Removed LDAP-lookup loop for new external users.
[usit-rt.git] / sbin / rt-fulltext-indexer
CommitLineData
84fb5b46
MKG
1#!/usr/bin/perl
2# BEGIN BPS TAGGED BLOCK {{{
3#
4# COPYRIGHT:
5#
6# This software is Copyright (c) 1996-2012 Best Practical Solutions, LLC
7# <sales@bestpractical.com>
8#
9# (Except where explicitly superseded by other copyright notices)
10#
11#
12# LICENSE:
13#
14# This work is made available to you under the terms of Version 2 of
15# the GNU General Public License. A copy of that license should have
16# been provided with this software, but in any event can be snarfed
17# from www.gnu.org.
18#
19# This work is distributed in the hope that it will be useful, but
20# WITHOUT ANY WARRANTY; without even the implied warranty of
21# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
22# General Public License for more details.
23#
24# You should have received a copy of the GNU General Public License
25# along with this program; if not, write to the Free Software
26# Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA
27# 02110-1301 or visit their web page on the internet at
28# http://www.gnu.org/licenses/old-licenses/gpl-2.0.html.
29#
30#
31# CONTRIBUTION SUBMISSION POLICY:
32#
33# (The following paragraph is not intended to limit the rights granted
34# to you to modify and distribute this software under the terms of
35# the GNU General Public License and is only of importance to you if
36# you choose to contribute your changes and enhancements to the
37# community by submitting them to Best Practical Solutions, LLC.)
38#
39# By intentionally submitting any modifications, corrections or
40# derivatives to this work, or any other work intended for use with
41# Request Tracker, to Best Practical Solutions, LLC, you confirm that
42# you are the copyright holder for those contributions and you grant
43# Best Practical Solutions, LLC a nonexclusive, worldwide, irrevocable,
44# royalty-free, perpetual, license to use, copy, create derivative
45# works based on those contributions, and sublicense and distribute
46# those contributions and any derivatives thereof.
47#
48# END BPS TAGGED BLOCK }}}
49use strict;
50use warnings;
51no warnings 'once';
52
53# fix lib paths, some may be relative
54BEGIN {
55 require File::Spec;
56 my @libs = ("lib", "local/lib");
57 my $bin_path;
58
59 for my $lib (@libs) {
60 unless ( File::Spec->file_name_is_absolute($lib) ) {
61 unless ($bin_path) {
62 if ( File::Spec->file_name_is_absolute(__FILE__) ) {
63 $bin_path = ( File::Spec->splitpath(__FILE__) )[1];
64 }
65 else {
66 require FindBin;
67 no warnings "once";
68 $bin_path = $FindBin::Bin;
69 }
70 }
71 $lib = File::Spec->catfile( $bin_path, File::Spec->updir, $lib );
72 }
73 unshift @INC, $lib;
74 }
75}
76
77BEGIN {
78 use RT;
79 RT::LoadConfig();
80 RT::Init();
81};
82use RT::Interface::CLI ();
83
84my %OPT = (
85 help => 0,
86 debug => 0,
87);
88my @OPT_LIST = qw(help|h! debug!);
89
90my $db_type = RT->Config->Get('DatabaseType');
91if ( $db_type eq 'Pg' ) {
92 %OPT = (
93 %OPT,
94 limit => 0,
95 all => 0,
96 );
97 push @OPT_LIST, 'limit=i', 'all!';
98}
99elsif ( $db_type eq 'mysql' ) {
100 %OPT = (
101 %OPT,
102 limit => 0,
103 all => 0,
104 xmlpipe2 => 0,
105 );
106 push @OPT_LIST, 'limit=i', 'all!', 'xmlpipe2!';
107}
108elsif ( $db_type eq 'Oracle' ) {
109 %OPT = (
110 %OPT,
111 memory => '2M',
112 );
113 push @OPT_LIST, qw(memory=s);
114}
115
116use Getopt::Long qw(GetOptions);
117GetOptions( \%OPT, @OPT_LIST );
118
119if ( $OPT{'help'} ) {
120 RT::Interface::CLI->ShowHelp(
121 Sections => 'NAME|DESCRIPTION|'. uc($db_type),
122 );
123}
124
125my $fts_config = RT->Config->Get('FullTextSearch') || {};
126unless ( $fts_config->{'Enable'} ) {
127 print STDERR <<EOT;
128
129Full text search is disabled in your RT configuration. Run
b5747ff2 130/www/var/rt/sbin/rt-setup-fulltext-index to configure and enable it.
84fb5b46
MKG
131
132EOT
133 exit 1;
134}
135unless ( $fts_config->{'Indexed'} ) {
136 print STDERR <<EOT;
137
138Full text search is enabled in your RT configuration, but not with any
139full-text database indexing -- hence this tool is not required. Read
140the documentation for %FullTextSearch in your RT_Config for more details.
141
142EOT
143 exit 1;
144}
145
146if ( $db_type eq 'Oracle' ) {
147 my $index = $fts_config->{'IndexName'} || 'rt_fts_index';
148 $RT::Handle->dbh->do(
149 "begin ctx_ddl.sync_index(?, ?); end;", undef,
150 $index, $OPT{'memory'}
151 );
152 exit;
153} elsif ( $db_type eq 'mysql' ) {
154 unless ($OPT{'xmlpipe2'}) {
155 print STDERR <<EOT;
156
157Updates to the external Sphinx index are done via running the sphinx
158`indexer` tool:
159
160 indexer rt
161
162EOT
163 exit 1;
164 }
165}
166
167my @types = qw(text html);
168foreach my $type ( @types ) {
169 REDO:
170 my $attachments = attachments($type);
171 $attachments->Limit(
172 FIELD => 'id',
173 OPERATOR => '>',
174 VALUE => last_indexed($type)
175 );
176 $attachments->OrderBy( FIELD => 'id', ORDER => 'asc' );
177 $attachments->RowsPerPage( $OPT{'limit'} || 100 );
178
179 my $found = 0;
180 while ( my $a = $attachments->Next ) {
181 next if filter( $type, $a );
182 debug("Found attachment #". $a->id );
183 my $txt = extract($type, $a) or next;
184 $found++;
185 process( $type, $a, $txt );
186 debug("Processed attachment #". $a->id );
187 }
188 finalize( $type, $attachments ) if $found;
189 clean( $type );
190 goto REDO if $OPT{'all'} and $attachments->Count == ($OPT{'limit'} || 100)
191}
192
193sub attachments {
194 my $type = shift;
195 my $res = RT::Attachments->new( RT->SystemUser );
196 my $txn_alias = $res->Join(
197 ALIAS1 => 'main',
198 FIELD1 => 'TransactionId',
199 TABLE2 => 'Transactions',
200 FIELD2 => 'id',
201 );
202 $res->Limit(
203 ALIAS => $txn_alias,
204 FIELD => 'ObjectType',
205 VALUE => 'RT::Ticket',
206 );
207 my $ticket_alias = $res->Join(
208 ALIAS1 => $txn_alias,
209 FIELD1 => 'ObjectId',
210 TABLE2 => 'Tickets',
211 FIELD2 => 'id',
212 );
213 $res->Limit(
214 ALIAS => $ticket_alias,
215 FIELD => 'Status',
216 OPERATOR => '!=',
217 VALUE => 'deleted'
218 );
219
220 return goto_specific(
221 suffix => $type,
222 error => "Don't know how to find $type attachments",
223 arguments => [$res],
224 );
225}
226
227sub last_indexed {
228 my ($type) = (@_);
229 return goto_specific(
230 suffix => $db_type,
231 error => "Don't know how to find last indexed $type attachment for $db_type DB",
232 arguments => \@_,
233 );
234}
235
236sub filter {
237 my $type = shift;
238 return goto_specific(
239 suffix => $type,
240 arguments => \@_,
241 );
242}
243
244sub extract {
245 my $type = shift;
246 return goto_specific(
247 suffix => $type,
248 error => "No way to convert $type attachment into text",
249 arguments => \@_,
250 );
251}
252
253sub process {
254 return goto_specific(
255 suffix => $db_type,
256 error => "No processer for $db_type DB",
257 arguments => \@_,
258 );
259}
260
261sub finalize {
262 return goto_specific(
263 suffix => $db_type,
264 arguments => \@_,
265 );
266}
267
268sub clean {
269 return goto_specific(
270 suffix => $db_type,
271 arguments => \@_,
272 );
273}
274
275{
276sub last_indexed_mysql {
277 my $type = shift;
278 my $attr = $RT::System->FirstAttribute('LastIndexedAttachments');
279 return 0 unless $attr;
280 return 0 unless exists $attr->{ $type };
281 return $attr->{ $type } || 0;
282}
283
284sub process_mysql {
285 my ($type, $attachment, $text) = (@_);
286
287 my $doc = sphinx_template();
288
289 my $element = $doc->createElement('sphinx:document');
290 $element->setAttribute( id => $attachment->id );
291 $element->appendTextChild( content => $$text );
292
293 $doc->documentElement->appendChild( $element );
294}
295
296my $doc = undef;
297sub sphinx_template {
298 return $doc if $doc;
299
300 require XML::LibXML;
301 $doc = XML::LibXML::Document->new('1.0', 'UTF-8');
302 my $root = $doc->createElement('sphinx:docset');
303 $doc->setDocumentElement( $root );
304
305 my $schema = $doc->createElement('sphinx:schema');
306 $root->appendChild( $schema );
307 foreach ( qw(content) ) {
308 my $field = $doc->createElement('sphinx:field');
309 $field->setAttribute( name => $_ );
310 $schema->appendChild( $field );
311 }
312
313 return $doc;
314}
315
316sub finalize_mysql {
317 my ($type, $attachments) = @_;
318 sphinx_template()->toFH(*STDOUT, 1);
319}
320
321sub clean_mysql {
322 $doc = undef;
323}
324
325}
326
327sub last_indexed_pg {
328 my $type = shift;
329 my $attachments = attachments( $type );
330 my $alias = 'main';
331 if ( $fts_config->{'Table'} && $fts_config->{'Table'} ne 'Attachments' ) {
332 $alias = $attachments->Join(
333 TYPE => 'left',
334 FIELD1 => 'id',
335 TABLE2 => $fts_config->{'Table'},
336 FIELD2 => 'id',
337 );
338 }
339 $attachments->Limit(
340 ALIAS => $alias,
341 FIELD => $fts_config->{'Column'},
342 OPERATOR => 'IS NOT',
343 VALUE => 'NULL',
344 );
345 $attachments->OrderBy( FIELD => 'id', ORDER => 'desc' );
346 $attachments->RowsPerPage( 1 );
347 my $res = $attachments->First;
348 return 0 unless $res;
349 return $res->id;
350}
351
352sub process_pg {
353 my ($type, $attachment, $text) = (@_);
354
355 my $dbh = $RT::Handle->dbh;
356 my $table = $fts_config->{'Table'};
357 my $column = $fts_config->{'Column'};
358
359 my $query;
360 if ( $table ) {
361 if ( my ($id) = $dbh->selectrow_array("SELECT id FROM $table WHERE id = ?", undef, $attachment->id) ) {
362 $query = "UPDATE $table SET $column = to_tsvector(?) WHERE id = ?";
363 } else {
364 $query = "INSERT INTO $table($column, id) VALUES(to_tsvector(?), ?)";
365 }
366 } else {
367 $query = "UPDATE Attachments SET $column = to_tsvector(?) WHERE id = ?";
368 }
369
370 my $status = eval { $dbh->do( $query, undef, $$text, $attachment->id ) };
371 unless ( $status ) {
372 if ($dbh->errstr =~ /string is too long for tsvector/) {
373 warn "Attachment @{[$attachment->id]} not indexed, as it contains too many unique words to be indexed";
374 } else {
375 die "error: ". $dbh->errstr;
376 }
377 }
378}
379
380sub attachments_text {
381 my $res = shift;
382 $res->Limit( FIELD => 'ContentType', VALUE => 'text/plain' );
383 return $res;
384}
385
386sub extract_text {
387 my $attachment = shift;
388 my $text = $attachment->Content;
389 return undef unless defined $text && length($text);
390 return \$text;
391}
392
393sub attachments_html {
394 my $res = shift;
395 $res->Limit( FIELD => 'ContentType', VALUE => 'text/html' );
396 return $res;
397}
398
399sub filter_html {
400 my $attachment = shift;
401 if ( my $parent = $attachment->ParentObj ) {
402# skip html parts that are alternatives
403 return 1 if $parent->id
404 && $parent->ContentType eq 'mulitpart/alternative';
405 }
406 return 0;
407}
408
409sub extract_html {
410 my $attachment = shift;
411 my $text = $attachment->Content;
412 return undef unless defined $text && length($text);
413# TODO: html -> text
414 return \$text;
415}
416
417sub goto_specific {
418 my %args = (@_);
419
420 my $func = (caller(1))[3];
421 $func =~ s/.*:://;
422 my $call = $func ."_". lc $args{'suffix'};
423 unless ( defined &$call ) {
424 return undef unless $args{'error'};
425 require Carp; Carp::croak( $args{'error'} );
426 }
427 @_ = @{ $args{'arguments'} };
428 goto &$call;
429}
430
431
432# helper functions
433sub debug { print @_, "\n" if $OPT{debug}; 1 }
434sub error { $RT::Logger->error(_(@_)); 1 }
435sub warning { $RT::Logger->warn(_(@_)); 1 }
436
437=head1 NAME
438
439rt-fulltext-indexer - Indexer for full text search
440
441=head1 DESCRIPTION
442
443This is a helper script to keep full text indexes in sync with data.
444Read F<docs/full_text_indexing.pod> for complete details on how and when
445to run it.
446
447=head1 AUTHOR
448
449Ruslan Zakirov E<lt>ruz@bestpractical.comE<gt>,
450Alex Vandiver E<lt>alexmv@bestpractical.comE<gt>
451
452=cut
453