| 1 | #!/usr/bin/perl -w
|
|---|
| 2 |
|
|---|
| 3 | ###########################################################################
|
|---|
| 4 | #
|
|---|
| 5 | # buildcol.pl -- This program will build a particular collection
|
|---|
| 6 | # A component of the Greenstone digital library software
|
|---|
| 7 | # from the New Zealand Digital Library Project at the
|
|---|
| 8 | # University of Waikato, New Zealand.
|
|---|
| 9 | #
|
|---|
| 10 | # Copyright (C) 1999 New Zealand Digital Library Project
|
|---|
| 11 | #
|
|---|
| 12 | # This program is free software; you can redistribute it and/or modify
|
|---|
| 13 | # it under the terms of the GNU General Public License as published by
|
|---|
| 14 | # the Free Software Foundation; either version 2 of the License, or
|
|---|
| 15 | # (at your option) any later version.
|
|---|
| 16 | #
|
|---|
| 17 | # This program is distributed in the hope that it will be useful,
|
|---|
| 18 | # but WITHOUT ANY WARRANTY; without even the implied warranty of
|
|---|
| 19 | # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
|---|
| 20 | # GNU General Public License for more details.
|
|---|
| 21 | #
|
|---|
| 22 | # You should have received a copy of the GNU General Public License
|
|---|
| 23 | # along with this program; if not, write to the Free Software
|
|---|
| 24 | # Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
|
|---|
| 25 | #
|
|---|
| 26 | ###########################################################################
|
|---|
| 27 |
|
|---|
| 28 | package buildcol;
|
|---|
| 29 |
|
|---|
| 30 | BEGIN {
|
|---|
| 31 | die "GSDLHOME not set\n" unless defined $ENV{'GSDLHOME'};
|
|---|
| 32 | die "GSDLOS not set\n" unless defined $ENV{'GSDLOS'};
|
|---|
| 33 | unshift (@INC, "$ENV{'GSDLHOME'}/perllib");
|
|---|
| 34 | unshift (@INC, "$ENV{'GSDLHOME'}/perllib/plugins");
|
|---|
| 35 | unshift (@INC, "$ENV{'GSDLHOME'}/perllib/classify");
|
|---|
| 36 | }
|
|---|
| 37 |
|
|---|
| 38 | use colcfg;
|
|---|
| 39 | use parsargv;
|
|---|
| 40 | use util;
|
|---|
| 41 | use FileHandle;
|
|---|
| 42 |
|
|---|
| 43 | &main();
|
|---|
| 44 |
|
|---|
| 45 | sub print_usage {
|
|---|
| 46 | print STDOUT "\n";
|
|---|
| 47 | print STDOUT "buildcol.pl: Builds the indexes of a Greenstone collection.\n\n";
|
|---|
| 48 | print STDOUT " usage: $0 [options] collection-name\n\n";
|
|---|
| 49 | print STDOUT " options:\n";
|
|---|
| 50 | print STDOUT " -verbosity number 0=none, 3=lots\n";
|
|---|
| 51 | print STDOUT " -archivedir directory Where the archives live\n";
|
|---|
| 52 | print STDOUT " -builddir directory Where to put the built indexes\n";
|
|---|
| 53 | print STDOUT " -maxdocs number Maximum number of documents to build\n";
|
|---|
| 54 | print STDOUT " -debug Print output to STDOUT\n";
|
|---|
| 55 | print STDOUT " -mode all|compress_text|build_index|infodb\n";
|
|---|
| 56 | print STDOUT " -index indexname Index to build (will build all in\n";
|
|---|
| 57 | print STDOUT " config file if not set)\n";
|
|---|
| 58 | print STDOUT " -keepold will not destroy the current contents of the\n";
|
|---|
| 59 | print STDOUT " building directory\n";
|
|---|
| 60 | print STDOUT " -no_text Don't store compressed text. This option is\n";
|
|---|
| 61 | print STDOUT " useful for minimizing the size of the built\n";
|
|---|
| 62 | print STDOUT " indexes if you intend always to display the\n";
|
|---|
| 63 | print STDOUT " original documents at run time (i.e. you won't\n";
|
|---|
| 64 | print STDOUT " be able to retrieve the compressed text version)\n";
|
|---|
| 65 | print STDOUT " -allclassifications Don't remove empty classifications\n";
|
|---|
| 66 | print STDOUT " -create_images Attempt to create default images for new\n";
|
|---|
| 67 | print STDOUT " collection. This relies on the Gimp being\n";
|
|---|
| 68 | print STDOUT " installed along with relevant perl modules\n";
|
|---|
| 69 | print STDOUT " to allow scripting from perl\n";
|
|---|
| 70 | print STDOUT " -collectdir directory Collection directory (defaults to " .
|
|---|
| 71 | &util::filename_cat ($ENV{'GSDLHOME'}, "collect") . ")\n";
|
|---|
| 72 | print STDOUT " -out Filename or handle to print output status to.\n";
|
|---|
| 73 | print STDOUT " The default is STDERR\n";
|
|---|
| 74 | print STDOUT " -buildtype mg|mgpp This will override the config file setting.\n";
|
|---|
| 75 | print STDOUT " (default is mg)\n";
|
|---|
| 76 | print STDOUT " -no_strip_html Do not strip the html tags from the indexed text\n";
|
|---|
| 77 | print STDOUT " (only used for mgpp collections).\n\n";
|
|---|
| 78 | print STDOUT " [Type \"perl -S buildcol.pl | more\" if this help text scrolled off your screen]";
|
|---|
| 79 | print STDOUT "\n" unless $ENV{'GSDLOS'} =~ /^windows$/i;
|
|---|
| 80 | }
|
|---|
| 81 |
|
|---|
| 82 |
|
|---|
| 83 | sub main
|
|---|
| 84 | {
|
|---|
| 85 | my ($verbosity, $archivedir, $cachedir, $builddir, $maxdocs,
|
|---|
| 86 | $debug, $mode, $indexname, $keepold, $allclassifications,
|
|---|
| 87 | $create_images, $collectdir, $out, $buildtype, $textindex,
|
|---|
| 88 | $no_strip_html, $no_text);
|
|---|
| 89 |
|
|---|
| 90 | # note that no defaults are passed for most options as they're set
|
|---|
| 91 | # later (after we check the collect.cfg file)
|
|---|
| 92 | if (!parsargv::parse(\@ARGV,
|
|---|
| 93 | 'verbosity/\d+/', \$verbosity,
|
|---|
| 94 | 'archivedir/.*/', \$archivedir,
|
|---|
| 95 | 'cachedir/.*/', \$cachedir,
|
|---|
| 96 | 'builddir/.*/', \$builddir,
|
|---|
| 97 | 'maxdocs/^\-?\d+/', \$maxdocs,
|
|---|
| 98 | 'debug', \$debug,
|
|---|
| 99 | 'mode/^(all|compress_text|build_index|infodb)$/', \$mode,
|
|---|
| 100 | 'index/.*/', \$indexname,
|
|---|
| 101 | 'no_text', \$no_text,
|
|---|
| 102 | 'keepold', \$keepold,
|
|---|
| 103 | 'allclassifications', \$allclassifications,
|
|---|
| 104 | 'create_images', \$create_images,
|
|---|
| 105 | 'collectdir/.*/', \$collectdir,
|
|---|
| 106 | 'out/.*/STDERR', \$out,
|
|---|
| 107 | 'no_strip_html', \$no_strip_html,
|
|---|
| 108 | 'buildtype/^(mg|mgpp)$/', \$buildtype)) {
|
|---|
| 109 | &print_usage();
|
|---|
| 110 | die "\n";
|
|---|
| 111 | }
|
|---|
| 112 |
|
|---|
| 113 | $textindex = "";
|
|---|
| 114 | my $close_out = 0;
|
|---|
| 115 | if ($out !~ /^(STDERR|STDOUT)$/i) {
|
|---|
| 116 | open (OUT, ">$out") || die "Couldn't open output file $out\n";
|
|---|
| 117 | $out = "buildcol::OUT";
|
|---|
| 118 | $close_out = 1;
|
|---|
| 119 | }
|
|---|
| 120 | $out->autoflush(1);
|
|---|
| 121 |
|
|---|
| 122 | # get and check the collection
|
|---|
| 123 | if (($collection = &util::use_collection(@ARGV, $collectdir)) eq "") {
|
|---|
| 124 | &print_usage();
|
|---|
| 125 | die "\n";
|
|---|
| 126 | }
|
|---|
| 127 |
|
|---|
| 128 | # read the configuration file
|
|---|
| 129 | $configfilename = &util::filename_cat ($ENV{'GSDLCOLLECTDIR'}, "etc", "collect.cfg");
|
|---|
| 130 | if (-e $configfilename) {
|
|---|
| 131 | $collectcfg = &colcfg::read_collect_cfg ($configfilename);
|
|---|
| 132 |
|
|---|
| 133 | if ($verbosity !~ /\d+/) {
|
|---|
| 134 | if (defined $collectcfg->{'verbosity'} && $collectcfg->{'verbosity'} =~ /\d+/) {
|
|---|
| 135 | $verbosity = $collectcfg->{'verbosity'};
|
|---|
| 136 | } else {
|
|---|
| 137 | $verbosity = 2; # the default
|
|---|
| 138 | }
|
|---|
| 139 | }
|
|---|
| 140 | if (defined $collectcfg->{'buildtype'} && $buildtype eq "") {
|
|---|
| 141 | $buildtype = $collectcfg->{'buildtype'};
|
|---|
| 142 | }
|
|---|
| 143 | if ($buildtype eq "") {
|
|---|
| 144 | $buildtype = "mg"; # mg is the default
|
|---|
| 145 | }
|
|---|
| 146 | if (defined $collectcfg->{'archivedir'} && $archivedir eq "") {
|
|---|
| 147 | $archivedir = $collectcfg->{'archivedir'};
|
|---|
| 148 | }
|
|---|
| 149 | if (defined $collectcfg->{'cachedir'} && $cachedir eq "") {
|
|---|
| 150 | $cachedir = $collectcfg->{'cachedir'};
|
|---|
| 151 | }
|
|---|
| 152 | if (defined $collectcfg->{'builddir'} && $builddir eq "") {
|
|---|
| 153 | $builddir = $collectcfg->{'builddir'};
|
|---|
| 154 | }
|
|---|
| 155 | if ($maxdocs !~ /\-?\d+/) {
|
|---|
| 156 | if (defined $collectcfg->{'maxdocs'} && $collectcfg->{'maxdocs'} =~ /\-?\d+/) {
|
|---|
| 157 | $maxdocs = $collectcfg->{'maxdocs'};
|
|---|
| 158 | } else {
|
|---|
| 159 | $maxdocs = -1; # the default
|
|---|
| 160 | }
|
|---|
| 161 | }
|
|---|
| 162 | if (defined $collectcfg->{'debug'} && $collectcfg->{'debug'} =~ /^true$/i) {
|
|---|
| 163 | $debug = 1;
|
|---|
| 164 | }
|
|---|
| 165 | if ($mode !~ /^(all|compress_text|build_index|infodb)$/) {
|
|---|
| 166 | if (defined $collectcfg->{'mode'} && $collectcfg->{'mode'} =~ /^(all|compress_text|build_index|infodb)$/) {
|
|---|
| 167 | $mode = $collectcfg->{'mode'};
|
|---|
| 168 | } else {
|
|---|
| 169 | $mode = "all"; # the default
|
|---|
| 170 | }
|
|---|
| 171 | }
|
|---|
| 172 | if (defined $collectcfg->{'index'} && $indexname eq "") {
|
|---|
| 173 | $indexname = $collectcfg->{'index'};
|
|---|
| 174 | }
|
|---|
| 175 | if (defined $collectcfg->{'no_text'} && $no_text == 0) {
|
|---|
| 176 | if ($collectcfg->{'no_text'} =~ /^true$/i) {
|
|---|
| 177 | $no_text = 1;
|
|---|
| 178 | }
|
|---|
| 179 | }
|
|---|
| 180 | if (defined $collectcfg->{'allclassifications'} && $allclassifications == 0) {
|
|---|
| 181 | if ($collectcfg->{'allclassifications'} =~ /^true$/i) {
|
|---|
| 182 | $allclassifications = 1;
|
|---|
| 183 | }
|
|---|
| 184 | }
|
|---|
| 185 | if (defined $collectcfg->{'keepold'} && $collectcfg->{'keepold'} =~ /^true$/i) {
|
|---|
| 186 | $keepold = 1;
|
|---|
| 187 | }
|
|---|
| 188 | if (defined $collectcfg->{'create_images'} && $collectcfg->{'create_images'} =~ /^true$/i) {
|
|---|
| 189 | $create_images = 1;
|
|---|
| 190 | }
|
|---|
| 191 | if ($buildtype eq "mgpp" && defined $collectcfg->{'textcompress'}) {
|
|---|
| 192 | $textindex = $collectcfg->{'textcompress'};
|
|---|
| 193 | }
|
|---|
| 194 |
|
|---|
| 195 | } else {
|
|---|
| 196 | die "Couldn't find the configuration file $configfilename\n";
|
|---|
| 197 | }
|
|---|
| 198 |
|
|---|
| 199 | #set the text index
|
|---|
| 200 | if ($buildtype eq "mgpp") {
|
|---|
| 201 | if ($textindex eq "") {
|
|---|
| 202 | $textindex = "text";
|
|---|
| 203 | }
|
|---|
| 204 | }
|
|---|
| 205 | else {
|
|---|
| 206 | $textindex = "section:text";
|
|---|
| 207 | }
|
|---|
| 208 |
|
|---|
| 209 | # create default images if required
|
|---|
| 210 | if ($create_images) {
|
|---|
| 211 | my $collection_name = $collection;
|
|---|
| 212 | $collection_name = $collectcfg->{'collectionmeta'}->{'collectionname'}
|
|---|
| 213 | if defined $collectcfg->{'collectionmeta'}->{'collectionname'};
|
|---|
| 214 |
|
|---|
| 215 | &create_images ($collection_name);
|
|---|
| 216 | }
|
|---|
| 217 |
|
|---|
| 218 | # fill in the default archives and building directories if none
|
|---|
| 219 | # were supplied, turn all \ into / and remove trailing /
|
|---|
| 220 | $archivedir = &util::filename_cat ($ENV{'GSDLCOLLECTDIR'}, "archives") if $archivedir eq "";
|
|---|
| 221 | $archivedir =~ s/[\\\/]+/\//g;
|
|---|
| 222 | $archivedir =~ s/\/$//;
|
|---|
| 223 | $builddir = &util::filename_cat ($ENV{'GSDLCOLLECTDIR'}, "building") if $builddir eq "";
|
|---|
| 224 | $builddir =~ s/[\\\/]+/\//g;
|
|---|
| 225 | $builddir =~ s/\/$//;
|
|---|
| 226 |
|
|---|
| 227 | # update the archive cache if needed
|
|---|
| 228 | if ($cachedir) {
|
|---|
| 229 | print $out "Updating archive cache\n" if ($verbosity >= 1);
|
|---|
| 230 |
|
|---|
| 231 | $cachedir =~ s/[\\\/]+$//;
|
|---|
| 232 | $cachedir .= "/collect/$collection" unless
|
|---|
| 233 | $cachedir =~ /collect\/$collection/;
|
|---|
| 234 |
|
|---|
| 235 | $realarchivedir = "$cachedir/archives";
|
|---|
| 236 | $realbuilddir = "$cachedir/building";
|
|---|
| 237 | &util::mk_all_dir ($realarchivedir);
|
|---|
| 238 | &util::mk_all_dir ($realbuilddir);
|
|---|
| 239 | &util::cachedir ($archivedir, $realarchivedir, $verbosity);
|
|---|
| 240 |
|
|---|
| 241 | } else {
|
|---|
| 242 | $realarchivedir = $archivedir;
|
|---|
| 243 | $realbuilddir = $builddir;
|
|---|
| 244 | }
|
|---|
| 245 |
|
|---|
| 246 | # build it in realbuilddir
|
|---|
| 247 | &util::mk_all_dir ($realbuilddir);
|
|---|
| 248 |
|
|---|
| 249 |
|
|---|
| 250 | # if a builder class has been created for this collection, use it
|
|---|
| 251 | # otherwise, use the mg or mgpp builder
|
|---|
| 252 | if (-e "$ENV{'GSDLCOLLECTDIR'}/perllib/${collection}builder.pm") {
|
|---|
| 253 | $builderdir = "$ENV{'GSDLCOLLECTDIR'}/perllib";
|
|---|
| 254 | $buildertype = "${collection}builder";
|
|---|
| 255 | } else {
|
|---|
| 256 | $builderdir = "$ENV{'GSDLHOME'}/perllib";
|
|---|
| 257 | if ($buildtype eq "mgpp") {
|
|---|
| 258 | $buildertype = "mgppbuilder";
|
|---|
| 259 | }
|
|---|
| 260 | else {
|
|---|
| 261 | $buildertype = "mgbuilder";
|
|---|
| 262 | }
|
|---|
| 263 | }
|
|---|
| 264 |
|
|---|
| 265 | require "$builderdir/$buildertype.pm";
|
|---|
| 266 |
|
|---|
| 267 | eval("\$builder = new $buildertype(\$collection, " .
|
|---|
| 268 | "\$realarchivedir, \$realbuilddir, \$verbosity, " .
|
|---|
| 269 | "\$maxdocs, \$debug, \$keepold, \$allclassifications, " .
|
|---|
| 270 | "\$out, \$no_text)");
|
|---|
| 271 | die "$@" if $@;
|
|---|
| 272 |
|
|---|
| 273 | $builder->init();
|
|---|
| 274 |
|
|---|
| 275 | if ($buildertype eq "mgppbuilder" && $no_strip_html) {
|
|---|
| 276 | $builder->set_strip_html(0);
|
|---|
| 277 | }
|
|---|
| 278 | if ($mode =~ /^all$/i) {
|
|---|
| 279 | $builder->compress_text($textindex);
|
|---|
| 280 | $builder->build_indexes($indexname);
|
|---|
| 281 | $builder->make_infodatabase();
|
|---|
| 282 | $builder->collect_specific();
|
|---|
| 283 | } elsif ($mode =~ /^compress_text$/i) {
|
|---|
| 284 | $builder->compress_text($textindex);
|
|---|
| 285 | } elsif ($mode =~ /^build_index$/i) {
|
|---|
| 286 | $builder->build_indexes($indexname);
|
|---|
| 287 | } elsif ($mode =~ /^infodb$/i) {
|
|---|
| 288 | $builder->make_infodatabase();
|
|---|
| 289 | } else {
|
|---|
| 290 | die "unknown mode: $mode\n";
|
|---|
| 291 | }
|
|---|
| 292 |
|
|---|
| 293 | $builder->make_auxiliary_files() if !$debug;
|
|---|
| 294 | $builder->deinit();
|
|---|
| 295 |
|
|---|
| 296 | if (($realbuilddir ne $builddir) && !$debug) {
|
|---|
| 297 | print $out "Copying back the cached build\n" if ($verbosity >= 1);
|
|---|
| 298 | &util::rm_r ($builddir);
|
|---|
| 299 | &util::cp_r ($realbuilddir, $builddir);
|
|---|
| 300 | }
|
|---|
| 301 |
|
|---|
| 302 | close OUT if $close_out;
|
|---|
| 303 | }
|
|---|
| 304 |
|
|---|
| 305 | sub create_images {
|
|---|
| 306 | my ($collection_name) = @_;
|
|---|
| 307 |
|
|---|
| 308 | my $image_script = &util::filename_cat ($ENV{'GSDLHOME'}, "bin", "script", "gimp", "title_icon.pl");
|
|---|
| 309 | if (!-e $image_script) {
|
|---|
| 310 | print $out "WARNING: Image making script ($image_script) could not be found\n";
|
|---|
| 311 | print $out " Default images will not be generated\n\n";
|
|---|
| 312 | return;
|
|---|
| 313 | }
|
|---|
| 314 |
|
|---|
| 315 | my $imagedir = &util::filename_cat ($ENV{'GSDLCOLLECTDIR'}, "images");
|
|---|
| 316 |
|
|---|
| 317 | &util::mk_all_dir ($imagedir);
|
|---|
| 318 |
|
|---|
| 319 | # create the images
|
|---|
| 320 | system ("$image_script -size 1.5 -image_dir \"$imagedir\" -filename $collection.gif -text \"$collection_name\"");
|
|---|
| 321 | system ("$image_script -image_dir \"$imagedir\" -filename ${collection}sm.gif -text \"$collection_name\"");
|
|---|
| 322 |
|
|---|
| 323 | # update the collect.cfg configuration file (this will need
|
|---|
| 324 | # to be changed when the config file format changes)
|
|---|
| 325 | if (!open (CFGFILE, $configfilename)) {
|
|---|
| 326 | print $out "WARNING: Couldn't open config file ($configfilename)\n";
|
|---|
| 327 | print $out " for updating so collection images may not be linked correctly\n";
|
|---|
| 328 | return;
|
|---|
| 329 | }
|
|---|
| 330 |
|
|---|
| 331 | my $line = ""; my $file = "";
|
|---|
| 332 | my $found = 0; my $foundsm = 0;
|
|---|
| 333 | while (defined ($line = <CFGFILE>)) {
|
|---|
| 334 | if ($line =~ /collectionmeta\s+iconcollection\s+/) {
|
|---|
| 335 | $line = "collectionmeta iconcollection _httpprefix_/collect/$collection/images/$collection.gif\n";
|
|---|
| 336 | $found = 1;
|
|---|
| 337 | } elsif ($line =~ /collectionmeta\s+iconcollectionsmall\s+/) {
|
|---|
| 338 | $line = "collectionmeta iconcollectionsmall _httpprefix_/collect/$collection/images/${collection}sm.gif\n";
|
|---|
| 339 | $foundsm = 1;
|
|---|
| 340 | }
|
|---|
| 341 | $file .= $line;
|
|---|
| 342 | }
|
|---|
| 343 | close CFGFILE;
|
|---|
| 344 |
|
|---|
| 345 | $file .= "collectionmeta iconcollection _httpprefix_/collect/$collection/images/$collection.gif\n" if !$found;
|
|---|
| 346 | $file .= "collectionmeta iconcollectionsmall _httpprefix_/collect/$collection/images/${collection}sm.gif\n" if !$foundsm;
|
|---|
| 347 |
|
|---|
| 348 | if (!open (CFGFILE, ">$configfilename")) {
|
|---|
| 349 | print $out "WARNING: Couldn't open config file ($configfilename)\n";
|
|---|
| 350 | print $out " for updating so collection images may not be linked correctly\n";
|
|---|
| 351 | return;
|
|---|
| 352 | }
|
|---|
| 353 | print CFGFILE $file;
|
|---|
| 354 | close CFGFILE;
|
|---|
| 355 | }
|
|---|
| 356 |
|
|---|
