4 |
use strict; |
use strict; |
5 |
use warnings; |
use warnings; |
6 |
|
|
7 |
our $VERSION = '0.06'; |
our $VERSION = '0.10'; |
8 |
|
|
9 |
use Carp; |
use Carp; |
10 |
use File::Temp qw/ tempdir /; |
use File::Temp qw/ tempdir /; |
11 |
|
use BerkeleyDB; |
12 |
#use YAML; |
#use YAML; |
13 |
|
|
14 |
=head1 NAME |
=head1 NAME |
15 |
|
|
16 |
SWISH::PlusPlus - Perl extension SWISH++ |
SWISH::PlusPlus - Perl extension for full-text indexer SWISH++ with properties support |
17 |
|
|
18 |
=head1 SYNOPSIS |
=head1 SYNOPSIS |
19 |
|
|
23 |
=head1 DESCRIPTION |
=head1 DESCRIPTION |
24 |
|
|
25 |
This is perl module to use SWISH++ indexer by Paul J. Lucas. SWISH++ is |
This is perl module to use SWISH++ indexer by Paul J. Lucas. SWISH++ is |
26 |
rewrite of swish-e in C++ which is extremly fast (thank to mmap), but without |
rewrite of swish-e in C++ which is extremely fast (due to mmap usage and |
27 |
support for properties (which this module tries to fix). |
clever language heuristics), but without support for properties (which this |
28 |
|
module tries to fix). |
29 |
Implementation of this module is crafted after L<Plucene::Simple> and it |
|
30 |
should be easy to replace Plucene with this module for increased |
Implementation of API is something in-between C<SWISH::API> and |
31 |
performance. However, this module is not plug-in replacement. |
C<Plucene::Simple>. It should be easy to replace Plucene or swish-e with |
32 |
|
this module for increased performance. However, this module is not plug-in |
33 |
|
replacement. |
34 |
|
|
35 |
=head1 METHODS |
=head1 METHODS |
36 |
|
|
37 |
=head2 new |
=head2 new |
38 |
|
|
39 |
Create new indexing object. |
Create new instance for index. |
40 |
|
|
41 |
my $i = SWISH::PlusPlus->new( |
my $i = SWISH::PlusPlus->new( |
42 |
index_dir => '/path/to/index', |
index_dir => '/path/to/index', |
47 |
use_stopwords => 1, |
use_stopwords => 1, |
48 |
); |
); |
49 |
|
|
50 |
Options to new are following: |
Options are described below: |
51 |
|
|
52 |
=over 5 |
=over 5 |
53 |
|
|
54 |
=item C<index_dir> |
=item C<index_dir> |
55 |
|
|
56 |
Path to directory in which index will be created. |
Path to directory in which index and meta database will be created. |
57 |
|
|
58 |
=item C<index> |
=item C<index> |
59 |
|
|
73 |
=item C<meta_in_body> |
=item C<meta_in_body> |
74 |
|
|
75 |
This option (off by default) enables to search content of meta fields |
This option (off by default) enables to search content of meta fields |
76 |
without specifing them (like they are in body of document). This will |
without specifying them (like they are in body of document). This will |
77 |
somewhat increate index size. |
somewhat increase index size. |
78 |
|
|
79 |
=item C<use_stopwords> |
=item C<use_stopwords> |
80 |
|
|
95 |
|
|
96 |
my $index_dir = $self->{'index_dir'}; |
my $index_dir = $self->{'index_dir'}; |
97 |
|
|
98 |
|
my $cwd; |
99 |
|
chomp($cwd = `pwd`); |
100 |
|
$self->{'cwd'} = $cwd || carp "can't get cwd!"; |
101 |
|
|
102 |
if ($index_dir !~ m#^/#) { |
if ($index_dir !~ m#^/#) { |
|
chomp(my $cwd = `pwd`); |
|
103 |
$index_dir = "$cwd/$index_dir"; |
$index_dir = "$cwd/$index_dir"; |
104 |
print STDERR "## full path to index_dir: $index_dir\n" if ($self->{'debug'}); |
print STDERR "## full path to index_dir: $index_dir\n" if ($self->{'debug'}); |
105 |
$self->{'index_dir'} = $index_dir; |
$self->{'index_dir'} = $index_dir; |
121 |
|
|
122 |
=head2 check_bin |
=head2 check_bin |
123 |
|
|
124 |
Check if swish++ binaries specified in L<new> are available and verify |
Check if SWISH++ binaries specified in L<new> are available and verify |
125 |
version signature. |
version signature. |
126 |
|
|
127 |
if ($i->check_bin) { |
if ($i->check_bin) { |
132 |
|
|
133 |
$i->{'version'} |
$i->{'version'} |
134 |
|
|
135 |
which you can examine to see version. |
which you can examined to see numeric version (something like C<6.0.4>). |
136 |
|
|
137 |
=cut |
=cut |
138 |
|
|
149 |
confess $self->{'search'}," binary is not SWISH++" unless ($s =~ m/^SWISH\+\+/); |
confess $self->{'search'}," binary is not SWISH++" unless ($s =~ m/^SWISH\+\+/); |
150 |
|
|
151 |
if ($i eq $s) { |
if ($i eq $s) { |
152 |
|
$i =~ s/^SWISH\+\+\s+// || confess "can't strip SWISH++ from version"; |
153 |
$self->{'version'} = $i; |
$self->{'version'} = $i; |
154 |
return 1; |
return 1; |
155 |
} else { |
} else { |
163 |
|
|
164 |
Quick way to add simple data to index. |
Quick way to add simple data to index. |
165 |
|
|
166 |
$i->index_document($key, $data); |
$i->index_document($path, $data); |
167 |
$i->index_document( 42 => 'meaning of life' ); |
$i->index_document( 42 => 'meaning of life' ); |
168 |
|
|
169 |
|
C<$path> value is really path, so you don't want to use directory |
170 |
|
separators (slashes, /) in it probably. |
171 |
|
|
172 |
=cut |
=cut |
173 |
|
|
174 |
sub index_document { |
sub index_document { |
188 |
|
|
189 |
=head2 add |
=head2 add |
190 |
|
|
191 |
Add document with metadata to index. |
Add document with meta-data to index. |
192 |
|
|
193 |
$i->add( |
$i->add( |
194 |
path => 'path/to/document', |
path => 'path/to/document', |
212 |
|
|
213 |
return 1; |
return 1; |
214 |
} |
} |
215 |
|
|
216 |
=head2 search |
=head2 search |
217 |
|
|
218 |
Search your index. |
Search your index using any valid SWISH++ query. |
219 |
|
|
220 |
my @results = $i->search("swhish query"); |
my @results = $i->search("swish query"); |
221 |
|
|
222 |
Returns array with result IDs. |
Returns array with elements like this: |
223 |
|
|
224 |
|
{ |
225 |
|
rank => 10, # rank of result |
226 |
|
path => 'path to result', # path to result |
227 |
|
size => 999, # size in bytes |
228 |
|
title => 'title of result' # title meta property |
229 |
|
} |
230 |
|
|
231 |
=cut |
=cut |
232 |
|
|
235 |
|
|
236 |
my $query = shift || return; |
my $query = shift || return; |
237 |
|
|
238 |
$self->_close_index; |
$self->finish_update; |
239 |
|
$self->_tie_meta_db(DB_RDONLY); |
240 |
|
|
241 |
my @results; |
my @results; |
242 |
|
|
243 |
# escape double quotes in query for shell |
# escape double quotes in query for shell |
244 |
$query =~ s/"/\\"/g; |
$query =~ s/"/\\"/g; |
245 |
|
|
246 |
my $open_cmd = $self->{'search'}." -i ".$self->{'index_dir'}.'/index "'.$query.'" |'; |
my $open_cmd = $self->{'search'} . |
247 |
print STDERR "## search $open_cmd\n" if ($self->{'debug'}); |
' -i ' . $self->{'index_dir'}.'/index' . |
248 |
|
' "' . $query . '"'. |
249 |
|
' |'; |
250 |
|
print STDERR "## search: $open_cmd\n" if ($self->{'debug'}); |
251 |
|
|
252 |
open(SEARCH, $open_cmd) || confess "can't start $open_cmd: $!"; |
open(SEARCH, $open_cmd) || confess "can't start $open_cmd: $!"; |
253 |
while(<SEARCH>) { |
my $l; |
254 |
next if (/^#/); |
while($l = <SEARCH>) { |
255 |
chomp; |
next if ($l =~ /^#/); |
256 |
print STDERR "## $_\n" if ($self->{'debug'}); |
chomp($l); |
257 |
my ($rank,$path,$size,$title) = split(/ /,$_,4); |
print STDERR "## $l\n" if ($self->{'debug'}); |
258 |
|
my ($rank,$path,$size,$title) = split(/ /,$l,4); |
259 |
|
$path =~ s#^\./##; # strip from path |
260 |
push @results, { |
push @results, { |
261 |
rank => $rank, |
rank => $rank, |
262 |
path => $path, |
path => $path, |
272 |
return @results; |
return @results; |
273 |
} |
} |
274 |
|
|
275 |
|
=head2 property |
276 |
|
|
277 |
|
Return stored meta property from result or result path. |
278 |
|
|
279 |
|
print $i->property('path', 'title'); |
280 |
|
print $i->property($res->{'path'}, 'title'); |
281 |
|
|
282 |
|
=cut |
283 |
|
|
284 |
|
sub property { |
285 |
|
my $self = shift; |
286 |
|
|
287 |
|
my ($path,$meta) = @_; |
288 |
|
|
289 |
|
if ($path =~ m/^HASH/) { |
290 |
|
$path = $path->{'path'} || confess "can't find path in input data"; |
291 |
|
} |
292 |
|
|
293 |
|
my $val = $self->{'meta_db'}->{"$path-$meta"}; |
294 |
|
|
295 |
|
print STDERR "## property $path-$meta: ",($val || 'undef'),"\n" if ($self->{'debug'}); |
296 |
|
return $val; |
297 |
|
} |
298 |
|
|
299 |
=head2 finish_update |
=head2 finish_update |
300 |
|
|
301 |
This method will close index. |
This method will close index binary and enable search. Searching is not |
302 |
|
available while indexing is in process. |
303 |
|
|
304 |
$i->finish_update; |
$i->finish_update; |
305 |
|
|
306 |
It will be called on DESTROY when $i goes out of scope. |
Usually, you don't need to call this method directly. It will be called on |
307 |
|
DESTROY when $i goes out of scope or when you first call search in session |
308 |
|
if indexing was started. |
309 |
|
|
310 |
=cut |
=cut |
311 |
|
|
312 |
sub finish_update { |
sub finish_update { |
313 |
my $self = shift; |
my $self = shift; |
314 |
|
|
315 |
$self->_close_index; |
print STDERR "## finish_update\n" if ($self->{'debug'}); |
316 |
|
|
317 |
|
$self->_close_index && $self->_untie_meta_db; |
318 |
} |
} |
319 |
|
|
320 |
sub DESTROY { |
sub DESTROY { |
324 |
|
|
325 |
=head1 PRIVATE METHODS |
=head1 PRIVATE METHODS |
326 |
|
|
327 |
Private methods implement internals for creating temporary file needed for |
Private methods implement internals for creating temporary files needed for |
328 |
swish++. You should have no need to call them directly, and they are here |
SWISH++. You should have no need to call them directly, and they are here |
329 |
just to have documentation. |
just to have documentation. |
330 |
|
|
331 |
=head2 _init_indexer |
=head2 _init_indexer |
342 |
sub _init_indexer { |
sub _init_indexer { |
343 |
my $self = shift; |
my $self = shift; |
344 |
|
|
345 |
$self->{'tmp_dir'} = tempdir( CLEANUP => 1 ) || confess "can't create temporary directory: $!"; |
return if ($self->{'_index_fh'}); |
346 |
|
|
347 |
|
my $tmp_dir = tempdir( CLEANUP => 1 ) || confess "can't create temporary directory: $!"; |
348 |
|
$self->{'tmp_dir'} = $tmp_dir; |
349 |
|
|
350 |
chdir $self->{'tmp_dir'} || confess "can't chdir to ".$self->{'tmp_dir'}.": $!"; |
chdir $tmp_dir || confess "can't chdir to ".$tmp_dir.": $!"; |
351 |
|
|
352 |
|
print STDERR "## tmp_dir: $tmp_dir" if ($self->{'debug'}); |
353 |
|
|
354 |
my $opt = "-v " . ($self->{'debug'} || '0'); |
my $opt = "-v " . ($self->{'debug'} || '0'); |
355 |
|
|
360 |
$opt .= " -s _stopwords_"; |
$opt .= " -s _stopwords_"; |
361 |
} |
} |
362 |
|
|
363 |
my $open_cmd = '| '.$self->{'index'}.' '.$opt.' -e "html:*" -i '.$self->{'index_dir'}.'/index -'; |
my $index_dir = $self->{'index_dir'} || confess "no index_dir?"; |
364 |
|
|
365 |
|
my $open_cmd = '| '.$self->{'index'}.' '.$opt.' -e "html:*" -i '.$index_dir.'/index -'; |
366 |
|
|
367 |
print STDERR "## open index $open_cmd\n" if ($self->{'index'}); |
print STDERR "## init_indexer: $open_cmd\n" if ($self->{'debug'}); |
368 |
|
|
369 |
open($self->{'index_fh'}, $open_cmd) || confess "can't start index with $open_cmd: $!"; |
open($self->{'_index_fh'}, $open_cmd) || confess "can't start index with $open_cmd: $!"; |
370 |
|
|
371 |
|
chdir $self->{'cwd'} || confess "can't chdir to ".$self->{'cwd'}.": $!"; |
372 |
|
|
373 |
return $self->{'index_fh'}; |
$self->_tie_meta_db(DB_CREATE); |
374 |
|
|
375 |
|
return $self->{'_index_fh'}; |
376 |
} |
} |
377 |
|
|
378 |
=head2 _create_doc |
=head2 _create_doc |
379 |
|
|
380 |
Create temporary file and pass it's name to swish++ |
Create temporary file and pass it's name to SWISH++ |
381 |
|
|
382 |
$i->_create_doc( |
$i->_create_doc( |
383 |
path => 'path/to/store/in/index', |
path => 'path/to/store/in/index', |
399 |
my $arg = {@_}; |
my $arg = {@_}; |
400 |
|
|
401 |
# open indexer if needed |
# open indexer if needed |
402 |
$self->{'index_fh'} ||= $self->_init_indexer; |
$self->_init_indexer; |
403 |
|
|
404 |
my $path = $self->{'tmp_dir'} || confess "no tmp_dir?"; |
my $path = $self->{'tmp_dir'} || confess "no tmp_dir?"; |
405 |
|
my $id = $arg->{'path'} || confess "no path?"; |
406 |
|
$path .= "/$id"; |
407 |
|
|
408 |
|
print STDERR "## _create_doc: $path\n" if ($self->{'debug'}); |
409 |
|
|
410 |
open(TMP, '>', $arg->{'path'}) || die "can't create temp file ".$arg->{'path'}.": $!"; |
open(TMP, '>', $path) || die "can't create temp file $path: $!"; |
411 |
|
|
412 |
print TMP '<html><head>'; |
print TMP '<html><head>'; |
413 |
|
|
418 |
my $content = $arg->{'meta'}->{$name}; |
my $content = $arg->{'meta'}->{$name}; |
419 |
print TMP qq{<meta name="$name" content="$content">}; |
print TMP qq{<meta name="$name" content="$content">}; |
420 |
$arg->{'body'} .= " $content" if ($self->{'meta_in_body'}); |
$arg->{'body'} .= " $content" if ($self->{'meta_in_body'}); |
421 |
|
$self->{'meta_db'}->{"$id-$name"} = $content; |
422 |
} |
} |
423 |
} |
} |
424 |
|
|
425 |
if (defined($arg->{'title'})) { |
my $title = $arg->{'title'}; |
426 |
print TMP '<title>' . ($arg->{'title'} || '') . '</title>'; |
if (defined($title)) { |
427 |
$arg->{'body'} .= " ".$arg->{'title'} if ($self->{'meta_in_body'}); |
print TMP "<title>$title</title>"; |
428 |
|
$arg->{'body'} .= " $title" if ($self->{'meta_in_body'}); |
429 |
|
$self->{'meta_db'}->{"$id-title"} = $title; |
430 |
} |
} |
431 |
|
|
432 |
print TMP '</head><body>' . $arg->{'body'} . '</body></html>'; |
print TMP '</head><body>' . $arg->{'body'} . '</body></html>'; |
433 |
|
|
434 |
close(TMP) || confess "can't close tmp file ".$arg->{'path'}.": $!"; |
close(TMP) || confess "can't close tmp file ".$arg->{'path'}.": $!"; |
435 |
|
|
436 |
print { $self->{'index_fh'} } $arg->{'path'}."\n"; |
print { $self->{'_index_fh'} } "$id\n"; |
437 |
} |
} |
438 |
|
|
439 |
=head2 _close_index |
=head2 _close_index |
449 |
sub _close_index { |
sub _close_index { |
450 |
my $self = shift; |
my $self = shift; |
451 |
|
|
452 |
return unless ($self->{'index_fh'}); |
return unless ($self->{'_index_fh'}); |
453 |
|
|
454 |
print STDERR "## close index\n" if ($self->{'debug'}); |
print STDERR "## close index\n" if ($self->{'debug'}); |
455 |
|
|
456 |
close($self->{'index_fh'}); |
close($self->{'_index_fh'}) || confess "can't close index: $!"; |
457 |
undef $self->{'index_fh'}; |
undef $self->{'_index_fh'}; |
458 |
|
|
459 |
|
return 1; |
460 |
|
} |
461 |
|
|
462 |
|
=head2 _tie_meta_db |
463 |
|
|
464 |
|
Open BerkeleyDB database with meta properties. |
465 |
|
|
466 |
|
$i->_tie_meta_db(DB_CREATE); |
467 |
|
$i->_tie_meta_db(DB_RDONLY); |
468 |
|
|
469 |
|
} |
470 |
|
|
471 |
|
=cut |
472 |
|
|
473 |
|
sub _tie_meta_db { |
474 |
|
my $self = shift; |
475 |
|
|
476 |
|
my $flags = shift || confess "need DB_CREATE or DB_RDONLY"; |
477 |
|
|
478 |
|
return if ($self->{'_meta_db_flags'} && $self->{'_meta_db_flags'} == $flags); |
479 |
|
|
480 |
|
print STDERR "## _tie_meta_db($flags)\n" if ($self->{'debug'}); |
481 |
|
|
482 |
|
$self->_untie_meta_db; |
483 |
|
$self->{'_meta_db_flags'} = $flags; |
484 |
|
|
485 |
|
my $file = $self->{'index_dir'}.'/meta.db'; |
486 |
|
|
487 |
|
tie %{$self->{'meta_db'}}, "BerkeleyDB::Hash", |
488 |
|
-Filename => $file, |
489 |
|
-Flags => $flags |
490 |
|
or confess "cannot open $file: $! $BerkeleyDB::Error\n" ; |
491 |
|
|
492 |
|
return 1; |
493 |
|
} |
494 |
|
|
495 |
|
=head2 _untie_meta_db |
496 |
|
|
497 |
|
Close BerkeleyDB database with meta properties. |
498 |
|
|
499 |
|
$i->_untie_meta_db; |
500 |
|
|
501 |
|
=cut |
502 |
|
|
503 |
|
sub _untie_meta_db { |
504 |
|
my $self = shift; |
505 |
|
|
506 |
|
return unless ($self->{'meta_db'}); |
507 |
|
|
508 |
|
print STDERR "## _untie_meta_db\n" if ($self->{'debug'}); |
509 |
|
untie %{$self->{'meta_db'}} || confess "can't untie!"; |
510 |
|
undef $self->{'meta_db'}; |
511 |
|
undef $self->{'_meta_db_flags'}; |
512 |
|
|
513 |
|
return 1; |
514 |
} |
} |
515 |
|
|
516 |
1; |
1; |
524 |
|
|
525 |
=head2 Debian |
=head2 Debian |
526 |
|
|
527 |
Debian version of swish++ is often old (version 5 at moment of this writing |
Debian version of SWISH++ is often old (version 5 at moment of this writing |
528 |
while version 6 is available in source code), so this module by default |
while version 6 is available in source code), so this module by default |
529 |
uses executable names B<index> and B<search> for self-compiled version |
uses executable names B<index> and B<search> for self-compiled version |
530 |
instead of one from Debian package. See L<new> how to specify Debian |
instead of one from Debian package. See L<new> how to specify Debian |
532 |
|
|
533 |
=head2 SWISH++ |
=head2 SWISH++ |
534 |
|
|
535 |
Aside from very good rewrite in C++, SWISH++ is fatster because it has |
Aside from very good rewrite in C++, SWISH++ is faster because it uses |
536 |
claver heuristics about which data in input files are words to index and |
claver heuristics about which data in input files are words to index and |
537 |
which are not. It's based on English language and might be best choice if |
which are not. It's based on English language and might be best choice if |
538 |
you plan to install large amount of long text documents. |
you plan to index large amount of long text documents. |
539 |
|
|
540 |
However, if you plan to index all data from structured storage (e.g. RDBMS) |
However, if you plan to index all data from structured storage (e.g. RDBMS) |
541 |
you might want B<all> words from data to end up in index as opposed to just |
you might want B<all> words from data to end up in index as opposed to just |
543 |
don't plan to index English texts with this module. |
don't plan to index English texts with this module. |
544 |
|
|
545 |
With distribution build versions of SWISH++ you might have problems with |
With distribution build versions of SWISH++ you might have problems with |
546 |
disepearing words. To overcome this problem, you will have to compile and |
disapearing words. To overcome this problem, you will have to compile and |
547 |
configure SWISH++ yourself (because language characteristics are |
configure SWISH++ yourself (because language characteristics are |
548 |
compilation-time option). |
compilation-time option). |
549 |
|
|
559 |
doesn't recognize 2004-12-05 as date. Have in mind that your index size |
doesn't recognize 2004-12-05 as date. Have in mind that your index size |
560 |
might explode. |
might explode. |
561 |
|
|
562 |
|
=head1 BUGS |
563 |
|
|
564 |
|
Currently there is no way to specify which meta data will be stored as |
565 |
|
properties. B<This will be fixed very soon>. |
566 |
|
|
567 |
|
There is no garbage collection on temporary files created for SWISH++. This |
568 |
|
means that one run of indexer will take additional disk space for temporary |
569 |
|
files, which will be removed at end. There should be some way to remove |
570 |
|
files after they are indexed by SWISH++. However, at this early stage of |
571 |
|
development it's just not supported yet. Have plenty of disk space! |
572 |
|
|
573 |
=head1 SEE ALSO |
=head1 SEE ALSO |
574 |
|
|
575 |
C<swish++> web site L<http://homepage.mac.com/pauljlucas/software/swish/> |
SWISH++ web site L<http://homepage.mac.com/pauljlucas/software/swish/> |
576 |
|
|
577 |
=head1 AUTHOR |
=head1 AUTHOR |
578 |
|
|