1 | #!/usr/bin/perl -w
|
---|
2 |
|
---|
3 | ###########################################################################
|
---|
4 | #
|
---|
5 | # buildcol.pl --
|
---|
6 | #
|
---|
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 | # This program will build a particular collection.
|
---|
30 | package buildcol;
|
---|
31 |
|
---|
32 | # Environment
|
---|
33 | BEGIN
|
---|
34 | {
|
---|
35 | die "GSDLHOME not set\n" unless defined $ENV{'GSDLHOME'};
|
---|
36 | die "GSDLOS not set\n" unless defined $ENV{'GSDLOS'};
|
---|
37 |
|
---|
38 | # Order is important. With unshift want our XMLParser to be
|
---|
39 | # found ahead of XML/XPath
|
---|
40 |
|
---|
41 | unshift (@INC, $ENV{'GSDLHOME'} . '/perllib/cpan/XML/XPath');
|
---|
42 | unshift (@INC, $ENV{'GSDLHOME'} . '/perllib/classify');
|
---|
43 | unshift (@INC, $ENV{'GSDLHOME'} . '/perllib/plugins');
|
---|
44 | unshift (@INC, $ENV{'GSDLHOME'} . '/perllib/cpan');
|
---|
45 | unshift (@INC, $ENV{'GSDLHOME'} . '/perllib');
|
---|
46 |
|
---|
47 | if (defined $ENV{'GSDL-RUN-SETUP'})
|
---|
48 | {
|
---|
49 | require util;
|
---|
50 | &util::setup_greenstone_env($ENV{'GSDLHOME'}, $ENV{'GSDLOS'});
|
---|
51 | }
|
---|
52 |
|
---|
53 | if (defined $ENV{'GSDLEXTS'})
|
---|
54 | {
|
---|
55 | my @extensions = split(/:/, $ENV{'GSDLEXTS'});
|
---|
56 | foreach my $e (@extensions)
|
---|
57 | {
|
---|
58 | my $ext_prefix = $ENV{'GSDLHOME'} . '/ext/' . $e;
|
---|
59 |
|
---|
60 | unshift(@INC, $ext_prefix . '/perllib');
|
---|
61 | unshift(@INC, $ext_prefix . '/perllib/cpan');
|
---|
62 | unshift(@INC, $ext_prefix . '/perllib/plugins');
|
---|
63 | unshift(@INC, $ext_prefix . '/perllib/classify');
|
---|
64 | }
|
---|
65 | }
|
---|
66 | if (defined $ENV{'GSDL3EXTS'})
|
---|
67 | {
|
---|
68 | my @extensions = split(/:/, $ENV{'GSDL3EXTS'});
|
---|
69 | foreach my $e (@extensions)
|
---|
70 | {
|
---|
71 | my $ext_prefix = $ENV{'GSDL3SRCHOME'} . '/ext/' . $e;
|
---|
72 |
|
---|
73 | unshift(@INC, $ext_prefix . '/perllib');
|
---|
74 | unshift(@INC, $ext_prefix . '/perllib/cpan');
|
---|
75 | unshift(@INC, $ext_prefix . '/perllib/plugins');
|
---|
76 | unshift(@INC, $ext_prefix . '/perllib/classify');
|
---|
77 | }
|
---|
78 | }
|
---|
79 | }
|
---|
80 |
|
---|
81 | # Pragma
|
---|
82 | use strict;
|
---|
83 | no strict 'refs'; # allow filehandles to be variables and vice versa
|
---|
84 | no strict 'subs'; # allow barewords (eg STDERR) as function arguments
|
---|
85 |
|
---|
86 | # Modules
|
---|
87 | use Symbol qw<qualify>; # Needed for runtime loading of modules [jmt12]
|
---|
88 |
|
---|
89 | # Greenstone Modules
|
---|
90 | use buildcolutils;
|
---|
91 | use FileUtils;
|
---|
92 | use util;
|
---|
93 |
|
---|
94 | # Globals
|
---|
95 | # - build up arguments list/control
|
---|
96 | my $mode_list =
|
---|
97 | [ { 'name' => "all",
|
---|
98 | 'desc' => "{buildcol.mode.all}" },
|
---|
99 | { 'name' => "compress_text",
|
---|
100 | 'desc' => "{buildcol.mode.compress_text}" },
|
---|
101 | { 'name' => "build_index",
|
---|
102 | 'desc' => "{buildcol.mode.build_index}" },
|
---|
103 | { 'name' => "infodb",
|
---|
104 | 'desc' => "{buildcol.mode.infodb}" } ];
|
---|
105 |
|
---|
106 | my $sec_index_list =
|
---|
107 | [ {'name' => "never",
|
---|
108 | 'desc' => "{buildcol.sections_index_document_metadata.never}" },
|
---|
109 | {'name' => "always",
|
---|
110 | 'desc' => "{buildcol.sections_index_document_metadata.always}" },
|
---|
111 | {'name' => "unless_section_metadata_exists",
|
---|
112 | 'desc' => "{buildcol.sections_index_document_metadata.unless_section_metadata_exists}" }
|
---|
113 | ];
|
---|
114 |
|
---|
115 | my $arguments =
|
---|
116 | [ { 'name' => "remove_empty_classifications",
|
---|
117 | 'desc' => "{buildcol.remove_empty_classifications}",
|
---|
118 | 'type' => "flag",
|
---|
119 | 'reqd' => "no",
|
---|
120 | 'modegli' => "2" },
|
---|
121 | { 'name' => "archivedir",
|
---|
122 | 'desc' => "{buildcol.archivedir}",
|
---|
123 | 'type' => "string",
|
---|
124 | 'reqd' => "no",
|
---|
125 | 'hiddengli' => "yes" },
|
---|
126 | { 'name' => "builddir",
|
---|
127 | 'desc' => "{buildcol.builddir}",
|
---|
128 | 'type' => "string",
|
---|
129 | 'reqd' => "no",
|
---|
130 | 'hiddengli' => "yes" },
|
---|
131 | # { 'name' => "cachedir",
|
---|
132 | # 'desc' => "{buildcol.cachedir}",
|
---|
133 | # 'type' => "string",
|
---|
134 | # 'reqd' => "no" },
|
---|
135 | { 'name' => "collectdir",
|
---|
136 | 'desc' => "{buildcol.collectdir}",
|
---|
137 | 'type' => "string",
|
---|
138 | # parsearg left "" as default
|
---|
139 | #'deft' => &FileUtils::filenameConcatenate($ENV{'GSDLHOME'}, "collect"),
|
---|
140 | 'reqd' => "no",
|
---|
141 | 'hiddengli' => "yes" },
|
---|
142 | { 'name' => "site",
|
---|
143 | 'desc' => "{buildcol.site}",
|
---|
144 | 'type' => "string",
|
---|
145 | 'deft' => "",
|
---|
146 | 'reqd' => "no",
|
---|
147 | 'hiddengli' => "yes" },
|
---|
148 | { 'name' => "debug",
|
---|
149 | 'desc' => "{buildcol.debug}",
|
---|
150 | 'type' => "flag",
|
---|
151 | 'reqd' => "no",
|
---|
152 | 'hiddengli' => "yes" },
|
---|
153 | { 'name' => "faillog",
|
---|
154 | 'desc' => "{buildcol.faillog}",
|
---|
155 | 'type' => "string",
|
---|
156 | # parsearg left "" as default
|
---|
157 | #'deft' => &FileUtils::filenameConcatenate("<collectdir>", "colname", "etc", "fail.log"),
|
---|
158 | 'reqd' => "no",
|
---|
159 | 'modegli' => "3" },
|
---|
160 | { 'name' => "index",
|
---|
161 | 'desc' => "{buildcol.index}",
|
---|
162 | 'type' => "string",
|
---|
163 | 'reqd' => "no",
|
---|
164 | 'modegli' => "3" },
|
---|
165 | { 'name' => "incremental",
|
---|
166 | 'desc' => "{buildcol.incremental}",
|
---|
167 | 'type' => "flag",
|
---|
168 | 'hiddengli' => "yes" },
|
---|
169 | { 'name' => "keepold",
|
---|
170 | 'desc' => "{buildcol.keepold}",
|
---|
171 | 'type' => "flag",
|
---|
172 | 'reqd' => "no",
|
---|
173 | #'modegli' => "3",
|
---|
174 | 'hiddengli' => "yes" },
|
---|
175 | { 'name' => "removeold",
|
---|
176 | 'desc' => "{buildcol.removeold}",
|
---|
177 | 'type' => "flag",
|
---|
178 | 'reqd' => "no",
|
---|
179 | #'modegli' => "3",
|
---|
180 | 'hiddengli' => "yes" },
|
---|
181 | { 'name' => "language",
|
---|
182 | 'desc' => "{scripts.language}",
|
---|
183 | 'type' => "string",
|
---|
184 | 'reqd' => "no",
|
---|
185 | 'modegli' => "3" },
|
---|
186 | { 'name' => "maxdocs",
|
---|
187 | 'desc' => "{buildcol.maxdocs}",
|
---|
188 | 'type' => "int",
|
---|
189 | 'reqd' => "no",
|
---|
190 | 'hiddengli' => "yes" },
|
---|
191 | { 'name' => "maxnumeric",
|
---|
192 | 'desc' => "{buildcol.maxnumeric}",
|
---|
193 | 'type' => "int",
|
---|
194 | 'reqd' => "no",
|
---|
195 | 'deft' => "4",
|
---|
196 | 'range' => "4,512",
|
---|
197 | 'modegli' => "3" },
|
---|
198 | { 'name' => "mode",
|
---|
199 | 'desc' => "{buildcol.mode}",
|
---|
200 | 'type' => "enum",
|
---|
201 | 'list' => $mode_list,
|
---|
202 | # parsearg left "" as default
|
---|
203 | # 'deft' => "all",
|
---|
204 | 'reqd' => "no",
|
---|
205 | 'modegli' => "3" },
|
---|
206 | { 'name' => "no_strip_html",
|
---|
207 | 'desc' => "{buildcol.no_strip_html}",
|
---|
208 | 'type' => "flag",
|
---|
209 | 'reqd' => "no",
|
---|
210 | 'modegli' => "3" },
|
---|
211 | { 'name' => "store_metadata_coverage",
|
---|
212 | 'desc' => "{buildcol.store_metadata_coverage}",
|
---|
213 | 'type' => "flag",
|
---|
214 | 'reqd' => "no",
|
---|
215 | 'modegli' => "3" },
|
---|
216 | { 'name' => "no_text",
|
---|
217 | 'desc' => "{buildcol.no_text}",
|
---|
218 | 'type' => "flag",
|
---|
219 | 'reqd' => "no",
|
---|
220 | 'modegli' => "2" },
|
---|
221 | { 'name' => "sections_index_document_metadata",
|
---|
222 | 'desc' => "{buildcol.sections_index_document_metadata}",
|
---|
223 | 'type' => "enum",
|
---|
224 | 'list' => $sec_index_list,
|
---|
225 | 'reqd' => "no",
|
---|
226 | 'modegli' => "2" },
|
---|
227 | { 'name' => "sections_sort_on_document_metadata",
|
---|
228 | 'desc' => "{buildcol.sections_sort_on_document_metadata}",
|
---|
229 | 'type' => "enum",
|
---|
230 | 'list' => $sec_index_list,
|
---|
231 | 'reqd' => "no",
|
---|
232 | 'modegli' => "2" },
|
---|
233 | { 'name' => "out",
|
---|
234 | 'desc' => "{buildcol.out}",
|
---|
235 | 'type' => "string",
|
---|
236 | 'deft' => "STDERR",
|
---|
237 | 'reqd' => "no",
|
---|
238 | 'hiddengli' => "yes" },
|
---|
239 | { 'name' => "verbosity",
|
---|
240 | 'desc' => "{buildcol.verbosity}",
|
---|
241 | 'type' => "int",
|
---|
242 | # parsearg left "" as default
|
---|
243 | #'deft' => "2",
|
---|
244 | 'reqd' => "no",
|
---|
245 | 'modegli' => "3" },
|
---|
246 | { 'name' => "gli",
|
---|
247 | 'desc' => "",
|
---|
248 | 'type' => "flag",
|
---|
249 | 'reqd' => "no",
|
---|
250 | 'hiddengli' => "yes" },
|
---|
251 | { 'name' => "xml",
|
---|
252 | 'desc' => "{scripts.xml}",
|
---|
253 | 'type' => "flag",
|
---|
254 | 'reqd' => "no",
|
---|
255 | 'hiddengli' => "yes" },
|
---|
256 | { 'name' => "activate",
|
---|
257 | 'desc' => "{buildcol.activate}",
|
---|
258 | 'type' => "flag",
|
---|
259 | 'reqd' => "no",
|
---|
260 | 'hiddengli' => "yes" },
|
---|
261 | { 'name' => "indexname",
|
---|
262 | 'desc' => "{buildcol.index}",
|
---|
263 | 'type' => "string",
|
---|
264 | 'reqd' => "no",
|
---|
265 | 'modegli' => "3" },
|
---|
266 | { 'name' => "indexlevel",
|
---|
267 | 'desc' => "{buildcol.indexlevel}",
|
---|
268 | 'type' => "string",
|
---|
269 | 'reqd' => "no",
|
---|
270 | 'modegli' => "3" },
|
---|
271 | ];
|
---|
272 |
|
---|
273 | my $options = { 'name' => "buildcol.pl",
|
---|
274 | 'desc' => "{buildcol.desc}",
|
---|
275 | 'args' => $arguments };
|
---|
276 |
|
---|
277 | # The hash maps between argument and the buildcolutils subclass supporting that
|
---|
278 | # argument - allowing for extensions to override the normal buildcolutils as
|
---|
279 | # necessary
|
---|
280 | my $function_to_subclass_mappings = {};
|
---|
281 |
|
---|
282 | # Lets get the party rolling... or ball started... hmmm
|
---|
283 | &main();
|
---|
284 |
|
---|
285 | exit;
|
---|
286 |
|
---|
287 | sub main
|
---|
288 | {
|
---|
289 | # Dynamically include arguments from any subclasses of buildcolutils we find
|
---|
290 | # in the extensions directory
|
---|
291 | if (defined $ENV{'GSDLEXTS'})
|
---|
292 | {
|
---|
293 | &_scanForSubclasses($ENV{'GSDLHOME'}, $ENV{'GSDLEXTS'});
|
---|
294 | }
|
---|
295 | if (defined $ENV{'GSDL3EXTS'})
|
---|
296 | {
|
---|
297 | &_scanForSubclasses($ENV{'GSDL3SRCHOME'}, $ENV{'GSDL3EXTS'});
|
---|
298 | }
|
---|
299 |
|
---|
300 | # Loop through arguments, checking to see if any depend on a specific
|
---|
301 | # subclass of buildcolutils. Note that we load the first subclass we
|
---|
302 | # encounter so only support a single 'override' ATM.
|
---|
303 | my $subclass;
|
---|
304 | foreach my $argument (@ARGV)
|
---|
305 | {
|
---|
306 | # proper arguments start with a hyphen
|
---|
307 | if ($argument =~ /^-/ && defined $function_to_subclass_mappings->{$argument})
|
---|
308 | {
|
---|
309 | my $required_subclass = $function_to_subclass_mappings->{$argument};
|
---|
310 | if (!defined $subclass)
|
---|
311 | {
|
---|
312 | $subclass = $required_subclass;
|
---|
313 | }
|
---|
314 | # Oh noes! The user has included specific arguments from two different
|
---|
315 | # subclasses... this isn't supported
|
---|
316 | elsif ($subclass ne $required_subclass)
|
---|
317 | {
|
---|
318 | print STDERR "Error! You cannot specify arguments from two different extention specific buildcolutils modules: " . $subclass . " != " . $required_subclass . "\n";
|
---|
319 | exit;
|
---|
320 | }
|
---|
321 | }
|
---|
322 | }
|
---|
323 |
|
---|
324 | my $buildcolutils;
|
---|
325 | if (defined $subclass)
|
---|
326 | {
|
---|
327 | print "* Loading overriding buildcolutils module: " . $subclass . "\n";
|
---|
328 | require $subclass . '.pm';
|
---|
329 | $buildcolutils = new $subclass(\@ARGV, $options);
|
---|
330 | }
|
---|
331 | # We don't have a overridden buildcolutils, or the above command failed
|
---|
332 | # somehow so load the base class
|
---|
333 | if (!defined $buildcolutils)
|
---|
334 | {
|
---|
335 | $buildcolutils = new buildcolutils(\@ARGV, $options);
|
---|
336 | }
|
---|
337 |
|
---|
338 | my $collection = $buildcolutils->get_collection();
|
---|
339 | if (defined $collection)
|
---|
340 | {
|
---|
341 | my ($config_filename,$collect_cfg) = $buildcolutils->read_collection_cfg($collection, $options);
|
---|
342 | $buildcolutils->set_collection_options($collect_cfg);
|
---|
343 |
|
---|
344 | my $builders_ref = $buildcolutils->prepare_builders($config_filename, $collect_cfg);
|
---|
345 | $buildcolutils->build_collection($builders_ref);
|
---|
346 | $buildcolutils->build_auxiliary_files($builders_ref);
|
---|
347 | $buildcolutils->complete_builders($builders_ref);
|
---|
348 |
|
---|
349 | # The user may have requested the collection be activated
|
---|
350 | $buildcolutils->activate_collection();
|
---|
351 | }
|
---|
352 |
|
---|
353 | # Cleanup
|
---|
354 | $buildcolutils->deinit();
|
---|
355 | }
|
---|
356 | # main()
|
---|
357 |
|
---|
358 | # @function _scanForSubclasses()
|
---|
359 | # @param $dir The extension directory to look within
|
---|
360 | # @param $exts A list of the available extensions (as a colon separated string)
|
---|
361 | # @return The number of subclasses of buildcolutils found as an Integer
|
---|
362 | sub _scanForSubclasses
|
---|
363 | {
|
---|
364 | my ($dir, $exts) = @_;
|
---|
365 | my $class_count = 0;
|
---|
366 | my $ext_prefix = &FileUtils::filenameConcatenate($dir, "ext");
|
---|
367 | my @extensions = split(/:/, $exts);
|
---|
368 | foreach my $e (@extensions)
|
---|
369 | {
|
---|
370 | # - any subclass must be prefixed with the name of the ext
|
---|
371 | my $package_name = $e . 'buildcolutils';
|
---|
372 | $package_name =~ s/[^a-z]//gi; # package names have limited characters
|
---|
373 | my $file_name = $package_name . '.pm';
|
---|
374 | my $file_path = &FileUtils::filenameConcatenate($ext_prefix, $e, 'perllib', $file_name);
|
---|
375 | # see if we have a subclass lurking in that extension folder
|
---|
376 | if (&FileUtils::fileExists($file_path))
|
---|
377 | {
|
---|
378 | # - note we load the filename (with pm) unlike normal modules
|
---|
379 | require $file_name;
|
---|
380 | # - make call to the newly created package
|
---|
381 | my $symbol = qualify('getSupportedArguments', $package_name);
|
---|
382 | # - strict prevents strings being used as function calls, so temporarily
|
---|
383 | # disable that pragma
|
---|
384 | no strict;
|
---|
385 | # - lets check that the function we are about to call actually exists
|
---|
386 | if ( defined &{$symbol} )
|
---|
387 | {
|
---|
388 | my $extra_arguments = &{$symbol}();
|
---|
389 | foreach my $argument (@{$extra_arguments})
|
---|
390 | {
|
---|
391 | # - record a mapping from each extra arguments to the subclass
|
---|
392 | # that supports it. We put the hyphen on here to make comparing
|
---|
393 | # with command line arguments even easier
|
---|
394 | $function_to_subclass_mappings->{'-' . $argument->{'name'}} = $package_name;
|
---|
395 | # - and them add them as acceptable arguments to import.pl
|
---|
396 | push(@{$options->{'args'}}, $argument);
|
---|
397 | }
|
---|
398 | $class_count++;
|
---|
399 | }
|
---|
400 | else
|
---|
401 | {
|
---|
402 | print "Warning! A subclass of buildcolutils module (named '" . $file_name . "') does not implement the required getSupportedArguments() function - ignoring. Found in: " . $file_path . "\n";
|
---|
403 | }
|
---|
404 | }
|
---|
405 | }
|
---|
406 | return $class_count;
|
---|
407 | }
|
---|
408 | # _scanForSubclasses()
|
---|