| 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 archiveinf-doc info database equivalent)
|
|---|
| 27 | # -- i.e. the file generated in the archives directory when an import is done),
|
|---|
| 28 | # processing each file it finds
|
|---|
| 29 |
|
|---|
| 30 | package ArchivesInfPlugin;
|
|---|
| 31 |
|
|---|
| 32 | use util;
|
|---|
| 33 | use FileUtils;
|
|---|
| 34 | use doc;
|
|---|
| 35 | use EncodingUtil;
|
|---|
| 36 | use plugin;
|
|---|
| 37 | use arcinfo;
|
|---|
| 38 | use gsprintf;
|
|---|
| 39 |
|
|---|
| 40 | use strict;
|
|---|
| 41 | no strict 'refs'; # allow filehandles to be variables and viceversa
|
|---|
| 42 |
|
|---|
| 43 |
|
|---|
| 44 | BEGIN {
|
|---|
| 45 | @ArchivesInfPlugin::ISA = ('EncodingUtil');
|
|---|
| 46 | }
|
|---|
| 47 |
|
|---|
| 48 | my $arguments = [
|
|---|
| 49 | { 'name' => "reversesort",
|
|---|
| 50 | 'desc' => "{ArchivesInfPlugin.reversesort}",
|
|---|
| 51 | 'type' => "flag",
|
|---|
| 52 | 'reqd' => "no",
|
|---|
| 53 | 'modegli' => "2" },
|
|---|
| 54 | { 'name' => "sort",
|
|---|
| 55 | 'desc' => "{ArchivesInfPlugin.sort}",
|
|---|
| 56 | 'type' => "flag",
|
|---|
| 57 | 'reqd' => "no",
|
|---|
| 58 | 'modegli' => "2" }
|
|---|
| 59 |
|
|---|
| 60 | ];
|
|---|
| 61 |
|
|---|
| 62 | my $options = { 'name' => "ArchivesInfPlugin",
|
|---|
| 63 | 'desc' => "{ArchivesInfPlugin.desc}",
|
|---|
| 64 | 'abstract' => "no",
|
|---|
| 65 | 'inherits' => "yes",
|
|---|
| 66 | 'args' => $arguments};
|
|---|
| 67 |
|
|---|
| 68 | sub gsprintf
|
|---|
| 69 | {
|
|---|
| 70 | return &gsprintf::gsprintf(@_);
|
|---|
| 71 | }
|
|---|
| 72 |
|
|---|
| 73 | sub new {
|
|---|
| 74 | my ($class) = shift (@_);
|
|---|
| 75 | my ($pluginlist,$inputargs,$hashArgOptLists) = @_;
|
|---|
| 76 | push(@$pluginlist, $class);
|
|---|
| 77 |
|
|---|
| 78 | push(@{$hashArgOptLists->{"ArgList"}},@{$arguments});
|
|---|
| 79 | push(@{$hashArgOptLists->{"OptList"}},$options);
|
|---|
| 80 |
|
|---|
| 81 | my $self = new EncodingUtil($pluginlist, $inputargs, $hashArgOptLists);
|
|---|
| 82 |
|
|---|
| 83 | return bless $self, $class;
|
|---|
| 84 | }
|
|---|
| 85 |
|
|---|
| 86 | # called once, at the start of processing
|
|---|
| 87 | sub init {
|
|---|
| 88 | my $self = shift (@_);
|
|---|
| 89 | my ($verbosity, $outhandle, $failhandle) = @_;
|
|---|
| 90 |
|
|---|
| 91 | # verbosity is passed through from the processor
|
|---|
| 92 | $self->{'verbosity'} = $verbosity;
|
|---|
| 93 |
|
|---|
| 94 | # as are the outhandle and failhandle
|
|---|
| 95 | $self->{'outhandle'} = $outhandle if defined $outhandle;
|
|---|
| 96 | $self->{'failhandle'} = $failhandle;
|
|---|
| 97 |
|
|---|
| 98 | }
|
|---|
| 99 |
|
|---|
| 100 | sub deinit {
|
|---|
| 101 | my ($self) = @_;
|
|---|
| 102 |
|
|---|
| 103 | my $archive_info = $self->{'archive_info'};
|
|---|
| 104 | my $verbosity = $self->{'verbosity'};
|
|---|
| 105 | my $outhandle = $self->{'outhandle'};
|
|---|
| 106 |
|
|---|
| 107 | if (defined $archive_info) {
|
|---|
| 108 | # Get the infodbtype value for this collection from the arcinfo object
|
|---|
| 109 | my $infodbtype = $archive_info->{'infodbtype'};
|
|---|
| 110 | my $archive_info_filename = $self->{'archive_info_filename'};
|
|---|
| 111 | my $infodb_file_handle = &dbutil::open_infodb_write_handle($infodbtype, $archive_info_filename, "append");
|
|---|
| 112 |
|
|---|
| 113 | my $file_list = $archive_info->get_file_list();
|
|---|
| 114 |
|
|---|
| 115 | foreach my $subfile (@$file_list) {
|
|---|
| 116 | my $doc_oid = $subfile->[1];
|
|---|
| 117 |
|
|---|
| 118 | my $index_status = $archive_info->get_status_info($doc_oid);
|
|---|
| 119 |
|
|---|
| 120 | if ($index_status eq "D") {
|
|---|
| 121 | # delete
|
|---|
| 122 | $archive_info->delete_info($doc_oid);
|
|---|
| 123 | &dbutil::delete_infodb_entry($infodbtype, $infodb_file_handle, $doc_oid);
|
|---|
| 124 |
|
|---|
| 125 | my $doc_file = $subfile->[0];
|
|---|
| 126 | my $base_dir =$self->{'base_dir'};
|
|---|
| 127 |
|
|---|
| 128 | my $doc_filename = &FileUtils::filenameConcatenate($base_dir,$doc_file);
|
|---|
| 129 |
|
|---|
| 130 | my ($doc_tailname, $doc_dirname, $suffix)
|
|---|
| 131 | = File::Basename::fileparse($doc_filename, "\\.[^\\.]+\$");
|
|---|
| 132 |
|
|---|
| 133 | print $outhandle "Removing $doc_dirname\n" if ($verbosity>2);
|
|---|
| 134 |
|
|---|
| 135 | &FileUtils::removeFilesRecursive($doc_dirname);
|
|---|
| 136 |
|
|---|
| 137 |
|
|---|
| 138 | }
|
|---|
| 139 | elsif ($index_status =~ m/^(I|R)$/) {
|
|---|
| 140 | # mark as "been indexed"
|
|---|
| 141 | $archive_info->set_status_info($doc_oid,"B");
|
|---|
| 142 | }
|
|---|
| 143 | }
|
|---|
| 144 |
|
|---|
| 145 | &dbutil::close_infodb_write_handle($infodbtype, $infodb_file_handle);
|
|---|
| 146 | $archive_info->save_info($archive_info_filename);
|
|---|
| 147 | }
|
|---|
| 148 | }
|
|---|
| 149 |
|
|---|
| 150 | # called at the beginning of each plugin pass (import has one, buildin has many)
|
|---|
| 151 | sub begin {
|
|---|
| 152 | my $self = shift (@_);
|
|---|
| 153 | my ($pluginfo, $base_dir, $processor, $maxdocs) = @_;
|
|---|
| 154 |
|
|---|
| 155 | $self->{'base_dir'} = $base_dir;
|
|---|
| 156 | }
|
|---|
| 157 |
|
|---|
| 158 | sub remove_all {
|
|---|
| 159 | my $self = shift (@_);
|
|---|
| 160 | my ($pluginfo, $base_dir, $processor, $maxdocs) = @_;
|
|---|
| 161 | }
|
|---|
| 162 |
|
|---|
| 163 | sub remove_one {
|
|---|
| 164 | my $self = shift (@_);
|
|---|
| 165 | my ($file, $oids, $archivedir) = @_;
|
|---|
| 166 | return undef; # only called during import at this stage, this will never be processing a file
|
|---|
| 167 |
|
|---|
| 168 | }
|
|---|
| 169 |
|
|---|
| 170 |
|
|---|
| 171 | # called at the end of each plugin pass
|
|---|
| 172 | sub end {
|
|---|
| 173 | my ($self) = shift (@_);
|
|---|
| 174 |
|
|---|
| 175 | }
|
|---|
| 176 |
|
|---|
| 177 |
|
|---|
| 178 | # return 1 if this class might recurse using $pluginfo
|
|---|
| 179 | sub is_recursive {
|
|---|
| 180 | my $self = shift (@_);
|
|---|
| 181 |
|
|---|
| 182 | return 1;
|
|---|
| 183 | }
|
|---|
| 184 |
|
|---|
| 185 |
|
|---|
| 186 | sub compile_stats {
|
|---|
| 187 | my $self = shift(@_);
|
|---|
| 188 | my ($stats) = @_;
|
|---|
| 189 | }
|
|---|
| 190 |
|
|---|
| 191 | # We don't do metadata_read
|
|---|
| 192 | sub metadata_read {
|
|---|
| 193 | my $self = shift (@_);
|
|---|
| 194 | my ($pluginfo, $base_dir, $file, $block_hash,
|
|---|
| 195 | $extrametakeys, $extrametadata, $extrametafile,
|
|---|
| 196 | $processor, $gli, $aux) = @_;
|
|---|
| 197 |
|
|---|
| 198 | return undef;
|
|---|
| 199 | }
|
|---|
| 200 |
|
|---|
| 201 | sub file_block_read {
|
|---|
| 202 |
|
|---|
| 203 | my $self = shift (@_);
|
|---|
| 204 | my ($pluginfo, $base_dir, $file, $block_hash, $metadata, $gli) = @_;
|
|---|
| 205 |
|
|---|
| 206 | if ($file eq "OIDcount") {
|
|---|
| 207 | my ($filename_full_path, $filename_no_path)
|
|---|
| 208 | = &util::get_full_filenames($base_dir, $file);
|
|---|
| 209 | $self->block_raw_filename($block_hash,$filename_full_path);
|
|---|
| 210 | return 1;
|
|---|
| 211 | }
|
|---|
| 212 |
|
|---|
| 213 | # otherwise, we don't do any file blocking
|
|---|
| 214 |
|
|---|
| 215 | return undef;
|
|---|
| 216 | }
|
|---|
| 217 |
|
|---|
| 218 |
|
|---|
| 219 | # return number of files processed, undef if can't process
|
|---|
| 220 | # Note that $base_dir might be "" and that $file might
|
|---|
| 221 | # include directories
|
|---|
| 222 | sub read {
|
|---|
| 223 | my $self = shift (@_);
|
|---|
| 224 | my ($pluginfo, $base_dir, $file, $block_hash, $metadata, $processor, $maxdocs,$total_count, $gli) = @_;
|
|---|
| 225 | my $outhandle = $self->{'outhandle'};
|
|---|
| 226 |
|
|---|
| 227 | my $count = 0;
|
|---|
| 228 |
|
|---|
| 229 | # This function only makes sense at build-time
|
|---|
| 230 | return if (ref($processor) !~ /buildproc$/i);
|
|---|
| 231 |
|
|---|
| 232 | # Get the infodbtype value for this collection from the buildproc object
|
|---|
| 233 | my $infodbtype = $processor->{'infodbtype'};
|
|---|
| 234 | $infodbtype = "gdbm" if $infodbtype eq "gdbm-txtgz";
|
|---|
| 235 |
|
|---|
| 236 | # see if this has a archives information file within it
|
|---|
| 237 | ## my $archive_info_filename = &FileUtils::filenameConcatenate($base_dir,$file,"archives.inf");
|
|---|
| 238 | my $archive_info_filename = &dbutil::get_infodb_file_path($infodbtype, "archiveinf-doc", &FileUtils::filenameConcatenate($base_dir, $file));
|
|---|
| 239 |
|
|---|
| 240 | if (-e $archive_info_filename) {
|
|---|
| 241 |
|
|---|
| 242 | # found an archives.inf file
|
|---|
| 243 | &gsprintf($outhandle, "ArchivesInfPlugin: {common.processing} $archive_info_filename\n") if $self->{'verbosity'} > 1;
|
|---|
| 244 |
|
|---|
| 245 | # read in the archives information file
|
|---|
| 246 | my $archive_info = new arcinfo($infodbtype);
|
|---|
| 247 | $self->{'archive_info'} = $archive_info;
|
|---|
| 248 | $self->{'archive_info_filename'} = $archive_info_filename;
|
|---|
| 249 | if ($self->{'reversesort'}) {
|
|---|
| 250 | $archive_info->reverse_sort();
|
|---|
| 251 | } elsif ($self->{'sort'}) {
|
|---|
| 252 | $archive_info->sort();
|
|---|
| 253 | }
|
|---|
| 254 |
|
|---|
| 255 | $archive_info->load_info ($archive_info_filename);
|
|---|
| 256 |
|
|---|
| 257 | my $file_list = $archive_info->get_file_list();
|
|---|
| 258 |
|
|---|
| 259 | # process each file
|
|---|
| 260 | foreach my $subfile (@$file_list) {
|
|---|
| 261 |
|
|---|
| 262 | last if ($maxdocs != -1 && ($total_count + $count) >= $maxdocs);
|
|---|
| 263 |
|
|---|
| 264 | my $tmp = &FileUtils::filenameConcatenate($file, $subfile->[0]);
|
|---|
| 265 | next if $tmp eq $file;
|
|---|
| 266 |
|
|---|
| 267 | my $doc_oid = $subfile->[1];
|
|---|
| 268 | my $index_status = $archive_info->get_status_info($doc_oid);
|
|---|
| 269 |
|
|---|
| 270 | my $curr_mode = $processor->get_mode();
|
|---|
| 271 | my $new_mode = $curr_mode;
|
|---|
| 272 | my $group_position = $archive_info->get_group_position($doc_oid);
|
|---|
| 273 |
|
|---|
| 274 | # Start by assuming we want to process the file...
|
|---|
| 275 | my $process_file = 1;
|
|---|
| 276 |
|
|---|
| 277 | # ... unless we have processed files into a group doc.xml, in which case we only process the xml for the first one
|
|---|
| 278 | if (defined $group_position && $group_position >1) {
|
|---|
| 279 | $process_file = 0;
|
|---|
| 280 | }
|
|---|
| 281 | # ...unless the build processor is incremental capable and -incremental was specified, in which case we need to check its index_status flag
|
|---|
| 282 | elsif ($processor->is_incremental_capable() && $self->{'incremental'})
|
|---|
| 283 | {
|
|---|
| 284 | # Check to see if the file needs indexing
|
|---|
| 285 | if ($index_status eq "B")
|
|---|
| 286 | {
|
|---|
| 287 | # Don't process this file as it has already been indexed
|
|---|
| 288 | $process_file = 0;
|
|---|
| 289 | }
|
|---|
| 290 | elsif ($index_status eq "D") {
|
|---|
| 291 | # Need to be delete it from the index.
|
|---|
| 292 | $new_mode = $curr_mode."delete";
|
|---|
| 293 | $process_file = 1;
|
|---|
| 294 | }
|
|---|
| 295 | elsif ($index_status eq "R") {
|
|---|
| 296 | # Need to be reindexed/replaced
|
|---|
| 297 | $new_mode = $curr_mode."reindex";
|
|---|
| 298 |
|
|---|
| 299 | $process_file = 1;
|
|---|
| 300 | }
|
|---|
| 301 | }
|
|---|
| 302 | # ... or we're being asked to delete it (in which case skip it)
|
|---|
| 303 | elsif ($index_status eq "D") {
|
|---|
| 304 | # Non-incremental Delete
|
|---|
| 305 | # It's already been deleted from the archives directory
|
|---|
| 306 | # (done during import.pl)
|
|---|
| 307 | # => All we need to do here is not process it
|
|---|
| 308 |
|
|---|
| 309 | $process_file = 0;
|
|---|
| 310 | }
|
|---|
| 311 |
|
|---|
| 312 | if (!$processor->is_incremental_capable() && $self->{'incremental'}) {
|
|---|
| 313 | # Nag feature
|
|---|
| 314 | if (!defined $self->{'incremental-warning'}) {
|
|---|
| 315 | print $outhandle "\n";
|
|---|
| 316 | print $outhandle "Warning: command-line option '-incremental' used with *non-incremental*\n";
|
|---|
| 317 | print $outhandle " processor '", ref $processor, "'. Some conflicts may arise.\n";
|
|---|
| 318 | print $outhandle "\n";
|
|---|
| 319 | sleep 10;
|
|---|
| 320 | $self->{'incremental-warning'} = 1;
|
|---|
| 321 | }
|
|---|
| 322 | }
|
|---|
| 323 |
|
|---|
| 324 | if ($process_file) {
|
|---|
| 325 | # note: metadata is not carried on to the next level
|
|---|
| 326 |
|
|---|
| 327 | $processor->set_mode($new_mode) if ($new_mode ne $curr_mode);
|
|---|
| 328 |
|
|---|
| 329 | $count += &plugin::read ($pluginfo, $base_dir, $tmp, $block_hash, {}, $processor, $maxdocs, ($total_count+$count), $gli);
|
|---|
| 330 |
|
|---|
| 331 | $processor->set_mode($curr_mode) if ($new_mode ne $curr_mode);
|
|---|
| 332 | }
|
|---|
| 333 | }
|
|---|
| 334 |
|
|---|
| 335 | return $count;
|
|---|
| 336 | }
|
|---|
| 337 |
|
|---|
| 338 |
|
|---|
| 339 | # wasn't an archives directory, someone else will have to process it
|
|---|
| 340 | return undef;
|
|---|
| 341 | }
|
|---|
| 342 |
|
|---|
| 343 | 1;
|
|---|
