fix timefmt load into session position
[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); # broken with Mojo::JSON, see https://rt.cpan.org/Public/Bug/Display.html?id=86592
9 use Data::Dumper;
10 use subs 'dump';
11 sub dump { Dumper(@_) };
12
13 use File::Slurp;
14 use Encode;
15 use locale;
16 use File::Find;
17 use Storable;
18 use Time::HiRes qw(time);
19 use File::Path qw(mkpath);
20 use Text::Unaccent::PurePerl;
21 use Digest::MD5;
22 use Statistics::Descriptive;
23
24 our $imports;
25 foreach my $module ( glob('lib/MojoFacets/Import/*.pm') ) {
26         $module =~ s{lib/(\w+)/(\w+)/(.*)\.pm}{$1::$2::$3};
27         eval "use $module";
28         die "$module: $!" if $!;
29         my ( $ext, $priority ) = $module->ext;
30         $imports->{$priority || 'file'}->{$ext} = $module;
31         warn "# import $ext $module\n";
32 }
33
34 warn "# import loaded ",dump( $imports );
35
36 sub import_module {
37         my $full_path = shift;
38
39 #       warn "# import_module $full_path\n";
40
41         return if $full_path =~ m/\.columns$/;
42
43         foreach my $ext ( keys %{ $imports->{file} } ) {
44                 if ( -f $full_path && $full_path =~ m/$ext/i ) {
45                         return $imports->{file}->{$ext};
46                         last;
47                 }
48         }
49
50         foreach my $ext ( keys %{ $imports->{directory} } ) {
51                 if ( -f $full_path && $full_path =~ m/$ext/i ) {
52                         return $imports->{directory}->{$ext};
53                         last;
54                 }
55         }
56 }
57
58 our $loaded;
59 our $filters;
60
61 sub index {
62         my $self = shift;
63
64         my $data_dir = $self->app->home->rel_file('data');
65         die "no data dir $data_dir" unless -d $data_dir;
66
67         my @files;
68         my $changes;
69
70         find( sub {
71                 my $file = $File::Find::name;
72
73                 next if $file =~ m/.timefmt$/;
74
75                 if ( -f $file && $file =~ m/([^\/]+)\.changes\/(\d+[\.,]\d+.+)/ ) {
76                         push @{ $changes->{$1} }, $2
77                 } elsif ( import_module( $file ) ) {
78                         my $mtime = (stat($file))[9]; # mtime
79                         $file =~ s/$data_dir\/*//;
80                         push @files, $file;
81                         $loaded->{$file}->{mtime} ||= $mtime;
82                 } else {
83                         #warn "IGNORE: $file\n";
84                 }
85         }, $data_dir);
86
87         no warnings qw(uninitialized); # mtime
88         @files = sort { $loaded->{$b}->{mtime} <=> $loaded->{$a}->{mtime} || lc $a cmp lc $b } @files,
89                         grep { defined $loaded->{$_}->{generated} } keys %$loaded;
90
91         my $size;
92         $size->{$_} = -s "$data_dir/$_" foreach @files;
93
94         $self->render(
95                 files => [ @files ],
96                 size => $size,
97                 loaded => $loaded,
98                 filters => $filters,
99                 dump_path => { map { $_ => $self->_dump_path($_) } @files },
100                 changes => $changes,
101         );
102 }
103
104 sub _dump_path {
105         my ( $self, $name ) = @_;
106         my $dir = $self->app->home->rel_file('data');
107         $name =~ s/^$dir//;
108         $name =~ s/\/+/_/g;
109         return '/tmp/mojo_facets.' . $name . '.storable';
110 }
111
112 sub _save {
113         my ( $self, $path ) = @_;
114
115         my $dump_path = $self->_dump_path( $path );
116         my $first_load = ! -e $dump_path;
117         warn "save loaded to $dump_path";
118         my $info = $loaded->{$path};
119         store $info, $dump_path;
120
121         if ( $first_load ) {
122                 my $mtime = $loaded->{$path}->{mtime};
123                 utime $mtime, $mtime, $dump_path;
124                 warn "sync time to $path at $mtime\n";
125         }
126
127         warn $dump_path, ' ', -s $dump_path, " bytes\n";
128         return $dump_path;
129 }
130
131
132 sub __stats {
133
134         my $stats;
135
136         my $nr_items = $#{ $_[0] } + 1;
137
138         warn "__stats $nr_items\n";
139
140 #       foreach my $e ( @{ $_[0] } ) {
141         foreach my $i ( 0 .. $#{$_[0]} ) {
142                 print STDERR " $i" if $i % 5000;
143                 my $e = $_[0]->[$i];
144                 foreach my $n ( keys %$e ) {
145                         $stats->{$n}->{count}++;
146                         my @v;
147                         if ( ref $e->{$n} eq 'ARRAY' ) {
148                                 $stats->{$n}->{array} += $#{ $e->{$n} } + 1;
149                                 @v = @{ $e->{$n} };
150                         } else {
151                                 @v = ( $e->{$n} );
152                         }
153
154                         foreach my $x ( @v ) {
155                                 if ( ! defined $x ) { # FIXME really null
156                                         $stats->{$n}->{empty}++;
157                                         next;
158                                 }
159
160                                 $stats->{$n}->{numeric}++
161                                         if $x =~ m/^[-+]?([0-9]*\.[0-9]+|[0-9]+)$/;
162                                 $stats->{$n}->{empty}++
163                                         if length $x == 0; # faster than $x =~ m/^\s*$/;
164                         }
165
166                 }
167         }
168
169         foreach my $n ( keys %$stats ) {
170                 my $s = $stats->{$n};
171                 next unless defined $s->{array};
172                 if ( $s->{array} == $s->{count} ) {
173                         delete $s->{array};
174                         if ( $s->{count} == $nr_items ) {
175                                 warn "check $n for uniqeness\n";
176                                 my $unique;
177                                 foreach my $e ( @{ $_[0] } ) {
178                                         if ( ++$unique->{ $e->{$n}->[0] } == 2 ) {
179                                                 $unique = 0;
180                                                 last;
181                                         }
182                                 }
183                                 if ( $unique ) {
184                                         $stats->{$n}->{unique} = 1;
185                                         #warn "# $n unique ",dump( $unique );
186                                 }
187                         }
188                 }
189         }
190
191         warn "# __stats ",dump($stats);
192
193         return $stats;
194 }
195
196 sub _param_or_session {
197         $_[0]->param( $_[1] ) || $_[0]->session( $_[1] )
198 }
199
200 sub stats {
201         my $self = shift;
202         my $path = $self->_param_or_session('path');
203         warn "stats $path\n";
204         delete $loaded->{$path}->{stats};
205         return $self->redirect_to( '/data/columns' );
206 }
207
208
209 sub _load_path {
210         my ( $self, $path ) = @_;
211
212         return if defined $loaded->{$path}->{generated};
213
214         my $full_path = $self->app->home->rel_file( 'data/' . $path );
215         return $self->redirect_to('/data/index') unless -r $full_path;
216
217         my $dump_path = $self->_dump_path( $path );
218
219         if ( defined $loaded->{$path}->{data} ) {
220                 my $mtime = (stat($full_path))[9];
221                 return if $loaded->{$path}->{mtime} == $mtime;
222                 warn "reload $full_path, modified ", time() - $mtime, " seconds ago\n";
223         } elsif ( -e $dump_path ) {
224                 warn "dump_path $dump_path ", -s $dump_path, " bytes loading...\n";
225                 my $info = retrieve $dump_path;
226                 $loaded->{ $path } = $info;
227                 return;
228         }
229
230         my $data;
231         if ( my $module = import_module( $full_path ) ) {
232                 $data = $module->new( full_path => $full_path )->data;
233         } else {
234                 die "can't load $full_path";
235         }
236
237         my @header;
238
239         if ( defined $data->{header} ) {
240                 if ( ref $data->{header} eq 'ARRAY' ) {
241                         @header = @{ $data->{header} };
242                 } else {
243                         warn "header not array ", dump( $data->{header} );
244                 }
245         }
246
247         my $stats = __stats( $data->{items} );
248
249         @header =
250                 sort { $stats->{$b}->{count} <=> $stats->{$a}->{count} }
251                 grep { defined $stats->{$_}->{count} } keys %$stats
252                 unless @header;
253
254         my $info = {
255                 header => [ @header ],
256                 stats  => $stats,
257                 full_path => $full_path,
258                 size => -s $full_path,
259                 mtime => (stat($full_path))[9],
260                 data => $data,
261                 defined $data->{generated} ? ( generated => 1 ) : (),
262         };
263
264         $loaded->{ $path } = $info;
265         $self->_save( $path ) unless $info->{generated};
266
267
268 }
269
270
271 sub load {
272         my $self = shift;
273
274         my @paths = @{ $self->every_param('paths') };
275         warn "# paths ", dump @paths;
276
277         foreach my $p ( keys %$loaded ) {
278                 next if grep { /^\Q$p\E$/ } @paths;
279                 warn "remove $p from memory\n";
280                 delete $loaded->{$p};
281         }
282
283         $self->_load_path( $_ ) foreach @paths;
284
285         my $path = $self->param('path') || $self->session('path') || $paths[0] || $self->redirect_to('/data/index');
286
287         warn "# path $path\n";
288         $self->_load_path( $path );
289
290         $self->session( 'path' => $path );
291
292         my $timefmt_path = $self->_permanent_path( 'timefmt' );
293         if ( -e $timefmt_path ) {
294                 my $timefmt = read_file $timefmt_path;
295                 $self->session( 'timefmt', $timefmt );
296                 warn "timefmt = ", $timefmt;
297         }
298
299         my $redirect_to = '/data/items';
300
301         $self->session( 'header' => $loaded->{$path}->{header} );
302         if ( ! defined $loaded->{$path}->{columns} ) {
303                 my $columns_path = $self->_permanent_path( 'columns' );
304                 if ( -e $columns_path ) {
305                         my @columns = map { s/[\r\n]+$//; $_ } read_file $columns_path, binmode => ':utf8';
306                         $loaded->{$path}->{columns} = [ @columns ];
307                         warn "# columns_path $columns_path ",dump(@columns);
308                 } else {
309                         $loaded->{$path}->{columns} = $loaded->{$path}->{header}
310                 }
311
312                 $redirect_to = '/data/columns';
313         }
314         $self->session( 'columns' => $loaded->{$path}->{columns} );
315         $self->session( 'order'   => $loaded->{$path}->{columns}->[0] );
316         return $self->redirect_to( $redirect_to );
317 }
318
319
320 sub _loaded {
321         my ( $self, $name ) = @_;
322         my $path = $self->session('path') || $self->param('path');
323         return $self->redirect_to('/data/index') unless $path;
324
325         if ( defined $loaded->{$path}->{modified} && $loaded->{$path}->{modified} > 1 ) {
326                 my $caller = (caller(1))[3];
327                 if ( $caller =~ m/::edit/ ) {
328                         warn "rebuild stats for $path ignored caller $caller\n";
329                 } else {
330                         warn "rebuild stats for $path FORCED by modified caller $caller\n";
331 #                       $loaded->{$path}->{stats} = __stats( $loaded->{$path}->{data}->{items} );
332                         $loaded->{$path}->{rebuild_stats} = 1;
333                         $loaded->{$path}->{modified} = 1;
334                 }
335         }
336
337         if ( defined $loaded->{$path}->{rebuild_stats} ) {
338                 warn "rebuild_stats $path";
339                 $loaded->{$path}->{stats} = __stats( $loaded->{$path}->{data}->{items} );
340                 delete $loaded->{$path}->{rebuild_stats};
341         }
342
343         if ( ! defined $loaded->{$path}->{$name} ) {
344                 warn "$path $name isn't loaded\n";
345                 $self->_load_path( $path );
346                 if ( ! defined $loaded->{$path}->{stats} ) {
347                         warn "rebuild stats for $path\n";
348                         $loaded->{$path}->{stats} = __stats( $loaded->{$path}->{data}->{items} );
349                 }
350                 if ( ! defined $loaded->{$path}->{$name} ) {
351                         warn "MISSING $name for $path\n";
352                         return $self->redirect_to('/data/index')
353                 }
354         }
355
356         $self->session( 'modified' => $loaded->{$path}->{modified} );
357
358         return $loaded->{$path}->{$name};
359 }
360
361
362 sub _checked {
363         my $self = shift;
364         my $checked;
365         $checked->{$_}++ foreach @_;
366 #       warn "# _checked ",dump($checked);
367         return $checked;
368 }
369
370 sub _permanent_path {
371         my $self = shift;
372         my $path = $self->_param_or_session('path');
373         $self->app->home->rel_file('data') . '/' . join('.', $path, @_);
374 }
375
376 sub __unac {
377         my $n = shift;
378         $n = unac_string('utf-8',$n);
379         $n =~ s/\W+/_/g;
380         return $n;
381 }
382
383 sub _column_from_unac {
384         my ($self,$name) = @_;
385
386         my $stats = $self->_loaded('stats');
387         my $cols_norm = { map { __unac( $_ ) => $_ } keys %$stats };
388
389         $cols_norm->{$name} || die "can't find column $name in ", dump($cols_norm);
390 }
391
392 sub _export_path {
393         my $self = shift;
394         my $path = $self->_param_or_session('path');
395         if ( ! $path ) {
396                 warn "no path in param or session";
397                 return;
398         }
399         my $dir = $self->app->home->rel_file('public') . "/export/$path";
400         mkpath $dir unless -e $dir;
401         my $name = __export_path_name( $path, @_ );
402         my $full = $dir . '/' . $name;
403         $full =~ s/\/+$// if -d $full; # strip trailing slash for dirs
404         return $full;
405 }
406
407 sub __export_path_name {
408         my $max_length = 80;
409
410         my $path = shift;
411         my $name = join('.', map { __unac($_) } @_ );
412         if ( length($name) > $max_length ) {
413                 $name = substr($name,0,$max_length) . Digest::MD5::md5_hex substr($name,$max_length);
414         }
415         return $name;
416 }
417
418 sub columns {
419     my $self = shift;
420
421         my $view_path = $self->_permanent_path( 'views' );
422
423         if ( $self->param('columns') ) {
424                 my @columns = $self->_param_array('columns');
425                 write_file( $self->_permanent_path( 'columns' ), { binmode => ':utf8' }, map { "$_\n" } @columns );
426                 if ( my $view = $self->param('view') ) {
427                         mkdir $view_path unless -e $view_path;
428                         write_file( "$view_path/$view", { binmode => ':utf8' }, map { "$_\n" } @columns );
429                         warn "view $view_path/$view ", -s "$view_path/$view", " bytes\n";
430                 }
431
432                 return $self->redirect_to('/data/items');
433
434         } elsif ( ! $self->session('header') ) {
435                 return $self->redirect_to('/');
436                 return $self->redirect_to('/data/load');
437         }
438
439         if ( my $id = $self->param('id') ) {
440                 my $view_full = "$view_path/$id";
441                 if ( -f $view_full ) {
442                         my @columns = map { chomp; $_ } read_file $view_full, binmode => ':utf8';
443                         warn "view $view_full loaded ", dump @columns;
444                         $self->session( 'columns' => [ @columns ] );
445                         return $self->redirect_to('/data/items');
446                 }
447         }
448
449         my $stats = $self->_loaded( 'stats' );
450
451         my @columns;
452         @columns = grep { defined $stats->{$_}->{count} } @{ $self->session('columns') } if $self->session('columns');
453
454         foreach my $c ( @{ $self->session( 'header' ) } ) {
455                 push @columns, $c unless grep { /^\Q$c\E$/ } @columns;
456         }
457
458         my @views;
459         if ( -d $view_path ) {
460                 @views = map { s{^\Q$view_path\E/*}{}; $_ } glob "$view_path/*";
461                 warn "# views ",dump @views;
462         }
463
464     $self->render(
465                 message => 'Select columns to display',
466                 stats => $stats,
467                 columns => \@columns,
468                 checked => $self->_checked( $self->_param_array('columns') ),
469                 views => \@views,
470         );
471 }
472
473 sub _param_array {
474     my ($self,$name) = @_;
475
476         my @array = @{ $self->every_param($name) };
477         my $path  = $self->session('path');
478
479         if ( @array ) {
480                 $self->session($name => [ @array ]);
481                 $loaded->{$path}->{$name} = [ @array ];
482         } elsif ( my $session = $self->session($name) ) {
483                 if ( ref $session eq 'ARRAY' ) {
484                         @array = @$session;
485                 } else {
486                         die "$name not array ",dump($session);
487                 }
488         }
489         #warn "# $name ",dump @array;
490         return @array;
491 }
492
493 sub _param_scalar {
494     my ($self,$name,$default) = @_;
495
496         my $scalar = $self->param($name);
497
498         if ( defined $scalar ) {
499                 $self->session($name => $scalar);
500         } else {
501                 $scalar = $self->session($name);
502         }
503
504         if ( ! defined $scalar ) {
505                 $scalar = $default;
506                 if ( defined $scalar ) {
507                         $self->session($name => $scalar);
508                 } else {
509                         warn "no default for $name";
510                 }
511         }
512
513         warn "# _param_scalar $name ",dump $scalar;
514         return $scalar;
515 }
516
517 sub filter {
518         my $self = shift;
519
520         my $name = $self->param('filter_name') || die "name?";
521         my @vals = @{ $self->every_param('filter_vals') };
522
523         $self->_remove_filter( $name );
524         if ( @vals ) {
525                 $self->_filter_on_data( $name, @vals );
526                 if ( my $permanent = $self->param('_permanent') ) {
527                         my $permanent_path = $self->_export_path( 'filter', $name, $permanent );
528                         write_file $permanent_path, { binmode => ':utf8' }, map { "$_\n" } @vals;
529                         warn "permanent filter $permanent_path ", -s $permanent_path;
530                 }
531         }
532
533         $self->session( 'offset' => 0 );
534
535         return $self->redirect_to('/data/items');
536 }
537
538 sub _filter_on_data {
539         my ( $self, $name, @vals ) = @_;
540
541         my $path = $self->session('path');
542
543         if ( ! defined $loaded->{$path}->{stats}->{ $name } ) {
544                 warn "filter $name not found in data set";
545                 return;
546         }
547
548         $filters->{$name} = [ @vals ];
549         warn "_filter_on_data $name ", $#vals + 1, " values on $path\n";
550
551         my $filter_hash;
552         $filter_hash->{$_}++ foreach @vals;
553
554         #warn "# filter_hash ",dump( $filter_hash );
555
556         my $items = $self->_loaded('data')->{items};
557
558         my $include_missing = defined $filter_hash->{_missing};
559         my $filtered_items;
560
561         foreach my $i ( 0 .. $#$items ) {
562
563                 if ( defined $items->[$i]->{$name} ) {
564                         my $row = $items->[$i]->{$name};
565                         $row = [ $row ] unless ref $row eq 'ARRAY'; # FIXME probably wrong place
566                         foreach my $v ( @$row ) {
567                                 if ( defined $filter_hash->{ $v } ) {
568                                         $filtered_items->{$i}++;
569                                 }
570                         }
571                 } elsif ( $include_missing ) {
572                         $filtered_items->{$i}++;
573                 }
574         }
575
576         #warn "# filter $name ",dump($filtered_items);
577
578         # invalidate filters on other datasets
579         foreach my $dataset ( grep { exists $loaded->{$_}->{filters}->{$name} } keys %$loaded ) {
580                 delete $loaded->{$dataset}->{filters}->{$name};
581                 delete $loaded->{$dataset}->{filtered};
582         }
583
584         $loaded->{$path}->{filters}->{$name} = $filtered_items;
585         warn "filter $name with ", scalar keys %$filtered_items, " items created\n";
586 }
587
588
589 sub _current_filters {
590         my $self = shift;
591         my $current_filters;
592         my $stats = $self->_loaded('stats');
593
594         $current_filters->{ $_ } = $filters->{ $_ }
595                 foreach ( grep { defined $filters->{ $_ } } keys %$stats )
596         ;
597         warn "# _current_filters ",dump( keys %$current_filters );
598         return $current_filters;
599 }
600
601 sub _data_sorted_by {
602         my ( $self, $order ) = @_;
603
604         my $path = $self->session('path');
605
606         warn "_data_sorted_by $order from $path";
607
608         if ( defined $loaded->{$path}->{sorted}->{$order} ) {
609                 return $loaded->{$path}->{sorted}->{$order};
610         }
611
612         my $data = $self->_loaded( 'data' );
613         my $numeric = $self->_is_numeric($order);
614         my $missing = $numeric ? 0 : '';
615         no warnings qw(numeric);
616         my $nr = 0;
617         my @sorted = map {
618                 $_->[0]
619         } sort {
620                 if ( $numeric ) {
621                         $a->[1] <=> $b->[1]
622                 } else {
623                         $a->[1] cmp $b->[1]
624                 }
625         } map {
626                 my $v;
627                 if ( ! exists $_->{$order} ) {
628                         $v = $missing;
629                 } elsif ( ref $_->{$order} eq 'ARRAY' ) {
630                         $v = join('', @{$_->{$order}});
631                 } else {
632                         $v = $_->{$order};
633                 }
634                 [ $nr++, $v ]
635         } @{ $data->{items} }
636         ;
637
638         warn "sorted: $order numeric: $numeric items: ", $#sorted + 1, "\n";
639         #warn "# sorted ",dump( @sorted );
640
641         $loaded->{$path}->{sorted}->{$order} = [ @sorted ];
642 }
643
644
645 sub __all_filters {
646         my $order = pop @_;
647         join(',', sort(@_), 'order', $order);
648 }
649
650 our ($out, $key,$value);
651
652 our $lookup_path_col;
653 our $on;
654
655 sub __commit_begin {
656         warn "__commit_begin";
657         $lookup_path_col = undef;
658         $on = undef;
659 }
660
661 sub __commit_end {
662         warn "__commit_end";
663         $lookup_path_col = undef; # cleanup memory
664         $on = undef;
665 }
666
667 sub lookup {
668         warn "# lookup ",dump @_;
669         my ( $vals, $on_path, $on_col, $code, $stat_code ) = @_;
670         die "code is not sub{ ... } but ", dump $code unless ref $code eq 'CODE';
671
672         if ( ! exists $loaded->{$on_path} ) {
673                 my @possible_paths = grep { /\Q$on_path\E/ } keys %$loaded;
674                 die "more than one dataset available for '$on_path' ",dump @possible_paths if $#possible_paths > 0;
675                 $on_path = shift @possible_paths;
676                 warn "## fuzzy selected path $on_path";
677         }
678
679         my $items = $loaded->{$on_path}->{data}->{items} || die "no items for $on_path";
680
681         if ( ! exists $lookup_path_col->{$on_path}->{$on_col} ) {
682                 warn "create lookup_path_col $on_path $on_col";
683                 foreach my $i ( 0 .. $#$items ) {
684                         my $item = $items->[$i];
685                         if ( exists $item->{$on_col} ) {
686                                 if ( ref $item->{$on_col} eq 'ARRAY' ) {
687                                         foreach my $v ( @{ $item->{$on_col} } ) {
688                                                 push @{ $lookup_path_col->{$on_path}->{$on_col}->{$v} }, $i;
689                                         }
690                                 } elsif ( ! ref $item->{$on_col} ) { # scalar
691                                         my $v = $item->{$on_col};
692                                         push @{ $lookup_path_col->{$on_path}->{$on_col}->{$v} }, $i;
693                                 } else {
694                                         die "unknown type of ",dump $item->{$on_col};
695                                 }
696                         }
697                 }
698                 warn "XXX ",dump $lookup_path_col->{$on_path}->{$on_col} if $ENV{DEBUG};
699         }
700
701         my $stat;
702         $stat = Statistics::Descriptive::Full->new() if $stat_code;
703
704         foreach my $v ( ref $vals eq 'ARRAY' ? @$vals : ( $vals ) ) {
705                 foreach my $i ( @{ $lookup_path_col->{$on_path}->{$on_col}->{$v} } ) {
706                         $on = $items->[$i];
707                         warn "XXX lookup code $v $i ",dump $on if $ENV{DEBUG};
708                         $code->($stat);
709                 }
710                 $stat_code->( $stat ) if $stat_code;
711         }
712 }
713
714 sub __commit_path_code {
715         my ( $path, $i, $code, $commit_changed ) = @_;
716
717         my $items = $loaded->{$path}->{data}->{items} || die "no items for $path";
718         my $row = $items->[$i];
719         my $update;
720         eval $code;
721         foreach ( keys %$update ) {
722                 $$commit_changed->{$_}++;
723                 $loaded->{$path}->{data}->{items}->[$i]->{$_} = $update->{$_};
724         }
725         #warn "__commit_path_code $path $i ",dump( $update );
726 }
727
728 # uses templates/admin.html.ep
729 sub _switch_dataset {
730         my $self = shift;
731
732         my $datasets;
733
734         foreach my $path ( keys %$loaded ) {
735                 next unless exists $loaded->{$path}->{data};
736                 push @$datasets, $path;
737         }
738
739         warn "# datasets ",dump($datasets);
740
741         $self->stash( 'datasets' => $datasets );
742 }
743
744 sub items {
745         my $self = shift;
746
747         $self->_switch_dataset;
748
749         if ( my $timefmt = $self->param('timefmt') ) {
750                 $self->session('timefmt', $timefmt);
751                 warn "session store timefmt $timefmt\n";
752                 my $timefmt_path = $self->_permanent_path( 'timefmt' );
753                 write_file $timefmt_path, $timefmt;
754                 warn "## $timefmt_path $timefmt"
755         }
756
757         if ( my $show = $self->param('id') ) {
758                 $self->param('show', $show);
759                 warn "show $show\n";
760         }
761
762         my $path = $self->_param_scalar('path');
763
764         my @columns = $self->_param_array('columns');
765         return $self->redirect_to('/data/columns') unless @columns;
766         my $order   = $self->_param_scalar('order', $columns[0]);
767         my $sort    = $self->_param_scalar('sort', 'a');
768         my $offset  = $self->_param_scalar('offset', 0);
769         my $limit   = $self->_param_scalar('limit', 20);
770         $self->_param_scalar('show', 'table');
771
772         # fix offset when changing limit
773         $offset = int( $offset / $limit ) * $limit;
774
775         if ( ! grep { /^\Q$order\E$/ } @columns ) {
776                 $order = $columns[0];
777                 $self->session( order => $order );
778         }
779         my $sorted = $self->_data_sorted_by( $order );
780
781         my @filter_names;
782         if ( $filters ) {
783                 foreach my $name ( keys %$filters ) {
784                         if ( ! defined $loaded->{$path}->{stats}->{ $name } ) {
785                                 warn "skip filter $name not found in $path\n";
786                                 next;
787                         }
788                         push @filter_names, $name;
789                 }
790                 warn "filter_names ",dump( @filter_names );
791                 foreach my $name ( @filter_names ) {
792                         next if ref $loaded->{$path}->{filters}->{$name} eq 'ARRAY';
793                         $self->_filter_on_data( $name, @{ $filters->{$name} } );
794                 }
795         }
796
797         my $all_filters = __all_filters( @filter_names,$order );
798
799 #       warn "# all_filters $all_filters ", dump( $loaded->{$path}->{filtered}->{$all_filters} );
800
801         if ( ! defined $loaded->{$path}->{filtered}->{$all_filters} ) {
802
803                 my $path_filters = $loaded->{$path}->{filters};
804
805                 warn "create combined filter for $all_filters from ", $#$sorted + 1, " items\n";
806
807                 my @filtered;
808                 foreach my $i ( 0 .. $#$sorted ) {
809                         my $pos = $sorted->[$i];
810
811                         if ( $#filter_names == -1 ) {
812                                 push @filtered, $pos;
813                                 next;
814                         }
815
816                         my $skip = 0;
817                         foreach ( @filter_names ) {
818                                 $skip ||= 1 if ! defined $path_filters->{$_}->{$pos};
819                         }
820                         next if $skip;
821
822                         push @filtered, $pos;
823                 }
824
825                 $loaded->{$path}->{filtered}->{$all_filters} = [ @filtered ];
826         }
827
828         my $filtered = $loaded->{$path}->{filtered}->{$all_filters}
829                 if defined $loaded->{$path}->{filtered}->{$all_filters};
830
831         warn "all_filters $all_filters produced ", $#$filtered + 1, " items\n" if $filtered;
832
833         my $data = $self->_loaded('data');
834
835         my $code = $self->_param_scalar('code','');
836         $code =~ s{\r}{}gs;
837         $code =~ s{\n+$}{\n}s;
838
839         # XXX convert @row->{foo} into @{$row->{foo}}
840         $code =~ s|\@(row->\{[^}]+\})|\@{\$$1}|gs;
841
842         my $commit = $self->param('commit');
843         my $test = $self->param('test');
844
845         my $commit_changed;
846         __commit_begin;
847
848         if ( $code && ( $test || $commit ) ) {
849                 # XXX find columns used in code snippet and show them to user
850                 my $order = 0;
851                 foreach my $column ( $code =~ m/\$row->\{([^}]+)\}/g ) {
852                         if ( $column =~ s/^(['"])// ) {
853                                 $column =~ s/$1$//;
854                         }
855                         next if $column =~ m/\$/; # hide columns with vars in them
856                         $commit_changed->{$column} = 0;
857                 }
858         }
859
860         my $code_path = $self->app->home->rel_file('public') . "/code";
861         if ( $commit ) {
862
863                 __path_modified( $path, 'commit' );
864
865                 warn "# commit on ", $#$filtered + 1, " items:\n$code\n";
866                 ( $key, $value, $out ) = ( 'key', 'value' );
867                 foreach ( 0 .. $#$filtered ) {
868                         my $i = $filtered->[$_];
869                         __commit_path_code( $path, $i, $code, \$commit_changed );
870                 }
871
872                 # this might move before $out to recalculate stats on source dataset?
873                 __path_rebuild_stats( $path );
874                 my $c = { map { $_ => 1 } @columns };
875                 my @added_columns = sort grep { ! $c->{$_} } keys %$commit_changed;
876                 warn "# added_columns ",dump( @added_columns );
877                 unshift @columns, @added_columns;
878
879                 $loaded->{$path}->{columns} = [ @columns ];
880                 warn "# new columns ",dump( @columns );
881
882                 __invalidate_path_column( $path, $_ ) foreach keys %$commit_changed;
883
884                 $self->_save_change({
885                         path => $path,
886                         time => $self->param('time') || time(),
887                         user => $self->param('user') || $ENV{'LOGNAME'},
888                         code => $code,
889                         commit_changed => $commit_changed,
890                 });
891
892                 if ( my $description = $self->param('code_description') ) {
893                         my $depends = $self->param('code_depends') || die "no code_depends?";
894                         my $path = "$code_path/$depends.$description.pl";
895                         if ( -e $path && ! $self->param('overwrite') ) {
896                                 warn "# code $path not saved\n";
897                         } else {
898                                 write_file(  $path, { binmode => ':utf8' }, "$code\n" );
899                                 warn "code $path ", -s $path, " bytes saved\n";
900                         }
901                 }
902
903                 # remove console
904                 $code = '';
905                 if ( $out ) {
906                         my $commit_dataset = join('.'
907                                 , $self->param('code_depends')
908                                 , $self->param('code_description')
909                                 , time()
910                         );
911                         $key ||= 'key';
912                         $value ||= 'value';
913                         warn "key $key value $value";
914                         my $items;
915                         foreach my $n ( keys %$out ) {
916                                 my $i = { $key => [ $n ] };
917                                 my $ref = ref $out->{$n};
918                                 if ( $ref eq 'HASH' ) {
919                                         $i->{$_} = [ $out->{$n}->{$_} ] foreach keys %{ $out->{$n} };
920                                 } elsif ( $ref eq 'ARRAY' ) {
921                                         $i->{$_} = $out->{$n};
922                                 } elsif ( ! $ref ) {
923                                         $i->{$value} = [ $out->{$n} ];
924                                 } else {
925                                         $i->{_error} = [ dump($out->{$n}) ];
926                                 }
927                                 push @$items, $i;
928                         };
929                         undef $out;
930                         my $stats = __stats( $items );
931                         my @columns = grep { ! m/^\Q$key\E$/ } sort keys %$stats;
932                         unshift @columns, $key;
933
934                         $loaded->{$commit_dataset} = {
935                                 header => [ @columns ],
936                                 columns => [ @columns ],
937                                 mtime => time(),
938                                 data => { items => $items },
939                                 stats => $stats,
940                                 generated => 1,
941                         };
942                         warn "# loaded out ", dump( $loaded->{$commit_dataset} );
943                         $self->session('path', $commit_dataset);
944                         $self->session('columns', [ @columns ]);
945                         $self->session('order', $key);
946                         return $self->redirect_to('/data/items');
947                 }
948
949                 $self->session('columns', [ @columns ]);
950         }
951
952         my $sorted_items;
953         my $from_end = $sort eq 'd' ? $#$filtered : 0;
954         my $test_changed;
955         my ( $key, $value, $out ) = ( 'key', 'value' ); # XXX make local
956         foreach ( 0 .. $limit ) {
957                 my $i = $_ + $offset;
958                 last unless defined $filtered->[$i];
959                 $i = $from_end - $i if $from_end;
960                 my $id = $filtered->[$i];
961                 my $row = Storable::dclone $data->{items}->[ $id ];
962                 if ( $code && $test ) {
963                         my $update;
964                         eval $code;
965                         if ( $@ ) {
966                                 warn "ERROR evaling $@", dump($code);
967                                 $self->stash('eval_error', $@) if $@;
968                         } else {
969                                 warn "EVAL ",dump($update);
970                                 foreach ( keys %$update ) {
971                                         $test_changed->{$_}++;
972                                         $row->{$_} = $update->{$_};
973                                 }
974                         }
975                 }
976                 $row->{_row_id} ||= $id;
977                 push @$sorted_items, $row;
978         }
979
980         if ( $self->param('export') ) {
981                 my $export_path = $self->_export_path( 'items', @columns);
982                 open(my $fh, '>', $export_path) || die "ERROR: can't open $export_path: $!";
983                 print $fh "#",join("\t",@columns),"\n";
984                 foreach my $f ( 0 .. $#$filtered ) {
985                         print $fh join("\t", map {
986                                 my $i = $data->{items}->[ $filtered->[$f] ];
987                                 my $v = '\N';
988                                 if ( ! defined $i->{$_} ) {
989                                         # nop
990                                 } elsif ( ref $i->{$_} eq 'ARRAY' ) {
991                                         $v =join(',', @{ $i->{$_} });
992                                         $v = '\N' if length($v) == 0;
993                                 } elsif ( ! ref $i->{$_} ) {
994                                         $v = $i->{$_};
995                                 } else {
996                                         $v = dump $i->{$_};
997                                 }
998                                 $v;
999                         } @columns),"\n";
1000                 }
1001                 close($fh);
1002                 warn "export $export_path ", -s $export_path, " bytes\n";
1003         }
1004
1005         my ( $code_depends, $code_description );
1006
1007         if ( $test ) {
1008
1009                 warn "# test_changed ",dump( $test_changed );
1010                 my $c = { map { $_ => 1 } @columns };
1011                 my @added_columns = sort grep { ! $c->{$_} } keys %$test_changed;
1012                 unshift @columns, @added_columns;
1013
1014                 warn "# sorted_items ", $#$sorted_items + 1, " offset $offset limit $limit order $sort";
1015
1016                 my $depends_on;
1017                 my $tmp = $code; $tmp =~ s/\$row->\{(['"]?)([\w\s]+)\1/$depends_on->{$2}++/gse;
1018                 warn "# depends_on ",dump $depends_on;
1019
1020                 my $test_added = Storable::dclone $test_changed;
1021                 delete $test_added->{$_} foreach keys %$depends_on;
1022
1023                 $code_depends = $self->param('code_depends')
1024                 || join(',', keys %$depends_on);
1025
1026                 $code_description = $self->param('code_description') ||
1027                 join(',', keys %$test_added);
1028
1029                 $code_depends ||= $code_description; # self-modifing
1030                 if ( ! $code_depends && $out ) {
1031                         $code_depends = $key;
1032                         $code_description = $value;
1033                 }
1034
1035                 warn "# test_changed ",dump( $test_changed, $code_depends, $code_description );
1036
1037         } # test?
1038
1039         __commit_end;
1040
1041         $self->render(
1042                 order => $order,
1043                 offset => $offset,
1044                 limit => $limit,
1045                 sorted => $sorted_items,
1046                 columns => [ @columns ],
1047                 rows => $#$filtered + 1,
1048                 numeric => { map { $_, $self->_is_numeric($_) } @columns },
1049                 unique  => { map { $_, $self->_is_unique( $_) } @columns },
1050                 filters => $self->_current_filters,
1051                 code => $code,
1052                 cols_changed => $commit ? $commit_changed : $test_changed,
1053                 code_depends => $code_depends,
1054                 code_description => $code_description,
1055                 code_path => $code_path,
1056                 out => $out,
1057         );
1058
1059 }
1060
1061
1062 sub order {
1063         my $self = shift;
1064         $self->session('order', $self->param('order'));
1065         $self->session('sort', $self->param('sort'));
1066         return $self->redirect_to('/data/items');
1067 }
1068
1069 sub _is_numeric {
1070         my ( $self, $name ) = @_;
1071
1072         my $stats = $self->_loaded( 'stats' );
1073
1074         # sort facet numerically if more >50% elements are numeric
1075         my $count = $stats->{$name}->{count};
1076         $count   -= $stats->{$name}->{empty} if defined $stats->{$name}->{empty};
1077         defined $stats->{$name}->{numeric} &&
1078                 $stats->{$name}->{numeric} > $count / 2;
1079 }
1080
1081 sub _is_unique {
1082         my ( $self, $name ) = @_;
1083         my $stats = $self->_loaded( 'stats' );
1084         defined $stats->{$name}->{unique};
1085 }
1086
1087 sub _remove_filter {
1088         my ($self,$name) = @_;
1089         warn "_remove_filter $name\n";
1090
1091         my $path = $self->session('path');
1092
1093         delete $filters->{$name};
1094         delete $loaded->{$path}->{filters}->{$name};
1095         warn "filters left: ", keys %{ $loaded->{$path}->{filters} };
1096
1097         foreach (
1098                         grep { /\Q$name\E/ }
1099                         keys %{ $loaded->{$path}->{filtered} }
1100         ) {
1101                 delete $loaded->{$path}->{filtered}->{$_};
1102                 warn "remove filtered cache $_";
1103         }
1104 }
1105
1106 sub facet {
1107         my $self = shift;
1108
1109         my $path = $self->session('path') || return $self->redirect_to( '/data/index' );
1110
1111         if ( my $name = $self->param('remove') ) {
1112                 $self->_remove_filter( $name );
1113                 return $self->redirect_to( '/data/items' );
1114         }
1115
1116         my $facet;
1117         my $name = $self->param('name') || die "no name";
1118
1119         my $all = $self->_param_scalar('all', 1);
1120         my $data = $self->_loaded('data');
1121
1122         my $filters = $self->_current_filters;
1123         my $all_filters = __all_filters( keys %$filters,$self->session('order') );
1124         my $filtered = $loaded->{$path}->{filtered}->{$all_filters}
1125                 if defined $loaded->{$path}->{filtered}->{$all_filters};
1126
1127         if ( ! $filtered || $all ) {
1128                 $filtered = [ 0 .. $#{ $data->{items} } ];
1129                 warn "filter all values\n";
1130         } else {
1131                 warn "filter using $all_filters\n";
1132         }
1133
1134         foreach my $i ( @$filtered ) {
1135                 my $item = $data->{items}->[$i];
1136                 if ( ! exists $item->{$name} || ! defined $item->{$name} ) {
1137                         $facet->{ _missing }++;
1138                 } elsif ( ref $item->{$name} eq 'ARRAY' ) {
1139                         $facet->{$_}++ foreach @{ $item->{$name} };
1140                 } else {
1141                         $facet->{ $item->{$name} }++;
1142                 }
1143         }
1144
1145         my $checked_values = $self->_checked( @{ $filters->{$name} } ) if defined $filters->{$name};
1146
1147         if ( my $code = $self->param('code') ) {
1148                 my $out;
1149                 foreach my $value ( keys %$facet ) {
1150                         my $count = $facet->{$value};
1151                         my $checked = $checked_values->{$value};
1152                         eval $code;
1153                         if ( $@ ) {
1154                                 $out = $@;
1155                                 warn "ERROR: $@\n$code\n";
1156                                 last;
1157                         } elsif ( $checked != $checked_values->{$value} ) {
1158                                 warn "checked $value $count -> $checked\n";
1159                                 $checked_values->{$value} = $checked;
1160                         }
1161                 }
1162                 warn "out ",dump( $out );
1163                 $self->stash( out => $out );
1164         }
1165
1166 #       warn "# facet $name ",dump $facet;
1167
1168         my @facet_names =
1169                   $all                      ? keys %$facet
1170                 : defined $filters->{$name} ? @{ $filters->{$name} }
1171                 : keys %$facet;
1172
1173         my $numeric = $self->_is_numeric($name);
1174
1175         my $sort = $self->param('sort');
1176         # sort numeric facets with more than 5 values ascending
1177         $sort ||= $numeric && $#facet_names > 4 ? 'a' : 'c';
1178
1179         @facet_names = sort {
1180                 my $result;
1181                 if ( $sort eq 'a' ) {
1182                         $result = $numeric ? $a <=> $b : lc $a cmp lc $b;
1183                 } elsif ( $sort eq 'd' ) {
1184                         $result = $numeric ? $b <=> $a : lc $b cmp lc $a;
1185                 } elsif ( $sort eq 'c' ) {
1186                         $result = ( $facet->{$b} || -1 ) <=> ( $facet->{$a} || -1 )
1187                 } else {
1188                         warn "unknown sort: $sort";
1189                         $result = $a cmp $b;
1190                 }
1191                 $result = $a cmp $b unless defined $result; # FIXME cludge for numeric facets with invalid data
1192                 $result;
1193         } @facet_names;
1194
1195         $self->render( name => $name, facet => $facet, checked => $checked_values,
1196                 facet_names => \@facet_names, sort => $sort, numeric => $numeric,
1197         );
1198 }
1199
1200
1201 sub __invalidate_path_column {
1202         my ( $path, $name ) = @_;
1203
1204         if ( defined $loaded->{$path}->{sorted}->{$name} ) {
1205                 delete $loaded->{$path}->{sorted}->{$name};
1206                 warn "# invalidate $path sorted $name\n";
1207         }
1208
1209         foreach ( grep { m/\Q$name\E/ } keys %{ $loaded->{$path}->{filtered} } ) {
1210                 delete $loaded->{$path}->{filtered}->{$_};
1211                 warn "# invalidate $path filtered $_\n";
1212         }
1213 }
1214
1215 sub __path_modified {
1216         my ( $path, $value ) = @_;
1217         $value = 1 unless defined $value;
1218         
1219         $loaded->{$path}->{modified}  = $value;
1220
1221         warn "# __path_modified $path $value\n";
1222 }
1223
1224 sub __path_rebuild_stats { $loaded->{ $_[0] }->{rebuild_stats} = 1 };
1225
1226 sub _save_change {
1227         my ($self,$change) = @_;
1228
1229         my $change_path = $self->_permanent_path( 'changes' );
1230         mkdir $change_path unless -d $change_path;
1231         $change_path .= '/' . $change->{time};
1232         store $change, $change_path;
1233         utime $change->{time}, $change->{time}, $change_path;
1234         warn "_save_change $change_path ", dump($change);
1235 }
1236
1237 sub edit {
1238         my $self = shift;
1239         my $new_content = $self->param('new_content');
1240         $new_content  ||= $self->param('content'); # backward compatibility with old actions
1241
1242         my $i = $self->param('_row_id');
1243         die "invalid _row_id ",dump($i) unless $i =~ m/^\d+$/;
1244         my $path = $self->param('path') || die "no path";
1245         my $name = $self->param('name') || die "no name";
1246         my $status = 200; # 200 = OK, 201 = Created
1247
1248         my $data = $self->_loaded('data');
1249
1250         if ( defined $loaded->{$path}->{data}->{items}->[$i] ) {
1251                 $new_content =~ s/^\s+//s;
1252                 $new_content =~ s/\s+$//s;
1253                 my $v;
1254                 if ( $new_content =~ /\xB6/ ) { # para
1255                         $v = [ split(/\s*\xB6\s*/, $new_content) ];
1256                 } else {
1257                         $v = [ $new_content ];
1258                 }
1259
1260                 my $old = dump $loaded->{$path}->{data}->{items}->[$i]->{$name};
1261                 my $new = dump $v;
1262                 if ( $old ne $new
1263                         && ! ( $old eq 'undef' && length($new_content) == 0 ) # new value empty, previous undef
1264                 ) {
1265                         $self->_save_change({
1266                                 path => $path,
1267                                 column => $name,
1268                                 pos => $i,
1269                                 old => $loaded->{$path}->{data}->{items}->[$i]->{$name},
1270                                 new => $v,
1271                                 time => $self->param('time') || time(),
1272                                 user => $self->param('user') || $ENV{'LOGNAME'},
1273                                 unique => {
1274                                         map { $_ => $loaded->{$path}->{data}->{items}->[$i]->{$_}->[0] }
1275                                         grep { defined $loaded->{$path}->{stats}->{$_}->{unique} }
1276                                         keys %{ $loaded->{$path}->{stats} }
1277                                 },
1278                         });
1279
1280                         warn "# change $path $i $old -> $new\n";
1281                         $loaded->{$path}->{data}->{items}->[$i]->{$name} = $v;
1282
1283                         __invalidate_path_column( $path, $name );
1284
1285                         $status = 201; # created
1286                         __path_rebuild_stats( $path );
1287         
1288                         $new_content = join("\xB6",@$v);
1289
1290                 } else {
1291                         warn "# unchanged $path $i $old\n";
1292                         $status = 304;
1293                 }
1294         } else {
1295                 $new_content = "$path $i $name doesn't exist\n";
1296                 $status = 404;
1297         }
1298
1299         warn "# edit $status ", dump $new_content;
1300
1301         $self->render(
1302                 status => $status,
1303                 new_content => scalar $new_content,
1304         );
1305 }
1306
1307
1308 sub save {
1309         my $self = shift;
1310         my $path = $self->_param_or_session('path');
1311         my $dump_path = $self->_save( $path );
1312         __path_modified( $path, 0 );
1313
1314         return $self->redirect_to( '/data/items' );
1315 }
1316
1317 sub export {
1318         my $self = shift;
1319
1320         my $dir = $self->app->home->rel_file('public');
1321
1322         if ( my $import = $self->param('import') ) {
1323
1324                 if ( $import =~ m{/filter\.(.+?)\..+} ) {
1325                         my $name = $self->_column_from_unac( $1 );
1326
1327                         my @vals = map { chomp; $_ }
1328                                 read_file "$dir/export/$import", binmode => ':utf8';
1329
1330                         $self->_remove_filter( $name );
1331                         $self->_filter_on_data( $name, @vals );
1332                         $self->session( 'offset' => 0 );
1333                         return $self->redirect_to('/data/items');
1334                 } else {
1335                         warn "UNKNOWN IMPORT $import";
1336                 }
1337         }
1338
1339         if ( my $remove = $self->param('remove') ) {
1340                 my $path = "$dir/export/$remove";
1341                 unlink $path if -e $path;
1342                 $path .= '.png';
1343                 unlink $path if -e $path;
1344         }
1345
1346         my $path = $self->_export_path || return $self->redirect_to('/data/index');
1347
1348         my @files = grep { ! /\.png$/ } glob "$path/*";
1349         my $mtime = { map { $_ => (stat($_))[9] } @files };
1350         @files = sort { $mtime->{$b} <=> $mtime->{$a} } @files;
1351         $self->render( export => [ @files ] );
1352 }
1353
1354 sub __loaded_paths {
1355         return
1356                 grep { defined $loaded->{$_}->{data} }
1357                 keys %$loaded;
1358 }
1359
1360 sub reload {
1361         my $self = shift;
1362         $self->stash( reload => 1 );
1363         $self->remove;
1364 #       $self->_load_path( $self->param('path') );
1365         $self->redirect_to('/data/load?path=' . $self->param('path') );
1366 }
1367
1368 sub remove {
1369         my $self = shift;
1370         my $path = $self->param('path');
1371         if ( $path =~ m{^/tmp/mojo_facets\.} ) {
1372                 unlink $path;
1373                 warn "# unlink $path";
1374                 if ( my $name = $self->param('name') ) {
1375                         delete $loaded->{$name};
1376                         warn "# remove $name from memory";
1377                 }
1378         } else {
1379                 warn "WARNING: $path unlink ignored";
1380         }
1381         return if $self->stash('reload');
1382         return $self->redirect_to( '/data/load' );
1383 }
1384
1385 1;