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