         $test_changed->{$_}++;
798                                         $row->{$_} = $update->{$_};
799                                 }
800                         }
801                 }
802                 $row->{_row_id} ||= $id;
803                 push @$sorted_items, $row;
804         }
805
806         if ( $self->param('export') ) {
807                 my $export_path = $self->_export_path( 'items', @columns);
808                 open(my $fh, '>', $export_path) || warn "ERROR: can't open $export_path: $!";
809                 foreach my $f ( 0 .. $#$filtered ) {
810                         print $fh join("\t", map {
811                                 my $i = $data->{items}->[ $filtered->[$f] ];
812                                 my $v = '\N';
813                                 if ( ! defined $i->{$_} ) {
814                                         # nop
815                                 } elsif ( ref $i->{$_} eq 'ARRAY' ) {
816                                         $v =join(',', @{ $i->{$_} });
817                                         $v = '\N' if length($v) == 0;
818                                 } elsif ( ! ref $i->{$_} ) {
819                                         $v = $i->{$_};
820                                 } else {
821                                         $v = dump $i->{$_};
822                                 }
823                                 $v;
824                         } @columns),"\n";
825                 }
826                 close($fh);
827                 warn "export $export_path ", -s $export_path, " bytes\n";
828         }
829
830         warn "# test_changed ",dump( $test_changed );
831         my $c = { map { $_ => 1 } @columns };
832         my @added_columns = sort grep { ! $c->{$_} } keys %$test_changed;
833         unshift @columns, @added_columns;
834
835         warn "# sorted_items ", $#$sorted_items + 1, " offset $offset limit $limit order $sort";
836
837         my $code_depends = $self->param('code_depends')||
838         join(',', sort grep { $test_changed->{$_} == 0 } keys %$test_changed );
839         my $code_description = $self->param('code_description') ||
840         join(',', @added_columns);
841
842         $code_depends ||= $code_description; # self-modifing
843         if ( ! $code_depends && $out ) {
844                 $code_depends = $key;
845                 $code_description = $value;
846         }
847
848         warn "# test_changed ",dump( $test_changed, $code_depends, $code_description );
849
850         $self->render(
851                 order => $order,
852                 offset => $offset,
853                 limit => $limit,
854                 sorted => $sorted_items,
855                 columns => [ @columns ],
856                 rows => $#$filtered + 1,
857                 numeric => { map { $_, $self->_is_numeric($_) } @columns },
858                 unique  => { map { $_, $self->_is_unique( $_) } @columns },
859                 filters => $self->_current_filters,
860                 code => $code,
861                 cols_changed => $commit ? $commit_changed : $test_changed,
862                 code_depends => $code_depends,
863                 code_description => $code_description,
864                 code_path => $code_path,
865                 out => $out,
866         );
867
868 }
869
870
871 sub order {
872         my $self = shift;
873         $self->session('order', $self->param('order'));
874         $self->session('sort', $self->param('sort'));
875         $self->redirect_to('/data/items');
876 }
877
878 sub _is_numeric {
879         my ( $self, $name ) = @_;
880
881         my $stats = $self->_loaded( 'stats' );
882
883         # sort facet numerically if more >50% elements are numeric
884         my $count = $stats->{$name}->{count};
885         $count   -= $stats->{$name}->{empty} if defined $stats->{$name}->{empty};
886         defined $stats->{$name}->{numeric} &&
887                 $stats->{$name}->{numeric} > $count / 2;
888 }
889
890 sub _is_unique {
891         my ( $self, $name ) = @_;
892         my $stats = $self->_loaded( 'stats' );
893         defined $stats->{$name}->{unique};
894 }
895
896 sub _remove_filter {
897         my ($self,$name) = @_;
898         warn "_remove_filter $name\n";
899
900         my $path = $self->session('path');
901
902         delete $filters->{$name};
903         delete $loaded->{$path}->{filters}->{$name};
904         warn "filters left: ", keys %{ $loaded->{$path}->{filters} };
905
906         foreach (
907                         grep { /\Q$name\E/ }
908                         keys %{ $loaded->{$path}->{filtered} }
909         ) {
910                 delete $loaded->{$path}->{filtered}->{$_};
911                 warn "remove filtered cache $_";
912         }
913 }
914
915 sub facet {
916         my $self = shift;
917
918         my $path = $self->session('path') || $self->redirect_to( '/data/index' );
919
920         if ( my $name = $self->param('remove') ) {
921                 $self->_remove_filter( $name );
922                 $self->redirect_to( '/data/items' );
923         }
924
925         my $facet;
926         my $name = $self->param('name') || die "no name";
927
928         my $all = $self->_param_scalar('all', 1);
929         my $data = $self->_loaded('data');
930
931         my $filters = $self->_current_filters;
932         my $all_filters = __all_filters( keys %$filters,$self->session('order') );
933         my $filtered = $loaded->{$path}->{filtered}->{$all_filters}
934                 if defined $loaded->{$path}->{filtered}->{$all_filters};
935
936         if ( ! $filtered || $all ) {
937                 $filtered = [ 0 .. $#{ $data->{items} } ];
938                 warn "filter all values\n";
939         } else {
940                 warn "filter using $all_filters\n";
941         }
942
943         foreach my $i ( @$filtered ) {
944                 my $item = $data->{items}->[$i];
945                 if ( ! exists $item->{$name} ) {
946                         $facet->{ _missing }++;
947                 } elsif ( ref $item->{$name} eq 'ARRAY' ) {
948                         $facet->{$_}++ foreach @{ $item->{$name} };
949                 } else {
950                         $facet->{ $item->{$name} }++;
951                 }
952         }
953
954         if ( my $code = $self->param('code') ) {
955                 my $out;
956                 foreach my $v ( keys %$facet ) {
957                         my $c = $facet->{$v};
958                         eval $code;
959                         if ( $@ ) {
960                                 $out = $@;
961                                 warn "ERROR: $@\n$code\n";
962                                 last;
963                         }
964                 }
965                 warn "out ",dump( $out );
966                 $self->stash( out => $out );
967         }
968
969 #       warn "# facet $name ",dump $facet;
970
971         my $checked;
972         my @facet_names =
973                   $all                      ? keys %$facet
974                 : defined $filters->{$name} ? @{ $filters->{$name} }
975                 : keys %$facet;
976
977         $checked = $self->_checked( @{ $filters->{$name} } ) if defined $filters->{$name};
978
979         my $numeric = $self->_is_numeric($name);
980
981         my $sort = $self->param('sort');
982         # sort numeric facets with more than 5 values ascending
983         $sort ||= $numeric && $#facet_names > 4 ? 'a' : 'c';
984
985         @facet_names = sort {
986                 my $result;
987                 if ( $sort eq 'a' ) {
988                         $result = $numeric ? $a <=> $b : lc $a cmp lc $b;
989                 } elsif ( $sort eq 'd' ) {
990                         $result = $numeric ? $b <=> $a : lc $b cmp lc $a;
991                 } elsif ( $sort eq 'c' ) {
992                         $result = ( $facet->{$b} || -1 ) <=> ( $facet->{$a} || -1 )
993                 } else {
994                         warn "unknown sort: $sort";
995                         $result = $a cmp $b;
996                 }
997                 $result = $a cmp $b unless defined $result; # FIXME cludge for numeric facets with invalid data
998                 $result;
999         } @facet_names;
1000
1001         $self->render( name => $name, facet => $facet, checked => $checked,
1002                 facet_names => \@facet_names, sort => $sort, numeric => $numeric,
1003         );
1004 }
1005
1006
1007 sub __invalidate_path_column {
1008         my ( $path, $name ) = @_;
1009
1010         if ( defined $loaded->{$path}->{sorted}->{$name} ) {
1011                 delete $loaded->{$path}->{sorted}->{$name};
1012                 warn "# invalidate $path sorted $name\n";
1013         }
1014
1015         foreach ( grep { m/\Q$name\E/ } keys %{ $loaded->{$path}->{filtered} } ) {
1016                 delete $loaded->{$path}->{filtered}->{$_};
1017                 warn "# invalidate $path filtered $_\n";
1018         }
1019 }
1020
1021 sub __path_modified {
1022         my ( $path, $value ) = @_;
1023         $value = 1 unless defined $value;
1024         
1025         $loaded->{$path}->{modified}  = $value;
1026
1027         warn "# __path_modified $path $value\n";
1028 }
1029
1030 sub __path_rebuild_stats { $loaded->{ $_[0] }->{rebuild_stats} = 1 };
1031
1032 sub _save_change {
1033         my ($self,$change) = @_;
1034
1035         my $change_path = $self->_permanent_path( 'changes' );
1036         mkdir $change_path unless -d $change_path;
1037         $change_path .= '/' . $change->{time};
1038         store $change, $change_path;
1039         utime $change->{time}, $change->{time}, $change_path;
1040         warn "_save_change $change_path ", dump($change);
1041 }
1042
1043 sub edit {
1044         my $self = shift;
1045         my $new_content = $self->param('new_content');
1046         $new_content  ||= $self->param('content'); # backward compatibility with old actions
1047
1048         my $i = $self->param('_row_id');
1049         die "invalid _row_id ",dump($i) unless $i =~ m/^\d+$/;
1050         my $path = $self->param('path') || die "no path";
1051         my $name = $self->param('name') || die "no name";
1052         my $status = 200; # 200 = OK, 201 = Created
1053
1054         my $data = $self->_loaded('data');
1055
1056         if ( defined $loaded->{$path}->{data}->{items}->[$i] ) {
1057                 $new_content =~ s/^\s+//s;
1058                 $new_content =~ s/\s+$//s;
1059                 my $v;
1060                 if ( $new_content =~ /\xB6/ ) { # para
1061                         $v = [ split(/\s*\xB6\s*/, $new_content) ];
1062                 } else {
1063                         $v = [ $new_content ];
1064                 }
1065
1066                 my $old = dump $loaded->{$path}->{data}->{items}->[$i]->{$name};
1067                 my $new = dump $v;
1068                 if ( $old ne $new
1069                         && ! ( $old eq 'undef' && length($new_content) == 0 ) # new value empty, previous undef
1070                 ) {
1071                         $self->_save_change({
1072                                 path => $path,
1073                                 column => $name,
1074                                 pos => $i,
1075                                 old => $loaded->{$path}->{data}->{items}->[$i]->{$name},
1076                                 new => $v,
1077                                 time => $self->param('time') || time(),
1078                                 user => $self->param('user') || $ENV{'LOGNAME'},
1079                                 unique => {
1080                                         map { $_ => $loaded->{$path}->{data}->{items}->[$i]->{$_}->[0] }
1081                                         grep { defined $loaded->{$path}->{stats}->{$_}->{unique} }
1082                                         keys %{ $loaded->{$path}->{stats} }
1083                                 },
1084                         });
1085
1086                         warn "# change $path $i $old -> $new\n";
1087                         $loaded->{$path}->{data}->{items}->[$i]->{$name} = $v;
1088
1089                         __invalidate_path_column( $path, $name );
1090
1091                         $status = 201; # created
1092                         __path_rebuild_stats( $path );
1093         
1094                         $new_content = join("\xB6",@$v);
1095
1096                 } else {
1097                         warn "# unchanged $path $i $old\n";
1098                         $status = 304;
1099                 }
1100         } else {
1101                 $new_content = "$path $i $name doesn't exist\n";
1102                 $status = 404;
1103         }
1104
1105         warn "# edit $status ", dump $new_content;
1106
1107         $self->render(
1108                 status => $status,
1109                 new_content => scalar $new_content,
1110         );
1111 }
1112
1113
1114 sub save {
1115         my $self = shift;
1116         my $path = $self->_param_or_session('path');
1117         my $dump_path = $self->_save( $path );
1118         __path_modified( $path, 0 );
1119
1120         $self->redirect_to( '/data/items' );
1121 }
1122
1123 sub export {
1124         my $self = shift;
1125
1126         my $dir = $self->app->home->rel_dir('public');
1127
1128         if ( my $import = $self->param('import') ) {
1129
1130                 if ( $import =~ m{/filter\.(.+?)\..+} ) {
1131                         my $name = $self->_column_from_unac( $1 );
1132
1133                         my @vals = map { chomp; $_ }
1134                                 read_file "$dir/export/$import", binmode => ':utf8';
1135
1136                         $self->_remove_filter( $name );
1137                         $self->_filter_on_data( $name, @vals );
1138                         $self->session( 'offset' => 0 );
1139                         $self->redirect_to('/data/items');
1140                 } else {
1141                         warn "UNKNOWN IMPORT $import";
1142                 }
1143         }
1144
1145         if ( my $remove = $self->param('remove') ) {
1146                 my $path = "$dir/export/$remove";
1147                 unlink $path if -e $path;
1148                 $path .= '.png';
1149                 unlink $path if -e $path;
1150         }
1151
1152         my $path = $self->_export_path || $self->redirect_to('/data/index');
1153
1154         my @files = grep { ! /\.png$/ } glob "$path/*";
1155         my $mtime = { map { $_ => (stat($_))[9] } @files };
1156         @files = sort { $mtime->{$b} <=> $mtime->{$a} } @files;
1157         $self->render( export => [ @files ] );
1158 }
1159
1160 sub __loaded_paths {
1161         return
1162                 grep { defined $loaded->{$_}->{data} }
1163                 keys %$loaded;
1164 }
1165
1166 1;