added </tt> for filename
[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 use Digest::MD5;
18
19 use MojoFacets::Import::File;
20 use MojoFacets::Import::HTMLTable;
21 use MojoFacets::Import::CSV;
22 use MojoFacets::Import::CouchDB;
23 use MojoFacets::Import::SQL;
24
25 our $loaded;
26 our $filters;
27
28 sub index {
29         my $self = shift;
30
31         my $data_dir = $self->app->home->rel_dir('data');
32         die "no data dir $data_dir" unless -d $data_dir;
33
34         my @files;
35         my $changes;
36         find( sub {
37                 my $file = $File::Find::name;
38                 if ( -f $file && $file =~ m/\.(js(on)?|txt)$/ ) {
39                         $file =~ s/$data_dir\/*//;
40                         push @files, $file;
41                 } elsif ( -f $file && $file =~ m/([^\/]+)\.changes\/(\d+\.\d+.+)/ ) {
42                         push @{ $changes->{$1} }, $2
43                 } elsif ( -d $file && $file =~ m/\.html$/ ) {
44                         $file =~ s/$data_dir\/*//;
45                         push @files, $file;
46                 } elsif ( -f $file && $file =~ m/\.(csv|storabe|couchdb|sql)$/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         return $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         return $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/.sql/i ) {
193                         $data = MojoFacets::Import::SQL->new( full_path => $full_path )->data;
194                 } elsif ( $full_path =~ m/.couchdb/i ) {
195                         $data = MojoFacets::Import::CouchDB->new( full_path => $full_path )->data;
196                 } elsif ( $full_path =~ m/.storable/ ) {
197                         warn "open $full_path ", -s $full_path, " bytes";
198                         open(my $pipe, "<", $full_path) || die $!;
199                         while ( my $o = eval { Storable::fd_retrieve $pipe } ) {
200                                 if ( defined $o->{item} ) {
201                                         push @{ $data->{items} }, $o->{item};
202                                 } else {
203                                         warn "SKIP ",dump($o);
204                                 }
205                         }
206                         close($pipe);
207                         warn "loaded ", $#{ $data->{items} } + 1, " items from $full_path\n";
208                         $data->{generated}++;
209                 } else {
210                         $data = MojoFacets::Import::File->new( full_path => $full_path, path => $path )->data;
211                 }
212         } elsif ( -d $full_path && $full_path =~ m/.html/ ) {
213                 $data = MojoFacets::Import::HTMLTable->new( dir => $full_path )->data;
214         } else {
215                 die "can't load $full_path";
216         }
217
218         my @header;
219
220         if ( defined $data->{header} ) {
221                 if ( ref $data->{header} eq 'ARRAY' ) {
222                         @header = @{ $data->{header} };
223                 } else {
224                         warn "header not array ", dump( $data->{header} );
225                 }
226         }
227
228         my $stats = __stats( $data->{items} );
229
230         @header =
231                 sort { $stats->{$b}->{count} <=> $stats->{$a}->{count} }
232                 grep { defined $stats->{$_}->{count} } keys %$stats
233                 unless @header;
234
235         my $info = {
236                 header => [ @header ],
237                 stats  => $stats,
238                 full_path => $full_path,
239                 size => -s $full_path,
240                 mtime => (stat($full_path))[9],
241                 data => $data,
242                 defined $data->{generated} ? ( generated => 1 ) : (),
243         };
244
245         $loaded->{ $path } = $info;
246         $self->_save( $path ) unless $info->{generated};
247
248 }
249
250
251 sub load {
252         my $self = shift;
253
254         my $path = $self->param('path') || return $self->redirect_to( '/data/index' );
255
256         my @paths = $self->param('paths');
257         warn "# paths ", dump @paths;
258
259         foreach my $p ( keys %$loaded ) {
260                 next if grep { /^\Q$p\E$/ } @paths;
261                 warn "remove $p from memory\n";
262                 delete $loaded->{$p};
263         }
264
265         $self->_load_path( $_ ) foreach @paths;
266
267         warn "# path $path\n";
268         $self->_load_path( $path );
269
270         $self->session( 'path' => $path );
271
272         my $redirect_to = '/data/items';
273
274         $self->session( 'header' => $loaded->{$path}->{header} );
275         if ( ! defined $loaded->{$path}->{columns} ) {
276                 my $columns_path = $self->_permanent_path( 'columns' );
277                 if ( -e $columns_path ) {
278                         my @columns = map { s/[\r\n]+$//; $_ } read_file $columns_path, binmode => ':utf8';
279                         $loaded->{$path}->{columns} = [ @columns ];
280                         warn "# columns_path $columns_path ",dump(@columns);
281                 } else {
282                         $loaded->{$path}->{columns} = $loaded->{$path}->{header}
283                 }
284
285                 $redirect_to = '/data/columns';
286         }
287         $self->session( 'columns' => $loaded->{$path}->{columns} );
288         $self->session( 'order'   => $loaded->{$path}->{columns}->[0] );
289         return $self->redirect_to( $redirect_to );
290 }
291
292
293 sub _loaded {
294         my ( $self, $name ) = @_;
295         my $path = $self->session('path') || $self->param('path');
296         return $self->redirect_to('/data/index') unless $path;
297
298         if ( defined $loaded->{$path}->{modified} && $loaded->{$path}->{modified} > 1 ) {
299                 my $caller = (caller(1))[3];
300                 if ( $caller =~ m/::edit/ ) {
301                         warn "rebuild stats for $path ignored caller $caller\n";
302                 } else {
303                         warn "rebuild stats for $path FORCED by modified caller $caller\n";
304 #                       $loaded->{$path}->{stats} = __stats( $loaded->{$path}->{data}->{items} );
305                         $loaded->{$path}->{rebuild_stats} = 1;
306                         $loaded->{$path}->{modified} = 1;
307                 }
308         }
309
310         if ( defined $loaded->{$path}->{rebuild_stats} ) {
311                 warn "rebuild_stats $path";
312                 $loaded->{$path}->{stats} = __stats( $loaded->{$path}->{data}->{items} );
313                 delete $loaded->{$path}->{rebuild_stats};
314         }
315
316         if ( ! defined $loaded->{$path}->{$name} ) {
317                 warn "$path $name isn't loaded\n";
318                 $self->_load_path( $path );
319                 if ( ! defined $loaded->{$path}->{stats} ) {
320                         warn "rebuild stats for $path\n";
321                         $loaded->{$path}->{stats} = __stats( $loaded->{$path}->{data}->{items} );
322                 }
323                 if ( ! defined $loaded->{$path}->{$name} ) {
324                         warn "MISSING $name for $path\n";
325                         return $self->redirect_to('/data/index')
326                 }
327         }
328
329         $self->session( 'modified' => $loaded->{$path}->{modified} );
330
331         return $loaded->{$path}->{$name};
332 }
333
334
335 sub _checked {
336         my $self = shift;
337         my $checked;
338         $checked->{$_}++ foreach @_;
339 #       warn "# _checked ",dump($checked);
340         return $checked;
341 }
342
343 sub _permanent_path {
344         my $self = shift;
345         my $path = $self->_param_or_session('path');
346         $self->app->home->rel_dir('data') . '/' . join('.', $path, @_);
347 }
348
349 sub __unac {
350         my $n = shift;
351         $n = unac_string($n);
352         $n =~ s/\W+/_/g;
353         return $n;
354 }
355
356 sub _column_from_unac {
357         my ($self,$name) = @_;
358
359         my $stats = $self->_loaded('stats');
360         my $cols_norm = { map { __unac( $_ ) => $_ } keys %$stats };
361
362         $cols_norm->{$name} || die "can't find column $name in ", dump($cols_norm);
363 }
364
365 sub _export_path {
366         my $self = shift;
367         my $path = $self->_param_or_session('path');
368         if ( ! $path ) {
369                 warn "no path in param or session";
370                 return;
371         }
372         my $dir = $self->app->home->rel_dir('public') . "/export/$path";
373         mkpath $dir unless -e $dir;
374         my $name = __export_path_name( $path, @_ );
375         my $full = $dir . '/' . $name;
376         $full =~ s/\/+$// if -d $full; # strip trailing slash for dirs
377         return $full;
378 }
379
380 sub __export_path_name {
381         my $max_length = 80;
382
383         my $path = shift;
384         my $name = join('.', map { __unac($_) } @_ );
385         if ( length($name) > $max_length ) {
386                 $name = substr($name,0,$max_length) . Digest::MD5::md5_hex substr($name,$max_length);
387         }
388         return $name;
389 }
390
391 sub columns {
392     my $self = shift;
393
394         my $view_path = $self->_permanent_path( 'views' );
395
396         if ( $self->param('columns') ) {
397                 my @columns = $self->_param_array('columns');
398                 write_file( $self->_permanent_path( 'columns' ), { binmode => ':utf8' }, map { "$_\n" } @columns );
399                 if ( my $view = $self->param('view') ) {
400                         mkdir $view_path unless -e $view_path;
401                         write_file( "$view_path/$view", { binmode => ':utf8' }, map { "$_\n" } @columns );
402                         warn "view $view_path/$view ", -s "$view_path/$view", " bytes\n";
403                 }
404
405                 return $self->redirect_to('/data/items');
406
407         } elsif ( ! $self->session('header') ) {
408                 return $self->redirect_to('/data/load');
409         }
410
411         if ( my $id = $self->param('id') ) {
412                 my $view_full = "$view_path/$id";
413                 if ( -f $view_full ) {
414                         my @columns = map { chomp; $_ } read_file $view_full, binmode => ':utf8';
415                         warn "view $view_full loaded ", dump @columns;
416                         $self->session( 'columns' => [ @columns ] );
417                         return $self->redirect_to('/data/items');
418                 }
419         }
420
421         my $stats = $self->_loaded( 'stats' );
422
423         my @columns;
424         @columns = grep { defined $stats->{$_}->{count} } @{ $self->session('columns') } if $self->session('columns');
425
426         foreach my $c ( @{ $self->session( 'header' ) } ) {
427                 push @columns, $c unless grep { /^\Q$c\E$/ } @columns;
428         }
429
430         my @views;
431         if ( -d $view_path ) {
432                 @views = map { s{^\Q$view_path\E/*}{}; $_ } glob "$view_path/*";
433                 warn "# views ",dump @views;
434         }
435
436     $self->render(
437                 message => 'Select columns to display',
438                 stats => $stats,
439                 columns => \@columns,
440                 checked => $self->_checked( $self->_param_array('columns') ),
441                 views => \@views,
442         );
443 }
444
445 sub _param_array {
446     my ($self,$name) = @_;
447
448         my @array = $self->param($name);
449         my $path  = $self->session('path');
450
451         if ( @array ) {
452                 $self->session($name => [ @array ]);
453                 $loaded->{$path}->{$name} = [ @array ];
454         } elsif ( my $session = $self->session($name) ) {
455                 if ( ref $session eq 'ARRAY' ) {
456                         @array = @$session;
457                 } else {
458                         die "$name not array ",dump($session);
459                 }
460         }
461         #warn "# $name ",dump @array;
462         return @array;
463 }
464
465 sub _param_scalar {
466     my ($self,$name,$default) = @_;
467
468         my $scalar = $self->param($name);
469
470         if ( defined $scalar ) {
471                 $self->session($name => $scalar);
472         } else {
473                 $scalar = $self->session($name);
474         }
475
476         if ( ! defined $scalar ) {
477                 $scalar = $default;
478                 die "no default for $name" unless defined $scalar;
479                 $self->session($name => $scalar);
480         }
481
482         warn "# _param_scalar $name ",dump $scalar;
483         return $scalar;
484 }
485
486 sub filter {
487         my $self = shift;
488
489         my $name = $self->param('filter_name') || die "name?";
490         my @vals = $self->param('filter_vals');
491
492         $self->_remove_filter( $name );
493         if ( @vals ) {
494                 $self->_filter_on_data( $name, @vals );
495                 if ( my $permanent = $self->param('_permanent') ) {
496                         my $permanent_path = $self->_export_path( 'filter', $name, $permanent );
497                         write_file $permanent_path, { binmode => ':utf8' }, map { "$_\n" } @vals;
498                         warn "permanent filter $permanent_path ", -s $permanent_path;
499                 }
500         }
501
502         $self->session( 'offset' => 0 );
503
504         return $self->redirect_to('/data/items');
505 }
506
507 sub _filter_on_data {
508         my ( $self, $name, @vals ) = @_;
509
510         my $path = $self->session('path');
511
512         if ( ! defined $loaded->{$path}->{stats}->{ $name } ) {
513                 warn "filter $name not found in data set";
514                 return;
515         }
516
517         $filters->{$name} = [ @vals ];
518         warn "_filter_on_data $name ", $#vals + 1, " values on $path\n";
519
520         my $filter_hash;
521         $filter_hash->{$_}++ foreach @vals;
522
523         #warn "# filter_hash ",dump( $filter_hash );
524
525         my $items = $self->_loaded('data')->{items};
526
527         my $include_missing = defined $filter_hash->{_missing};
528         my $filtered_items;
529
530         foreach my $i ( 0 .. $#$items ) {
531
532                 if ( defined $items->[$i]->{$name} ) {
533                         my $row = $items->[$i]->{$name};
534                         $row = [ $row ] unless ref $row eq 'ARRAY'; # FIXME probably wrong place
535                         foreach my $v ( @$row ) {
536                                 if ( defined $filter_hash->{ $v } ) {
537                                         $filtered_items->{$i}++;
538                                 }
539                         }
540                 } elsif ( $include_missing ) {
541                         $filtered_items->{$i}++;
542                 }
543         }
544
545         #warn "# filter $name ",dump($filtered_items);
546
547         $loaded->{$path}->{filters}->{$name} = $filtered_items;
548         warn "filter $name with ", scalar keys %$filtered_items, " items created\n";
549 }
550
551
552 sub _current_filters {
553         my $self = shift;
554         my $current_filters;
555         my $stats = $self->_loaded('stats');
556
557         $current_filters->{ $_ } = $filters->{ $_ }
558                 foreach ( grep { defined $filters->{ $_ } } keys %$stats )
559         ;
560         warn "# _current_filters ",dump( keys %$current_filters );
561         return $current_filters;
562 }
563
564 sub _data_sorted_by {
565         my ( $self, $order ) = @_;
566
567         my $path = $self->session('path');
568
569         warn "_data_sorted_by $order from $path";
570
571         if ( defined $loaded->{$path}->{sorted}->{$order} ) {
572                 return $loaded->{$path}->{sorted}->{$order};
573         }
574
575         my $data = $self->_loaded( 'data' );
576         my $numeric = $self->_is_numeric($order);
577         my $missing = $numeric ? 0 : '';
578         no warnings qw(numeric);
579         my $nr = 0;
580         my @sorted = map {
581                 $_->[0]
582         } sort {
583                 if ( $numeric ) {
584                         $a->[1] <=> $b->[1]
585                 } else {
586                         $a->[1] cmp $b->[1]
587                 }
588         } map {
589                 my $v;
590                 if ( ! exists $_->{$order} ) {
591                         $v = $missing;
592                 } elsif ( ref $_->{$order} eq 'ARRAY' ) {
593                         $v = join('', @{$_->{$order}});
594                 } else {
595                         $v = $_->{$order};
596                 }
597                 [ $nr++, $v ]
598         } @{ $data->{items} }
599         ;
600
601         warn "sorted: $order numeric: $numeric items: ", $#sorted + 1, "\n";
602         #warn "# sorted ",dump( @sorted );
603
604         $loaded->{$path}->{sorted}->{$order} = [ @sorted ];
605 }
606
607
608 sub __all_filters {
609         my $order = pop @_;
610         join(',', sort(@_), 'order', $order);
611 }
612
613 our ($out, $key,$value);
614
615 sub __commit_path_code {
616         my ( $path, $i, $code, $commit_changed ) = @_;
617
618         my $items = $loaded->{$path}->{data}->{items} || die "no items for $path";
619         my $row = $items->[$i];
620         my $update;
621         eval $code;
622         foreach ( keys %$update ) {
623                 $$commit_changed->{$_}++;
624                 $loaded->{$path}->{data}->{items}->[$i]->{$_} = $update->{$_};
625         }
626         #warn "__commit_path_code $path $i ",dump( $update );
627 }
628
629 # uses templates/admin.html.ep
630 sub _switch_dataset {
631         my $self = shift;
632
633         my $datasets;
634
635         foreach my $path ( keys %$loaded ) {
636                 next unless exists $loaded->{$path}->{data};
637                 push @$datasets, $path;
638         }
639
640         warn "# datasets ",dump($datasets);
641
642         $self->stash( 'datasets' => $datasets );
643 }
644
645 sub items {
646         my $self = shift;
647
648         $self->_switch_dataset;
649
650         if ( my $show = $self->param('id') ) {
651                 $self->param('show', $show);
652                 warn "show $show\n";
653         }
654
655         my $path = $self->_param_scalar('path');
656
657         my @columns = $self->_param_array('columns');
658         return $self->redirect_to('/data/columns') unless @columns;
659         my $order   = $self->_param_scalar('order', $columns[0]);
660         my $sort    = $self->_param_scalar('sort', 'a');
661         my $offset  = $self->_param_scalar('offset', 0);
662         my $limit   = $self->_param_scalar('limit', 20);
663         $self->_param_scalar('show', 'table');
664
665         # fix offset when changing limit
666         $offset = int( $offset / $limit ) * $limit;
667
668         if ( ! grep { /^\Q$order\E$/ } @columns ) {
669                 $order = $columns[0];
670                 $self->session( order => $order );
671         }
672         my $sorted = $self->_data_sorted_by( $order );
673
674         my @filter_names;
675         if ( $filters ) {
676                 foreach my $name ( keys %$filters ) {
677                         if ( ! defined $loaded->{$path}->{stats}->{ $name } ) {
678                                 warn "skip filter $name not found in $path\n";
679                                 next;
680                         }
681                         push @filter_names, $name;
682                 }
683                 warn "filter_names ",dump( @filter_names );
684                 foreach my $name ( @filter_names ) {
685                         next if ref $loaded->{$path}->{filters}->{$name} eq 'ARRAY';
686                         $self->_filter_on_data( $name, @{ $filters->{$name} } );
687                 }
688         }
689
690         my $all_filters = __all_filters( @filter_names,$order );
691
692 #       warn "# all_filters $all_filters ", dump( $loaded->{$path}->{filtered}->{$all_filters} );
693
694         if ( ! defined $loaded->{$path}->{filtered}->{$all_filters} ) {
695
696                 my $path_filters = $loaded->{$path}->{filters};
697
698                 warn "create combined filter for $all_filters from ", $#$sorted + 1, " items\n";
699
700                 my @filtered;
701                 foreach my $i ( 0 .. $#$sorted ) {
702                         my $pos = $sorted->[$i];
703
704                         if ( $#filter_names == -1 ) {
705                                 push @filtered, $pos;
706                                 next;
707                         }
708
709                         my $skip = 0;
710                         foreach ( @filter_names ) {
711                                 $skip ||= 1 if ! defined $path_filters->{$_}->{$pos};
712                         }
713                         next if $skip;
714
715                         push @filtered, $pos;
716                 }
717
718                 $loaded->{$path}->{filtered}->{$all_filters} = [ @filtered ];
719         }
720
721         my $filtered = $loaded->{$path}->{filtered}->{$all_filters}
722                 if defined $loaded->{$path}->{filtered}->{$all_filters};
723
724         warn "all_filters $all_filters produced ", $#$filtered + 1, " items\n" if $filtered;
725
726         my $data = $self->_loaded('data');
727
728         my $code = $self->_param_scalar('code','');
729         $code =~ s{\r}{}gs;
730         $code =~ s{\n+$}{\n}s;
731
732         my $commit = $self->param('commit');
733         my $test = $self->param('test');
734
735         my $commit_changed;
736
737         if ( $code && ( $test || $commit ) ) {
738                 # XXX find columns used in code snippet and show them to user
739                 my $order = 0;
740                 foreach my $column ( $code =~ m/\$row->{([^}]+)}/g ) {
741                         if ( $column =~ s/^(['"])// ) {
742                                 $column =~ s/$1$//;
743                         }
744                         next if $column =~ m/\$/; # hide columns with vars in them
745                         $commit_changed->{$column} = 0;
746                 }
747         }
748
749         my $code_path = $self->app->home->rel_dir('public') . "/code";
750         if ( $commit ) {
751
752                 warn "# commit on ", $#$filtered + 1, " items:\n$code\n";
753                 ( $key, $value, $out ) = ( 'key', 'value' );
754                 foreach ( 0 .. $#$filtered ) {
755                         my $i = $filtered->[$_];
756                         __commit_path_code( $path, $i, $code, \$commit_changed );
757                 }
758
759                 # this might move before $out to recalculate stats on source dataset?
760                 __path_rebuild_stats( $path );
761                 my $c = { map { $_ => 1 } @columns };
762                 my @added_columns = sort grep { ! $c->{$_} } keys %$commit_changed;
763                 warn "# added_columns ",dump( @added_columns );
764                 unshift @columns, @added_columns;
765
766                 $loaded->{$path}->{columns} = [ @columns ];
767                 warn "# new columns ",dump( @columns );
768
769                 __invalidate_path_column( $path, $_ ) foreach keys %$commit_changed;
770
771                 $self->_save_change({
772                         path => $path,
773                         time => $self->param('time') || time(),
774                         user => $self->param('user') || $ENV{'LOGNAME'},
775                         code => $code,
776                         commit_changed => $commit_changed,
777                 });
778
779                 if ( my $description = $self->param('code_description') ) {
780                         my $depends = $self->param('code_depends') || die "no code_depends?";
781                         my $path = "$code_path/$depends.$description.pl";
782                         if ( -e $path && ! $self->param('overwrite') ) {
783                                 warn "# code $path not saved\n";
784                         } else {
785                                 write_file(  $path, { binmode => ':utf8' }, "$code\n" );
786                                 warn "code $path ", -s $path, " bytes saved\n";
787                         }
788                 }
789
790                 # remove console
791                 $code = '';
792                 if ( $out ) {
793                         my $commit_dataset = join('.'
794                                 , $self->param('code_depends')
795                                 , $self->param('code_description')
796                                 , time()
797                         );
798                         $key ||= 'key';
799                         $value ||= 'value';
800                         warn "key $key value $value";
801                         my $items;
802                         foreach my $n ( keys %$out ) {
803                                 my $i = { $key => [ $n ] };
804                                 my $ref = ref $out->{$n};
805                                 if ( $ref eq 'HASH' ) {
806                                         $i->{$_} = [ $out->{$n}->{$_} ] foreach keys %{ $out->{$n} };
807                                 } elsif ( $ref eq 'ARRAY' ) {
808                                         $i->{$_} = $out->{$n};
809                                 } elsif ( ! $ref ) {
810                                         $i->{$value} = [ $out->{$n} ];
811                                 } else {
812                                         $i->{_error} = [ dump($out->{$n}) ];
813                                 }
814                                 push @$items, $i;
815                         };
816                         undef $out;
817                         my $stats = __stats( $items );
818                         my @columns = grep { ! m/^\Q$key\E$/ } sort keys %$stats;
819                         unshift @columns, $key;
820
821                         $loaded->{$commit_dataset} = {
822                                 header => [ @columns ],
823                                 columns => [ @columns ],
824                                 mtime => time(),
825                                 data => { items => $items },
826                                 stats => $stats,
827                                 generated => 1,
828                         };
829                         warn "# loaded out ", dump( $loaded->{$commit_dataset} );
830                         $self->session('path', $commit_dataset);
831                         $self->session('columns', [ @columns ]);
832                         $self->session('order', $key);
833                         return $self->redirect_to('/data/items');
834                 }
835
836                 $self->session('columns', [ @columns ]);
837         }
838
839         my $sorted_items;
840         my $from_end = $sort eq 'd' ? $#$filtered : 0;
841         my $test_changed;
842         my ( $key, $value, $out ) = ( 'key', 'value' ); # XXX make local
843         foreach ( 0 .. $limit ) {
844                 my $i = $_ + $offset;
845                 last unless defined $filtered->[$i];
846                 $i = $from_end - $i if $from_end;
847                 my $id = $filtered->[$i];
848                 my $row = Storable::dclone $data->{items}->[ $id ];
849                 if ( $code && $test ) {
850                         my $update;
851                         eval $code;
852                         if ( $@ ) {
853                                 warn "ERROR evaling $@", dump($code);
854                                 $self->stash('eval_error', $@) if $@;
855                         } else {
856                                 warn "EVAL ",dump($update);
857                                 foreach ( keys %$update ) {
858                                         $test_changed->{$_}++;
859                                         $row->{$_} = $update->{$_};
860                                 }
861                         }
862                 }
863                 $row->{_row_id} ||= $id;
864                 push @$sorted_items, $row;
865         }
866
867         if ( $self->param('export') ) {
868                 my $export_path = $self->_export_path( 'items', @columns);
869                 open(my $fh, '>', $export_path) || die "ERROR: can't open $export_path: $!";
870                 print $fh "#",join("\t",@columns),"\n";
871                 foreach my $f ( 0 .. $#$filtered ) {
872                         print $fh join("\t", map {
873                                 my $i = $data->{items}->[ $filtered->[$f] ];
874                                 my $v = '\N';
875                                 if ( ! defined $i->{$_} ) {
876                                         # nop
877                                 } elsif ( ref $i->{$_} eq 'ARRAY' ) {
878                                         $v =join(',', @{ $i->{$_} });
879                                         $v = '\N' if length($v) == 0;
880                                 } elsif ( ! ref $i->{$_} ) {
881                                         $v = $i->{$_};
882                                 } else {
883                                         $v = dump $i->{$_};
884                                 }
885                                 $v;
886                         } @columns),"\n";
887                 }
888                 close($fh);
889                 warn "export $export_path ", -s $export_path, " bytes\n";
890         }
891
892         warn "# test_changed ",dump( $test_changed );
893         my $c = { map { $_ => 1 } @columns };
894         my @added_columns = sort grep { ! $c->{$_} } keys %$test_changed;
895         unshift @columns, @added_columns;
896
897         warn "# sorted_items ", $#$sorted_items + 1, " offset $offset limit $limit order $sort";
898
899         my $depends_on;
900         my $tmp = $code; $tmp =~ s/\$row->{(['"]?)(\w+)\1/$depends_on->{$2}++/gse;
901         warn "# depends_on ",dump $depends_on;
902
903         my $test_added = Storable::dclone $test_changed;
904         delete $test_added->{$_} foreach keys %$depends_on;
905
906         my $code_depends = $self->param('code_depends')
907         || join(',', keys %$depends_on);
908
909         my $code_description = $self->param('code_description') ||
910         join(',', keys %$test_added);
911
912         $code_depends ||= $code_description; # self-modifing
913         if ( ! $code_depends && $out ) {
914                 $code_depends = $key;
915                 $code_description = $value;
916         }
917
918         warn "# test_changed ",dump( $test_changed, $code_depends, $code_description );
919
920         $self->render(
921                 order => $order,
922                 offset => $offset,
923                 limit => $limit,
924                 sorted => $sorted_items,
925                 columns => [ @columns ],
926                 rows => $#$filtered + 1,
927                 numeric => { map { $_, $self->_is_numeric($_) } @columns },
928                 unique  => { map { $_, $self->_is_unique( $_) } @columns },
929                 filters => $self->_current_filters,
930                 code => $code,
931                 cols_changed => $commit ? $commit_changed : $test_changed,
932                 code_depends => $code_depends,
933                 code_description => $code_description,
934                 code_path => $code_path,
935                 out => $out,
936         );
937
938 }
939
940
941 sub order {
942         my $self = shift;
943         $self->session('order', $self->param('order'));
944         $self->session('sort', $self->param('sort'));
945         return $self->redirect_to('/data/items');
946 }
947
948 sub _is_numeric {
949         my ( $self, $name ) = @_;
950
951         my $stats = $self->_loaded( 'stats' );
952
953         # sort facet numerically if more >50% elements are numeric
954         my $count = $stats->{$name}->{count};
955         $count   -= $stats->{$name}->{empty} if defined $stats->{$name}->{empty};
956         defined $stats->{$name}->{numeric} &&
957                 $stats->{$name}->{numeric} > $count / 2;
958 }
959
960 sub _is_unique {
961         my ( $self, $name ) = @_;
962         my $stats = $self->_loaded( 'stats' );
963         defined $stats->{$name}->{unique};
964 }
965
966 sub _remove_filter {
967         my ($self,$name) = @_;
968         warn "_remove_filter $name\n";
969
970         my $path = $self->session('path');
971
972         delete $filters->{$name};
973         delete $loaded->{$path}->{filters}->{$name};
974         warn "filters left: ", keys %{ $loaded->{$path}->{filters} };
975
976         foreach (
977                         grep { /\Q$name\E/ }
978                         keys %{ $loaded->{$path}->{filtered} }
979         ) {
980                 delete $loaded->{$path}->{filtered}->{$_};
981                 warn "remove filtered cache $_";
982         }
983 }
984
985 sub facet {
986         my $self = shift;
987
988         my $path = $self->session('path') || return $self->redirect_to( '/data/index' );
989
990         if ( my $name = $self->param('remove') ) {
991                 $self->_remove_filter( $name );
992                 return $self->redirect_to( '/data/items' );
993         }
994
995         my $facet;
996         my $name = $self->param('name') || die "no name";
997
998         my $all = $self->_param_scalar('all', 1);
999         my $data = $self->_loaded('data');
1000
1001         my $filters = $self->_current_filters;
1002         my $all_filters = __all_filters( keys %$filters,$self->session('order') );
1003         my $filtered = $loaded->{$path}->{filtered}->{$all_filters}
1004                 if defined $loaded->{$path}->{filtered}->{$all_filters};
1005
1006         if ( ! $filtered || $all ) {
1007                 $filtered = [ 0 .. $#{ $data->{items} } ];
1008                 warn "filter all values\n";
1009         } else {
1010                 warn "filter using $all_filters\n";
1011         }
1012
1013         foreach my $i ( @$filtered ) {
1014                 my $item = $data->{items}->[$i];
1015                 if ( ! exists $item->{$name} ) {
1016                         $facet->{ _missing }++;
1017                 } elsif ( ref $item->{$name} eq 'ARRAY' ) {
1018                         $facet->{$_}++ foreach @{ $item->{$name} };
1019                 } else {
1020                         $facet->{ $item->{$name} }++;
1021                 }
1022         }
1023
1024         my $checked_values = $self->_checked( @{ $filters->{$name} } ) if defined $filters->{$name};
1025
1026         if ( my $code = $self->param('code') ) {
1027                 my $out;
1028                 foreach my $value ( keys %$facet ) {
1029                         my $count = $facet->{$value};
1030                         my $checked = $checked_values->{$value};
1031                         eval $code;
1032                         if ( $@ ) {
1033                                 $out = $@;
1034                                 warn "ERROR: $@\n$code\n";
1035                                 last;
1036                         } elsif ( $checked != $checked_values->{$value} ) {
1037                                 warn "checked $value $count -> $checked\n";
1038                                 $checked_values->{$value} = $checked;
1039                         }
1040                 }
1041                 warn "out ",dump( $out );
1042                 $self->stash( out => $out );
1043         }
1044
1045 #       warn "# facet $name ",dump $facet;
1046
1047         my @facet_names =
1048                   $all                      ? keys %$facet
1049                 : defined $filters->{$name} ? @{ $filters->{$name} }
1050                 : keys %$facet;
1051
1052         my $numeric = $self->_is_numeric($name);
1053
1054         my $sort = $self->param('sort');
1055         # sort numeric facets with more than 5 values ascending
1056         $sort ||= $numeric && $#facet_names > 4 ? 'a' : 'c';
1057
1058         @facet_names = sort {
1059                 my $result;
1060                 if ( $sort eq 'a' ) {
1061                         $result = $numeric ? $a <=> $b : lc $a cmp lc $b;
1062                 } elsif ( $sort eq 'd' ) {
1063                         $result = $numeric ? $b <=> $a : lc $b cmp lc $a;
1064                 } elsif ( $sort eq 'c' ) {
1065                         $result = ( $facet->{$b} || -1 ) <=> ( $facet->{$a} || -1 )
1066                 } else {
1067                         warn "unknown sort: $sort";
1068                         $result = $a cmp $b;
1069                 }
1070                 $result = $a cmp $b unless defined $result; # FIXME cludge for numeric facets with invalid data
1071                 $result;
1072         } @facet_names;
1073
1074         $self->render( name => $name, facet => $facet, checked => $checked_values,
1075                 facet_names => \@facet_names, sort => $sort, numeric => $numeric,
1076         );
1077 }
1078
1079
1080 sub __invalidate_path_column {
1081         my ( $path, $name ) = @_;
1082
1083         if ( defined $loaded->{$path}->{sorted}->{$name} ) {
1084                 delete $loaded->{$path}->{sorted}->{$name};
1085                 warn "# invalidate $path sorted $name\n";
1086         }
1087
1088         foreach ( grep { m/\Q$name\E/ } keys %{ $loaded->{$path}->{filtered} } ) {
1089                 delete $loaded->{$path}->{filtered}->{$_};
1090                 warn "# invalidate $path filtered $_\n";
1091         }
1092 }
1093
1094 sub __path_modified {
1095         my ( $path, $value ) = @_;
1096         $value = 1 unless defined $value;
1097         
1098         $loaded->{$path}->{modified}  = $value;
1099
1100         warn "# __path_modified $path $value\n";
1101 }
1102
1103 sub __path_rebuild_stats { $loaded->{ $_[0] }->{rebuild_stats} = 1 };
1104
1105 sub _save_change {
1106         my ($self,$change) = @_;
1107
1108         my $change_path = $self->_permanent_path( 'changes' );
1109         mkdir $change_path unless -d $change_path;
1110         $change_path .= '/' . $change->{time};
1111         store $change, $change_path;
1112         utime $change->{time}, $change->{time}, $change_path;
1113         warn "_save_change $change_path ", dump($change);
1114 }
1115
1116 sub edit {
1117         my $self = shift;
1118         my $new_content = $self->param('new_content');
1119         $new_content  ||= $self->param('content'); # backward compatibility with old actions
1120
1121         my $i = $self->param('_row_id');
1122         die "invalid _row_id ",dump($i) unless $i =~ m/^\d+$/;
1123         my $path = $self->param('path') || die "no path";
1124         my $name = $self->param('name') || die "no name";
1125         my $status = 200; # 200 = OK, 201 = Created
1126
1127         my $data = $self->_loaded('data');
1128
1129         if ( defined $loaded->{$path}->{data}->{items}->[$i] ) {
1130                 $new_content =~ s/^\s+//s;
1131                 $new_content =~ s/\s+$//s;
1132                 my $v;
1133                 if ( $new_content =~ /\xB6/ ) { # para
1134                         $v = [ split(/\s*\xB6\s*/, $new_content) ];
1135                 } else {
1136                         $v = [ $new_content ];
1137                 }
1138
1139                 my $old = dump $loaded->{$path}->{data}->{items}->[$i]->{$name};
1140                 my $new = dump $v;
1141                 if ( $old ne $new
1142                         && ! ( $old eq 'undef' && length($new_content) == 0 ) # new value empty, previous undef
1143                 ) {
1144                         $self->_save_change({
1145                                 path => $path,
1146                                 column => $name,
1147                                 pos => $i,
1148                                 old => $loaded->{$path}->{data}->{items}->[$i]->{$name},
1149                                 new => $v,
1150                                 time => $self->param('time') || time(),
1151                                 user => $self->param('user') || $ENV{'LOGNAME'},
1152                                 unique => {
1153                                         map { $_ => $loaded->{$path}->{data}->{items}->[$i]->{$_}->[0] }
1154                                         grep { defined $loaded->{$path}->{stats}->{$_}->{unique} }
1155                                         keys %{ $loaded->{$path}->{stats} }
1156                                 },
1157                         });
1158
1159                         warn "# change $path $i $old -> $new\n";
1160                         $loaded->{$path}->{data}->{items}->[$i]->{$name} = $v;
1161
1162                         __invalidate_path_column( $path, $name );
1163
1164                         $status = 201; # created
1165                         __path_rebuild_stats( $path );
1166         
1167                         $new_content = join("\xB6",@$v);
1168
1169                 } else {
1170                         warn "# unchanged $path $i $old\n";
1171                         $status = 304;
1172                 }
1173         } else {
1174                 $new_content = "$path $i $name doesn't exist\n";
1175                 $status = 404;
1176         }
1177
1178         warn "# edit $status ", dump $new_content;
1179
1180         $self->render(
1181                 status => $status,
1182                 new_content => scalar $new_content,
1183         );
1184 }
1185
1186
1187 sub save {
1188         my $self = shift;
1189         my $path = $self->_param_or_session('path');
1190         my $dump_path = $self->_save( $path );
1191         __path_modified( $path, 0 );
1192
1193         return $self->redirect_to( '/data/items' );
1194 }
1195
1196 sub export {
1197         my $self = shift;
1198
1199         my $dir = $self->app->home->rel_dir('public');
1200
1201         if ( my $import = $self->param('import') ) {
1202
1203                 if ( $import =~ m{/filter\.(.+?)\..+} ) {
1204                         my $name = $self->_column_from_unac( $1 );
1205
1206                         my @vals = map { chomp; $_ }
1207                                 read_file "$dir/export/$import", binmode => ':utf8';
1208
1209                         $self->_remove_filter( $name );
1210                         $self->_filter_on_data( $name, @vals );
1211                         $self->session( 'offset' => 0 );
1212                         return $self->redirect_to('/data/items');
1213                 } else {
1214                         warn "UNKNOWN IMPORT $import";
1215                 }
1216         }
1217
1218         if ( my $remove = $self->param('remove') ) {
1219                 my $path = "$dir/export/$remove";
1220                 unlink $path if -e $path;
1221                 $path .= '.png';
1222                 unlink $path if -e $path;
1223         }
1224
1225         my $path = $self->_export_path || return $self->redirect_to('/data/index');
1226
1227         my @files = grep { ! /\.png$/ } glob "$path/*";
1228         my $mtime = { map { $_ => (stat($_))[9] } @files };
1229         @files = sort { $mtime->{$b} <=> $mtime->{$a} } @files;
1230         $self->render( export => [ @files ] );
1231 }
1232
1233 sub __loaded_paths {
1234         return
1235                 grep { defined $loaded->{$_}->{data} }
1236                 keys %$loaded;
1237 }
1238
1239 sub remove {
1240         my $self = shift;
1241         my $path = $self->param('path');
1242         if ( $path =~ m{^/tmp/mojo_facets\.} ) {
1243                 unlink $path;
1244                 warn "# unlink $path";
1245                 if ( my $name = $self->param('name') ) {
1246                         delete $loaded->{$name};
1247                         warn "# remove $name from memory";
1248                 }
1249         } else {
1250                 warn "WARNING: $path unlink ignored";
1251         }
1252         return $self->redirect_to( '/data/load' );
1253 }
1254
1255 1;