| 1 | #!/usr/bin/perl -w
|
|---|
| 2 |
|
|---|
| 3 | ###########################################################################
|
|---|
| 4 | #
|
|---|
| 5 | # gs2_mkcol.pl -- create the framework for a gs2 compatible collection in
|
|---|
| 6 | # gs3
|
|---|
| 7 | # A component of the Greenstone digital library software
|
|---|
| 8 | # from the New Zealand Digital Library Project at the
|
|---|
| 9 | # University of Waikato, New Zealand.
|
|---|
| 10 | #
|
|---|
| 11 | # Copyright (C) 1999 New Zealand Digital Library Project
|
|---|
| 12 | #
|
|---|
| 13 | # This program is free software; you can redistribute it and/or modify
|
|---|
| 14 | # it under the terms of the GNU General Public License as published by
|
|---|
| 15 | # the Free Software Foundation; either version 2 of the License, or
|
|---|
| 16 | # (at your option) any later version.
|
|---|
| 17 | #
|
|---|
| 18 | # This program is distributed in the hope that it will be useful,
|
|---|
| 19 | # but WITHOUT ANY WARRANTY; without even the implied warranty of
|
|---|
| 20 | # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
|---|
| 21 | # GNU General Public License for more details.
|
|---|
| 22 | #
|
|---|
| 23 | # You should have received a copy of the GNU General Public License
|
|---|
| 24 | # along with this program; if not, write to the Free Software
|
|---|
| 25 | # Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
|
|---|
| 26 | #
|
|---|
| 27 | ###########################################################################
|
|---|
| 28 |
|
|---|
| 29 |
|
|---|
| 30 | # This program will setup a new collection from a model one. It does this by
|
|---|
| 31 | # copying the model, moving files to have the correct names, and replacing
|
|---|
| 32 | # text within the files to match the parameters.
|
|---|
| 33 |
|
|---|
| 34 | package gs2_mkcol;
|
|---|
| 35 |
|
|---|
| 36 | BEGIN {
|
|---|
| 37 | die "GSDLHOME not set\n" unless defined $ENV{'GSDLHOME'};
|
|---|
| 38 | die "GSDLOS not set\n" unless defined $ENV{'GSDLOS'};
|
|---|
| 39 | unshift (@INC, "$ENV{'GSDLHOME'}/perllib");
|
|---|
| 40 | }
|
|---|
| 41 |
|
|---|
| 42 | use parsargv;
|
|---|
| 43 | use util;
|
|---|
| 44 | use cfgread;
|
|---|
| 45 |
|
|---|
| 46 | sub print_usage {
|
|---|
| 47 | print STDOUT "\n";
|
|---|
| 48 | print STDOUT "gs2_mkcol.pl: Creates the directory structure for a new\n";
|
|---|
| 49 | print STDOUT " Greenstone2 compatible collection in Greenstone3.\n\n";
|
|---|
| 50 | print STDOUT " usage: $0 -site site-home -creator email [options] collection-name\n\n";
|
|---|
| 51 | print STDOUT " options:\n";
|
|---|
| 52 | print STDOUT " -optionfile file Get options from file, useful on systems where\n";
|
|---|
| 53 | print STDOUT " long command lines may cause problems\n";
|
|---|
| 54 | print STDOUT " -collectdir Directory where new collection will be created.\n";
|
|---|
| 55 | print STDOUT " Default is " .
|
|---|
| 56 | &util::filename_cat("site-home", "collect") . "\n";
|
|---|
| 57 | print STDOUT " -maintainer email The collection maintainer's email address (if\n";
|
|---|
| 58 | print STDOUT " different from the creator)\n";
|
|---|
| 59 | print STDOUT " -public true|false If this collection has anonymous access\n";
|
|---|
| 60 | print STDOUT " -buildtype mg|mgpp Whether to use mg or mgpp to build the collection\n";
|
|---|
| 61 | print STDOUT " Default is mgpp.\n";
|
|---|
| 62 | print STDOUT " -title text The title for the collection\n";
|
|---|
| 63 | print STDOUT " -about text The about text for the collection\n";
|
|---|
| 64 | print STDOUT " -plugin text perl plugin module to use (there may be multiple\n";
|
|---|
| 65 | print STDOUT " plugin entries)\n";
|
|---|
| 66 | print STDOUT " -quiet Operate quietly\n";
|
|---|
| 67 | print STDOUT " Note that -creator and -site must be specified. You can make changes to all\n";
|
|---|
| 68 | print STDOUT " options later by editing the collect.cfg configuration file for your\n";
|
|---|
| 69 | print STDOUT " new collection (it'll be in the \"etc\" directory).\n\n";
|
|---|
| 70 | print STDOUT " [Type \"perl -S mkcol.pl | more\" if this help text scrolled off your screen]";
|
|---|
| 71 | print STDOUT "\n" unless $ENV{'GSDLOS'} =~ /^windows$/i;
|
|---|
| 72 | }
|
|---|
| 73 |
|
|---|
| 74 | sub traverse_dir
|
|---|
| 75 | {
|
|---|
| 76 | my ($modeldir, $coldir) = @_;
|
|---|
| 77 | my ($newfile, @filetext);
|
|---|
| 78 |
|
|---|
| 79 | if (!(-e $coldir)) {
|
|---|
| 80 |
|
|---|
| 81 | my $store_umask = umask(0002);
|
|---|
| 82 | my $mkdir_ok = mkdir ($coldir, 0777);
|
|---|
| 83 | umask($store_umask);
|
|---|
| 84 |
|
|---|
| 85 | if (!$mkdir_ok)
|
|---|
| 86 | {
|
|---|
| 87 | die "$!";
|
|---|
| 88 | }
|
|---|
| 89 | }
|
|---|
| 90 |
|
|---|
| 91 | opendir(DIR, $modeldir) || die "Can't read $modeldir";
|
|---|
| 92 | my @files = grep(!/^(\.\.?|CVS)$/, readdir(DIR));
|
|---|
| 93 | closedir(DIR);
|
|---|
| 94 |
|
|---|
| 95 | foreach $file (@files)
|
|---|
| 96 | {
|
|---|
| 97 | my $thisfile = &util::filename_cat ($modeldir, $file);
|
|---|
| 98 | if (-d $thisfile) {
|
|---|
| 99 | my $colfiledir = &util::filename_cat ($coldir, $file);
|
|---|
| 100 | traverse_dir ($thisfile, $colfiledir);
|
|---|
| 101 |
|
|---|
| 102 | } else {
|
|---|
| 103 | my $destfile = $file;
|
|---|
| 104 | $destfile =~ s/^modelcol/$collection/;
|
|---|
| 105 | $destfile =~ s/^MODELCOL/$capcollection/;
|
|---|
| 106 | print STDOUT " doing replacements for $destfile\n" unless $quiet;
|
|---|
| 107 | $destfile = &util::filename_cat ($coldir, $destfile);
|
|---|
| 108 |
|
|---|
| 109 | open (INFILE, $thisfile) ||
|
|---|
| 110 | die "ERROR: Can't read file $thisfile";
|
|---|
| 111 | open (OUTFILE, ">$destfile") ||
|
|---|
| 112 | die "ERROR: Can't create file $destfile";
|
|---|
| 113 |
|
|---|
| 114 | while (defined ($line = <INFILE>)) {
|
|---|
| 115 | $line =~ s/\*\*collection\*\*/$collection/g;
|
|---|
| 116 | $line =~ s/\*\*COLLECTION\*\*/$capcollection/g;
|
|---|
| 117 | $line =~ s/\*\*creator\*\*/$creator/g;
|
|---|
| 118 | $line =~ s/\*\*maintainer\*\*/$maintainer/g;
|
|---|
| 119 | $line =~ s/\*\*public\*\*/$public/g;
|
|---|
| 120 | $line =~ s/\*\*title\*\*/$title/g;
|
|---|
| 121 | $line =~ s/\*\*about\*\*/$about/g;
|
|---|
| 122 | $line =~ s/\*\*plugins\*\*/$pluginstring/g;
|
|---|
| 123 | $line =~ s/\*\*buildtype\*\*/$buildtype/g;
|
|---|
| 124 | $line =~ s/\*\*searchtype\*\*/$searchtype/g;
|
|---|
| 125 | $line =~ s/\*\*indexes\*\*/$indexes/g;
|
|---|
| 126 | $line =~ s/\*\*defaultindex\*\*/$defaultindex/g;
|
|---|
| 127 | $line =~ s/\*\*indexmeta\*\*/$indexmeta/g;
|
|---|
| 128 | $line =~ s/\*\*xmlindexes\*\*/$xmlindexes/g;
|
|---|
| 129 | #$line =~ s/\*\*xmlplugins\*\*/$xmlpluginstring/g;
|
|---|
| 130 |
|
|---|
| 131 | print OUTFILE $line;
|
|---|
| 132 | }
|
|---|
| 133 |
|
|---|
| 134 | close (OUTFILE);
|
|---|
| 135 | close (INFILE);
|
|---|
| 136 | }
|
|---|
| 137 | }
|
|---|
| 138 | }
|
|---|
| 139 |
|
|---|
| 140 | # get and check options
|
|---|
| 141 | sub parse_args {
|
|---|
| 142 | my ($argref) = @_;
|
|---|
| 143 | if (!&parsargv::parse($argref,
|
|---|
| 144 | 'optionfile/.*/', \$optionfile,
|
|---|
| 145 | 'collectdir/.*/', \$collectdir,
|
|---|
| 146 | 'site/.*/', \$sitehome,
|
|---|
| 147 | 'creator/\w+\@[\w\.]+/', \$creator,
|
|---|
| 148 | 'maintainer/\w+\@[\w\.]+/', \$maintainer,
|
|---|
| 149 | 'public/true|false/true', \$public,
|
|---|
| 150 | 'buildtype/mg|mgpp/mgpp', \$buildtype,
|
|---|
| 151 | 'title/.+/', \$title,
|
|---|
| 152 | 'about/.+/', \$about,
|
|---|
| 153 | 'plugin/.+', \@plugin,
|
|---|
| 154 | 'quiet', \$quiet,
|
|---|
| 155 | )) {
|
|---|
| 156 | &print_usage();
|
|---|
| 157 | die "\n";
|
|---|
| 158 | }
|
|---|
| 159 | }
|
|---|
| 160 |
|
|---|
| 161 | sub main {
|
|---|
| 162 |
|
|---|
| 163 | &parse_args (\@ARGV);
|
|---|
| 164 | if ($optionfile =~ /\w/) {
|
|---|
| 165 | open (OPTIONS, $optionfile) || die "Couldn't open $optionfile\n";
|
|---|
| 166 | my $line = [];
|
|---|
| 167 | my $options = [];
|
|---|
| 168 | while (defined ($line = &cfgread::read_cfg_line ('mkcol::OPTIONS'))) {
|
|---|
| 169 | push (@$options, @$line);
|
|---|
| 170 | }
|
|---|
| 171 | close OPTIONS;
|
|---|
| 172 | &parse_args ($options);
|
|---|
| 173 |
|
|---|
| 174 | }
|
|---|
| 175 |
|
|---|
| 176 | # load default plugins if none were on command line
|
|---|
| 177 | if (!scalar(@plugin)) {
|
|---|
| 178 | @plugin = (ZIPPlug,GAPlug,TEXTPlug,HTMLPlug,EMAILPlug,
|
|---|
| 179 | PDFPlug,RTFPlug,WordPlug,PSPlug,ArcPlug,RecPlug);
|
|---|
| 180 | }
|
|---|
| 181 |
|
|---|
| 182 | # get and check the collection name
|
|---|
| 183 | ($collection) = @ARGV;
|
|---|
| 184 | if (!defined($collection)) {
|
|---|
| 185 | print STDOUT "ERROR: no collection name was specified\n";
|
|---|
| 186 | &print_usage();
|
|---|
| 187 | die "\n";
|
|---|
| 188 | }
|
|---|
| 189 |
|
|---|
| 190 | if (length($collection) > 8) {
|
|---|
| 191 | print STDOUT "ERROR: The collection name must be less than 8 characters\n";
|
|---|
| 192 | print STDOUT " so compatibility with earlier filesystems can be\n";
|
|---|
| 193 | print STDOUT " maintained.\n";
|
|---|
| 194 | die "\n";
|
|---|
| 195 | }
|
|---|
| 196 |
|
|---|
| 197 | if ($collection eq "gs2model" || $collection eq "gs3model") {
|
|---|
| 198 | print STDOUT "ERROR: No collection can be named gs2model or gs3model as these are the\n";
|
|---|
| 199 | print STDOUT " names of the model collections.\n";
|
|---|
| 200 | die "\n";
|
|---|
| 201 | }
|
|---|
| 202 |
|
|---|
| 203 | if ($collection eq "CVS") {
|
|---|
| 204 | print STDOUT "ERROR: No collection can be named CVS as this may interfere\n";
|
|---|
| 205 | print STDOUT " with directories created by the CVS versioning system\n";
|
|---|
| 206 | die "\n";
|
|---|
| 207 | }
|
|---|
| 208 |
|
|---|
| 209 | #check that -site has been specified
|
|---|
| 210 | if (!defined($sitehome) || $sitehome eq "") {
|
|---|
| 211 | print STDOUT "ERROR: The site was not defined. This variable is\n";
|
|---|
| 212 | print STDOUT " needed to locate the collect directory.\n";
|
|---|
| 213 | die "\n";
|
|---|
| 214 | }
|
|---|
| 215 | #check that its a valid directory
|
|---|
| 216 | if (!-d $sitehome) {
|
|---|
| 217 | print STDOUT "ERROR: $sitehome doesn't exist\n";
|
|---|
| 218 | die "\n";
|
|---|
| 219 | }
|
|---|
| 220 | if (!defined($creator) || $creator eq "") {
|
|---|
| 221 | print STDOUT "ERROR: The creator was not defined. This variable is\n";
|
|---|
| 222 | print STDOUT " needed to recognise duplicate collection names.\n";
|
|---|
| 223 | die "\n";
|
|---|
| 224 | }
|
|---|
| 225 |
|
|---|
| 226 | if (!defined($maintainer) || $maintainer eq "") {
|
|---|
| 227 | $maintainer = $creator;
|
|---|
| 228 | }
|
|---|
| 229 |
|
|---|
| 230 | $public = "true" unless defined $public;
|
|---|
| 231 | $buildtype = "mgpp" unless defined $buildtype;
|
|---|
| 232 |
|
|---|
| 233 | $searchtype = "";
|
|---|
| 234 | if ($buildtype eq "mgpp") {
|
|---|
| 235 | $searchtype = "searchtype plain form";
|
|---|
| 236 | }
|
|---|
| 237 |
|
|---|
| 238 | if (!defined($title) || $title eq "") {
|
|---|
| 239 | $title = $collection;
|
|---|
| 240 | }
|
|---|
| 241 |
|
|---|
| 242 | # get capitalised version of the collection
|
|---|
| 243 | $capcollection = $collection;
|
|---|
| 244 | $capcollection =~ tr/a-z/A-Z/;
|
|---|
| 245 |
|
|---|
| 246 | # get the strings to include.
|
|---|
| 247 | $pluginstring = "";
|
|---|
| 248 | foreach $plugin (@plugin) {
|
|---|
| 249 | if ($plugin eq RecPlug) {
|
|---|
| 250 | $pluginstring .= "plugin $plugin -use_metadata_files\n";
|
|---|
| 251 | } else {
|
|---|
| 252 | $pluginstring .= "plugin $plugin\n";
|
|---|
| 253 | }
|
|---|
| 254 | }
|
|---|
| 255 |
|
|---|
| 256 | $mdir = &util::filename_cat ($sitehome, "collect", "gs2model");
|
|---|
| 257 |
|
|---|
| 258 | if (defined $collectdir && $collectdir =~ /\w/) {
|
|---|
| 259 | if (!-d $collectdir) {
|
|---|
| 260 | print STDOUT "ERROR: $collectdir doesn't exist\n";
|
|---|
| 261 | die "\n";
|
|---|
| 262 | }
|
|---|
| 263 | $cdir = &util::filename_cat ($collectdir, $collection);
|
|---|
| 264 | } else {
|
|---|
| 265 | $cdir = &util::filename_cat ($sitehome, "collect", $collection);
|
|---|
| 266 | }
|
|---|
| 267 |
|
|---|
| 268 | # make sure the model collection exists
|
|---|
| 269 | die "ERROR: Cannot find the model collection $mdir" unless (-d $mdir);
|
|---|
| 270 |
|
|---|
| 271 | # make sure this collection does not already exist
|
|---|
| 272 | if (-e $cdir) {
|
|---|
| 273 | print STDOUT "ERROR: This collection already exists\n";
|
|---|
| 274 | die "\n";
|
|---|
| 275 | }
|
|---|
| 276 |
|
|---|
| 277 | # set up the default indexes - this could be a command line option at some stage
|
|---|
| 278 | # the names are added in here for the xml ones, but they should be added after building once the names are known.
|
|---|
| 279 | if ($buildtype eq "mg") {
|
|---|
| 280 | $indexes = "document:text document:Title document:Source";
|
|---|
| 281 | $defaultindex = "defaultindex document:text";
|
|---|
| 282 | $indexmeta = "collectionmeta .document:text \"text\"\ncollectionmeta .document:Title \"titles\"\ncollectionmeta .document:Source \"filenames\"\n";
|
|---|
| 283 | } elsif ($buildtype eq "mgpp") {
|
|---|
| 284 | $indexes = "allfields text metadata";
|
|---|
| 285 | $defaultindex = "";
|
|---|
| 286 | $indexmeta = "collectionmeta .text \"text\"\ncollectionmeta .allfields \"entire documents\"\n";
|
|---|
| 287 | } else {
|
|---|
| 288 | print STDOUT "Error: buildtype should be mg or mgpp, but its $buildtype\n";
|
|---|
| 289 | }
|
|---|
| 290 | # start creating the collection
|
|---|
| 291 | print STDOUT "\nCreating the collection $collection...\n" unless $quiet;
|
|---|
| 292 | &traverse_dir ($mdir, $cdir);
|
|---|
| 293 | print STDOUT "\nThe new collection was created successfully at\n" unless $quiet;
|
|---|
| 294 | print STDOUT "$cdir\n" unless $quiet;
|
|---|
| 295 | }
|
|---|
| 296 |
|
|---|
| 297 | &main ();
|
|---|
