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 |
|
---|
29 | package colcfg;
|
---|
30 |
|
---|
31 | use cfgread;
|
---|
32 | use gsprintf 'gsprintf';
|
---|
33 | use util;
|
---|
34 | use FileUtils;
|
---|
35 |
|
---|
36 | use strict;
|
---|
37 |
|
---|
38 | # the collection configuration file data is stored in the form
|
---|
39 | #
|
---|
40 | # {'infodbtype'}->string
|
---|
41 | # {'creator'}->string
|
---|
42 | # {'public'}->string
|
---|
43 | # {'complexmeta'}->string (true, false)
|
---|
44 | # {'defaultindex'}->string
|
---|
45 | # {'importdir'}->string
|
---|
46 | # {'archivedir'}->string
|
---|
47 | # {'cachedir'}->string
|
---|
48 | # {'builddir'}->string
|
---|
49 | # {'removeold'}->string
|
---|
50 | # {'textcompress'}->string
|
---|
51 | # {'buildtype'}->string
|
---|
52 | # {'orthogonalbuildtypes'}->array of strings
|
---|
53 | # {'maxnumeric'}->string
|
---|
54 | # {'separate_cjk'}->string
|
---|
55 | # {'sections_index_document_metadata'}->string (never, always, unless_section_metadata_exists)
|
---|
56 | # {'sections_sort_on_document_metadata'}->string (never, always, unless_section_metadata_exists)
|
---|
57 | # {'languagemetadata'} -> string
|
---|
58 | # {'maintainer'}->array of strings
|
---|
59 | # {'languages'}->array of strings
|
---|
60 | # {'indexsubcollections'}->array of strings
|
---|
61 | # {'indexes'}->array of strings
|
---|
62 | # {'indexoptions'}->array of strings (stem, casefold, accentfold, separate_cjk)
|
---|
63 | # {'dontbuild'}->array of strings
|
---|
64 | # {'dontgdbm'}->array of strings
|
---|
65 | # {'mirror'}->array of strings
|
---|
66 | # {'phind'}->array of strings
|
---|
67 | # {'plugout'}->array of strings
|
---|
68 | # {'levels'}->array of strings (for mgpp eg Section, Paragraph)
|
---|
69 | # {'searchtype'}->array of strings (for mgpp, form or plain)
|
---|
70 | # {'sortfields'}->array of strings (for lucene)
|
---|
71 | # {'subcollection'}->hash of key-value pairs
|
---|
72 |
|
---|
73 | # {'acquire'}->array of arrays of strings
|
---|
74 | # {'plugin'}->array of arrays of strings
|
---|
75 | # {'classify'}->array of arrays of strings
|
---|
76 |
|
---|
77 | # {'collectionmeta'}->hash of key->hash of param-value -used
|
---|
78 | # for language specification
|
---|
79 | # for example, collectionmeta->collectionname->default->demo
|
---|
80 | # ->mi->maori demo
|
---|
81 |
|
---|
82 | # convenience method for reading in either collect.cfg/collectionConfig.xml
|
---|
83 | sub read_collection_cfg {
|
---|
84 | my ($filename,$gs_mode) = @_;
|
---|
85 |
|
---|
86 | my $collectcfg = undef;
|
---|
87 |
|
---|
88 | if ($gs_mode eq "gs2") {
|
---|
89 | $collectcfg = &colcfg::read_collect_cfg ($filename);
|
---|
90 | } elsif ($gs_mode eq "gs3") {
|
---|
91 | $collectcfg = &colcfg::read_collection_cfg_xml ($filename);
|
---|
92 | }
|
---|
93 | else {
|
---|
94 | print STDERR "Failed to read collection configuration file\n";
|
---|
95 | print STDERR " Unrecognized mode: $gs_mode\n";
|
---|
96 | }
|
---|
97 |
|
---|
98 | return $collectcfg;
|
---|
99 | }
|
---|
100 |
|
---|
101 | # convenience method for writing out either collect.cfg/collectionConfig.xml
|
---|
102 | # is this ever used??
|
---|
103 | sub write_collection_cfg {
|
---|
104 | my ($filename, $collectcfg_data, $gs_mode) = @_;
|
---|
105 |
|
---|
106 | if ($gs_mode eq "gs2") {
|
---|
107 | &colcfg::write_collect_cfg ($filename, $collectcfg_data );
|
---|
108 | } elsif ($gs_mode eq "gs3") {
|
---|
109 | &colcfg::write_collection_cfg_xml ($filename, $collectcfg_data);
|
---|
110 | }
|
---|
111 | else {
|
---|
112 | print STDERR "Failed to write collection configuration file\n";
|
---|
113 | print STDERR " Unrecognized mode: $gs_mode\n";
|
---|
114 | }
|
---|
115 | }
|
---|
116 |
|
---|
117 | # the build configuration file data is stored in the form
|
---|
118 | #
|
---|
119 | # {'infodbtype'}->string
|
---|
120 | # {'builddate'}->string
|
---|
121 | # {'buildtype'}->string
|
---|
122 | # {'orthogonalbuildtypes'}->array of strings
|
---|
123 | # {'metadata'}->array of strings
|
---|
124 | # {'languages'}->array of strings
|
---|
125 | # {'numdocs'}->string
|
---|
126 | # {'numsections'}->string
|
---|
127 | # {'numwords'}->string
|
---|
128 | # {'numbytes'}->string
|
---|
129 | # {'maxnumeric'}->string
|
---|
130 | # {'indexfields'}->array of strings
|
---|
131 | # {'indexfieldmap'}->array of strings in the form "field->FI"
|
---|
132 | # {'indexmap'} -> array of strings
|
---|
133 | # {'indexlevels'} -> array of strings
|
---|
134 | # {'indexsortfields'} -> array of strings
|
---|
135 | # {'indexsortfieldmap'} -> array of strings in the form "field->byFI"
|
---|
136 | # {'stemindexes'} -> string (int)
|
---|
137 | # {'textlevel'}->string
|
---|
138 | # {'levelmap'} -> array of strings in the form "level->shortname"
|
---|
139 |
|
---|
140 | # convenience method for reading in either build.cfg/buildConfig.xml
|
---|
141 | sub read_building_cfg {
|
---|
142 | my ($filename,$gs_mode) = @_;
|
---|
143 |
|
---|
144 | my $buildcfg = undef;
|
---|
145 |
|
---|
146 | if ($gs_mode eq "gs2") {
|
---|
147 | $buildcfg = &colcfg::read_build_cfg ($filename);
|
---|
148 | } elsif ($gs_mode eq "gs3") {
|
---|
149 | $buildcfg = &colcfg::read_build_cfg_xml ($filename);
|
---|
150 | }
|
---|
151 | else {
|
---|
152 | print STDERR "Failed to read building configuration file\n";
|
---|
153 | print STDERR " Unrecognized mode: $gs_mode\n";
|
---|
154 | }
|
---|
155 |
|
---|
156 | return $buildcfg;
|
---|
157 | }
|
---|
158 |
|
---|
159 | # convenience method for writing out either build.cfg/buildConfig.xml
|
---|
160 | # haven't got one, as gs3 version needs extra parameters
|
---|
161 | #sub write_building_cfg {}
|
---|
162 |
|
---|
163 | ##############################
|
---|
164 | ### gs2/gs3 specific methods
|
---|
165 | ###############################
|
---|
166 |
|
---|
167 | #####################################
|
---|
168 | ### collect.cfg/collectionConfig.xml
|
---|
169 | #####################################
|
---|
170 |
|
---|
171 | # gs2 read in collect.cfg
|
---|
172 | sub read_collect_cfg {
|
---|
173 | my ($filename) = @_;
|
---|
174 |
|
---|
175 | return &cfgread::read_cfg_file_unicode ($filename,
|
---|
176 | q/^(infodbtype|creator|public|complexmeta|defaultindex|importdir|/ .
|
---|
177 | q/archivedir|exportdir|cachedir|builddir|removeold|/ .
|
---|
178 | q/textcompress|buildtype|othogonalbuildtypes|no_text|keepold|NO_IMPORT|gzip|/ .
|
---|
179 | q/verbosity|remove_empty_classifications|OIDtype|OIDmetadata|oidtype|oidmetadata|/ .
|
---|
180 | q/groupsize|maxdocs|debug|mode|saveas|saveas_options|/ .
|
---|
181 | q/sortmeta|removesuffix|removeprefix|create_images|/ .
|
---|
182 | q/maxnumeric|languagemetadata|/ .
|
---|
183 | q/no_strip_html|index|sections_index_document_metadata|sections_sort_on_document_metadata|/ .
|
---|
184 | q/store_metadata_coverage|indexname|indexlevel)$/,
|
---|
185 | q/(maintainer|languages|indexsubcollections|orthogonalbuildtypes|/ .
|
---|
186 | q/indexes|indexoptions|dontbuild|dontgdbm|mirror|levels|sortfields|plugout|/ .
|
---|
187 | q/searchtype|searchtypes)$/,
|
---|
188 | q/^(subcollection|format)$/,
|
---|
189 | q/^(acquire|plugin|classify)$/,
|
---|
190 | q/^(collectionmeta)$/);
|
---|
191 | }
|
---|
192 |
|
---|
193 | # gs2 write out collect.cfg
|
---|
194 | sub write_collect_cfg {
|
---|
195 | my ($filename, $data) = @_;
|
---|
196 |
|
---|
197 | &cfgread::write_cfg_file($filename, $data,
|
---|
198 | q/^(infodbtype|creator|public|complexmeta|defaultindex|importdir|/ .
|
---|
199 | q/archivedir|cachedir|builddir|removeold|/ .
|
---|
200 | q/textcompress|buildtype|no_text|keepold|NO_IMPORT|gzip|/ .
|
---|
201 | q/verbosity|remove_empty_classifications|OIDtype|OIDmetadata|/.
|
---|
202 | q/groupsize|maxdocs|debug|mode|saveas|/ .
|
---|
203 | q/sortmeta|removesuffix|removeprefix|create_images|/ .
|
---|
204 | q/maxnumeric|languagemetadata/ .
|
---|
205 | q/no_strip_html|index|sections_index_document_metadata|sections_sort_on_document_metadata)$/.
|
---|
206 | q/store_metadata_coverage)$/,
|
---|
207 | q/^(maintainer|languages|indexsubcollections|orthogonalbuildtypes|/ .
|
---|
208 | q/indexes|indexoptions|dontbuild|dontgdbm|mirror|levels|/.
|
---|
209 | q/searchtype|searchtypes)$/,
|
---|
210 | q/^(subcollection|format)$/,
|
---|
211 | q/^(acquire|plugin|classify)$/,
|
---|
212 | q/^(collectionmeta)$/);
|
---|
213 | }
|
---|
214 |
|
---|
215 | # gs3 read in collectionConfig.xml
|
---|
216 | sub read_collection_cfg_xml {
|
---|
217 | my ($filename) = @_;
|
---|
218 |
|
---|
219 | require collConfigxml;
|
---|
220 | return &collConfigxml::read_cfg_file ($filename);
|
---|
221 | }
|
---|
222 |
|
---|
223 | # gs3 write out collectionConfig.xml
|
---|
224 | sub write_collection_cfg_xml {
|
---|
225 |
|
---|
226 | }
|
---|
227 |
|
---|
228 | #####################################
|
---|
229 | ### build.cfg/buildConfig.xml
|
---|
230 | ######################################
|
---|
231 |
|
---|
232 | # gs2 read build.cfg
|
---|
233 | sub read_build_cfg {
|
---|
234 | my ($filename) = @_;
|
---|
235 |
|
---|
236 | return &cfgread::read_cfg_file ($filename,
|
---|
237 | q/^(earliestdatestamp|infodbtype|builddate|buildtype|numdocs|numsections|numwords|numbytes|maxnumeric|textlevel|indexstem|stemindexes|separate_cjk)$/,
|
---|
238 | q/^(indexmap|subcollectionmap|languagemap|orthogonalbuildtypes|notbuilt|indexfields|indexfieldmap|indexlevels|levelmap|indexsortfields|indexsortfieldmap)$/);
|
---|
239 |
|
---|
240 | }
|
---|
241 |
|
---|
242 | # gs2 write build.cfg
|
---|
243 | sub write_build_cfg {
|
---|
244 | my ($filename, $data) = @_;
|
---|
245 |
|
---|
246 | &cfgread::write_cfg_file($filename, $data,
|
---|
247 | q/^(earliestdatestamp|infodbtype|builddate|buildtype|numdocs|numsections|numwords|numbytes|maxnumeric|textlevel|indexstem|stemindexes|separate_cjk)$/,
|
---|
248 | q/^(indexmap|subcollectionmap|languagemap|orthogonalbuildtypes|notbuilt|indexfields|indexfieldmap|indexlevels|levelmap|indexsortfields|indexsortfieldmap)$/);
|
---|
249 | }
|
---|
250 |
|
---|
251 | # gs3 read buildConfig.xml
|
---|
252 | sub read_build_cfg_xml {
|
---|
253 |
|
---|
254 | my ($filename) = @_;
|
---|
255 |
|
---|
256 | require buildConfigxml;
|
---|
257 | return &buildConfigxml::read_cfg_file($filename);
|
---|
258 | }
|
---|
259 |
|
---|
260 | # gs3 write buildConfig.xml
|
---|
261 | sub write_build_cfg_xml {
|
---|
262 | my ($buildoutfile, $buildcfg, $collectcfg) = @_;
|
---|
263 |
|
---|
264 | require buildConfigxml;
|
---|
265 | return &buildConfigxml::write_cfg_file ($buildoutfile, $buildcfg, $collectcfg);
|
---|
266 | }
|
---|
267 |
|
---|
268 |
|
---|
269 | # method to check for filename of collect.cfg, and gs mode.
|
---|
270 | sub get_collect_cfg_name_old {
|
---|
271 | my ($out) = @_;
|
---|
272 |
|
---|
273 | # First check if there's a
|
---|
274 | # gsdl/collect/COLLECTION/custom/COLLECTION/etc/custom.cfg file. This
|
---|
275 | # customization was added for DLC by Stefan, 30/6/2007.
|
---|
276 | my $configfilename = &FileUtils::filenameConcatenate ($ENV{'GSDLCOLLECTDIR'}, "custom", $ENV{'GSDLCOLLECTION'}, "etc", "custom.cfg");
|
---|
277 |
|
---|
278 | if (-e $configfilename) {
|
---|
279 | return ($configfilename, "gs2");
|
---|
280 | }
|
---|
281 |
|
---|
282 | # Check if there is a collectionConfig.xml file. If there is one, it's gs3
|
---|
283 | $configfilename = &FileUtils::filenameConcatenate ($ENV{'GSDLCOLLECTDIR'}, "etc", "collectionConfig.xml");
|
---|
284 | if (-e $configfilename) {
|
---|
285 | return ($configfilename, "gs3");
|
---|
286 | }
|
---|
287 |
|
---|
288 | # If we get to here we check if there is a collect.cfg file in the usual place, i.e. it is gs2.
|
---|
289 | $configfilename = &FileUtils::filenameConcatenate ($ENV{'GSDLCOLLECTDIR'}, "etc", "collect.cfg");
|
---|
290 | if (-e $configfilename) {
|
---|
291 | return ($configfilename, "gs2");
|
---|
292 | }
|
---|
293 |
|
---|
294 | # Error. No collection configuration file.
|
---|
295 | (&gsprintf($out, "{common.cannot_find_cfg_file}\n", $configfilename) && die);
|
---|
296 | }
|
---|
297 |
|
---|
298 | # method to check for filename of collect.cfg
|
---|
299 | # needs to be given gs_version, since we can have a GS2 collection ported into
|
---|
300 | # GS3 which could potentially have collect.cfg AND collectionConfig.xml
|
---|
301 | # in which case the older version of this subroutine (get_collect_cfg_name_old)
|
---|
302 | # will return the wrong answer for the gs version we're using.
|
---|
303 | sub get_collect_cfg_name {
|
---|
304 | my ($out, $gs_version) = @_;
|
---|
305 |
|
---|
306 | # First check if there's a
|
---|
307 | # gsdl/collect/COLLECTION/custom/COLLECTION/etc/custom.cfg file. This
|
---|
308 | # customization was added for DLC by Stefan, 30/6/2007.
|
---|
309 | my $configfilename;
|
---|
310 |
|
---|
311 | if($gs_version eq "gs2") {
|
---|
312 | $configfilename = &FileUtils::filenameConcatenate ($ENV{'GSDLCOLLECTDIR'}, "custom", $ENV{'GSDLCOLLECTION'}, "etc", "custom.cfg");
|
---|
313 |
|
---|
314 | if (-e $configfilename) {
|
---|
315 | return $configfilename;
|
---|
316 | }
|
---|
317 | }
|
---|
318 |
|
---|
319 | # Check if there is a collectionConfig.xml file if it's gs3
|
---|
320 | if($gs_version eq "gs3") {
|
---|
321 | $configfilename = &FileUtils::filenameConcatenate ($ENV{'GSDLCOLLECTDIR'}, "etc", "collectionConfig.xml");
|
---|
322 | if (-e $configfilename) {
|
---|
323 | return $configfilename;
|
---|
324 | }
|
---|
325 | }
|
---|
326 |
|
---|
327 | # Check if there is a collect.cfg file in the usual place for gs2.
|
---|
328 | if($gs_version eq "gs2") {
|
---|
329 | $configfilename = &FileUtils::filenameConcatenate ($ENV{'GSDLCOLLECTDIR'}, "etc", "collect.cfg");
|
---|
330 | if (-e $configfilename) {
|
---|
331 | return $configfilename;
|
---|
332 | }
|
---|
333 | }
|
---|
334 |
|
---|
335 | # Error. No collection configuration file.
|
---|
336 | (&gsprintf($out, "{common.cannot_find_cfg_file}\n", $configfilename) && die);
|
---|
337 | }
|
---|
338 |
|
---|
339 |
|
---|
340 |
|
---|
341 | sub use_collection {
|
---|
342 | my ($site, $collection, $collectdir) = @_;
|
---|
343 |
|
---|
344 | if ((defined $site) && ($site ne ""))
|
---|
345 | {
|
---|
346 | return &util::use_site_collection($site, $collection, $collectdir);
|
---|
347 | }
|
---|
348 | else
|
---|
349 | {
|
---|
350 | return &util::use_collection($collection, $collectdir);
|
---|
351 | }
|
---|
352 | }
|
---|
353 |
|
---|
354 |
|
---|
355 | 1;
|
---|
356 |
|
---|
357 |
|
---|