1 | ###########################################################################
|
---|
2 | #
|
---|
3 | # solrbuildproc.pm -- perl wrapper for building index with Solr
|
---|
4 | # A component of the Greenstone digital library software
|
---|
5 | # from the New Zealand Digital Library Project at the
|
---|
6 | # University of Waikato, New Zealand.
|
---|
7 | #
|
---|
8 | # Copyright (C) 1999 New Zealand Digital Library Project
|
---|
9 | #
|
---|
10 | # This program is free software; you can redistribute it and/or modify
|
---|
11 | # it under the terms of the GNU General Public License as published by
|
---|
12 | # the Free Software Foundation; either version 2 of the License, or
|
---|
13 | # (at your option) any later version.
|
---|
14 | #
|
---|
15 | # This program is distributed in the hope that it will be useful,
|
---|
16 | # but WITHOUT ANY WARRANTY; without even the implied warranty of
|
---|
17 | # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
---|
18 | # GNU General Public License for more details.
|
---|
19 | #
|
---|
20 | # You should have received a copy of the GNU General Public License
|
---|
21 | # along with this program; if not, write to the Free Software
|
---|
22 | # Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
|
---|
23 | #
|
---|
24 | ###########################################################################
|
---|
25 |
|
---|
26 | package solrbuildproc;
|
---|
27 |
|
---|
28 | # This document processor outputs a document for solr to process
|
---|
29 |
|
---|
30 | # Rather then use the XML structure developed for mgppbuilder/mgppbuildproc
|
---|
31 | # whose use was then extended to Lucene, Solr has its own XML syntax:
|
---|
32 | #
|
---|
33 | # http://wiki.apache.org/solr/UpdateXmlMessages
|
---|
34 | #
|
---|
35 | # Using this means we don't need to write SolrWrapper.jar, as had to be
|
---|
36 | # done for Lucene, translating the XML syntax piped to it into appropriate
|
---|
37 | # calls to the Lucene API
|
---|
38 |
|
---|
39 |
|
---|
40 | use lucenebuildproc;
|
---|
41 | use ghtml;
|
---|
42 | use strict;
|
---|
43 | no strict 'refs'; # allow filehandles to be variables and viceversa
|
---|
44 |
|
---|
45 |
|
---|
46 | use IncrementalBuildUtils;
|
---|
47 |
|
---|
48 | sub BEGIN {
|
---|
49 | @solrbuildproc::ISA = ('lucenebuildproc');
|
---|
50 | }
|
---|
51 |
|
---|
52 |
|
---|
53 | sub new {
|
---|
54 | my $class = shift @_;
|
---|
55 | my $self = new lucenebuildproc (@_);
|
---|
56 |
|
---|
57 | $self->{'actualfacetfields'} = {};
|
---|
58 | $self->{'facetfieldnamemap'} = {};
|
---|
59 | return bless $self, $class;
|
---|
60 | }
|
---|
61 |
|
---|
62 | sub set_facetfields {
|
---|
63 | my $self = shift (@_);
|
---|
64 |
|
---|
65 | my ($facetfields) = @_;
|
---|
66 | $self->{'facetfields'} = ();
|
---|
67 | # lets just go through and check for text, allfields, metadata which are only valid for indexes, not for facetfields
|
---|
68 | foreach my $s (@$facetfields) {
|
---|
69 | if ($s !~ /^(text|allfields|metadata)$/) {
|
---|
70 | push (@{$self->{'facetfields'}}, $s);
|
---|
71 | }
|
---|
72 | }
|
---|
73 | }
|
---|
74 |
|
---|
75 | #----
|
---|
76 |
|
---|
77 | sub index_field_mapping_edit {
|
---|
78 | my $self = shift (@_);
|
---|
79 | my ($doc_obj,$file,$edit_mode) = @_;
|
---|
80 |
|
---|
81 | # Only add/update gets to here
|
---|
82 | # Currently there is no need to distinguish between these edit modes
|
---|
83 |
|
---|
84 | my $outhandle = $self->{'outhandle'};
|
---|
85 |
|
---|
86 | # only study this document if it is one to be indexed
|
---|
87 | return if ($doc_obj->get_doc_type() ne "indexed_doc");
|
---|
88 |
|
---|
89 | my $indexed_doc = $self->is_subcollection_doc($doc_obj);
|
---|
90 |
|
---|
91 | # get the parameters for the output
|
---|
92 | # split on : just in case there is subcoll and lang stuff
|
---|
93 | my ($fields) = split (/:/, $self->{'index'});
|
---|
94 |
|
---|
95 | my $doc_section = 0; # just for this document
|
---|
96 |
|
---|
97 | # get the text for this document
|
---|
98 | my $section = $doc_obj->get_top_section();
|
---|
99 |
|
---|
100 | while (defined $section)
|
---|
101 | {
|
---|
102 | $doc_section++;
|
---|
103 |
|
---|
104 | # if we are doing subcollections, then some docs shouldn't be
|
---|
105 | # considered for indexing
|
---|
106 |
|
---|
107 | my $indexed_section
|
---|
108 | = $doc_obj->get_metadata_element($section, "gsdldoctype")
|
---|
109 | || "indexed_section";
|
---|
110 |
|
---|
111 | if (($indexed_doc == 0)
|
---|
112 | || ($indexed_section ne "indexed_section" && $indexed_section ne "indexed_doc")) {
|
---|
113 | $section = $doc_obj->get_next_section($section);
|
---|
114 | next;
|
---|
115 | }
|
---|
116 |
|
---|
117 | # has the user added a 'metadata' index?
|
---|
118 | my $all_metadata_specified = 0;
|
---|
119 |
|
---|
120 | # which fields have already been indexed?
|
---|
121 | # (same as fields, but in a map)
|
---|
122 | my $specified_fields = {};
|
---|
123 |
|
---|
124 | # do we have an allfields index??
|
---|
125 | my $allfields_index = 0;
|
---|
126 |
|
---|
127 | # collect up all the text for it in here
|
---|
128 | my $allfields_text = "";
|
---|
129 |
|
---|
130 | foreach my $field (split (/;/, $fields)) {
|
---|
131 | if ($field eq "allfields") {
|
---|
132 | $allfields_index = 1;
|
---|
133 | } elsif ($field eq "metadata") {
|
---|
134 | $all_metadata_specified = 1;
|
---|
135 | }
|
---|
136 | }
|
---|
137 |
|
---|
138 | foreach my $field (split (/;/, $fields)) {
|
---|
139 |
|
---|
140 | # only deal with this field if it doesn't start with top or
|
---|
141 | # this is the first section
|
---|
142 | my $real_field = $field;
|
---|
143 | next if (($real_field =~ s/^top//) && ($doc_section != 1));
|
---|
144 |
|
---|
145 | # process these two later
|
---|
146 | next if ($real_field eq "allfields" || $real_field eq "metadata");
|
---|
147 |
|
---|
148 | # individual metadata and or text specified
|
---|
149 | # -- could be a comma separated list
|
---|
150 | $specified_fields->{$real_field} = 1;
|
---|
151 |
|
---|
152 | if (!defined $self->{'indexfieldmap'}->{$real_field}) {
|
---|
153 | my $shortname = $self->get_or_create_shortname($real_field);
|
---|
154 | $self->{'indexfieldmap'}->{$real_field} = $shortname;
|
---|
155 | $self->{'indexfieldmap'}->{$shortname} = 1;
|
---|
156 | }
|
---|
157 | } # foreach field
|
---|
158 |
|
---|
159 |
|
---|
160 | if ($all_metadata_specified) {
|
---|
161 |
|
---|
162 | my $new_text = "";
|
---|
163 | my $shortname = "";
|
---|
164 | my $metadata = $doc_obj->get_all_metadata ($section);
|
---|
165 |
|
---|
166 | foreach my $pair (@$metadata) {
|
---|
167 | my ($mfield, $mvalue) = (@$pair);
|
---|
168 |
|
---|
169 | # no value
|
---|
170 | next unless defined $mvalue && $mvalue ne "";
|
---|
171 |
|
---|
172 | # we have already indexed this
|
---|
173 | next if defined ($specified_fields->{$mfield});
|
---|
174 |
|
---|
175 | # check fields here, maybe others dont want - change to use dontindex!!
|
---|
176 | next if ($mfield eq "Identifier" || $mfield eq "classifytype" || $mfield eq "assocfilepath");
|
---|
177 | next if ($mfield =~ /^gsdl/);
|
---|
178 |
|
---|
179 | if (defined $self->{'indexfieldmap'}->{$mfield}) {
|
---|
180 | $shortname = $self->{'indexfieldmap'}->{$mfield};
|
---|
181 | }
|
---|
182 | else {
|
---|
183 | $shortname = $self->get_or_create_shortname($mfield);
|
---|
184 | $self->{'indexfieldmap'}->{$mfield} = $shortname;
|
---|
185 | $self->{'indexfieldmap'}->{$shortname} = 1;
|
---|
186 | }
|
---|
187 |
|
---|
188 | if (!defined $self->{'indexfields'}->{$mfield}) {
|
---|
189 | $self->{'indexfields'}->{$mfield} = 1;
|
---|
190 | }
|
---|
191 | }
|
---|
192 | }
|
---|
193 |
|
---|
194 | if ($allfields_index) {
|
---|
195 | # add the index name mapping
|
---|
196 | $self->{'indexfieldmap'}->{"allfields"} = "ZZ";
|
---|
197 | $self->{'indexfieldmap'}->{"ZZ"} = 1;
|
---|
198 | }
|
---|
199 |
|
---|
200 | $section = $doc_obj->get_next_section($section);
|
---|
201 |
|
---|
202 | } # while defined section
|
---|
203 |
|
---|
204 |
|
---|
205 | }
|
---|
206 |
|
---|
207 | # UNUSED now by default.
|
---|
208 | # Georgy overrode the mgppbuildproc::create_shortname() method in commit 32441 to create the method below, to override the inherited
|
---|
209 | # behaviour so that create_shortname() worked appropriately for his use cases involving multiple analyzers.
|
---|
210 | # As a result, create_shortname() for solr no longer did a lookup into the %mgppbuildproc::static_indexfield_map for registered shortnames.
|
---|
211 | # For the rest, this method is a copy mgppbuildproc::create_shortname().
|
---|
212 | # But we want the original mgppbuildproc::create_shortname() behaviour restored, as it does the lookups into %static_indexfield_map that's necessary for us.
|
---|
213 | # So we've renamed this function to create_shortname_multi_solr_analyzer below so it won't get called as default beahviour any more.
|
---|
214 | # Rename to create_shortname() when requiring Georgy's behaviour.
|
---|
215 | sub create_shortname_multi_solr_analyzer {
|
---|
216 | my $self = shift(@_);
|
---|
217 |
|
---|
218 | my ($realname) = @_;
|
---|
219 | my @realnamelist = split(",", $realname);
|
---|
220 | map {$_=~ s/^[a-zA-Z]+\.//;} @realnamelist; #remove namespaces
|
---|
221 | my ($singlename) = $realnamelist[0];
|
---|
222 |
|
---|
223 | # try our predefined static mapping
|
---|
224 | my $name;
|
---|
225 | # we can't use the quick map, so join all fields back together (without namespaces), and try sets of two characters.
|
---|
226 | $realname = join ("", @realnamelist);
|
---|
227 | #try the first two chars
|
---|
228 | my $shortname;
|
---|
229 | if ($realname =~ /^[^\w]*(\w)[^\w]*(\w)/) {
|
---|
230 | $shortname = "$1$2";
|
---|
231 | } else {
|
---|
232 | # there aren't two letdig's in the field - try arbitrary combinations
|
---|
233 | $realname = "ABCDEFGHIJKLMNOPQRSTUVWXYZ";
|
---|
234 | $shortname = "AB";
|
---|
235 | }
|
---|
236 | $shortname =~ tr/a-z/A-Z/;
|
---|
237 |
|
---|
238 | #if already used, take the first and third letdigs and so on
|
---|
239 | my $count = 1;
|
---|
240 | while (defined $self->{'indexfieldmap'}->{$shortname}) {
|
---|
241 | if ($realname =~ /^[^\w]*(\w)([^\w]*\w){$count}[^\w]*(\w)/) {
|
---|
242 | $shortname = "$1$3";
|
---|
243 | $count++;
|
---|
244 | $shortname =~ tr/a-z/A-Z/;
|
---|
245 |
|
---|
246 | }
|
---|
247 | else {
|
---|
248 | #remove up to and incl the first letdig
|
---|
249 | $realname =~ s/^[^\w]*\w//;
|
---|
250 | $count = 0;
|
---|
251 | }
|
---|
252 | }
|
---|
253 |
|
---|
254 | return $shortname ;
|
---|
255 | }
|
---|
256 |
|
---|
257 |
|
---|
258 | sub index_field_mapping {
|
---|
259 | my $self = shift (@_);
|
---|
260 | my ($doc_obj,$file) = @_;
|
---|
261 |
|
---|
262 | $self->index_field_mapping_edit($doc_obj,$file,"add");
|
---|
263 | }
|
---|
264 |
|
---|
265 | sub index_field_mappingreindex
|
---|
266 | {
|
---|
267 | my $self = shift (@_);
|
---|
268 | my ($doc_obj,$file) = @_;
|
---|
269 |
|
---|
270 | $self->index_field_mapping_edit($doc_obj,$file,"update");
|
---|
271 | }
|
---|
272 |
|
---|
273 | sub index_field_mappingdelete
|
---|
274 | {
|
---|
275 | my $self = shift (@_);
|
---|
276 | my ($doc_obj,$file) = @_;
|
---|
277 |
|
---|
278 | return; # nothing to be done
|
---|
279 | }
|
---|
280 |
|
---|
281 |
|
---|
282 | #----
|
---|
283 |
|
---|
284 | sub textedit {
|
---|
285 | my $self = shift (@_);
|
---|
286 | my ($doc_obj,$file,$edit_mode) = @_;
|
---|
287 |
|
---|
288 |
|
---|
289 | if (!$self->get_indexing_text()) {
|
---|
290 | # In text-compress mode:
|
---|
291 | # => want document to be output in the simple <Doc>..</Doc> as is
|
---|
292 | # done by its super-class
|
---|
293 | return $self->SUPER::textedit(@_);
|
---|
294 | }
|
---|
295 |
|
---|
296 | # "update" for $edit_mode near identical to "add" as we use Solr in its
|
---|
297 | # default mode of replacing an existing document if the new document
|
---|
298 | # has the same doc id. Main area of difference between "add" and "update"
|
---|
299 | # is that we do not update our 'stats' for number of documents or number
|
---|
300 | # of bytes processed. The latter is inaccurate, but considered better
|
---|
301 | # than allowing the value to steadily climb.
|
---|
302 |
|
---|
303 |
|
---|
304 | my $solrhandle = $self->{'output_handle'};
|
---|
305 | my $outhandle = $self->{'outhandle'};
|
---|
306 |
|
---|
307 | # only output this document if it is one to be indexed
|
---|
308 | return if ($doc_obj->get_doc_type() ne "indexed_doc");
|
---|
309 |
|
---|
310 | # skip this document if in "compress-text" mode and asked to delete it
|
---|
311 | return if (!$self->get_indexing_text() && ($edit_mode eq "delete"));
|
---|
312 |
|
---|
313 | my $indexed_doc = $self->is_subcollection_doc($doc_obj);
|
---|
314 |
|
---|
315 | # this is another document
|
---|
316 | if ($edit_mode eq "add") {
|
---|
317 | $self->{'num_docs'} += 1;
|
---|
318 | }
|
---|
319 | elsif ($edit_mode eq "delete") {
|
---|
320 | $self->{'num_docs'} -= 1;
|
---|
321 | }
|
---|
322 |
|
---|
323 | # get the parameters for the output
|
---|
324 | # split on : just in case there is subcoll and lang stuff
|
---|
325 | my ($fields) = split (/:/, $self->{'index'});
|
---|
326 |
|
---|
327 | my $levels = $self->{'levels'};
|
---|
328 | my $ldoc_level = $levels->{'document'};
|
---|
329 | my $lsec_level = $levels->{'section'};
|
---|
330 |
|
---|
331 | my $gs2_docOID = $doc_obj->get_OID();
|
---|
332 |
|
---|
333 | my $start_doc;
|
---|
334 | my $end_doc;
|
---|
335 |
|
---|
336 | if ($edit_mode eq "add") {
|
---|
337 | $start_doc = " <add>\n";
|
---|
338 | $start_doc .= " <doc>\n";
|
---|
339 | $start_doc .= " <field name=\"docOID\">$gs2_docOID</field>\n";
|
---|
340 |
|
---|
341 | $end_doc = " </doc>\n";
|
---|
342 | $end_doc .= " </add>\n";
|
---|
343 | }
|
---|
344 | else {
|
---|
345 | $start_doc = " <delete>\n";
|
---|
346 | $start_doc .= " <id>$gs2_docOID</id>\n";
|
---|
347 |
|
---|
348 | $end_doc = " </delete>\n";
|
---|
349 |
|
---|
350 | # for delete mode, we need to specify just the docOID to delete and we're done
|
---|
351 | my $text = $start_doc;
|
---|
352 | $text .= $end_doc;
|
---|
353 | print $solrhandle $text;
|
---|
354 | return;
|
---|
355 | }
|
---|
356 |
|
---|
357 | # add/update, delete
|
---|
358 |
|
---|
359 | my $sec_tag_name = "";
|
---|
360 | if ($lsec_level)
|
---|
361 | {
|
---|
362 | $sec_tag_name = $mgppbuildproc::level_map{'section'};
|
---|
363 | }
|
---|
364 |
|
---|
365 | my $doc_section = 0; # just for this document
|
---|
366 |
|
---|
367 | # only output if working with doc level
|
---|
368 | # my $text = undef;
|
---|
369 |
|
---|
370 | my $text = ($sec_tag_name eq "") ? $start_doc : "";
|
---|
371 |
|
---|
372 | # my $text = $start_doc if ($sec_tag_name eq "");
|
---|
373 |
|
---|
374 | # get the text for this document
|
---|
375 | my $section = $doc_obj->get_top_section();
|
---|
376 |
|
---|
377 | while (defined $section)
|
---|
378 | {
|
---|
379 | # update a few statistics
|
---|
380 | $doc_section++;
|
---|
381 | $self->{'num_sections'}++;
|
---|
382 |
|
---|
383 | my $sec_gs2_id = $self->{'num_sections'};
|
---|
384 | my $sec_gs2_docOID = $gs2_docOID;
|
---|
385 | $sec_gs2_docOID .= ".$section" if ($section ne "");
|
---|
386 |
|
---|
387 | my $start_sec;
|
---|
388 | my $end_sec;
|
---|
389 |
|
---|
390 | if ($edit_mode eq "add") {
|
---|
391 | $start_sec = " <add>\n";
|
---|
392 | $start_sec .= " <doc>\n";
|
---|
393 | $start_sec .= " <field name=\"docOID\">$sec_gs2_docOID</field>\n";
|
---|
394 |
|
---|
395 | $end_sec = " </doc>\n";
|
---|
396 | $end_sec .= " </add>\n";
|
---|
397 | }
|
---|
398 | else {
|
---|
399 | $start_sec = " <delete>\n";
|
---|
400 | $start_sec .= " <id>$sec_gs2_docOID</id>\n";
|
---|
401 |
|
---|
402 | $end_sec = " </delete>\n";
|
---|
403 |
|
---|
404 | # for delete mode, should specify only this section's docOID to delete, then move on to the next section
|
---|
405 | my $text = $start_sec;
|
---|
406 | $text .= $end_sec;
|
---|
407 | print $solrhandle $text;
|
---|
408 | $section = $doc_obj->get_next_section($section);
|
---|
409 | next;
|
---|
410 | }
|
---|
411 |
|
---|
412 |
|
---|
413 | # if we are doing subcollections, then some docs shouldn't be indexed.
|
---|
414 | # but we need to put the section tag placeholders in there so the
|
---|
415 | # sections match up with database
|
---|
416 | my $indexed_section = $doc_obj->get_metadata_element($section, "gsdldoctype") || "indexed_section";
|
---|
417 | if (($indexed_doc == 0) || ($indexed_section ne "indexed_section" && $indexed_section ne "indexed_doc")) {
|
---|
418 | if ($sec_tag_name ne "") {
|
---|
419 | $text .= $start_sec;
|
---|
420 | $text .= $end_sec;
|
---|
421 | }
|
---|
422 | $section = $doc_obj->get_next_section($section);
|
---|
423 | next;
|
---|
424 | }
|
---|
425 |
|
---|
426 | # add in start section tag if indexing at the section level
|
---|
427 | $text .= $start_sec if ($sec_tag_name ne "");
|
---|
428 |
|
---|
429 | if ($edit_mode eq "add") {
|
---|
430 | $self->{'num_bytes'} += $doc_obj->get_text_length ($section);
|
---|
431 | }
|
---|
432 | elsif ($edit_mode eq "delete") {
|
---|
433 | $self->{'num_bytes'} -= $doc_obj->get_text_length ($section);
|
---|
434 | }
|
---|
435 |
|
---|
436 |
|
---|
437 | # has the user added a 'metadata' index?
|
---|
438 | my $all_metadata_specified = 0;
|
---|
439 | # which fields have already been indexed? (same as fields, but in a map)
|
---|
440 | my $specified_fields = {};
|
---|
441 |
|
---|
442 | # do we have an allfields index??
|
---|
443 | my $allfields_index = 0;
|
---|
444 | # collect up all the text for it in here
|
---|
445 | my $allfields_text = "";
|
---|
446 | foreach my $field (split (/;/, $fields)) {
|
---|
447 | if ($field eq "allfields") {
|
---|
448 | $allfields_index = 1;
|
---|
449 | } elsif ($field eq "metadata") {
|
---|
450 | $all_metadata_specified = 1;
|
---|
451 | }
|
---|
452 | }
|
---|
453 |
|
---|
454 | foreach my $field (split (/;/, $fields)) {
|
---|
455 |
|
---|
456 | # only deal with this field if it doesn't start with top or
|
---|
457 | # this is the first section
|
---|
458 | my $real_field = $field;
|
---|
459 | next if (($real_field =~ s/^top//) && ($doc_section != 1));
|
---|
460 |
|
---|
461 | # process these two later
|
---|
462 | next if ($real_field eq "allfields" || $real_field eq "metadata");
|
---|
463 |
|
---|
464 | #individual metadata and or text specified - could be a comma separated list
|
---|
465 | $specified_fields->{$real_field} = 1;
|
---|
466 | my $shortname="";
|
---|
467 | my $new_field = 0; # have we found a new field name?
|
---|
468 | if (defined $self->{'indexfieldmap'}->{$real_field}) {
|
---|
469 | $shortname = $self->{'indexfieldmap'}->{$real_field};
|
---|
470 | }
|
---|
471 | else {
|
---|
472 | $shortname = $self->get_or_create_shortname($real_field);
|
---|
473 | $new_field = 1;
|
---|
474 | }
|
---|
475 |
|
---|
476 | my @metadata_list = (); # put any metadata values in here
|
---|
477 | my $section_text = ""; # put the text in here
|
---|
478 | foreach my $submeta (split /,/, $real_field) {
|
---|
479 | if ($submeta eq "text") {
|
---|
480 | # no point in indexing text more than once
|
---|
481 | if ($section_text eq "") {
|
---|
482 | $section_text = $doc_obj->get_text($section);
|
---|
483 | if ($self->{'indexing_text'}) {
|
---|
484 | # we always strip html
|
---|
485 | &ghtml::htmlsafe($section_text);
|
---|
486 | #$section_text = $self->preprocess_text($section_text, 1, "");
|
---|
487 | }
|
---|
488 | else {
|
---|
489 | # leave html stuff in, but escape the tags
|
---|
490 | &ghtml::htmlsafe($section_text);
|
---|
491 | }
|
---|
492 | }
|
---|
493 | }
|
---|
494 | else {
|
---|
495 | $submeta =~ s/^ex\.//; #strip off ex.
|
---|
496 |
|
---|
497 | # its a metadata element
|
---|
498 | my @section_metadata = @{$doc_obj->get_metadata ($section, $submeta)};
|
---|
499 | if ($section ne $doc_obj->get_top_section() && $self->{'indexing_text'} && defined ($self->{'sections_index_document_metadata'})) {
|
---|
500 | if ($self->{'sections_index_document_metadata'} eq "always" || ( scalar(@section_metadata) == 0 && $self->{'sections_index_document_metadata'} eq "unless_section_metadata_exists")) {
|
---|
501 | push (@section_metadata, @{$doc_obj->get_metadata ($doc_obj->get_top_section(), $submeta)});
|
---|
502 | }
|
---|
503 | }
|
---|
504 | push (@metadata_list, @section_metadata);
|
---|
505 | }
|
---|
506 | } # for each field in this one index
|
---|
507 |
|
---|
508 | # now we add the text and/or metadata into new_text
|
---|
509 | if ($section_text ne "" || scalar(@metadata_list)) {
|
---|
510 | my $new_text = "";
|
---|
511 |
|
---|
512 | if ($section_text ne "") {
|
---|
513 |
|
---|
514 | if ($allfields_index) {
|
---|
515 | $allfields_text .= "$section_text ";
|
---|
516 | }
|
---|
517 |
|
---|
518 | # Remove any leading or trailing white space
|
---|
519 | $section_text =~ s/\s+$//;
|
---|
520 | $section_text =~ s/^\s+//;
|
---|
521 |
|
---|
522 | if ($self->{'indexing_text'}) {
|
---|
523 | # add the tag
|
---|
524 | $new_text .= "<field name=\"$shortname\" >$section_text</field>\n";
|
---|
525 | } else {
|
---|
526 | $new_text .= "$section_text ";
|
---|
527 | }
|
---|
528 | }
|
---|
529 |
|
---|
530 | foreach my $item (@metadata_list) {
|
---|
531 | &ghtml::htmlsafe($item);
|
---|
532 |
|
---|
533 | if ($allfields_index) {
|
---|
534 | $allfields_text .= "$item ";
|
---|
535 | }
|
---|
536 |
|
---|
537 | # Remove any leading or trailing white space
|
---|
538 | $item =~ s/\s+$//;
|
---|
539 | $item =~ s/^\s+//;
|
---|
540 |
|
---|
541 | if ($self->{'indexing_text'}) {
|
---|
542 | # add the tag
|
---|
543 | $new_text .= "<field name=\"$shortname\" >$item</field>\n";
|
---|
544 | } else {
|
---|
545 | $new_text .= "$item ";
|
---|
546 | }
|
---|
547 | } # end for loop processing @metadata_list
|
---|
548 |
|
---|
549 | # filter the text
|
---|
550 | $new_text = $self->filter_text ($field, $new_text);
|
---|
551 |
|
---|
552 | if ($edit_mode eq "add") {
|
---|
553 | $self->{'num_processed_bytes'} += length ($new_text);
|
---|
554 | $text .= "$new_text";
|
---|
555 | }
|
---|
556 | elsif ($edit_mode eq "update") {
|
---|
557 | $text .= "$new_text";
|
---|
558 | }
|
---|
559 | elsif ($edit_mode eq "delete") {
|
---|
560 | $self->{'num_processed_bytes'} -= length ($new_text);
|
---|
561 | }
|
---|
562 |
|
---|
563 |
|
---|
564 | if ($self->{'indexing_text'} && $new_field) {
|
---|
565 | # we need to add to the list in indexfields
|
---|
566 |
|
---|
567 | $self->{'indexfieldmap'}->{$real_field} = $shortname;
|
---|
568 | $self->{'indexfieldmap'}->{$shortname} = 1;
|
---|
569 | }
|
---|
570 |
|
---|
571 | }
|
---|
572 |
|
---|
573 | } # foreach field
|
---|
574 |
|
---|
575 |
|
---|
576 | if ($all_metadata_specified) {
|
---|
577 |
|
---|
578 | my $new_text = "";
|
---|
579 | my $shortname = "";
|
---|
580 | my $metadata = $doc_obj->get_all_metadata ($section);
|
---|
581 | foreach my $pair (@$metadata) {
|
---|
582 | my ($mfield, $mvalue) = (@$pair);
|
---|
583 |
|
---|
584 | # no value
|
---|
585 | next unless defined $mvalue && $mvalue ne "";
|
---|
586 |
|
---|
587 | # we have already indexed this
|
---|
588 | next if defined ($specified_fields->{$mfield});
|
---|
589 |
|
---|
590 | # check fields here, maybe others dont want - change to use dontindex!!
|
---|
591 | next if ($mfield eq "Identifier" || $mfield eq "classifytype" || $mfield eq "assocfilepath");
|
---|
592 | next if ($mfield =~ /^gsdl/);
|
---|
593 |
|
---|
594 | &ghtml::htmlsafe($mvalue);
|
---|
595 |
|
---|
596 | if (defined $self->{'indexfieldmap'}->{$mfield}) {
|
---|
597 | $shortname = $self->{'indexfieldmap'}->{$mfield};
|
---|
598 | }
|
---|
599 | else {
|
---|
600 | $shortname = $self->get_or_create_shortname($mfield);
|
---|
601 | $self->{'indexfieldmap'}->{$mfield} = $shortname;
|
---|
602 | $self->{'indexfieldmap'}->{$shortname} = 1;
|
---|
603 | }
|
---|
604 | $new_text .= "<field name=\"$shortname\">$mvalue</field>\n";
|
---|
605 | if ($allfields_index) {
|
---|
606 | $allfields_text .= "$mvalue ";
|
---|
607 | }
|
---|
608 |
|
---|
609 | if (!defined $self->{'indexfields'}->{$mfield}) {
|
---|
610 | $self->{'indexfields'}->{$mfield} = 1;
|
---|
611 | }
|
---|
612 |
|
---|
613 | }
|
---|
614 | # filter the text
|
---|
615 | $new_text = $self->filter_text ("metadata", $new_text);
|
---|
616 |
|
---|
617 | if ($edit_mode eq "add") {
|
---|
618 | $self->{'num_processed_bytes'} += length ($new_text);
|
---|
619 | $text .= "$new_text";
|
---|
620 | }
|
---|
621 | elsif ($edit_mode eq "update") {
|
---|
622 | $text .= "$new_text";
|
---|
623 | }
|
---|
624 | elsif ($edit_mode eq "delete") {
|
---|
625 | $self->{'num_processed_bytes'} -= length ($new_text);
|
---|
626 | }
|
---|
627 | }
|
---|
628 |
|
---|
629 | if ($allfields_index) {
|
---|
630 | # add the index name mapping
|
---|
631 | $self->{'indexfieldmap'}->{"allfields"} = "ZZ";
|
---|
632 | $self->{'indexfieldmap'}->{"ZZ"} = 1;
|
---|
633 |
|
---|
634 | my $new_text = "<field name=\"ZZ\">$allfields_text</field>\n";
|
---|
635 | # filter the text
|
---|
636 | $new_text = $self->filter_text ("allfields", $new_text);
|
---|
637 |
|
---|
638 | if ($edit_mode eq "add") {
|
---|
639 | $self->{'num_processed_bytes'} += length ($new_text);
|
---|
640 | $text .= "$new_text";
|
---|
641 | }
|
---|
642 | elsif ($edit_mode eq "update") {
|
---|
643 | $text .= "$new_text";
|
---|
644 | }
|
---|
645 | elsif ($edit_mode eq "delete") {
|
---|
646 | $self->{'num_processed_bytes'} -= length ($new_text);
|
---|
647 | }
|
---|
648 | }
|
---|
649 |
|
---|
650 | # only add sort and facet fields for this section if we are indexing this section, we are doing section level indexing or this is the top section
|
---|
651 | if ($self->{'indexing_text'} && ($sec_tag_name ne "" || $doc_section == 1 )) {
|
---|
652 | # add sort fields if there are any
|
---|
653 | foreach my $sfield (@{$self->{'sortfields'}}) {
|
---|
654 | # ignore special field rank/none
|
---|
655 | next if $sfield eq "rank" || $sfield eq "none";
|
---|
656 | my $sf_shortname = $self->get_or_create_sortfield_shortname($sfield);
|
---|
657 | my @metadata_list = (); # put any metadata values in here
|
---|
658 | foreach my $submeta (split /,/, $sfield) {
|
---|
659 | $submeta =~ s/^ex\.([^.]+)$/$1/; #strip off ex. iff it's the only metadata set prefix (will leave ex.dc.* intact)
|
---|
660 |
|
---|
661 | my @section_metadata = @{$doc_obj->get_metadata ($section, $submeta)};
|
---|
662 | if ($section ne $doc_obj->get_top_section() && defined ($self->{'sections_sort_on_document_metadata'})) {
|
---|
663 | if ($self->{'sections_sort_on_document_metadata'} eq "always" || ( scalar(@section_metadata) == 0 && $self->{'sections_sort_on_document_metadata'} eq "unless_section_metadata_exists")) {
|
---|
664 | push (@section_metadata, @{$doc_obj->get_metadata ($doc_obj->get_top_section(), $submeta)});
|
---|
665 | }
|
---|
666 | }
|
---|
667 | push (@metadata_list, @section_metadata);
|
---|
668 | }
|
---|
669 | my $new_text = "";
|
---|
670 | foreach my $item (@metadata_list) {
|
---|
671 | &ghtml::htmlsafe($item);
|
---|
672 | $new_text .= "$item ";
|
---|
673 | }
|
---|
674 | if ($new_text =~ /\S/) {
|
---|
675 | $new_text = "<field name=\"$sf_shortname\">$new_text</field>\n";
|
---|
676 | # filter the text???
|
---|
677 | $text .= "$new_text"; # add it to the main text block
|
---|
678 | #print "#### new_text: $new_text\n";
|
---|
679 |
|
---|
680 | $self->{'actualsortfields'}->{$sfield} = 1;
|
---|
681 | }
|
---|
682 | # print "#### TEXT: $text\n";
|
---|
683 |
|
---|
684 |
|
---|
685 | }# end add in sort fields
|
---|
686 |
|
---|
687 | # add facet fields if there are any
|
---|
688 | foreach my $sfield (@{$self->{'facetfields'}}) {
|
---|
689 | my $sf_shortname = $self->get_or_create_facetfield_shortname($sfield);
|
---|
690 | my @metadata_list = (); # put any metadata values in here
|
---|
691 | foreach my $submeta (split /,/, $sfield) {
|
---|
692 | $submeta =~ s/^ex\.([^.]+)$/$1/; #strip off ex. iff it's the only metadata set prefix (will leave ex.dc.* intact)
|
---|
693 |
|
---|
694 | my @section_metadata = @{$doc_obj->get_metadata ($section, $submeta)};
|
---|
695 | if ($section ne $doc_obj->get_top_section() && defined ($self->{'sections_sort_on_document_metadata'})) {
|
---|
696 | if ($self->{'sections_sort_on_document_metadata'} eq "always" || ( scalar(@section_metadata) == 0 && $self->{'sections_sort_on_document_metadata'} eq "unless_section_metadata_exists")) {
|
---|
697 | push (@section_metadata, @{$doc_obj->get_metadata ($doc_obj->get_top_section(), $submeta)});
|
---|
698 | }
|
---|
699 | }
|
---|
700 | push (@metadata_list, @section_metadata);
|
---|
701 | }
|
---|
702 | my $found_facet_value = 0;
|
---|
703 | foreach my $item (@metadata_list) {
|
---|
704 | &ghtml::htmlsafe($item);
|
---|
705 | if ($item =~ /\S/) {
|
---|
706 | $item = "<field name=\"$sf_shortname\">$item</field>\n";
|
---|
707 | # filter the text???
|
---|
708 | $text .= "$item"; # add it to the main text block
|
---|
709 | $found_facet_value = 1;
|
---|
710 | #print "#### new_text: $item\n";
|
---|
711 | }
|
---|
712 | }
|
---|
713 | if($found_facet_value) {
|
---|
714 | $self->{'actualfacetfields'}->{$sfield} = 1;
|
---|
715 | }
|
---|
716 |
|
---|
717 | }
|
---|
718 | } # end add in facet fields
|
---|
719 |
|
---|
720 | # add in end tag if at top-level doc root, or indexing at the section level
|
---|
721 | $text .= $end_sec if ($sec_tag_name ne "");
|
---|
722 |
|
---|
723 | $section = $doc_obj->get_next_section($section);
|
---|
724 | } # while defined section
|
---|
725 |
|
---|
726 |
|
---|
727 | # only output if working with doc level
|
---|
728 | $text .= $end_doc if ($sec_tag_name eq "");
|
---|
729 |
|
---|
730 | ## $text .= "<commit/>\n";
|
---|
731 |
|
---|
732 | # The following code looks like it's for debugging purposes, but
|
---|
733 | # committed by accident. Commenting out for now ...
|
---|
734 |
|
---|
735 | # open(TEXTOUT, '>:utf8', "text.out");
|
---|
736 | # print TEXTOUT "$text";
|
---|
737 | # close TEXTOUT;
|
---|
738 |
|
---|
739 | print $solrhandle $text;
|
---|
740 |
|
---|
741 | }
|
---|
742 |
|
---|
743 |
|
---|
744 |
|
---|
745 |
|
---|
746 | sub textreindex
|
---|
747 | {
|
---|
748 | my $self = shift (@_);
|
---|
749 | my ($doc_obj,$file) = @_;
|
---|
750 |
|
---|
751 | # the update command does not exist in solrbuildproc
|
---|
752 | # reindexing consists of deleting and then adding the same file
|
---|
753 | #$self->textedit($doc_obj,$file,"update");
|
---|
754 | $self->textedit($doc_obj,$file,"delete");
|
---|
755 | $self->textedit($doc_obj,$file,"add");
|
---|
756 | }
|
---|
757 |
|
---|
758 | sub get_or_create_facetfield_shortname {
|
---|
759 | my $self = shift(@_);
|
---|
760 |
|
---|
761 | my ($realname) = @_;
|
---|
762 |
|
---|
763 | if (defined $self->{'facetfieldnamemap'}->{$realname}) {
|
---|
764 | return $self->{'facetfieldnamemap'}->{$realname};
|
---|
765 | }
|
---|
766 |
|
---|
767 | # get the shortname made for this index field
|
---|
768 |
|
---|
769 | my $shortname = $self->get_or_create_shortname($realname);
|
---|
770 | $shortname = "fc".$shortname;
|
---|
771 |
|
---|
772 | $self->{'facetfieldnamemap'}->{$realname} = $shortname;
|
---|
773 | $self->{'facetfieldnamemap'}->{$shortname} = 1;
|
---|
774 | return $shortname;
|
---|
775 | }
|
---|
776 |
|
---|
777 |
|
---|
778 |
|
---|
779 | 1;
|
---|
780 |
|
---|
781 |
|
---|