root/gsdl/trunk/perllib/basebuildproc.pm @ 15708

Revision 15708, 17.5 KB (checked in by mdewsnip, 12 years ago)

Minor code rearrange.

  • Property svn:keywords set to Author Date Id Revision
Line 
1###########################################################################
2#
3# basebuildproc.pm --
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# This document processor outputs a document for indexing (should be
27# implemented by subclass) and storing in the database
28
29package basebuildproc;
30
31eval {require bytes};
32
33use classify;
34use dbutil;
35use doc;
36use docproc;
37use strict; no strict 'subs';
38use util;
39
40BEGIN {
41    @basebuildproc::ISA = ('docproc');
42}
43
44sub new()
45  {
46    my ($class, $collection, $source_dir, $build_dir, $keepold, $verbosity, $outhandle) = @_;
47    my $self = new docproc ();
48
49    # outhandle is where all the debugging info goes
50    # output_handle is where the output of the plugins is piped
51    # to (i.e. mg, database etc.)
52    $outhandle = STDERR unless defined $outhandle;
53
54    $self->{'collection'} = $collection;
55    $self->{'source_dir'} = $source_dir;
56    $self->{'build_dir'}  = $build_dir;
57    $self->{'keepold'}    = $keepold;
58    $self->{'verbosity'}  = $verbosity;
59    $self->{'outhandle'}  = $outhandle;
60
61    $self->{'classifiers'} = [];
62    $self->{'mode'} = "text";
63    $self->{'assocdir'} = $build_dir;
64    $self->{'dontdb'} = {};
65
66    $self->{'index'} = "section:text";
67    $self->{'indexexparr'} = [];
68
69    my $found_num_data = 0;
70    my $buildconfigfile = undef;
71
72    if ($keepold) {
73    # For incremental building need to seed num_docs etc from values
74    # stored in build.cfg (if present)
75      print STDERR "Keepold!\n";
76    $buildconfigfile = &util::filename_cat($build_dir, "build.cfg");
77      print STDERR "Build cfg: $buildconfigfile\n";
78    if (-e $buildconfigfile) {
79        $found_num_data = 1;
80    }
81    else {
82        # try the index dir
83        $buildconfigfile = &util::filename_cat($ENV{'GSDLCOLLECTDIR'},
84                           "index", "build.cfg");
85            print STDERR "Index cfg: $buildconfigfile\n";
86        if (-e $buildconfigfile) {
87        $found_num_data = 1;
88        }
89    }
90
91    }
92    #else
93    #  {
94    #    print STDERR "Removeold!\n";
95    #  }
96
97    if ($found_num_data)
98      {
99        #print STDERR "Found_Num_Data!\n";
100    my $buildcfg = &colcfg::read_build_cfg($buildconfigfile);
101    $self->{'starting_num_docs'}     = $buildcfg->{'numdocs'};
102        #print STDERR "- num_docs:     $self->{'starting_num_docs'}\n";
103    $self->{'starting_num_sections'} = $buildcfg->{'numsections'};
104        #print STDERR "- num_sections: $self->{'starting_num_sections'}\n";
105    $self->{'starting_num_bytes'}    = $buildcfg->{'numbytes'};
106        #print STDERR "- num_bytes:    $self->{'starting_num_bytes'}\n";
107    }
108    else
109      {
110        #print STDERR "NOT Found_Num_Data!\n";
111        $self->{'starting_num_docs'}     = 0;
112    $self->{'starting_num_sections'} = 0;
113    $self->{'starting_num_bytes'}    = 0;
114      }
115
116    $self->{'output_handle'} = "STDOUT";
117    $self->{'num_docs'}      = $self->{'starting_num_docs'};
118    $self->{'num_sections'}  = $self->{'starting_num_sections'};
119    $self->{'num_bytes'}     = $self->{'starting_num_bytes'};
120
121    $self->{'num_processed_bytes'} = 0;
122    $self->{'store_text'} = 1;
123
124    # what level (section/document) the database - indexer intersection is
125    $self->{'db_level'} = "section";
126    #used by browse interface
127    $self->{'doclist'} = [];
128
129    $self->{'indexing_text'} = 0;
130
131    return bless $self, $class;
132
133}
134
135sub reset {
136    my $self = shift (@_);
137
138    $self->{'num_docs'}      = $self->{'starting_num_docs'};
139    $self->{'num_sections'}  = $self->{'starting_num_sections'};
140    $self->{'num_bytes'}     = $self->{'starting_num_bytes'};
141   
142    $self->{'num_processed_bytes'} = 0;
143}
144
145sub zero_reset {
146    my $self = shift (@_);
147
148    $self->{'num_docs'}      = 0;
149    $self->{'num_sections'}  = 0;
150    $self->{'num_bytes'}     = 0;
151   
152    $self->{'num_processed_bytes'} = 0;
153}
154
155sub is_incremental_capable
156{
157    # By default we return 'no' as the answer
158    # Safer to assume non-incremental to start with, and then override in
159    # inherited classes that are.
160
161    return 0;
162}
163
164sub get_num_docs {
165    my $self = shift (@_);
166
167    return $self->{'num_docs'};
168}
169
170sub get_num_sections {
171    my $self = shift (@_);
172
173    return $self->{'num_sections'};
174}
175
176# num_bytes is the actual number of bytes in the collection
177# this is normally the same as what's processed during text compression
178sub get_num_bytes {
179    my $self = shift (@_);
180
181    return $self->{'num_bytes'};
182}
183
184# num_processed_bytes is the number of bytes actually passed
185# to mg for the current index
186sub get_num_processed_bytes {
187    my $self = shift (@_);
188
189    return $self->{'num_processed_bytes'};
190}
191
192sub set_output_handle {
193    my $self = shift (@_);
194    my ($handle) = @_;
195
196    $self->{'output_handle'} = $handle;
197}
198
199
200sub set_mode {
201    my $self = shift (@_);
202    my ($mode) = @_;
203
204    $self->{'mode'} = $mode;
205}
206
207sub get_mode {
208    my $self = shift (@_);
209
210    return $self->{'mode'};
211}
212
213sub set_assocdir {
214    my $self = shift (@_);
215    my ($assocdir) = @_;
216
217    $self->{'assocdir'} = $assocdir;
218}
219
220sub set_dontdb {
221    my $self = shift (@_);
222    my ($dontdb) = @_;
223
224    $self->{'dontdb'} = $dontdb;
225}
226
227sub set_index {
228    my $self = shift (@_);
229    my ($index, $indexexparr) = @_;
230
231    $self->{'index'} = $index;
232    $self->{'indexexparr'} = $indexexparr if defined $indexexparr;
233}
234
235sub set_index_languages {
236    my $self = shift (@_);
237    my ($lang_meta, $langarr) = @_;
238    $self->{'lang_meta'} = $lang_meta;
239    $self->{'langarr'} = $langarr;
240}
241
242sub get_index {
243    my $self = shift (@_);
244
245    return $self->{'index'};
246}
247
248sub set_classifiers {
249    my $self = shift (@_);
250    my ($classifiers) = @_;
251
252    $self->{'classifiers'} = $classifiers;
253}
254
255sub set_indexing_text {
256    my $self = shift (@_);
257    my ($indexing_text) = @_;
258
259    $self->{'indexing_text'} = $indexing_text;
260}
261
262sub get_indexing_text {
263    my $self = shift (@_);
264
265    return $self->{'indexing_text'};
266}
267
268sub set_store_text {
269    my $self = shift (@_);
270    my ($store_text) = @_;
271
272    $self->{'store_text'} = $store_text;
273}
274sub get_doc_list {
275    my $self = shift(@_);
276   
277    return @{$self->{'doclist'}};
278}
279
280# the standard database level is section, but you may want to change it to document
281sub set_db_level {
282    my $self= shift (@_);
283    my ($db_level) = @_;
284
285    $self->{'db_level'} = $db_level;
286}
287
288sub set_sections_index_document_metadata {
289    my $self= shift (@_);
290    my ($index_type) = @_;
291   
292    $self->{'sections_index_document_metadata'} = $index_type;
293}
294sub process {
295    my $self = shift (@_);
296    my $method = $self->{'mode'};
297
298    $self->$method(@_);
299}
300
301
302
303sub infodb_metadata_stats
304{
305    my $self = shift (@_);
306    my ($field) = @_;
307
308    # Keep some statistics relating to metadata sets used and
309    # frequency of particular metadata fields within each set
310
311    # Union of metadata prefixes and frequency of fields
312    # (both scoped for this document alone, and across whole collection)
313   
314    if ($field =~ m/^(.+)\.(.*)$/) {
315    my $prefix = $1;
316    my $core_field = $2;
317
318    $self->{'doc_mdprefix_fields'}->{$prefix}->{$core_field}++;
319    $self->{'mdprefix_fields'}->{$prefix}->{$core_field}++;
320    }
321    elsif ($field =~ m/^[[:upper:]]/) {
322    # implicit 'ex' metadata set
323
324    $self->{'doc_mdprefix_fields'}->{'ex'}->{$field}++;
325    $self->{'mdprefix_fields'}->{'ex'}->{$field}++;
326    }
327
328}
329
330
331sub infodb {
332    my $self = shift (@_);
333    my ($doc_obj, $filename) = @_;
334
335    # only output this document if it is a "indexed_doc" or "info_doc" (database only) document
336    my $doctype = $doc_obj->get_doc_type();
337    return if ($doctype ne "indexed_doc" && $doctype ne "info_doc");
338
339    my $archivedir = "";
340    if (defined $filename)
341    {
342    # doc_obj derived directly from file
343    my ($dir) = $filename =~ /^(.*?)(?:\/|\\)[^\/\\]*$/;
344    $dir = "" unless defined $dir;
345    $dir =~ s/\\/\//g;
346    $dir =~ s/^\/+//;
347    $dir =~ s/\/+$//;
348
349    $archivedir = $dir;
350
351    # resolve the final filenames of the files associated with this document
352    $self->assoc_files ($doc_obj, $archivedir);
353    }
354    else
355    {
356    # doc_obj reconstructed from database (has metadata, doc structure but no text)
357    my $top_section = $doc_obj->get_top_section();
358    $archivedir = $doc_obj->get_metadata_element($top_section,"archivedir");
359    }
360
361    #add this document to the browse structure
362    push(@{$self->{'doclist'}},$doc_obj->get_OID())
363    unless ($doctype eq "classification");
364
365    # classify this document
366    &classify::classify_doc ($self->{'classifiers'}, $doc_obj);
367
368    # this is another document
369    $self->{'num_docs'} += 1 unless ($doctype eq "classification");
370
371    # is this a paged or a hierarchical document
372    my ($thistype, $childtype) = $self->get_document_type ($doc_obj);
373
374    my $section = $doc_obj->get_top_section ();
375    my $doc_OID = $doc_obj->get_OID();
376    my $first = 1;
377    my $infodb_handle = $self->{'output_handle'};
378
379    $self->{'doc_mdprefix_fields'} = {};
380
381    while (defined $section)
382    {
383    my $section_OID = $doc_OID;
384    if ($section ne "")
385    {
386        $section_OID = $doc_OID . "." . $section;
387    }
388    my %section_infodb = ();
389
390    # update a few statistics
391    $self->{'num_bytes'} += $doc_obj->get_text_length ($section);
392    $self->{'num_sections'} += 1 unless ($doctype eq "classification");
393
394    # output the fact that this document is a document (unless doctype
395    # has been set to something else from within a plugin
396    my $dtype = $doc_obj->get_metadata_element ($section, "doctype");
397    if (!defined $dtype || $dtype !~ /\w/) {
398        $section_infodb{"doctype"} = [ "doc" ];
399    }
400
401    # Output whether this node contains text
402    #
403    # If doc_obj reconstructed from database file then no need to
404    # explicitly add <hastxt> as this is preserved as metadata when
405    # the database file is loaded in
406    if (defined $filename)
407    {
408        # doc_obj derived directly from file
409        if ($doc_obj->get_text_length($section) > 0) {
410        $section_infodb{"hastxt"} = [ "1" ];
411        } else {
412        $section_infodb{"hastxt"} = [ "0" ];
413        }
414    }
415
416    # output all the section metadata
417    my $metadata = $doc_obj->get_all_metadata ($section);
418    foreach my $pair (@$metadata) {
419        my ($field, $value) = (@$pair);
420
421        if ($field ne "Identifier" && $field !~ /^gsdl/ &&
422        defined $value && $value ne "") {       
423
424        # escape problematic stuff
425        $value =~ s/\\/\\\\/g;
426        $value =~ s/\n/\\n/g;
427        $value =~ s/\r/\\r/g;
428
429        # special case for URL metadata
430        if ($field =~ /^URL$/i) {
431            &dbutil::write_infodb_entry($infodb_handle, $value, { 'section' => [ $section_OID ] });
432        }
433
434        if (!defined $self->{'dontdb'}->{$field}) {
435            push(@{$section_infodb{$field}}, $value);
436
437            if ($section eq "")
438            {
439            $self->infodb_metadata_stats($field);
440            }
441        }
442        }
443    }
444
445    if ($section eq "")
446    {
447        my $doc_mdprefix_fields = $self->{'doc_mdprefix_fields'};
448
449        foreach my $prefix (keys %$doc_mdprefix_fields)
450        {
451        push(@{$section_infodb{"metadataset"}}, $prefix);
452
453        foreach my $field (keys %{$doc_mdprefix_fields->{$prefix}})
454        {
455            push(@{$section_infodb{"metadatalist-$prefix"}}, $field);
456
457            my $val = $doc_mdprefix_fields->{$prefix}->{$field};
458            push(@{$section_infodb{"metadatafreq-$prefix-$field"}}, $val);
459        }
460        }
461    }
462
463    # If doc_obj reconstructed from database file then no need to
464    # explicitly add <archivedir> as this is preserved as metadata when
465    # the database file is loaded in
466    if (defined $filename)
467    {
468        # output archivedir if at top level
469        if ($section eq $doc_obj->get_top_section()) {
470        $section_infodb{"archivedir"} = [ $archivedir ];
471        }
472    }
473
474    # output document display type
475    if ($first) {
476        $section_infodb{"thistype"} = [ $thistype ];
477    }
478
479    if ($self->{'db_level'} eq "document") {
480        # doc num is num_docs not num_sections
481        # output the matching document number
482        $section_infodb{"docnum"} = [ $self->{'num_docs'} ];
483    }
484    else {
485        # output a list of children
486        my $children = $doc_obj->get_children ($section);
487        if (scalar(@$children) > 0) {
488        $section_infodb{"childtype"} = [ $childtype ];
489        my $contains = "";
490        foreach my $child (@$children)
491        {
492            $contains .= ";" unless ($contains eq "");
493            if ($child =~ /^.*?\.(\d+)$/)
494            {
495            $contains .= "\".$1";
496            }
497            else
498            {
499            $contains .= "\".$child";
500            }
501        }
502        $section_infodb{"contains"} = [ $contains ];
503        }
504        # output the matching doc number
505        $section_infodb{"docnum"} = [ $self->{'num_sections'} ];
506    }
507   
508    &dbutil::write_infodb_entry($infodb_handle, $section_OID, \%section_infodb);
509   
510    # output a database entry for the document number
511    if ($self->{'db_level'} eq "document") {
512        &dbutil::write_infodb_entry($infodb_handle, $self->{'num_docs'}, { 'section' => [ $doc_OID ] });
513    }
514    else {
515        &dbutil::write_infodb_entry($infodb_handle, $self->{'num_sections'}, { 'section' => [ $section_OID ] });
516    }
517
518    $first = 0;
519    $section = $doc_obj->get_next_section($section);
520    last if ($self->{'db_level'} eq "document"); # if no sections wanted, only add the docs
521    }
522}
523
524
525sub text {
526    my $self = shift (@_);
527    my ($doc_obj) = @_;
528   
529    my $handle = $self->{'outhandle'};
530    print $handle "basebuildproc::text function must be implemented in sub classes\n";
531    die "\n";
532}
533
534# should the document be indexed - according to the subcollection and language
535# specification.
536sub is_subcollection_doc {
537    my $self = shift (@_);
538    my ($doc_obj) = @_;
539   
540    my $indexed_doc = 1;
541    foreach my $indexexp (@{$self->{'indexexparr'}}) {
542    $indexed_doc = 0;
543    my ($field, $exp, $options) = split /\//, $indexexp;
544    if (defined ($field) && defined ($exp)) {
545        my ($bool) = $field =~ /^(.)/;
546        $field =~ s/^.// if $bool eq '!';
547        my @metadata_values;
548        if ($field =~ /^filename$/i) {
549        push(@metadata_values, $doc_obj->get_source_filename());
550        }
551        else {
552        @metadata_values = @{$doc_obj->get_metadata($doc_obj->get_top_section(), $field)};
553        }
554        next unless @metadata_values;
555        foreach my $metadata_value (@metadata_values) {
556        if ($bool eq '!') {
557            if ($options =~ /^i$/i) {
558            if ($metadata_value !~ /$exp/i) {$indexed_doc = 1; last;}
559            } else {
560            if ($metadata_value !~ /$exp/) {$indexed_doc = 1; last;}
561            }
562        } else {
563            if ($options =~ /^i$/i) {
564            if ($metadata_value =~ /$exp/i) {$indexed_doc = 1; last;}
565            } else {
566            if ($metadata_value =~ /$exp/) {$indexed_doc = 1; last;}
567            }
568        }
569        }
570
571        last if ($indexed_doc == 1);
572    }
573    }
574   
575    # if this doc is so far in the sub collection, and we have lang info,
576    # now we check the languages to see if it matches
577    if($indexed_doc && defined $self->{'lang_meta'}) {
578    $indexed_doc = 0;
579    my $field = $doc_obj->get_metadata_element($doc_obj->get_top_section(), $self->{'lang_meta'});
580    if (defined $field) {
581        foreach my $lang (@{$self->{'langarr'}}) {
582        my ($bool) = $lang =~ /^(.)/;
583        if ($bool eq '!') {
584            $lang =~ s/^.//;
585            if ($field !~ /$lang/) {
586            $indexed_doc = 1; last;
587            }
588        } else {
589            if ($field =~ /$lang/) {
590            $indexed_doc = 1; last;
591            }
592        }
593        }
594    }
595    }
596    return $indexed_doc;
597   
598}
599
600# use 'Paged' if document has no more than 2 levels
601# and each section at second level has a number for
602# Title metadata
603# also use Paged if gsdlthistype metadata is set to Paged
604sub get_document_type {
605    my $self = shift (@_);
606    my ($doc_obj) = @_;
607
608    my $thistype = "VList";
609    my $childtype = "VList";
610    my $title;
611    my @tmp = ();
612   
613    my $section = $doc_obj->get_top_section ();
614   
615    my $gsdlthistype = $doc_obj->get_metadata_element ($section, "gsdlthistype");
616    if (defined $gsdlthistype) {
617    if ($gsdlthistype eq "Paged") {
618        $childtype = "Paged";
619        if ($doc_obj->get_text_length ($doc_obj->get_top_section())) {
620        $thistype = "Paged";
621        } else {
622        $thistype = "Invisible";
623        }
624       
625        return ($thistype, $childtype);
626    } elsif ($gsdlthistype eq "Hierarchy") {
627        return ($thistype, $childtype); # use VList, VList
628    }
629    }
630    my $first = 1;
631    while (defined $section) {
632    @tmp = split /\./, $section;
633    if (scalar(@tmp) > 1) {
634        return ($thistype, $childtype);
635    }
636    if (!$first) {
637        $title = $doc_obj->get_metadata_element ($section, "Title");
638        if (!defined $title || $title !~ /^\d+$/) {
639        return ($thistype, $childtype);
640        }
641    }
642    $first = 0;
643    $section = $doc_obj->get_next_section($section);
644    }
645    if ($doc_obj->get_text_length ($doc_obj->get_top_section())) {
646    $thistype = "Paged";
647    } else {
648    $thistype = "Invisible";
649    }
650    $childtype = "Paged";
651    return ($thistype, $childtype);
652}
653
654sub assoc_files() {
655    my $self = shift (@_);
656    my ($doc_obj, $archivedir) = @_;
657    my ($afile);
658   
659    foreach my $assoc_file (@{$doc_obj->get_assoc_files()}) {
660      #rint STDERR "Processing associated file - copy " . $assoc_file->[0] . " to " . $assoc_file->[1] . "\n";
661    # if assoc file starts with a slash, we put it relative to the assoc
662    # dir, otherwise it is relative to the HASH... directory
663    if ($assoc_file->[1] =~ m@^[/\\]@) {
664        $afile = &util::filename_cat($self->{'assocdir'}, $assoc_file->[1]);
665    } else {
666        $afile = &util::filename_cat($self->{'assocdir'}, $archivedir, $assoc_file->[1]);
667    }
668    &util::hard_link ($assoc_file->[0], $afile);
669    }
670}
671
Note: See TracBrowser for help on using the browser.