source: main/trunk/greenstone2/perllib/colcfg.pm@ 21308

Last change on this file since 21308 was 20416, checked in by kjdon, 15 years ago

language_metadata was sometimes with an underscore, and sometimes without. Have standardised to no underscore: languagemetadata
  • Property svn:executable set to *
  • Property svn:keywords set to Author Date Id Revision
File size: 9.6 KB
Line 
1###########################################################################
2#
3# colcfg.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# reads/writes the collection configuration files:
27# collect.cfg/collectionConfig.xml and build.cfg/buildConfig.xml
28
29package colcfg;
30
31use cfgread;
32use collConfigxml;
33use buildConfigxml;
34use gsprintf 'gsprintf';
35use strict;
36
37# the collection configuration file data is stored in the form
38#
39# {'infodbtype'}->string
40# {'creator'}->string
41# {'public'}->string
42# {'defaultindex'}->string
43# {'importdir'}->string
44# {'archivedir'}->string
45# {'cachedir'}->string
46# {'builddir'}->string
47# {'removeold'}->string
48# {'textcompress'}->string
49# {'buildtype'}->string
50# {'maxnumeric'}->string
51# {'separate_cjk'}->string
52# {'sections_index_document_metadata'}->string (always, unless_section_metadata_exists)
53# {'languagemetadata'} -> string
54# {'maintainer'}->array of strings
55# {'languages'}->array of strings
56# {'indexsubcollections'}->array of strings
57# {'indexes'}->array of strings
58# {'indexoptions'}->array of strings (stem, casefold, accentfold)
59# {'dontbuild'}->array of strings
60# {'dontgdbm'}->array of strings
61# {'mirror'}->array of strings
62# {'phind'}->array of strings
63# {'plugout'}->array of strings
64# {'levels'}->array of strings (for mgpp eg Section, Paragraph)
65# {'searchtype'}->array of strings (for mgpp, form or plain)
66
67# {'subcollection'}->hash of key-value pairs
68
69# {'acquire'}->array of arrays of strings
70# {'plugin'}->array of arrays of strings
71# {'classify'}->array of arrays of strings
72
73# {'collectionmeta'}->hash of key->hash of param-value -used
74# for language specification
75# for example, collectionmeta->collectionname->default->demo
76# ->mi->maori demo
77
78# convenience method for reading in either collect.cfg/collectionConfig.xml
79sub read_collection_cfg {
80 my ($filename,$gs_mode) = @_;
81
82 my $collectcfg = undef;
83
84 if ($gs_mode eq "gs2") {
85 $collectcfg = &colcfg::read_collect_cfg ($filename);
86 } elsif ($gs_mode eq "gs3") {
87 $collectcfg = &colcfg::read_collection_cfg_xml ($filename);
88 }
89 else {
90 print STDERR "Failed to read collection configuration file\n";
91 print STDERR " Unrecognized mode: $gs_mode\n";
92 }
93
94 return $collectcfg;
95}
96
97# convenience method for writing out either collect.cfg/collectionConfig.xml
98# is this ever used??
99sub write_collection_cfg {
100 my ($filename, $collectcfg_data, $gs_mode) = @_;
101
102 if ($gs_mode eq "gs2") {
103 &colcfg::write_collect_cfg ($filename, $collectcfg_data );
104 } elsif ($gs_mode eq "gs3") {
105 &colcfg::write_collection_cfg_xml ($filename, $collectcfg_data);
106 }
107 else {
108 print STDERR "Failed to write collection configuration file\n";
109 print STDERR " Unrecognized mode: $gs_mode\n";
110 }
111}
112
113# the build configuration file data is stored in the form
114#
115# {'infodbtype'}->string
116# {'builddate'}->string
117# {'buildtype'}->string
118# {'metadata'}->array of strings
119# {'languages'}->array of strings
120# {'numdocs'}->string
121# {'numsections'}->string
122# {'numwords'}->string
123# {'numbytes'}->string
124# {'maxnumeric'}->string
125# {'indexfields'}->array of strings
126# {'indexfieldmap'}->array of strings in the form "field->FI"
127# {'indexmap'} -> array of strings
128# {'indexlevels'} -> array of strings
129# {'stemindexes'} -> string (int)
130# {'textlevel'}->string
131# {'levelmap'} -> array of strings in the form "level->shortname"
132
133# convenience method for reading in either build.cfg/buildConfig.xml
134sub read_building_cfg {
135 my ($filename,$gs_mode) = @_;
136
137 my $buildcfg = undef;
138
139 if ($gs_mode eq "gs2") {
140 $buildcfg = &colcfg::read_build_cfg ($filename);
141 } elsif ($gs_mode eq "gs3") {
142 $buildcfg = &colcfg::read_build_cfg_xml ($filename);
143 }
144 else {
145 print STDERR "Failed to read building configuration file\n";
146 print STDERR " Unrecognized mode: $gs_mode\n";
147 }
148
149 return $buildcfg;
150}
151
152# convenience method for writing out either build.cfg/buildConfig.xml
153# haven't got one, as gs3 version needs extra parameters
154#sub write_building_cfg {}
155
156##############################
157### gs2/gs3 specific methods
158###############################
159
160#####################################
161### collect.cfg/collectionConfig.xml
162#####################################
163
164# gs2 read in collect.cfg
165sub read_collect_cfg {
166 my ($filename) = @_;
167
168 return &cfgread::read_cfg_file ($filename,
169 q/^(infodbtype|creator|public|defaultindex|importdir|/ .
170 q/archivedir|cachedir|builddir|removeold|/ .
171 q/textcompress|buildtype|no_text|keepold|gzip|/ .
172 q/verbosity|remove_empty_classifications|OIDtype|OIDmetadata|/ .
173 q/groupsize|maxdocs|debug|mode|saveas|/ .
174 q/sortmeta|removesuffix|removeprefix|create_images|/ .
175 q/maxnumeric|languagemetadata|/ .
176 q/no_strip_html|index|sections_index_document_metadata|/ .
177 q/store_metadata_coverage)$/,
178 q/(maintainer|languages|indexsubcollections|/ .
179 q/indexes|indexoptions|dontbuild|dontgdbm|mirror|levels|plugout|/ .
180 q/searchtype|searchtypes)$/,
181 q/^(subcollection|format)$/,
182 q/^(acquire|plugin|classify)$/,
183 q/^(collectionmeta)$/);
184}
185
186# gs2 write out collect.cfg
187sub write_collect_cfg {
188 my ($filename, $data) = @_;
189
190 &cfgread::write_cfg_file($filename, $data,
191 q/^(infodbtype|creator|public|defaultindex|importdir|/ .
192 q/archivedir|cachedir|builddir|removeold|/ .
193 q/textcompress|buildtype|no_text|keepold|gzip|/ .
194 q/verbosity|remove_empty_classifications|OIDtype|OIDmetadata|/.
195 q/groupsize|maxdocs|debug|mode|saveas|/ .
196 q/sortmeta|removesuffix|removeprefix|create_images|/ .
197 q/maxnumeric|languagemetadata/ .
198 q/no_strip_html|index|sections_index_document_metadata)$/,
199 q/^(maintainer|languages|indexsubcollections|/ .
200 q/indexes|indexoptions|dontbuild|dontgdbm|mirror|levels|/.
201 q/searchtype|searchtypes)$/,
202 q/^(subcollection|format)$/,
203 q/^(acquire|plugin|classify)$/,
204 q/^(collectionmeta)$/);
205}
206
207# gs3 read in collectionConfig.xml
208sub read_collection_cfg_xml {
209 my ($filename) = @_;
210
211 return &collConfigxml::read_cfg_file ($filename);
212}
213
214# gs3 write out collectionConfig.xml
215sub write_collection_cfg_xml {
216
217}
218
219#####################################
220### build.cfg/buildConfig.xml
221######################################
222
223# gs2 read build.cfg
224sub read_build_cfg {
225 my ($filename) = @_;
226
227 return &cfgread::read_cfg_file ($filename,
228 q/^(infodbtype|builddate|buildtype|numdocs|numsections|numwords|numbytes|maxnumeric|textlevel|indexstem|stemindexes|separate_cjk)$/,
229 q/^(indexmap|subcollectionmap|languagemap|notbuilt|indexfields|indexfieldmap|indexlevels|levelmap)$/);
230
231}
232
233# gs2 write build.cfg
234sub write_build_cfg {
235 my ($filename, $data) = @_;
236
237 &cfgread::write_cfg_file($filename, $data,
238 q/^(infodbtype|builddate|buildtype|numdocs|numsections|numwords|numbytes|maxnumeric|textlevel|indexstem|stemindexes|separate_cjk)$/,
239 q/^(indexmap|subcollectionmap|languagemap|notbuilt|indexfields|indexfieldmap|indexlevels|levelmap)$/);
240}
241
242# gs3 read buildConfig.xml
243sub read_build_cfg_xml {
244
245 my ($filename) = @_;
246
247 return &buildConfigxml::read_cfg_file($filename);
248}
249
250# gs3 write buildConfig.xml
251sub write_build_cfg_xml {
252 my ($buildoutfile, $buildcfg, $collectcfg, $disable_OAI) = @_;
253
254 return &buildConfigxml::write_cfg_file ($buildoutfile, $buildcfg, $collectcfg, $disable_OAI);
255}
256
257
258# method to check for filename of collect.cfg, and gs mode.
259sub get_collect_cfg_name {
260 my ($out) = @_;
261
262 # First check if there's a
263 # gsdl/collect/COLLECTION/custom/COLLECTION/etc/custom.cfg file. This
264 # customization was added for DLC by Stefan, 30/6/2007.
265 my $configfilename = &util::filename_cat ($ENV{'GSDLCOLLECTDIR'}, "custom", $ENV{'GSDLCOLLECTION'}, "etc", "custom.cfg");
266
267 if (-e $configfilename) {
268 return ($configfilename, "gs2");
269 }
270
271 # Now check if there is a collect.cfg file in the usual place, i.e. it is gs2.
272 $configfilename = &util::filename_cat ($ENV{'GSDLCOLLECTDIR'}, "etc", "collect.cfg");
273 if (-e $configfilename) {
274 return ($configfilename, "gs2");
275 }
276
277 # If we get to here we check if there is a collectionConfig.xml file,
278 # i.e. it is gs3.
279 $configfilename = &util::filename_cat ($ENV{'GSDLCOLLECTDIR'}, "etc", "collectionConfig.xml");
280 if (-e $configfilename) {
281 return ($configfilename, "gs3");
282 }
283
284 # Error. No collection configuration file.
285 (&gsprintf($out, "{common.cannot_find_cfg_file}\n", $configfilename) && die);
286}
287
288
289sub use_collection {
290 my ($site, $collection, $collectdir) = @_;
291
292 if ((defined $site) && ($site ne ""))
293 {
294 return &util::use_site_collection($site, $collection, $collectdir);
295 }
296 else
297 {
298 return &util::use_collection($collection, $collectdir);
299 }
300}
301
302
3031;
304
305
Note: See TracBrowser for help on using the repository browser.