root/gsdl/trunk/perllib/plugins/ArchivesInfPlugin.pm @ 18508

Revision 18508, 8.0 KB (checked in by davidb, 11 years ago)

Had to move location of where deletion of archive files was done

  • Property svn:executable set to *
  • Property svn:keywords set to Author Date Id Revision
Line 
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
31package ArchivesInfPlugin;
32
33use util;
34use PrintInfo;
35use plugin;
36use arcinfo;
37use gsprintf;
38
39use strict;
40no strict 'refs'; # allow filehandles to be variables and viceversa
41
42BEGIN {
43    @ArchivesInfPlugin::ISA = ('PrintInfo');
44}
45
46my $arguments = [
47         ];
48
49my $options = { 'name'     => "ArchivesInfPlugin",
50        'desc'     => "{ArchivesInfPlugin.desc}",
51        'abstract' => "no",
52        'inherits' => "yes" };
53         
54sub gsprintf
55{
56    return &gsprintf::gsprintf(@_);
57}
58
59sub new {
60    my ($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
73sub 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'} = $failhandle;
83
84}
85
86sub deinit {
87    my ($self) = @_;
88
89    my $archive_info = $self->{'archive_info'};
90    my $verbosity = $self->{'verbosity'};
91    my $outhandle = $self->{'outhandle'};
92
93    if (defined $archive_info) {
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
103        if ($index_status eq "D") {
104        # delete
105        $archive_info->delete_info($doc_oid);
106        &GDBMUtils::gdbmDatabaseRemove($archive_info_filename,$doc_oid);
107
108        my $doc_file = $subfile->[0];
109        my $base_dir =$self->{'base_dir'};
110
111        my $doc_filename = &util::filename_cat($base_dir,$doc_file);
112
113        my ($doc_tailname, $doc_dirname, $suffix)
114            = File::Basename::fileparse($doc_filename, "\\.[^\\.]+\$");
115
116        print STDERR "Removing $doc_dirname\n" if ($verbosity>2);
117
118        &util::rm_r($doc_dirname);
119
120
121        }
122        elsif ($index_status =~ m/^(I|R)$/) {
123        # mark as "been indexed"
124        $archive_info->set_status_info($doc_oid,"B");
125        }
126    }
127
128    $archive_info->save_info($archive_info_filename);
129    }
130}
131
132# called at the beginning of each plugin pass (import has one, buildin has many)
133sub begin {
134    my $self = shift (@_);
135    my ($pluginfo, $base_dir, $processor, $maxdocs) = @_;
136
137    $self->{'base_dir'} = $base_dir;
138}
139
140# called at the end of each plugin pass
141sub end {
142    my ($self) = shift (@_);
143
144}
145
146# called if we are doing incremental building
147sub set_incremental {
148    my $self = shift(@_);
149    my ($incremental) = @_;
150
151    $self->{'incremental'} = $incremental;
152}
153
154# return 1 if this class might recurse using $pluginfo
155sub is_recursive {
156    my $self = shift (@_);
157
158    return 1;
159}
160
161
162sub compile_stats {
163    my $self = shift(@_);
164    my ($stats) = @_;
165}
166
167# We don't do metadata_read
168sub metadata_read {
169    my $self = shift (@_);
170    my ($pluginfo, $base_dir, $file, $block_hash, $extrametakeys, $extrametadata, $processor, $maxdocs, $gli) = @_;
171
172    return undef;
173}
174
175# we don't do any file blocking
176sub file_block_read {
177
178    my $self = shift (@_); 
179    my ($pluginfo, $base_dir, $file, $block_hash, $metadata, $gli) = @_;
180
181    return undef;
182}
183
184
185# return number of files processed, undef if can't process
186# Note that $base_dir might be "" and that $file might
187# include directories
188sub read {
189    my $self = shift (@_);
190    my ($pluginfo, $base_dir, $file, $block_hash, $metadata, $processor, $maxdocs,$total_count, $gli) = @_;
191    my $outhandle = $self->{'outhandle'};
192
193    my $count = 0;
194
195    # see if this has a archives information file within it
196##    my $archive_info_filename = &util::filename_cat($base_dir,$file,"archives.inf");
197    my $db_ext = &util::is_little_endian() ? ".ldb" : ".bdb";
198    my $doc_db = "archiveinf-doc$db_ext";
199    my $archive_info_filename = &util::filename_cat($base_dir,$file,$doc_db);
200
201    if (-e $archive_info_filename) {
202
203    # found an archives.inf file
204    &gsprintf($outhandle, "ArchivesInfPlugin: {common.processing} $archive_info_filename\n") if $self->{'verbosity'} > 1;
205
206    # read in the archives information file
207    my $archive_info = new arcinfo ();
208    $self->{'archive_info'} = $archive_info;
209    $self->{'archive_info_filename'} = $archive_info_filename;
210
211    $archive_info->load_info ($archive_info_filename);
212   
213    my $file_list = $archive_info->get_file_list();
214
215    # process each file
216    foreach my $subfile (@$file_list) {
217
218        last if ($maxdocs != -1 && ($total_count + $count) >= $maxdocs);
219
220        my $tmp = &util::filename_cat ($file, $subfile->[0]);
221        next if $tmp eq $file;
222
223        my $doc_oid = $subfile->[1];
224        my $index_status = $archive_info->get_status_info($doc_oid);
225
226        my $curr_mode = $processor->get_mode();
227        my $new_mode = $curr_mode;
228
229        # Start by assuming we want to process the file...
230        my $process_file = 1;
231
232        # ...unless the build processor is incremental capable and -incremental was specified, in which case we need to check its index_status flag
233        if ($processor->is_incremental_capable() && $self->{'incremental'})
234        {
235            # Check to see if the file needs indexing
236        if ($index_status eq "B")
237        {
238            # Don't process this file as it has already been indexed
239            $process_file = 0;
240        }
241        elsif ($index_status eq "D") {
242            # Need to be delete it from the index.
243            $new_mode = $curr_mode."delete";
244            $process_file = 1;
245        }
246        elsif ($index_status eq "R") {
247            # Need to be reindexed/replaced
248            $new_mode = $curr_mode."reindex";
249
250            $process_file = 1;
251        }
252        }
253        # ... or we're being asked to delete it (in which case skip it)
254        elsif ($index_status eq "D") {
255        # Non-incremental Delete
256        # It's already been deleted from the archives directory
257        # (done during import.pl)
258        # => All we need to do here is not process it
259
260        $process_file = 0;
261        }
262
263        if (!$processor->is_incremental_capable() && $self->{'incremental'}) {
264        # Nag feature
265        if (!defined $self->{'incremental-warning'}) {
266            print $outhandle "\n";
267            print $outhandle "Warning: command-line option '-incremental' used with *non-incremental*\n";
268            print $outhandle "         processor '", ref $processor, "'. Some conflicts may arise.\n";
269            print $outhandle "\n";
270            sleep 10;
271            $self->{'incremental-warning'} = 1;
272        }
273        }
274
275        if ($process_file) {
276        # note: metadata is not carried on to the next level
277       
278        $processor->set_mode($new_mode) if ($new_mode ne $curr_mode);
279
280        $count += &plugin::read ($pluginfo, $base_dir, $tmp, $block_hash, {}, $processor, $maxdocs, ($total_count+$count), $gli);
281
282        $processor->set_mode($curr_mode) if ($new_mode ne $curr_mode);
283        }
284    }
285
286    return $count;
287    }
288
289    # wasn't an archives directory, someone else will have to process it
290    return undef;
291}
292
2931;
Note: See TracBrowser for help on using the browser.