[15600] | 1 | ###########################################################################
|
---|
| 2 | #
|
---|
[20096] | 3 | # buildConfigxml.pm --
|
---|
[15600] | 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 | ###########################################################################
|
---|
[14741] | 25 |
|
---|
[20105] | 26 | # reads in buildConfig.xml
|
---|
| 27 | # Note, only implemented the bits that are currently used, eg by incremental
|
---|
| 28 | # build code.
|
---|
| 29 | # The resulting data is not a full representation on buildConfig.xml.
|
---|
[15600] | 30 |
|
---|
[20096] | 31 | package buildConfigxml;
|
---|
[20105] | 32 |
|
---|
[15600] | 33 | use strict;
|
---|
| 34 | no strict 'refs';
|
---|
| 35 | no strict 'subs';
|
---|
| 36 |
|
---|
| 37 | # Wrapper that ensures the right version of XML::Parser is loaded given
|
---|
| 38 | # the version of Perl being used. Need to distinguish between Perl 5.6 and
|
---|
| 39 | # Perl 5.8
|
---|
| 40 | sub BEGIN {
|
---|
| 41 | my $perl_dir;
|
---|
| 42 |
|
---|
| 43 | # Note: $] encodes the version number of perl
|
---|
[21822] | 44 | if ($]>=5.010) {
|
---|
| 45 | $perl_dir = "perl-5.10";
|
---|
| 46 | }
|
---|
| 47 | elsif ($]>5.008) {
|
---|
[15600] | 48 | # perl 5.8.1 or above
|
---|
| 49 | $perl_dir = "perl-5.8";
|
---|
| 50 | }
|
---|
| 51 | elsif ($]<5.008) {
|
---|
| 52 | # assume perl 5.6
|
---|
| 53 | $perl_dir = "perl-5.6";
|
---|
| 54 | }
|
---|
| 55 | else {
|
---|
| 56 | print STDERR "Warning: Perl 5.8.0 is not a maintained release.\n";
|
---|
| 57 | print STDERR " Please upgrade to a newer version of Perl.\n";
|
---|
| 58 | $perl_dir = "perl-5.8";
|
---|
| 59 | }
|
---|
| 60 |
|
---|
| 61 | if ($ENV{'GSDLOS'} !~ /^windows$/i) {
|
---|
| 62 | # Use push to put this on the end, so an existing XML::Parser will be used by default
|
---|
| 63 | push (@INC, "$ENV{'GSDLHOME'}/perllib/cpan/$perl_dir");
|
---|
| 64 | }
|
---|
| 65 | }
|
---|
| 66 |
|
---|
| 67 | use XML::Parser;
|
---|
| 68 |
|
---|
[20105] | 69 |
|
---|
[17895] | 70 | # A mapping hash to resolve name discrepancy between gs2 and gs3.
|
---|
[20105] | 71 | my $nameMap = {"numDocs" => "numdocs",
|
---|
[19898] | 72 | "buildType" => "buildtype"
|
---|
[15600] | 73 | };
|
---|
[20105] | 74 |
|
---|
| 75 |
|
---|
[15600] | 76 | # A hash structure which is returned by sub read_cfg_file.
|
---|
| 77 | my $data = {};
|
---|
| 78 |
|
---|
| 79 | # use those unique attribute values to locate the text within the elements
|
---|
| 80 | my $currentLocation = "";
|
---|
[20105] | 81 | my $stringexp = q/^(buildType|numDocs)$/;
|
---|
[15619] | 82 |
|
---|
[20105] | 83 | my $indexmap_name = "";
|
---|
| 84 | my $haveindexfields = 0;
|
---|
[15600] | 85 |
|
---|
[20105] | 86 | # Reads in the model collection configuration file, collectionConfig.xml,
|
---|
| 87 | # into a structure which complies with the one used by gs2 (i.e. one read
|
---|
| 88 | # in by &cfgread::read_cfg_file).
|
---|
| 89 | sub read_cfg_file {
|
---|
| 90 | my ($filename) = @_;
|
---|
| 91 | $data = {};
|
---|
| 92 | if ($filename !~ /buildConfig\.xml$/ || !-f $filename) {
|
---|
| 93 | return undef;
|
---|
| 94 | }
|
---|
[15600] | 95 |
|
---|
[20105] | 96 | # create XML::Parser object for parsing metadata.xml files
|
---|
| 97 | my $parser;
|
---|
| 98 | if ($]<5.008) {
|
---|
| 99 | # Perl 5.6
|
---|
| 100 | $parser = new XML::Parser('Style' => 'Stream',
|
---|
| 101 | 'Handlers' => {'Char' => \&Char,
|
---|
| 102 | 'Doctype' => \&Doctype
|
---|
| 103 | });
|
---|
| 104 | }
|
---|
| 105 | else {
|
---|
| 106 | # Perl 5.8
|
---|
| 107 | $parser = new XML::Parser('Style' => 'Stream',
|
---|
| 108 | 'ProtocolEncoding' => 'ISO-8859-1',
|
---|
| 109 | 'Handlers' => {'Char' => \&Char,
|
---|
| 110 | 'Doctype' => \&Doctype
|
---|
| 111 | });
|
---|
| 112 | }
|
---|
[15600] | 113 |
|
---|
[20105] | 114 | if (!open (COLCFG, $filename)) {
|
---|
| 115 | print STDERR "buildConfigxml::read_cfg_file couldn't read the cfg file $filename\n";
|
---|
| 116 | } else {
|
---|
[15600] | 117 |
|
---|
[20105] | 118 | $parser->parsefile ($filename);# (COLCFG);
|
---|
| 119 | close (COLCFG);
|
---|
| 120 | }
|
---|
| 121 |
|
---|
| 122 | #&Display;
|
---|
| 123 | return $data;
|
---|
| 124 | }
|
---|
| 125 |
|
---|
[15600] | 126 | sub StartTag {
|
---|
| 127 | # Those marked with #@ will not be executed at the same time when this sub is being called
|
---|
| 128 | # so that if/elsif is used to avoid unnecessary tests
|
---|
| 129 | my ($expat, $element) = @_;
|
---|
[15619] | 130 |
|
---|
[15600] | 131 | my $name = $_{'name'};
|
---|
[20105] | 132 | my $shortname = $_{'shortname'};
|
---|
[15600] | 133 |
|
---|
[20105] | 134 |
|
---|
[15600] | 135 | #@ handling block metadataList
|
---|
[20105] | 136 | if (defined $name and $name =~ /$stringexp/){
|
---|
[15600] | 137 | $currentLocation = $name;
|
---|
[20105] | 138 | # the value will be retrieved later in Text sub
|
---|
[15600] | 139 | }
|
---|
| 140 |
|
---|
[20105] | 141 | #@ handle indexes - store indexmap (mg) or indexfields and indexfieldmap (mgpp/lucene)
|
---|
| 142 | elsif ($element =~ /^indexList$/) {
|
---|
| 143 | # set up the data arrays
|
---|
| 144 | # this assumes that the build type has been read already, which is
|
---|
| 145 | # currently the order we save the file in.
|
---|
| 146 | if ($data->{'buildtype'} eq "mg") {
|
---|
| 147 | $indexmap_name = "indexmap";
|
---|
| 148 | if (!defined $data->{"indexmap"}) {
|
---|
| 149 | $data->{"indexmap"} = [];
|
---|
| 150 | }
|
---|
[15619] | 151 | }
|
---|
[20105] | 152 | else {
|
---|
| 153 | $indexmap_name = "indexfieldmap";
|
---|
| 154 | $haveindexfields = 1;
|
---|
| 155 | if (!defined $data->{"indexfieldmap"}) {
|
---|
| 156 | $data->{"indexfieldmap"} = [];
|
---|
| 157 | }
|
---|
| 158 | if (!defined $data->{"indexfields"}) {
|
---|
| 159 | $data->{"indexfields"} = [];
|
---|
| 160 | }
|
---|
[17895] | 161 |
|
---|
[20105] | 162 | }
|
---|
| 163 |
|
---|
[17895] | 164 | }
|
---|
[15619] | 165 |
|
---|
[20105] | 166 | elsif ($element =~ /index/) {
|
---|
| 167 | # store each index in the map
|
---|
| 168 | if (defined $name && defined $shortname) {
|
---|
| 169 | push @{$data->{$indexmap_name}}, "$name->$shortname";
|
---|
| 170 | if ($haveindexfields) {
|
---|
| 171 | push @{$data->{'indexfields'}}, $name;
|
---|
| 172 | }
|
---|
[15600] | 173 | }
|
---|
| 174 | }
|
---|
| 175 |
|
---|
| 176 |
|
---|
| 177 | }
|
---|
| 178 |
|
---|
| 179 | sub EndTag {
|
---|
| 180 | my ($expat, $element) = @_;
|
---|
| 181 | }
|
---|
| 182 |
|
---|
| 183 | sub Text {
|
---|
[15619] | 184 | if (defined $currentLocation) {
|
---|
[20105] | 185 | #@ Handling block metadataList(numDocs, buildType)
|
---|
[15619] | 186 | if($currentLocation =~ /$stringexp/){
|
---|
| 187 | #print $currentLocation;
|
---|
| 188 | my $key = $nameMap->{$currentLocation};
|
---|
| 189 | $data->{$key} = $_;
|
---|
| 190 | undef $currentLocation;
|
---|
[20105] | 191 | }
|
---|
[15619] | 192 | }
|
---|
[15600] | 193 | }
|
---|
[15619] | 194 |
|
---|
[15600] | 195 | # This sub is for debugging purposes
|
---|
| 196 | sub Display {
|
---|
[20105] | 197 |
|
---|
| 198 | print "NumDocs = ".$data->{'numdocs'}."\n" if (defined $data->{'numdocs'});
|
---|
| 199 | print "BuildType = ".$data->{'buildtype'}."\n" if (defined $data->{'buildtype'});
|
---|
| 200 | print "IndexMap = ". join(" ",@{$data->{'indexmap'}})."\n" if (defined $data->{'indexmap'});
|
---|
| 201 | print "IndexFieldMap = ". join(" ",@{$data->{'indexfieldmap'}})."\n" if (defined $data->{'indexfieldmap'});
|
---|
| 202 | print "IndexFields = ". join(" ",@{$data->{'indexfields'}})."\n" if (defined $data->{'indexfields'});
|
---|
| 203 |
|
---|
[15600] | 204 | }
|
---|
[20105] | 205 |
|
---|
| 206 | # is this actually used??
|
---|
[15600] | 207 | sub Doctype {
|
---|
| 208 | my ($expat, $name, $sysid, $pubid, $internal) = @_;
|
---|
| 209 |
|
---|
[20105] | 210 | die if ($name !~ /^buildConfig$/);
|
---|
[15600] | 211 | }
|
---|
| 212 |
|
---|
| 213 | # This Char function overrides the one in XML::Parser::Stream to overcome a
|
---|
| 214 | # problem where $expat->{Text} is treated as the return value, slowing
|
---|
| 215 | # things down significantly in some cases.
|
---|
| 216 | sub Char {
|
---|
| 217 | if ($]<5.008) {
|
---|
| 218 | use bytes; # Necessary to prevent encoding issues with XML::Parser 2.31+ and Perl 5.6
|
---|
| 219 | }
|
---|
| 220 | $_[0]->{'Text'} .= $_[1];
|
---|
| 221 | return undef;
|
---|
| 222 | }
|
---|
[15619] | 223 |
|
---|
[15600] | 224 |
|
---|
| 225 |
|
---|
| 226 | sub write_line {
|
---|
| 227 | my ($filehandle, $line) = @_;
|
---|
| 228 | print $filehandle join ("", @$line), "\n";
|
---|
| 229 | }
|
---|
| 230 |
|
---|
| 231 | # Create the buildConfig.xml file for a specific collection
|
---|
| 232 | sub write_cfg_file {
|
---|
| 233 | # this sub is called in make_auxiliary_files() in basebuilder.pm
|
---|
| 234 | # the received args: $buildoutfile - destination file: buildConfig.xml
|
---|
[21785] | 235 | # $buildcfg - all build options,
|
---|
[20096] | 236 | # $collectcfg - contents of collectionConfig.xml read in by read_cfg_file sub in buildConfigxml.pm.
|
---|
[21785] | 237 | my ($buildoutfile, $buildcfg, $collectcfg) = @_;
|
---|
[15600] | 238 | my $line = [];
|
---|
| 239 |
|
---|
| 240 | if (!open (COLCFG, ">$buildoutfile")) {
|
---|
[20096] | 241 | print STDERR "buildConfigxml::write_cfg_file couldn't write the build config file $buildoutfile\n";
|
---|
[15600] | 242 | die;
|
---|
| 243 | }
|
---|
| 244 |
|
---|
| 245 | &write_line('COLCFG', ["<buildConfig xmlns:gsf=\"http://www.greenstone.org/greenstone3/schema/ConfigFormat\">"]);
|
---|
| 246 |
|
---|
| 247 | # output building metadata to build config file
|
---|
| 248 | my $buildtype;
|
---|
| 249 | if (defined $buildcfg->{"buildtype"}) {
|
---|
| 250 | $buildtype = $buildcfg->{"buildtype"};
|
---|
| 251 | } else {
|
---|
| 252 | $buildtype = "mgpp";
|
---|
| 253 | }
|
---|
| 254 | my $numdocs;
|
---|
| 255 | if (defined $buildcfg->{"numdocs"}) {
|
---|
| 256 | $numdocs = $buildcfg->{"numdocs"};
|
---|
| 257 | }
|
---|
| 258 | &write_line('COLCFG', ["<metadataList>"]);
|
---|
| 259 | &write_line('COLCFG', ["<metadata name=\"numDocs\">", $numdocs, "</metadata>"]);
|
---|
| 260 | &write_line('COLCFG', ["<metadata name=\"buildType\">", $buildtype, "</metadata>"]);
|
---|
[21783] | 261 | if (defined $buildcfg->{'indexstem'}) {
|
---|
| 262 | &write_line('COLCFG', ["<metadata name=\"indexStem\">", $buildcfg->{"indexstem"}, "</metadata>"]);
|
---|
| 263 | }
|
---|
| 264 | if (defined $buildcfg->{'infodbtype'}) {
|
---|
| 265 | &write_line('COLCFG', ["<metadata name=\"infodbType\">", $buildcfg->{"infodbtype"}, "</metadata>"]);
|
---|
| 266 | }
|
---|
[15600] | 267 | &write_line('COLCFG', ["</metadataList>"]);
|
---|
| 268 |
|
---|
| 269 | my $service_type = "MGPP";
|
---|
| 270 | if ($buildtype eq "mg") {
|
---|
| 271 | $service_type = "MG";
|
---|
| 272 | } elsif ($buildtype eq "lucene") {
|
---|
| 273 | $service_type = "Lucene";
|
---|
| 274 | }
|
---|
| 275 |
|
---|
| 276 | # output serviceRackList
|
---|
| 277 | &write_line('COLCFG', ["<serviceRackList>"]);
|
---|
| 278 |
|
---|
| 279 | # do the search service
|
---|
| 280 | &write_line('COLCFG', ["<serviceRack name=\"GS2", $service_type, "Search\">"]);
|
---|
| 281 | if (defined $buildcfg->{'indexstem'}) {
|
---|
| 282 | my $indexstem = $buildcfg->{'indexstem'};
|
---|
| 283 | &write_line('COLCFG', ["<indexStem name=\"", $indexstem, "\" />"]);
|
---|
| 284 | }
|
---|
[21439] | 285 | if (defined $buildcfg->{'infodbtype'}) {
|
---|
| 286 | my $infodbtype = $buildcfg->{'infodbtype'};
|
---|
| 287 | &write_line('COLCFG', ["<databaseType name=\"", $infodbtype, "\" />"]);
|
---|
| 288 | }
|
---|
[15600] | 289 |
|
---|
| 290 | #indexes
|
---|
| 291 | # maps index name to shortname
|
---|
| 292 | my $indexmap = {};
|
---|
| 293 | # keeps the order for indexes
|
---|
| 294 | my @indexlist = ();
|
---|
| 295 |
|
---|
| 296 | my $defaultindex = "";
|
---|
| 297 | my $first = 1;
|
---|
| 298 | my $maptype = "indexfieldmap";
|
---|
| 299 | if ($buildtype eq "mg") {
|
---|
| 300 | $maptype = "indexmap";
|
---|
| 301 | }
|
---|
| 302 |
|
---|
| 303 | #map {print $_."\n"} keys %$buildcfg;
|
---|
| 304 |
|
---|
| 305 | if (defined $buildcfg->{$maptype}) {
|
---|
| 306 | my $indexmap_t = $buildcfg->{$maptype};
|
---|
| 307 | foreach my $i (@$indexmap_t) {
|
---|
| 308 | my ($k, $v) = $i =~ /^(.*)\-\>(.*)$/;
|
---|
| 309 | $indexmap->{$k} = $v;
|
---|
| 310 | push @indexlist, $k;
|
---|
| 311 | if ($first) {
|
---|
| 312 | $defaultindex = $v;
|
---|
| 313 | $first = 0;
|
---|
| 314 | }
|
---|
| 315 | }
|
---|
| 316 | # now if the user has assigned a default index, we use it
|
---|
| 317 | if (defined $collectcfg->{"defaultindex"}) {
|
---|
| 318 | $defaultindex = $indexmap->{$collectcfg->{"defaultindex"}};
|
---|
| 319 | }
|
---|
| 320 |
|
---|
| 321 | } else {
|
---|
| 322 | print STDERR "$maptype not defined";
|
---|
| 323 | }
|
---|
| 324 | #for each index in indexList, write them out
|
---|
| 325 | &write_line('COLCFG', ["<indexList>"]);
|
---|
| 326 | foreach my $i (@indexlist) {
|
---|
| 327 | my $index = $indexmap->{$i};
|
---|
| 328 | &write_line('COLCFG', ["<index name=\"", $i, "\" ", "shortname=\"", $index, "\" />"]);
|
---|
| 329 | }
|
---|
| 330 | &write_line('COLCFG', ["</indexList>"]);
|
---|
| 331 |
|
---|
[20065] | 332 |
|
---|
[22485] | 333 | #$defaultindex = "ZZ" if (!$defaultindex); # index allfields by default
|
---|
| 334 | if($defaultindex) {
|
---|
| 335 | &write_line('COLCFG', ["<defaultIndex shortname=\"", $defaultindex, "\" />"]);
|
---|
| 336 | }
|
---|
[15600] | 337 |
|
---|
[20065] | 338 |
|
---|
[15600] | 339 | # do indexOptionList
|
---|
| 340 | if ($buildtype eq "mg" || $buildtype eq "mgpp") {
|
---|
| 341 | &write_line('COLCFG', ["<indexOptionList>"]);
|
---|
| 342 | my $stemindexes = 3; # default is stem and casefold
|
---|
| 343 | if (defined $buildcfg->{'stemindexes'} && $buildcfg->{'stemindexes'} =~ /^\d+$/ ) {
|
---|
| 344 | $stemindexes = $buildcfg->{'stemindexes'};
|
---|
| 345 | }
|
---|
| 346 | &write_line('COLCFG', ["<indexOption name=\"stemIndexes\" value=\"", $stemindexes, "\" />"]);
|
---|
| 347 |
|
---|
| 348 | my $maxnumeric = 4; # default
|
---|
| 349 | if (defined $buildcfg->{'maxnumeric'} && $buildcfg->{'maxnumeric'} =~ /^\d+$/) {
|
---|
| 350 | $maxnumeric = $buildcfg->{'maxnumeric'};
|
---|
| 351 | }
|
---|
| 352 | &write_line('COLCFG', ["<indexOption name=\"maxnumeric\" value=\"", $maxnumeric, "\" />"]);
|
---|
| 353 | &write_line('COLCFG', ["</indexOptionList>"]);
|
---|
| 354 | }
|
---|
| 355 |
|
---|
| 356 | # levelList
|
---|
| 357 | my $levelmap = {};
|
---|
| 358 | my @levellist = ();
|
---|
| 359 | my $default_search_level = "Doc";
|
---|
| 360 | my $default_retrieve_level = "Doc";
|
---|
[15685] | 361 | my $default_db_level = "Doc";
|
---|
[15600] | 362 | $first = 1;
|
---|
| 363 | if ($buildtype eq "mgpp" || $buildtype eq "lucene") {
|
---|
| 364 | if (defined $buildcfg->{'levelmap'}) {
|
---|
| 365 | my $levelmap_t = $buildcfg->{'levelmap'};
|
---|
| 366 | foreach my $l (@$levelmap_t) {
|
---|
| 367 | my ($key, $val) = $l =~ /^(.*)\-\>(.*)$/;
|
---|
| 368 | $levelmap->{$key} = $val;
|
---|
| 369 | push @levellist, $key;
|
---|
| 370 | if ($first) {
|
---|
| 371 | # let default search level follow the first level in the level list
|
---|
| 372 | $default_search_level = $val;
|
---|
[15685] | 373 | # retrieve/database levels may get modified later if text level is defined
|
---|
[15600] | 374 | $default_retrieve_level = $val;
|
---|
[15685] | 375 | $default_db_level = $val;
|
---|
[15600] | 376 | $first = 0;
|
---|
| 377 | }
|
---|
| 378 | }
|
---|
| 379 | }
|
---|
| 380 | # the default level assigned by the user is no longer ignored [Shaoqun], but the retrievel level stays the same.
|
---|
| 381 | #if (defined $collectcfg->{"defaultlevel"}) {
|
---|
| 382 | $default_search_level = $levelmap->{$collectcfg->{"defaultlevel"}};
|
---|
| 383 | # $default_retrieve_level = $default_search_level;
|
---|
| 384 | #}
|
---|
| 385 |
|
---|
| 386 | if (defined $buildcfg->{'textlevel'}) {
|
---|
[15685] | 387 | # let the retrieve/database levels always follow the textlevel
|
---|
[15600] | 388 | $default_retrieve_level = $buildcfg->{'textlevel'};
|
---|
[15685] | 389 | $default_db_level = $buildcfg->{'textlevel'};
|
---|
[15600] | 390 |
|
---|
| 391 | }
|
---|
| 392 | }
|
---|
| 393 | #for each level in levelList, write them out
|
---|
[21783] | 394 | if ($buildtype ne "mg") {
|
---|
[15600] | 395 | &write_line('COLCFG', ["<levelList>"]);
|
---|
| 396 | foreach my $lv (@levellist) {
|
---|
| 397 | my $level = $levelmap->{$lv};
|
---|
| 398 | &write_line('COLCFG', ["<level name=\"", $lv, "\" shortname=\"", $level, "\" />"]);
|
---|
| 399 | }
|
---|
| 400 | &write_line('COLCFG', ["</levelList>"]);
|
---|
| 401 | }
|
---|
| 402 | # add in defaultLevel as the same level as indexLevelList, making the reading job easier
|
---|
| 403 | if ($buildtype eq "lucene" || $buildtype eq "mgpp") {
|
---|
| 404 | &write_line('COLCFG', ["<defaultLevel shortname=\"", $default_search_level, "\" />"]);
|
---|
| 405 | }
|
---|
| 406 | if ($buildtype eq "lucene" || $buildtype eq "mgpp") {
|
---|
[15685] | 407 | &write_line('COLCFG', ["<defaultDBLevel shortname=\"", $default_db_level, "\" />"]);
|
---|
[15600] | 408 | }
|
---|
| 409 | # do searchTypeList
|
---|
| 410 | if ($buildtype eq "mgpp" || $buildtype eq "lucene") {
|
---|
| 411 | &write_line('COLCFG', ["<searchTypeList>"]);
|
---|
| 412 |
|
---|
| 413 | if (defined $buildcfg->{"searchtype"}) {
|
---|
| 414 | my $searchtype_t = $buildcfg->{"searchtype"};
|
---|
| 415 | foreach my $s (@$searchtype_t) {
|
---|
| 416 | &write_line('COLCFG', ["<searchType name=\"", $s, "\" />"]);
|
---|
| 417 | }
|
---|
| 418 | } else {
|
---|
| 419 | &write_line('COLCFG', ["<searchType name=\"plain\" />"]);
|
---|
| 420 | &write_line('COLCFG', ["<searchType name=\"form\" />"]);
|
---|
| 421 | }
|
---|
| 422 | &write_line('COLCFG', ["</searchTypeList>"]);
|
---|
| 423 | }
|
---|
| 424 |
|
---|
| 425 | # do indexLanguageList [in collect.cfg: languages; in build.cfg: languagemap]
|
---|
| 426 | $first = 1;
|
---|
| 427 | my $default_lang = "";
|
---|
| 428 | my $default_lang_short = "";
|
---|
| 429 | if (defined $buildcfg->{"languagemap"}) {
|
---|
| 430 | &write_line('COLCFG', ["<indexLanguageList>"]);
|
---|
| 431 |
|
---|
| 432 | my $langmap_t = $buildcfg->{"languagemap"};
|
---|
| 433 | foreach my $l (@$langmap_t) {
|
---|
| 434 | my ($k, $v) = $l =~ /^(.*)\-\>(.*)$/;
|
---|
| 435 |
|
---|
| 436 | &write_line('COLCFG', ["<indexLanguage name=\"", $k, "\" shortname=\"", $v, "\" />"]);
|
---|
| 437 | if ($first) {
|
---|
| 438 | $default_lang = $k; #name
|
---|
| 439 | $default_lang_short = $v; #shortname
|
---|
| 440 | $first = 0;
|
---|
| 441 | }
|
---|
| 442 | }
|
---|
| 443 |
|
---|
| 444 | &write_line('COLCFG', ["</indexLanguageList>"]);
|
---|
| 445 | # now if the user has assigned a default language (as "en", "ru" etc.)
|
---|
| 446 | if (defined $collectcfg->{"defaultlanguage"}) {
|
---|
| 447 | $default_lang = $collectcfg->{"defaultlanguage"};
|
---|
| 448 | }
|
---|
| 449 | &write_line('COLCFG', ["<defaultIndexLanguage name=\"", $default_lang, "\" shortname=\"", $default_lang_short, "\" />"]);
|
---|
| 450 | }
|
---|
| 451 |
|
---|
| 452 |
|
---|
| 453 | # do indexSubcollectionList
|
---|
| 454 | my $default_subcol = "";# make it in sub scope to be used in the concatenation
|
---|
| 455 | if (defined $buildcfg->{'subcollectionmap'}) {
|
---|
| 456 | &write_line('COLCFG', ["<indexSubcollectionList>"]);
|
---|
| 457 | my $subcolmap = {};
|
---|
| 458 | my @subcollist = ();
|
---|
| 459 | $first = 1;
|
---|
| 460 | my $subcolmap_t = $buildcfg->{'subcollectionmap'};
|
---|
| 461 | foreach my $l (@$subcolmap_t) {
|
---|
| 462 | my ($k, $v) = $l =~ /^(.*)\-\>(.*)$/;
|
---|
| 463 | $subcolmap->{$k} = $v;
|
---|
| 464 | push @subcollist, $k;
|
---|
| 465 | if ($first) {
|
---|
| 466 | $default_subcol = $v;
|
---|
| 467 | $first = 0;
|
---|
| 468 | }
|
---|
| 469 | }
|
---|
| 470 | foreach my $sl (@subcollist) {
|
---|
| 471 | my $subcol = $subcolmap->{$sl};
|
---|
| 472 | &write_line('COLCFG', ["<indexSubcollection name=\"", $sl, "\" shortname=\"", $subcol, "\" />"]);
|
---|
| 473 | }
|
---|
| 474 |
|
---|
| 475 | &write_line('COLCFG', ["</indexSubcollectionList>"]);
|
---|
| 476 | &write_line('COLCFG', ["<defaultIndexSubcollection shortname=\"", $default_subcol, "\" />"]);
|
---|
| 477 | }
|
---|
| 478 |
|
---|
| 479 | # close off search service
|
---|
| 480 | &write_line('COLCFG', ["</serviceRack>"]);
|
---|
| 481 |
|
---|
| 482 | # do the retrieve service
|
---|
| 483 | &write_line('COLCFG', ["<serviceRack name=\"GS2", $service_type, "Retrieve\">"]);
|
---|
| 484 |
|
---|
| 485 | # do default index
|
---|
| 486 | if (defined $buildcfg->{"languagemap"}) {
|
---|
| 487 | &write_line('COLCFG', ["<defaultIndexLanguage shortname=\"", $default_lang, "\" />"]);
|
---|
| 488 | }
|
---|
| 489 | if (defined $buildcfg->{'subcollectionmap'}) {
|
---|
| 490 | &write_line('COLCFG', ["<defaultIndexSubcollection shortname=\"", $default_subcol, "\" />"]);
|
---|
| 491 | }
|
---|
| 492 | if ($buildtype eq "mg") {
|
---|
| 493 | &write_line('COLCFG', ["<defaultIndex shortname=\"", $defaultindex, "\" />"]);
|
---|
| 494 | }
|
---|
| 495 |
|
---|
| 496 | if (defined $buildcfg->{'indexstem'}) {
|
---|
| 497 | my $indexstem = $buildcfg->{'indexstem'};
|
---|
| 498 | &write_line('COLCFG', ["<indexStem name=\"", $indexstem, "\" />"]);
|
---|
| 499 | }
|
---|
| 500 | if ($buildtype eq "mgpp" || $buildtype eq "lucene") {
|
---|
| 501 | &write_line('COLCFG', ["<defaultLevel shortname=\"", $default_retrieve_level, "\" />"]);
|
---|
| 502 | }
|
---|
[21439] | 503 | if (defined $buildcfg->{'infodbtype'}) {
|
---|
| 504 | my $infodbtype = $buildcfg->{'infodbtype'};
|
---|
| 505 | &write_line('COLCFG', ["<databaseType name=\"", $infodbtype, "\" />"]);
|
---|
| 506 | }
|
---|
| 507 |
|
---|
[15600] | 508 | &write_line('COLCFG', ["</serviceRack>"]);
|
---|
| 509 |
|
---|
| 510 | # do the browse service
|
---|
| 511 | my $count = 1;
|
---|
| 512 | my $phind = 0;
|
---|
| 513 | my $started_classifiers = 0;
|
---|
| 514 |
|
---|
| 515 | my $classifiers = $collectcfg->{"classify"};
|
---|
| 516 | foreach my $cl (@$classifiers) {
|
---|
| 517 | my $name = "CL$count";
|
---|
| 518 | $count++;
|
---|
| 519 | my ($classname) = @$cl[0];
|
---|
| 520 | if ($classname =~ /^phind$/i) {
|
---|
| 521 | $phind=1;
|
---|
| 522 | #should add it into coll config classifiers
|
---|
| 523 | next;
|
---|
| 524 | }
|
---|
| 525 |
|
---|
| 526 | if (not $started_classifiers) {
|
---|
| 527 | &write_line('COLCFG', ["<serviceRack name=\"GS2Browse\">"]);
|
---|
| 528 | if (defined $buildcfg->{'indexstem'}) {
|
---|
| 529 | my $indexstem = $buildcfg->{'indexstem'};
|
---|
| 530 | &write_line('COLCFG', ["<indexStem name=\"", $indexstem, "\" />"]);
|
---|
| 531 | }
|
---|
[21439] | 532 | if (defined $buildcfg->{'infodbtype'}) {
|
---|
| 533 | my $infodbtype = $buildcfg->{'infodbtype'};
|
---|
| 534 | &write_line('COLCFG', ["<databaseType name=\"", $infodbtype, "\" />"]);
|
---|
| 535 | }
|
---|
[15600] | 536 | &write_line('COLCFG', ["<classifierList>"]);
|
---|
| 537 | $started_classifiers = 1;
|
---|
| 538 | }
|
---|
| 539 | my $content = ''; #use buttonname first, then metadata
|
---|
| 540 | if ($classname eq "DateList") {
|
---|
| 541 | $content = "Date";
|
---|
| 542 | } else {
|
---|
| 543 | for (my $j=0; $j<scalar(@$cl); $j++) {
|
---|
| 544 | my $arg = @$cl[$j];
|
---|
| 545 | if ($arg eq "-buttonname"){
|
---|
| 546 | $content = @$cl[$j+1];
|
---|
| 547 | last;
|
---|
| 548 | } elsif ($arg eq "-metadata") {
|
---|
| 549 | $content = @$cl[$j+1];
|
---|
| 550 | }
|
---|
| 551 |
|
---|
| 552 | }
|
---|
| 553 | }
|
---|
| 554 | &write_line('COLCFG', ["<classifier name=\"", $name, "\" content=\"", $content, "\" />"]);
|
---|
| 555 | }
|
---|
| 556 | if ($started_classifiers) {
|
---|
| 557 | # end the classifiers
|
---|
| 558 | &write_line('COLCFG', ["</classifierList>"]);
|
---|
| 559 | # close off the Browse service
|
---|
| 560 | &write_line('COLCFG', ["</serviceRack>"]);
|
---|
| 561 | }
|
---|
| 562 |
|
---|
| 563 | # the phind classifier is a separate service
|
---|
| 564 | if ($phind) {
|
---|
| 565 | # if phind classifier
|
---|
| 566 | &write_line('COLCFG', ["<serviceRack name=\"PhindPhraseBrowse\" />"]);
|
---|
| 567 | }
|
---|
| 568 |
|
---|
| 569 |
|
---|
| 570 | &write_line('COLCFG', ["</serviceRackList>"]);
|
---|
| 571 | &write_line('COLCFG', ["</buildConfig>"]);
|
---|
| 572 |
|
---|
| 573 | close (COLCFG);
|
---|
| 574 | }
|
---|
| 575 |
|
---|
| 576 |
|
---|
| 577 | #########################################################
|
---|
| 578 |
|
---|
| 579 | 1;
|
---|