/[webpac2]/Webpacus/lib/Webpacus/Model/WebPAC.pm
This is repository of my old source code which isn't updated any more. Go to git.rot13.org for current projects!
ViewVC logotype

Contents of /Webpacus/lib/Webpacus/Model/WebPAC.pm

Parent Directory Parent Directory | Revision Log Revision Log


Revision 399 - (show annotations)
Sun Feb 19 12:37:27 2006 UTC (18 years, 2 months ago) by dpavlin
File size: 16225 byte(s)
 r461@llin:  dpavlin | 2006-02-19 13:46:03 +0100
 each site now uses default index which is named same as site, added site_url to view,
 refactor site handling

1 package Webpacus::Model::WebPAC;
2
3 use strict;
4 use warnings;
5 use lib '/data/webpac2/lib';
6 use base qw/
7 Catalyst::Model
8 /;
9 use WebPAC::Store 0.08;
10 use Search::Estraier 0.04;
11 use File::Slurp;
12 use Time::HiRes qw/time/;
13 use Encode qw/encode decode from_to/;
14 use Template;
15 use Data::Dumper;
16
17 =head1 NAME
18
19 Webpacus::Model::WebPAC - Catalyst Model
20
21 =head1 SYNOPSIS
22
23 See L<Webpacus> and L<WebPAC>.
24
25 =head1 DESCRIPTION
26
27 Catalyst Model for access to WebPAC data.
28
29 =head2 new
30
31 Configuration for hyperestraier in C<config.yaml> like this:
32
33 --- #YAML:1.0
34 # DO NOT USE TABS FOR INDENTATION OR label/value SEPARATION!!!
35
36 # configuration for hyper estraier full text search engine
37 hyperestraier:
38 masterurl: 'http://localhost:1978/node/webpac2'
39 defaultnode: 'webpac2'
40 defaultdepth: 1
41 user: 'admin'
42 passwd: 'admin'
43 hits_on_page: 100
44 hits_for_pager: 1000
45
46 webpac:
47 db_path: '/data/webpac2/db'
48 template_path: '/data/webpac2/conf/output/tt'
49 template: 'html_ffzg_results_short.tt'
50 # encoding comming from webpac
51 webpac_encoding: 'iso-8859-2'
52
53 =cut
54
55 sub new {
56 my ( $self, $c, $config ) = @_;
57
58 $self = $self->NEXT::new($c, $config);
59 $self->config($config);
60
61 my $log = $c->log;
62 $self->{log} = $log;
63
64 my $est_cfg = $c->config->{hyperestraier};
65 $est_cfg->{'log'} = $log;
66
67 $est_cfg->{encoding} = $est_cfg->{catalyst_encoding} || $c->config->{catalyst_encoding} or $c->log->fatal("can't find catalyst_encoding");
68
69 $log->debug("using config:" . Dumper($est_cfg) );
70
71 if (! $est_cfg->{database}) {
72 my $defaultnode = $est_cfg->{defaultnode} || $log->logdie("can't find defaultnode in estraier configuration");
73 $log->info("using default node $defaultnode");
74 $est_cfg->{database} = $defaultnode;
75 }
76
77 my $url = $est_cfg->{masterurl} . '/node/' . $est_cfg->{database};
78
79 $log->info("opening Hyper Estraier index $url as $est_cfg->{'user'}");
80
81 $self->{est_node} = Search::Estraier::Node->new(
82 url => $url,
83 user => $est_cfg->{user},
84 passwd => $est_cfg->{passwd},
85 );
86
87 $log->fatal("can't create Search::Estraier::Node $url") unless ($self->{est_node});
88
89 # save config parametars in object
90 foreach my $f (qw/
91 db_path template_path hits_on_page webpac_encoding defaultdepth
92 masterurl
93 /) {
94 $self->{$f} = $c->config->{hyperestraier}->{$f} ||
95 $c->config->{webpac}->{$f};
96 $log->debug("self->{$f} = " . $self->{$f});
97 }
98 my $db_path = $self->{db_path};
99 my $template_path = $self->{template_path};
100
101 $log->debug("using db path '$db_path', template path '$template_path'");
102
103 $self->{db} = new WebPAC::Store(
104 path => $db_path,
105 read_only => 1,
106 database => $est_cfg->{database},
107 );
108
109 # default template from config.yaml
110 $self->{template} ||= $c->config->{webpac}->{template};
111
112 $log->debug("converting encoding from webpac_encoding '" .
113 $c->config->{webpac}->{webpac_encoding} .
114 "'"
115 );
116
117 $self->{databases} = $c->config->{databases} || $log->fatal("can't find databases in config");
118
119 # create Template toolkit instance
120 $self->{'tt'} = Template->new(
121 INCLUDE_PATH => $template_path,
122 FILTERS => {
123 dump_html => sub {
124 return unless (@_);
125 my $out;
126 my $i = 1;
127 foreach my $v (@_) {
128 $out .= qq{<div id="dump_$i">} .
129 Data::HTMLDumper->Dump([ $v ],[ "v$i" ]) .
130 qq{</div>};
131 $i++;
132 }
133 $out =~ s!<table[^>/]*>!<table class="dump">!gis if ($out);
134 return $out;
135 }
136 },
137 EVAL_PERL => 1,
138 );
139
140 return $self;
141
142 }
143
144 =head2 setup_site
145
146 $self->setup_site('site_name');
147
148 Change node URL and database name according to site name (if available)
149
150 =cut
151
152 sub setup_site {
153 my $self = shift;
154
155 my $site = shift || return;
156
157 my $url = $self->{masterurl} . '/node/' . $site;
158 $self->{est_node}->set_url( $url );
159 $self->{log}->debug("setup_site $site");
160 }
161
162 =head2 search
163
164 my $m->search(
165 phrase => 'query phrase',
166 add_attr => \@add_attr
167 get_attr => [ '@uri' ],
168 max => 42,
169 template => 'result_template.tt',
170 depth => 1,
171 );
172
173 All fields are standard C<WebPAC::Search::Estraier> parametars except
174 C<template> which will (if specified) return results in HTML using
175 selected template.
176
177 =cut
178
179 sub search {
180 my $self = shift;
181
182 my $search_start_t = time();
183
184 my $args = {@_};
185
186 my $log = $self->{log};
187
188 $log->debug("search args: " . Dumper( $args ));
189
190 my $query = $args->{phrase} || $log->warn("no query phrase") && return;
191
192 my $template_filename = $args->{template} || $self->{template};
193
194 $args->{max} ||= $self->{'hits_for_pager'};
195 if (! $args->{max}) {
196 $args->{max} = 100;
197 $log->warn("max not set when calling model. Using default of $args->{max}");
198 }
199
200 my $times; # store some times for benchmarking
201
202 my $t = time();
203
204 # transfer depth of search
205 if (! $args->{depth}) {
206 my $default = $self->{defaultdepth} || $log->logdie("can't find defaultdepth in estraier configuration");
207 $args->{depth} = $default;
208 $log->warn("using default search depth $default");
209 }
210 $args->{depth} ||= 0;
211
212 $log->debug("searching for maximum $args->{max} results using depth $args->{depth} phrase: ", $query || '[none]');
213
214 #
215 # construct condition for Hyper Estraier
216 #
217 my $cond = Search::Estraier::Condition->new();
218 if ( ref($args->{add_attr}) eq 'ARRAY' ) {
219 $log->debug("adding search attributes: " . join(", ", @{ $args->{add_attr} }) );
220 map {
221 $cond->add_attr( $_ );
222 $log->debug(" + $_");
223 } @{ $args->{add_attr} };
224 };
225
226 $cond->set_phrase( $query ) if ($query);
227 $cond->set_options( $args->{options} ) if ($args->{options});
228 $cond->set_order( $args->{order} ) if ($args->{order});
229
230 my $max = $args->{max} || 7;
231 my $page = $args->{page} || 1;
232 if ($page < 1) {
233 $log->warn("page number $page < 1");
234 $page = 1;
235 }
236
237 $cond->set_max( $page * $max );
238
239 my $result = $self->{est_node}->search($cond, $args->{depth});
240 my $hits = $result->doc_num;
241
242 $times->{est} += time() - $t;
243
244 $log->debug( sprintf("search took %.6fs and returned $hits hits.", $times->{est}) );
245
246 $log->debug( "hints: " . Dumper($result->{hints}) );
247
248 #
249 # fetch results
250 #
251
252 my @results;
253
254 for my $i ( (($page - 1) * $max) .. ( $hits - 1 ) ) {
255
256 $t = time();
257
258 #$log->debug("get_doc($i)");
259 my $doc = $result->get_doc( $i );
260 if (! $doc) {
261 $log->warn("can't find result $i");
262 next;
263 }
264
265 my $hash;
266
267 foreach my $attr (@{ $args->{get_attr} }) {
268 my $val = $doc->attr( $attr );
269 #$log->debug("attr $attr = ", $val || 'undef');
270 $hash->{$attr} = $val if (defined($val));
271 }
272
273 $times->{hash} += time() - $t;
274
275 next unless ($hash);
276
277 if (! $args->{'template'}) {
278 push @results, $hash;
279 } else {
280 my ($database, $prefix, $id);
281
282 if ( $hash->{'@uri'} =~ m!/([^/]+)/([^/]+)/(\d+)$!) {
283 ($database, $prefix,$id) = ($1,$2,$3);
284 } else {
285 $log->warn("can't decode database/prefix/id from " . $hash->{'@uri'});
286 next;
287 }
288
289 #$log->debug("load_ds( id => $id, prefix => '$prefix' )");
290
291 $t = time();
292
293 my $ds = $self->{db}->load_ds( database => $database, prefix => $prefix, id => $id );
294 if (! $ds) {
295 $log->error("can't load_ds( ${database}/${prefix}/${id} )");
296 next;
297 }
298
299 $times->{db} += time() - $t;
300
301 #$log->debug( "ds = " . Dumper( \@html_results ) );
302
303 $t = time();
304
305 my $html = $self->apply(
306 template => $template_filename,
307 data => $ds,
308 record_uri => "${database}/${prefix}/${id}",
309 config => $self->{databases}->{$database},
310 );
311
312 $times->{apply} += time() - $t;
313
314 $t = time();
315
316 $html = decode($self->{webpac_encoding}, $html);
317
318 $times->{decode} += time() - $t;
319
320 push @results, $html;
321 }
322
323 }
324
325 #$log->debug( '@results = ' . Dumper( \@results ) );
326
327 $log->debug( sprintf(
328 "duration breakdown: estraier %.6fs, hash %.6fs, store %.6fs, apply %.6fs, decode %.06f, total: %.6fs",
329 $times->{est}, $times->{hash}, $times->{db}, $times->{apply}, $times->{decode}, time() - $search_start_t,
330 ) );
331
332 return \@results;
333 }
334
335 =head2 record
336
337 my $html = $m->record(
338 mfn => 42,
339 template => 'foo.tt',
340 );
341
342 This will load one record, convert it to html using C<template> and return
343 it.
344
345 =cut
346
347 sub record {
348 my $self = shift;
349
350 my $args = {@_};
351 my $log = $self->{log};
352 $log->debug("record args: " . Dumper( $args ));
353
354 foreach my $f (qw/record_uri template/) {
355 $log->fatal("need $f") unless ($args->{$f});
356 }
357
358 my ($database, $prefix, $id);
359
360 if ($args->{record_uri} =~ m#^([^/]+)/([^/]+)/([^/]+)$#) {
361 ($database, $prefix, $id) = ($1,$2,$3);
362 } else {
363 $log->error("can't parse $args->{record_uri} into prefix, database and uri");
364 return;
365 }
366
367 my $ds = $self->{db}->load_ds( id => $id, prefix => $prefix, database => $database );
368 if (! $ds) {
369 $log->error("can't load_ds( $database/$prefix/$id )");
370 return;
371 }
372
373 my $html = $self->apply(
374 template => $args->{template},
375 data => $ds,
376 record_uri => $args->{record_uri},
377 config => $self->{databases}->{$database},
378 );
379
380 $html = decode($self->{webpac_encoding}, $html);
381
382 return $html;
383 }
384
385
386 =head2 save_html
387
388 $m->save_html( '/full/path/to/file', $content );
389
390 It will use C<Encode> to convert content encoding back to
391 Webpac codepage, recode JavaScript Unicode entities (%u1234),
392 strip extra newlines at beginning and end, and save to
393 C</full/path/to/file.new> and if that succeeds, just rename
394 it over original file which should be atomic on filesystem level.
395
396 =cut
397
398 sub save_html {
399 my ($self, $path, $content) = @_;
400
401 # FIXME Should this be UTF-8 or someting?
402 my $js_encoding = $self->{webpac_encoding};
403 $js_encoding = 'UTF-16';
404
405 sub _conv_js {
406 return '0x' . $_[1];
407 return encode($_[0], chr(hex($_[1])));
408 }
409 #$content =~ s/%u([a-fA-F0-9]{4})/_conv_js($js_encoding,$1)/gex;
410 $content =~ s/^[\n\r]+//s;
411 $content =~ s/[\n\r]+$/\n/s;
412 $content =~ s/\n\r/\n/gs;
413
414 my $disk_encoding = $self->{webpac_encoding} || 'utf-8';
415 $self->{log}->debug("convert encoding to $disk_encoding");
416 from_to($content, 'utf-8', $disk_encoding) || $self->{log}->warn("encoding from utf-8 to $disk_encoding failed for: $content");
417
418 write_file($path . '.new', {binmode => ':raw' }, $content) || die "can't save ${path}.new $!";
419 rename $path . '.new', $path || die "can't rename to $path: $!";
420 }
421
422 =head2 load_html
423
424 my $html = $m->load_html('/full/path/to/file');
425
426 This will convert file from Webpac encoding to Catalyst and
427 convert that data to escaped HTML (for sending into
428 C<< <textarea/> >> tags in html.
429
430 =cut
431
432 sub load_html {
433 my ($self, $path) = @_;
434
435 die "no path?" unless ($path);
436
437 my $content = read_file($path, {binmode => ':raw' }) || die "can't read $path: $!";
438
439 return decode($self->{webpac_encoding}, $content);
440 }
441
442
443 =head2 apply
444
445 Create output from in-memory data structure using Template Toolkit template.
446
447 my $text = $tt->apply(
448 template => 'text.tt',
449 data => $ds,
450 record_uri => 'database/prefix/mfn',
451 );
452
453 It also has follwing template toolikit filter routies defined:
454
455 =cut
456
457 sub apply {
458 my $self = shift;
459
460 my $args = {@_};
461
462 my $log = $self->{log} || die "no log?";
463
464 foreach my $a (qw/template data/) {
465 $log->fatal("need $a") unless ($args->{$a});
466 }
467
468 =head3 tt_filter_type
469
470 filter to return values of specified from $ds, usage from TT template is in form
471 C<d('FieldName','delimiter')>, where C<delimiter> is optional, like this:
472
473 [% d('Title') %]
474 [% d('Author',', ' %]
475
476 =cut
477
478 sub tt_filter_type {
479 my ($data,$type) = @_;
480
481 die "no data?" unless ($data);
482 $type ||= 'display';
483
484 my $default_delimiter = {
485 'display' => '&#182;<br/>',
486 'index' => '\n',
487 };
488
489 return sub {
490
491 my ($name,$join) = @_;
492
493 die "no data hash" unless ($data->{'data'} && ref($data->{'data'}) eq 'HASH');
494 # Hm? Should we die here?
495 return unless ($name);
496
497 my $item = $data->{'data'}->{$name} || return;
498
499 my $v = $item->{$type} || return;
500
501 if (ref($v) eq 'ARRAY') {
502 if ($#{$v} == 0) {
503 $v = $v->[0];
504 } else {
505 $join = $default_delimiter->{$type} unless defined($join);
506 $v = join($join, @{$v});
507 }
508 } else {
509 warn("TT filter $type(): field $name values aren't ARRAY, ignoring");
510 }
511
512 return $v;
513 }
514 }
515
516 $args->{'d'} = tt_filter_type($args, 'display');
517 $args->{'display'} = tt_filter_type($args, 'display');
518
519 =head3 tt_filter_search
520
521 filter to return links to search, usage in TT:
522
523 [% search('FieldToDisplay','FieldToSearch','optional delimiter', 'optional_template.tt') %]
524
525 =cut
526
527 sub tt_filter_search {
528
529 my ($data) = @_;
530
531 die "no data?" unless ($data);
532
533 return sub {
534
535 my ($display,$search,$delimiter,$template) = @_;
536
537 # default delimiter
538 $delimiter ||= '&#182;<br/>',
539
540 die "no data hash" unless ($data->{'data'} && ref($data->{'data'}) eq 'HASH');
541 # Hm? Should we die here?
542 return unless ($display);
543
544 my $item = $data->{'data'}->{$display} || return;
545
546 return unless($item->{'display'});
547 if (! $item->{'search'}) {
548 warn "error in TT template: field $display didn't insert anything into search, use d('$display') and not search('$display'...)";
549 return;
550 }
551
552 my @warn;
553 foreach my $type (qw/display search/) {
554 push @warn, "field $display type $type values aren't ARRAY" unless (ref($item->{$type}) eq 'ARRAY');
555 }
556
557 if (@warn) {
558 warn("TT filter search(): " . join(",", @warn) . ", skipping");
559 return;
560 }
561 my @html;
562
563 my $d_el = $#{ $item->{'display'} };
564 my $s_el = $#{ $item->{'search'} };
565
566 # easy, both fields have same number of elements or there is just
567 # one search and multiple display
568 if ( $d_el == $s_el || $s_el == 0 ) {
569
570 foreach my $i ( 0 .. $d_el ) {
571
572 my $s;
573 if ($s_el > 0) {
574 $s = $item->{'search'}->[$i] or warn "can't find value $i for type search in field $search";
575 } else {
576 $s = $item->{'search'}->[0];
577 }
578 #$s =~ s/([^\w.-])/sprintf("%%%02X",ord($1))/eg;
579 $s = __quotemeta( $s );
580
581 my $d = $item->{'display'}->[$i] or warn "can't find value $i for type display in field $display";
582
583 my $template_arg = '';
584 $template_arg = qq{,'$template'} if ($template);
585
586 push @html, qq{<a href="#" onclick="return search_via_link('$search','$s'${template_arg})">$d</a>};
587 }
588
589 return join($delimiter, @html);
590 } else {
591 my $html = qq{<div class="notice">WARNING: we should really support if there is $d_el display elements and $s_el search elements, but currently there is no nice way to do so, so we will just display values</div>};
592 my $v = $item->{'display'};
593
594 if ($#{$v} == 0) {
595 $html .= $v->[0];
596 } else {
597 $html .= join($delimiter, @{$v});
598 }
599 return $html;
600 }
601 }
602 }
603
604 $args->{'search'} = tt_filter_search($args);
605
606 =head3 load_rec
607
608 Used mostly for onClick events like this:
609
610 <a href="#" onClick="[% load_rec( record_uri, 'template_name.tt') %]>foo</a>
611
612 It will automatically do sanity checking and create correct JavaScript code.
613
614 =cut
615
616 $args->{'load_rec'} = sub {
617 my @errors;
618
619 my $record_uri = shift or push @errors, "record_uri missing";
620 my $template = shift or push @errors, "template missing";
621
622 if ($record_uri !~ m#^[^/]+/[^/]+/[^/]+$#) {
623 push @errors, "invalid format of record_uri: $record_uri";
624 }
625
626 if (@errors) {
627 return "Logger.error('errors in load_rec: " . join(", ", @errors) . "'); return false;";
628 } else {
629 return "load_rec('$record_uri','$template'); return false;";
630 }
631 };
632
633 =head3 load_template
634
635 Used to re-submit search request and load results in different template
636
637 <a href="#" onClick="[% load_template( 'template_name.tt' ) %]">bar</a>
638
639 =cut
640
641 $args->{'load_template'} = sub {
642 my $template = shift or return "Logger.error('load_template missing template name!'); return false;";
643 return "load_template($template); return false;";
644 };
645
646 my $out;
647
648 $self->{'tt'}->process(
649 $args->{'template'},
650 $args,
651 \$out
652 ) || $log->error( "apply can't process template: ", $self->{'tt'}->error() );
653
654 return $out;
655 }
656
657
658 =head2 __quotemeta
659
660 Helper to quote JavaScript-friendly characters
661
662 =cut
663
664 sub __quotemeta {
665 local $_ = shift;
666 $_ = decode('iso-8859-2', $_);
667
668 s<([\x{0080}-\x{fffd}]+)>{sprintf '\u%0*v4X', '\u', $1}ge if ( Encode::is_utf8($_) );
669 {
670 use bytes;
671 s<((?:[^ \x21-\x7E]|(?:\\(?!u)))+)>{sprintf '\x%0*v2X', '\x', $1}ge;
672 }
673
674 s/\\x09/\\t/g;
675 s/\\x0A/\\n/g;
676 s/\\x0D/\\r/g;
677 s/"/\\"/g;
678 s/\\x5C/\\\\/g;
679
680 return $_;
681 }
682
683
684
685 =head1 AUTHOR
686
687 Dobrica Pavlinusic C<< <dpavlin@rot13.org> >>
688
689 =head1 LICENSE
690
691 This library is free software, you can redistribute it and/or modify
692 it under the same terms as Perl itself.
693
694 =cut
695
696 1;

  ViewVC Help
Powered by ViewVC 1.1.26