root/main/trunk/greenstone2/bin/script/import.pl @ 28703

Revision 28703, 11.1 KB (checked in by kjdon, 6 years ago)

can set the defaults now in the args structure, as we have default_argname to see if it was set by the user or not.

  • Property svn:executable set to *
  • Property svn:keywords set to Author Date Id Revision
Line 
1#!/usr/bin/perl -w
2
3###########################################################################
4#
5# import.pl --
6# A component of the Greenstone digital library software
7# from the New Zealand Digital Library Project at the
8# University of Waikato, New Zealand.
9#
10# Copyright (C) 1999 New Zealand Digital Library Project
11#
12# This program is free software; you can redistribute it and/or modify
13# it under the terms of the GNU General Public License as published by
14# the Free Software Foundation; either version 2 of the License, or
15# (at your option) any later version.
16#
17# This program is distributed in the hope that it will be useful,
18# but WITHOUT ANY WARRANTY; without even the implied warranty of
19# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
20# GNU General Public License for more details.
21#
22# You should have received a copy of the GNU General Public License
23# along with this program; if not, write to the Free Software
24# Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
25#
26###########################################################################
27
28
29# This program will import a number of files into a particular collection
30
31package import;
32
33BEGIN {
34    die "GSDLHOME not set\n" unless defined $ENV{'GSDLHOME'};
35    die "GSDLOS not set\n" unless defined $ENV{'GSDLOS'};
36    unshift (@INC, "$ENV{'GSDLHOME'}/perllib");
37    unshift (@INC, "$ENV{'GSDLHOME'}/perllib/cpan");
38    unshift (@INC, "$ENV{'GSDLHOME'}/perllib/plugins");
39    unshift (@INC, "$ENV{'GSDLHOME'}/perllib/plugouts");
40
41    if (defined $ENV{'GSDLEXTS'}) {
42    my @extensions = split(/:/,$ENV{'GSDLEXTS'});
43    foreach my $e (@extensions) {
44        my $ext_prefix = "$ENV{'GSDLHOME'}/ext/$e";
45
46        unshift (@INC, "$ext_prefix/perllib");
47        unshift (@INC, "$ext_prefix/perllib/cpan");
48        unshift (@INC, "$ext_prefix/perllib/plugins");
49        unshift (@INC, "$ext_prefix/perllib/plugouts");
50    }
51    }
52    if (defined $ENV{'GSDL3EXTS'}) {
53    my @extensions = split(/:/,$ENV{'GSDL3EXTS'});
54    foreach my $e (@extensions) {
55        my $ext_prefix = "$ENV{'GSDL3SRCHOME'}/ext/$e";
56
57        unshift (@INC, "$ext_prefix/perllib");
58        unshift (@INC, "$ext_prefix/perllib/cpan");
59        unshift (@INC, "$ext_prefix/perllib/plugins");
60        unshift (@INC, "$ext_prefix/perllib/plugouts");
61    }
62    }
63
64    if ((defined $ENV{'DEBUG_UNICODE'}) && (defined $ENV{'DEBUG_UNICODE'})) {
65    binmode(STDERR,":utf8");
66    }
67}
68
69# Pragma
70use strict;
71use warnings;
72
73# Modules
74use Symbol qw<qualify>; # Needed for runtime loading of modules [jmt12]
75
76# Greenstone Modules
77use FileUtils;
78use inexport;
79use util;
80use gsprintf 'gsprintf';
81
82
83# used to control output file format
84my $saveas_list =
85    [ { 'name' => "GreenstoneXML",
86        'desc' => "{export.saveas.GreenstoneXML}"},
87      { 'name' => "GreenstoneMETS",
88        'desc' => "{export.saveas.GreenstoneMETS}"},
89      ];
90
91
92# Possible attributes for each argument
93# name: The name of the argument
94# desc: A description (or more likely a reference to a description) for this argument
95# type: The type of control used to represent the argument. Options include: string, int, flag, regexp, metadata, language, enum etc
96# reqd: Is this argument required?
97# hiddengli: Is this argument hidden in GLI?
98# modegli: The lowest detail mode this argument is visible at in GLI
99
100my $saveas_argument
101    = { 'name' => "saveas",
102    'desc' => "{import.saveas}",
103    'type' => "enum",
104    'list' => $saveas_list,
105    'deft' => "GreenstoneXML",
106    'reqd' => "no",
107    'modegli' => "3" };
108
109
110my $arguments =
111    [
112      $saveas_argument,
113    { 'name' => "sortmeta",
114    'desc' => "{import.sortmeta}",
115    'type' => "string",
116    #'type' => "metadata", #doesn't work properly in GLI
117    'reqd' => "no",
118    'modegli' => "2" },
119      { 'name' => "removeprefix",
120    'desc' => "{BasClas.removeprefix}",
121    'type' => "regexp",
122    'deft' => "",
123    'reqd' => "no",
124    'modegli' => "3" },
125      { 'name' => "removesuffix",
126    'desc' => "{BasClas.removesuffix}",
127    'type' => "regexp",
128    'deft' => "",
129    'reqd' => "no",
130    'modegli' => "3" },
131      { 'name' => "groupsize",
132    'desc' => "{import.groupsize}",
133    'type' => "int",
134    'deft' => "1",
135    'reqd' => "no",
136    'modegli' => "2" },
137      { 'name' => "archivedir",
138    'desc' => "{import.archivedir}",
139    'type' => "string",
140    'reqd' => "no",
141    'deft' => "archives",
142        'hiddengli' => "yes" },
143      @$inexport::directory_arguments,
144      { 'name' => "gzip",
145    'desc' => "{import.gzip}",
146    'type' => "flag",
147    'reqd' => "no",
148    'modegli' => "3" },
149     @$inexport::arguments
150];
151
152my $options = { 'name' => "import.pl",
153        'desc' => "{import.desc}",
154        'args' => $arguments };
155
156my $function_to_inexport_subclass_mappings = {};
157
158sub main
159{
160  # Dynamically include arguments from any subclasses of inexport we find
161  # in the extensions directory
162  if (defined $ENV{'GSDLEXTS'})
163  {
164    &_scanForSubclasses($ENV{'GSDLHOME'}, $ENV{'GSDLEXTS'});
165  }
166  if (defined $ENV{'GSDL3EXTS'})
167  {
168    &_scanForSubclasses($ENV{'GSDL3SRCHOME'}, $ENV{'GSDL3EXTS'});
169  }
170
171  # Loop through arguments, checking to see if any depend on a specific
172  # subclass of InExport. Note that we load the first subclass we encounter
173  # so only support a single 'override' ATM.
174  my $inexport_subclass;
175  foreach my $argument (@ARGV)
176  {
177    # proper arguments start with a hyphen
178    if ($argument =~ /^-/ && defined $function_to_inexport_subclass_mappings->{$argument})
179    {
180      my $required_inexport_subclass = $function_to_inexport_subclass_mappings->{$argument};
181      if (!defined $inexport_subclass)
182      {
183        $inexport_subclass = $required_inexport_subclass;
184      }
185      # Oh noes! The user has included specific arguments from two different
186      # inexport subclasses... this isn't supported
187      elsif ($inexport_subclass ne $required_inexport_subclass)
188      {
189        print STDERR "Error! You cannot specify arguments from two different extention specific inexport modules: " . $inexport_subclass . " != " . $required_inexport_subclass . "\n";
190        exit;
191      }
192    }
193  }
194
195  my $inexport;
196  if (defined $inexport_subclass)
197  {
198    print "* Loading Overriding InExport Module: " . $inexport_subclass . "\n";
199    require $inexport_subclass . '.pm';
200    $inexport = new $inexport_subclass("import",\@ARGV,$options);
201  }
202  # We don't have a overridden inexport, or the above command failed somehow
203  # so load the base inexport class
204  if (!defined $inexport)
205  {
206    $inexport = new inexport("import",\@ARGV,$options);
207  }
208
209  my $collection = $inexport->get_collection();
210
211  if (defined $collection)
212  {
213    my ($config_filename,$collect_cfg) = $inexport->read_collection_cfg($collection,$options);
214
215    #$inexport->set_collection_options($collect_cfg);
216    &set_collection_options($inexport, $collect_cfg);
217
218    my $pluginfo = $inexport->process_files($config_filename,$collect_cfg);
219
220    $inexport->generate_statistics($pluginfo);
221  }
222
223  $inexport->deinit();
224}
225# main()
226
227# @function _scanForSubclasses()
228# @param $dir The extension directory to look within
229# @param $exts A list of the available extensions (as a colon separated string)
230# @return The number of subclasses of InExport found as an Integer
231sub _scanForSubclasses
232{
233  my ($dir, $exts) = @_;
234  my $inexport_class_count = 0;
235  my $ext_prefix = &FileUtils::filenameConcatenate($dir, "ext");
236  my @extensions = split(/:/, $exts);
237  foreach my $e (@extensions)
238  {
239    # - any subclass of InExport must be prefixed with the name of the ext
240    my $package_name = $e . 'inexport';
241    $package_name =~ s/[^a-z]//gi; # package names have limited characters
242    my $inexport_filename = $package_name . '.pm';
243    my $inexport_path = &FileUtils::filenameConcatenate($ext_prefix, $e, 'perllib', $inexport_filename);
244    # see if we have a subclass of InExport lurking in that extension folder
245    if (-f $inexport_path)
246    {
247      # - note we load the filename (with pm) unlike normal modules
248      require $inexport_filename;
249      # - make call to the newly created package
250      my $symbol = qualify('getSupportedArguments', $package_name);
251      # - strict prevents strings being used as function calls, so temporarily
252      #   disable that pragma
253      no strict;
254      # - lets check that the function we are about to call actually exists
255      if ( defined &{$symbol} )
256      {
257        my $extra_arguments = &{$symbol}();
258        foreach my $argument (@{$extra_arguments})
259        {
260          # - record a mapping from each extra arguments to the inexport class
261          #   that supports it. We put the hyphen on here to make comparing
262          #   with command line arguments even easier
263          $function_to_inexport_subclass_mappings->{'-' . $argument->{'name'}} = $package_name;
264          # - and them add them as acceptable arguments to import.pl
265          push(@{$options->{'args'}}, $argument);
266        }
267        $inexport_class_count++;
268      }
269      else
270      {
271        print "Warning! A subclass of InExport module (named '" . $inexport_filename . "') does not implement the required getSupportedArguments() function - ignoring. Found in: " . $inexport_path . "\n";
272      }
273    }
274  }
275  return $inexport_class_count;
276}
277# _scanForInExportModules()
278
279# look up collect.cfg for import options, then all inexport version for the
280# common ones
281sub set_collection_options
282{
283
284    my ($inexport, $collectcfg) = @_;
285    my $out        = $inexport->{'out'};
286
287    # check all options for default_optname - this will be set if the parsing
288    # code has just set the value based on the arg default. In this case,
289    # check in collect.cfg for the option
290   
291    # groupsize can only be defined for import, not export, and actually only
292    # applies to GreenstoneXML format.
293    if (defined $inexport->{'default_groupsize'}) {
294    if (defined $collectcfg->{'groupsize'} && $collectcfg->{'groupsize'} =~ /\d+/) {
295        $inexport->{'groupsize'} = $collectcfg->{'groupsize'};
296    }
297
298    }
299    if (defined $inexport->{'default_saveas'}) {
300    if (defined $collectcfg->{'saveas'}
301        && $collectcfg->{'saveas'} =~ /^(GreenstoneXML|GreenstoneMETS)$/) {
302        $inexport->{'saveas'} = $collectcfg->{'saveas'};
303    } else {
304        $inexport->{'saveas'} = "GreenstoneXML"; # the default
305    }
306    }
307   
308    my $sortmeta = $inexport->{'sortmeta'};
309    if (defined $collectcfg->{'sortmeta'} && $sortmeta eq "") {
310    $sortmeta = $collectcfg->{'sortmeta'};
311    }
312    # sortmeta cannot be used with group size
313    $sortmeta = undef unless defined $sortmeta && $sortmeta =~ /\S/;
314    if (defined $sortmeta && $inexport->{'groupsize'} > 1) {
315    &gsprintf($out, "{import.cannot_sort}\n\n");
316    $sortmeta = undef;
317    }
318    if (defined $sortmeta) {
319    &gsprintf($out, "{import.sortmeta_paired_with_ArchivesInfPlugin}\n\n");
320    }
321    $inexport->{'sortmeta'} = $sortmeta;
322
323    if (defined $collectcfg->{'removeprefix'} && $inexport->{'removeprefix'} eq "") {
324    $inexport->{'removeprefix'} = $collectcfg->{'removeprefix'};
325    }
326   
327    if (defined $collectcfg->{'removesuffix'} && $inexport->{'removesuffix'} eq "") {
328    $inexport->{'removesuffix'} = $collectcfg->{'removesuffix'};
329    }
330
331    $inexport->set_collection_options($collectcfg);
332 
333}
334&main();
Note: See TracBrowser for help on using the browser.