1 | ###########################################################################
|
---|
2 | #
|
---|
3 | # ArchivesInfPlugin.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 | # plugin which reads through an archives.inf (or GDBM equivalent,
|
---|
27 | # archiveinf-doc.{ldb,bdb} file (i.e. the file generated in the
|
---|
28 | # archives directory when an import is done), processing each file it
|
---|
29 | # finds
|
---|
30 |
|
---|
31 | package ArchivesInfPlugin;
|
---|
32 |
|
---|
33 | use util;
|
---|
34 | use PrintInfo;
|
---|
35 | use plugin;
|
---|
36 | use arcinfo;
|
---|
37 | use gsprintf;
|
---|
38 |
|
---|
39 | use strict;
|
---|
40 | no strict 'refs'; # allow filehandles to be variables and viceversa
|
---|
41 |
|
---|
42 | BEGIN {
|
---|
43 | @ArchivesInfPlugin::ISA = ('PrintInfo');
|
---|
44 | }
|
---|
45 |
|
---|
46 | my $arguments = [
|
---|
47 | ];
|
---|
48 |
|
---|
49 | my $options = { 'name' => "ArchivesInfPlugin",
|
---|
50 | 'desc' => "{ArchivesInfPlugin.desc}",
|
---|
51 | 'abstract' => "no",
|
---|
52 | 'inherits' => "yes" };
|
---|
53 |
|
---|
54 | sub gsprintf
|
---|
55 | {
|
---|
56 | return &gsprintf::gsprintf(@_);
|
---|
57 | }
|
---|
58 |
|
---|
59 | sub new {
|
---|
60 | my ($class) = shift (@_);
|
---|
61 | my ($pluginlist,$inputargs,$hashArgOptLists) = @_;
|
---|
62 | push(@$pluginlist, $class);
|
---|
63 |
|
---|
64 | push(@{$hashArgOptLists->{"ArgList"}},@{$arguments});
|
---|
65 | push(@{$hashArgOptLists->{"OptList"}},$options);
|
---|
66 |
|
---|
67 | my $self = new PrintInfo($pluginlist, $inputargs, $hashArgOptLists);
|
---|
68 |
|
---|
69 | return bless $self, $class;
|
---|
70 | }
|
---|
71 |
|
---|
72 | # called once, at the start of processing
|
---|
73 | sub init {
|
---|
74 | my $self = shift (@_);
|
---|
75 | my ($verbosity, $outhandle, $failhandle) = @_;
|
---|
76 |
|
---|
77 | # verbosity is passed through from the processor
|
---|
78 | $self->{'verbosity'} = $verbosity;
|
---|
79 |
|
---|
80 | # as are the outhandle and failhandle
|
---|
81 | $self->{'outhandle'} = $outhandle if defined $outhandle;
|
---|
82 | $self->{'failhandle'} = $failhandle;
|
---|
83 |
|
---|
84 | }
|
---|
85 |
|
---|
86 | sub deinit {
|
---|
87 | my ($self) = @_;
|
---|
88 |
|
---|
89 | my $archive_info = $self->{'archive_info'};
|
---|
90 |
|
---|
91 | if (defined $archive_info) {
|
---|
92 | print STDERR "********* have parsed and processed an archive info file\n";
|
---|
93 |
|
---|
94 | my $archive_info_filename = $self->{'archive_info_filename'};
|
---|
95 |
|
---|
96 | my $file_list = $archive_info->get_file_list();
|
---|
97 |
|
---|
98 | foreach my $subfile (@$file_list) {
|
---|
99 | my $doc_oid = $subfile->[1];
|
---|
100 |
|
---|
101 | my $index_status = $archive_info->get_status_info($doc_oid);
|
---|
102 | print STDERR "*** Updating $doc_oid $index_status\n";
|
---|
103 |
|
---|
104 | if ($index_status eq "D") {
|
---|
105 | # delete
|
---|
106 | $archive_info->delete_info($doc_oid);
|
---|
107 | }
|
---|
108 | elsif ($index_status =~ m/^(I|R)$/) {
|
---|
109 | # mark as "been indexed"
|
---|
110 | $archive_info->set_status_info($doc_oid,"B");
|
---|
111 | }
|
---|
112 | }
|
---|
113 |
|
---|
114 | $archive_info->save_info($archive_info_filename);
|
---|
115 | }
|
---|
116 | }
|
---|
117 |
|
---|
118 | # called at the beginning of each plugin pass (import has one, buildin has many)
|
---|
119 | sub begin {
|
---|
120 | my $self = shift (@_);
|
---|
121 | my ($pluginfo, $base_dir, $processor, $maxdocs) = @_;
|
---|
122 |
|
---|
123 | }
|
---|
124 |
|
---|
125 | # called at the end of each plugin pass
|
---|
126 | sub end {
|
---|
127 | my ($self) = shift (@_);
|
---|
128 |
|
---|
129 | }
|
---|
130 |
|
---|
131 | # called if we are doing incremental building
|
---|
132 | sub set_incremental {
|
---|
133 | my $self = shift(@_);
|
---|
134 | my ($incremental) = @_;
|
---|
135 |
|
---|
136 | $self->{'incremental'} = $incremental;
|
---|
137 | }
|
---|
138 |
|
---|
139 | # return 1 if this class might recurse using $pluginfo
|
---|
140 | sub is_recursive {
|
---|
141 | my $self = shift (@_);
|
---|
142 |
|
---|
143 | return 1;
|
---|
144 | }
|
---|
145 |
|
---|
146 |
|
---|
147 | sub compile_stats {
|
---|
148 | my $self = shift(@_);
|
---|
149 | my ($stats) = @_;
|
---|
150 | }
|
---|
151 |
|
---|
152 | # We don't do metadata_read
|
---|
153 | sub metadata_read {
|
---|
154 | my $self = shift (@_);
|
---|
155 | my ($pluginfo, $base_dir, $file, $block_hash, $extrametakeys, $extrametadata, $processor, $maxdocs, $gli) = @_;
|
---|
156 |
|
---|
157 | return undef;
|
---|
158 | }
|
---|
159 |
|
---|
160 | # we don't do any file blocking
|
---|
161 | sub file_block_read {
|
---|
162 |
|
---|
163 | my $self = shift (@_);
|
---|
164 | my ($pluginfo, $base_dir, $file, $block_hash, $metadata, $gli) = @_;
|
---|
165 |
|
---|
166 | return undef;
|
---|
167 | }
|
---|
168 |
|
---|
169 |
|
---|
170 | # return number of files processed, undef if can't process
|
---|
171 | # Note that $base_dir might be "" and that $file might
|
---|
172 | # include directories
|
---|
173 | sub read {
|
---|
174 | my $self = shift (@_);
|
---|
175 | my ($pluginfo, $base_dir, $file, $block_hash, $metadata, $processor, $maxdocs,$total_count, $gli) = @_;
|
---|
176 | my $outhandle = $self->{'outhandle'};
|
---|
177 |
|
---|
178 | my $count = 0;
|
---|
179 |
|
---|
180 | # see if this has a archives information file within it
|
---|
181 | ## my $archive_info_filename = &util::filename_cat($base_dir,$file,"archives.inf");
|
---|
182 | my $db_ext = &util::is_little_endian() ? ".ldb" : ".bdb";
|
---|
183 | my $doc_db = "archiveinf-doc$db_ext";
|
---|
184 | my $archive_info_filename = &util::filename_cat($base_dir,$file,$doc_db);
|
---|
185 |
|
---|
186 | if (-e $archive_info_filename) {
|
---|
187 |
|
---|
188 | # found an archives.inf file
|
---|
189 | &gsprintf($outhandle, "ArchivesInfPlugin: {common.processing} $archive_info_filename\n") if $self->{'verbosity'} > 1;
|
---|
190 |
|
---|
191 | # read in the archives information file
|
---|
192 | my $archive_info = new arcinfo ();
|
---|
193 | $self->{'archive_info'} = $archive_info;
|
---|
194 | $self->{'archive_info_filename'} = $archive_info_filename;
|
---|
195 |
|
---|
196 | $archive_info->load_info ($archive_info_filename);
|
---|
197 |
|
---|
198 | my $file_list = $archive_info->get_file_list();
|
---|
199 |
|
---|
200 | # process each file
|
---|
201 | foreach my $subfile (@$file_list) {
|
---|
202 |
|
---|
203 | last if ($maxdocs != -1 && ($total_count + $count) >= $maxdocs);
|
---|
204 |
|
---|
205 | my $tmp = &util::filename_cat ($file, $subfile->[0]);
|
---|
206 | next if $tmp eq $file;
|
---|
207 |
|
---|
208 | my $doc_oid = $subfile->[1];
|
---|
209 | my $index_status = $archive_info->get_status_info($doc_oid);
|
---|
210 |
|
---|
211 | my $curr_mode = $processor->get_mode();
|
---|
212 | my $new_mode = $curr_mode;
|
---|
213 |
|
---|
214 | # Start by assuming we want to process the file...
|
---|
215 | my $process_file = 1;
|
---|
216 |
|
---|
217 | # ...unless the build processor is incremental capable and -incremental was specified
|
---|
218 | if ($processor->is_incremental_capable() && $self->{'incremental'})
|
---|
219 | {
|
---|
220 | # Check to see if the file needs indexing
|
---|
221 | if ($index_status eq "B")
|
---|
222 | {
|
---|
223 | # Don't process this file as it has already been indexed
|
---|
224 | $process_file = 0;
|
---|
225 | }
|
---|
226 | elsif ($index_status eq "D") {
|
---|
227 | # Need to be delete it from the index.
|
---|
228 | $new_mode = $curr_mode."delete";
|
---|
229 | $process_file = 1;
|
---|
230 | }
|
---|
231 | elsif ($index_status eq "R") {
|
---|
232 | # Need to be delete it from the index.
|
---|
233 | $new_mode = $curr_mode."reindex";
|
---|
234 | $process_file = 1;
|
---|
235 | }
|
---|
236 | }
|
---|
237 | # ... or we're being asked to delete it (in which case skip it)
|
---|
238 | elsif ($index_status eq "D") {
|
---|
239 | # Delete it somehow from archives dir!!
|
---|
240 | # => get short name, lop off filename, concat archivedir
|
---|
241 | # move to recyle bin
|
---|
242 |
|
---|
243 | $process_file = 0;
|
---|
244 | }
|
---|
245 |
|
---|
246 | if ($process_file) {
|
---|
247 | # note: metadata is not carried on to the next level
|
---|
248 |
|
---|
249 | $processor->set_mode($new_mode) if ($new_mode ne $curr_mode);
|
---|
250 |
|
---|
251 | $count += &plugin::read ($pluginfo, $base_dir, $tmp, $block_hash, {}, $processor, $maxdocs, ($total_count+$count), $gli);
|
---|
252 |
|
---|
253 | $processor->set_mode($curr_mode) if ($new_mode ne $curr_mode);
|
---|
254 | }
|
---|
255 | }
|
---|
256 |
|
---|
257 | return $count;
|
---|
258 | }
|
---|
259 |
|
---|
260 | # wasn't an archives directory, someone else will have to process it
|
---|
261 | return undef;
|
---|
262 | }
|
---|
263 |
|
---|
264 | 1;
|
---|