source: main/trunk/greenstone2/perllib/plugins/BasePlugin.pm@ 31445

Last change on this file since 31445 was 31445, checked in by ak19, 7 years ago

added a method guessing_filesystem_encoding. use this to try and work out what encoding to use when turning import folder subdirs into perl native unicode strings
  • Property svn:keywords set to Author Date Id Revision
File size: 48.3 KB
Line 
1###########################################################################
2#
3# BasePlugin.pm -- base class for all the import plugins
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) 1999-2005 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
26package BasePlugin;
27
28use strict;
29no strict 'subs';
30no strict 'refs'; # allow filehandles to be variables and viceversa
31
32use File::Basename;
33use Encode;
34use Unicode::Normalize 'normalize';
35
36use encodings;
37use unicode;
38use doc;
39use ghtml;
40use gsprintf 'gsprintf';
41use util;
42use FileUtils;
43
44use PrintInfo;
45
46BEGIN {
47 @BasePlugin::ISA = ( 'PrintInfo' );
48}
49
50# the different methods that can be applied when renaming
51# imported documents and their associated files
52our $file_rename_method_list =
53 [ { 'name' => "url",
54 'desc' => "{BasePlugin.rename_method.url}" },
55 { 'name' => "base64",
56 'desc' => "{BasePlugin.rename_method.base64}" },
57 { 'name' => "none",
58 'desc' => "{BasePlugin.rename_method.none}",
59 'hiddengli' => "yes" } ];
60
61our $encoding_list =
62 [ { 'name' => "ascii",
63 'desc' => "{BasePlugin.encoding.ascii}" },
64 { 'name' => "utf8",
65 'desc' => "{BasePlugin.encoding.utf8}" },
66 { 'name' => "unicode",
67 'desc' => "{BasePlugin.encoding.unicode}" } ];
68
69
70my $e = $encodings::encodings;
71foreach my $enc (sort {$e->{$a}->{'name'} cmp $e->{$b}->{'name'}} keys (%$e))
72{
73 my $hashEncode =
74 {'name' => $enc,
75 'desc' => $e->{$enc}->{'name'}};
76
77 push(@{$encoding_list},$hashEncode);
78}
79
80our $encoding_plus_auto_list =
81 [ { 'name' => "auto",
82 'desc' => "{BasePlugin.filename_encoding.auto}" },
83 { 'name' => "auto-language-analysis",
84 'desc' => "{BasePlugin.filename_encoding.auto_language_analysis}" }, # textcat
85 { 'name' => "auto-filesystem-encoding",
86 'desc' => "{BasePlugin.filename_encoding.auto_filesystem_encoding}" }, # locale
87 { 'name' => "auto-fl",
88 'desc' => "{BasePlugin.filename_encoding.auto_fl}" }, # locale followed by textcat
89 { 'name' => "auto-lf",
90 'desc' => "{BasePlugin.filename_encoding.auto_lf}" } ]; # texcat followed by locale
91
92push(@{$encoding_plus_auto_list},@{$encoding_list});
93
94our $oidtype_list =
95 [ { 'name' => "auto",
96 'desc' => "{BasePlugin.OIDtype.auto}" },
97 { 'name' => "hash",
98 'desc' => "{import.OIDtype.hash}" },
99 { 'name' => "hash_on_ga_xml",
100 'desc' => "{import.OIDtype.hash_on_ga_xml}" },
101 { 'name' => "hash_on_full_filename",
102 'desc' => "{import.OIDtype.hash_on_full_filename}" },
103 { 'name' => "assigned",
104 'desc' => "{import.OIDtype.assigned}" },
105 { 'name' => "incremental",
106 'desc' => "{import.OIDtype.incremental}" },
107 { 'name' => "filename",
108 'desc' => "{import.OIDtype.filename}" },
109 { 'name' => "dirname",
110 'desc' => "{import.OIDtype.dirname}" },
111 { 'name' => "full_filename",
112 'desc' => "{import.OIDtype.full_filename}" } ];
113
114my $arguments =
115 [ { 'name' => "process_exp",
116 'desc' => "{BasePlugin.process_exp}",
117 'type' => "regexp",
118 'deft' => "",
119 'reqd' => "no" },
120 { 'name' => "no_blocking",
121 'desc' => "{BasePlugin.no_blocking}",
122 'type' => "flag",
123 'reqd' => "no"},
124 { 'name' => "block_exp",
125 'desc' => "{BasePlugin.block_exp}",
126 'type' => "regexp",
127 'deft' => "",
128 'reqd' => "no" },
129 { 'name' => "store_original_file",
130 'desc' => "{BasePlugin.store_original_file}",
131 'type' => "flag",
132 'reqd' => "no" },
133 { 'name' => "associate_ext",
134 'desc' => "{BasePlugin.associate_ext}",
135 'type' => "string",
136 'reqd' => "no" },
137 { 'name' => "associate_tail_re",
138 'desc' => "{BasePlugin.associate_tail_re}",
139 'type' => "string",
140 'reqd' => "no" },
141 { 'name' => "OIDtype",
142 'desc' => "{import.OIDtype}",
143 'type' => "enum",
144 'list' => $oidtype_list,
145 # leave default empty so we can tell if its been set or not - if not set will use option from import.pl
146 'deft' => "auto",
147 'reqd' => "no" },
148 { 'name' => "OIDmetadata",
149 'desc' => "{import.OIDmetadata}",
150 'type' => "metadata",
151 'deft' => "dc.Identifier",
152 'reqd' => "no" },
153# { 'name' => "use_as_doc_identifier",
154# 'desc' => "{BasePlugin.use_as_doc_identifier}",
155# 'type' => "string",
156# 'reqd' => "no" ,
157# 'deft' => "" } ,
158 { 'name' => "no_cover_image",
159 'desc' => "{BasePlugin.no_cover_image}",
160 'type' => "flag",
161 'reqd' => "no" },
162 { 'name' => "filename_encoding",
163 'desc' => "{BasePlugin.filename_encoding}",
164 'type' => "enum",
165 'deft' => "auto",
166 'list' => $encoding_plus_auto_list,
167 'reqd' => "no" },
168 { 'name' => "smart_block",
169 'desc' => "{common.deprecated}. {BasePlugin.smart_block}",
170 'type' => "flag",
171 'reqd' => "no",
172 'hiddengli' => "yes" }, # deprecated, but leave in for old collections
173 { 'name' => "file_rename_method",
174 'desc' => "{BasePlugin.file_rename_method}",
175 'type' => "enum",
176 'deft' => &get_default_file_rename_method(), # by default rename imported files and assoc files using this encoding
177 'list' => $file_rename_method_list,
178 'reqd' => "no"
179 }
180
181 ];
182
183
184my $options = { 'name' => "BasePlugin",
185 'desc' => "{BasePlugin.desc}",
186 'abstract' => "yes",
187 'inherits' => "no",
188 'args' => $arguments };
189
190sub new {
191
192 my ($class) = shift (@_);
193 my ($pluginlist,$inputargs,$hashArgOptLists,$auxiliary) = @_;
194 push(@$pluginlist, $class);
195
196 push(@{$hashArgOptLists->{"ArgList"}},@{$arguments});
197 push(@{$hashArgOptLists->{"OptList"}},$options);
198
199 my $self = new PrintInfo($pluginlist, $inputargs, $hashArgOptLists,$auxiliary);
200
201 if ($self->{'info_only'}) {
202 # don't worry about any options etc
203 return bless $self, $class;
204 }
205
206 if ($self->{'smart_block'}) {
207 print STDERR "WARNING: -smart_block option has been deprecated and is no longer useful\n";
208 }
209 $self->{'smart_block'} = undef;
210
211 my $plugin_name = (defined $pluginlist->[0]) ? $pluginlist->[0] : $class;
212 $self->{'plugin_type'} = $plugin_name;
213
214 # remove ex. from OIDmetadata iff it's the only namespace prefix
215 $self->{'OIDmetadata'} =~ s/^ex\.([^.]+)$/$1/ if defined $self->{'OIDmetadata'};
216 $self->{'num_processed'} = 0;
217 $self->{'num_not_processed'} = 0;
218 $self->{'num_blocked'} = 0;
219 $self->{'num_archives'} = 0;
220 $self->{'cover_image'} = 1; # cover image is on by default
221 $self->{'cover_image'} = 0 if ($self->{'no_cover_image'});
222 $self->{'can_process_directories'} = 0;
223 #$self->{'option_list'} = $hashArgOptLists->{"OptList"};
224
225 my $associate_ext = $self->{'associate_ext'};
226 if ((defined $associate_ext) && ($associate_ext ne "")) {
227
228 my $associate_tail_re = $self->{'associate_tail_re'};
229 if ((defined $associate_tail_re) && ($associate_tail_re ne "")) {
230 my $outhandle = $self->{'outhandle'};
231 print $outhandle "Warning: can only specify 'associate_ext' or 'associate_tail_re'\n";
232 print $outhandle " defaulting to 'associate_tail_re'\n";
233 }
234 else {
235 my @exts = split(/,/,$associate_ext);
236
237 my @exts_bracketed = map { $_ = "(?:\\.$_)" } @exts;
238 my $associate_tail_re = join("|",@exts_bracketed);
239 $self->{'associate_tail_re'} = $associate_tail_re;
240 }
241
242 delete $self->{'associate_ext'};
243 }
244
245 return bless $self, $class;
246
247}
248
249sub merge_inheritance
250{
251 my $self = {};
252 my @child_selfs = @_;
253
254 foreach my $child_self (@child_selfs) {
255 foreach my $key (keys %$child_self) {
256 if (defined $self->{$key}) {
257 if ($self->{$key} ne $child_self->{$key}) {
258# print STDERR "Warning: Conflicting value in multiple inheritance for '$key'\n";
259# print STDERR "Existing stored value = $self->{$key}\n";
260# print STDERR "New (child) value = $child_self->{$key}\n";
261# print STDERR "Keeping existing value\n";
262 # Existing value seems to be option specified in collect.cfg
263
264 ### $self->{$key} = $child_self->{$key};
265
266 }
267 else {
268## print STDERR "****Info: Value $self->{$key} for $key already defined through multiple inheritance as the same value\n";
269 }
270
271 }
272 else {
273 $self->{$key} = $child_self->{$key};
274 }
275 }
276 }
277
278 return $self;
279}
280
281# initialize BasePlugin options
282# if init() is overridden in a sub-class, remember to call BasePlugin::init()
283sub init {
284 my $self = shift (@_);
285 my ($verbosity, $outhandle, $failhandle) = @_;
286
287 # verbosity is passed through from the processor
288 $self->{'verbosity'} = $verbosity;
289
290 # as are the outhandle and failhandle
291 $self->{'outhandle'} = $outhandle if defined $outhandle;
292 $self->{'failhandle'} = $failhandle;
293# $self->SUPER::init(@_);
294
295 # set process_exp and block_exp to defaults unless they were
296 # explicitly set
297
298 if ((!$self->is_recursive()) and
299 (!defined $self->{'process_exp'}) || ($self->{'process_exp'} eq "")) {
300
301 $self->{'process_exp'} = $self->get_default_process_exp ();
302 if ($self->{'process_exp'} eq "") {
303 warn ref($self) . " Warning: Non-recursive plugin has no process_exp\n";
304 }
305 }
306
307 if ((!defined $self->{'block_exp'}) || ($self->{'block_exp'} eq "")) {
308 $self->{'block_exp'} = $self->get_default_block_exp ();
309 }
310
311}
312
313sub begin {
314 my $self = shift (@_);
315 my ($pluginfo, $base_dir, $processor, $maxdocs) = @_;
316
317 if ($self->{'OIDtype'} eq "auto") {
318 # hasn't been set in the plugin, use the processor values
319 $self->{'OIDtype'} = $processor->{'OIDtype'};
320 $self->{'OIDmetadata'} = $processor->{'OIDmetadata'};
321 }
322 if ($self->{'OIDtype'} eq "hash") {
323 # should we hash on the file or on the doc xml??
324 $self->{'OIDtype'} = $self->get_oid_hash_type();
325 if ($self->{'OIDtype'} !~ /^(hash_on_file|hash_on_ga_xml)$/) {
326 $self->{'OIDtype'} = "hash_on_file";
327 }
328 }
329}
330
331# This is called once if removeold is set with import.pl. Most plugins will do
332# nothing but if a plugin does any stuff outside of creating doc obj, then
333# it may need to clear something.
334sub remove_all {
335 my $self = shift (@_);
336 my ($pluginfo, $base_dir, $processor, $maxdocs) = @_;
337}
338
339# This is called per document for docs that have been deleted from the
340# collection. Most plugins will do nothing
341# but if a plugin does any stuff outside of creating doc obj, then it may need
342# to clear something.
343sub remove_one {
344 my $self = shift (@_);
345
346 my ($file, $oids, $archivedir) = @_;
347 return 0 if $self->can_process_this_file($file);
348 return undef;
349}
350
351sub end {
352 # potentially called at the end of each plugin pass
353 # import.pl only has one plugin pass, but buildcol.pl has multiple ones
354
355 my ($self) = shift (@_);
356}
357
358sub deinit {
359 # called only once, after all plugin passes have been done
360
361 my ($self) = @_;
362}
363
364# default hashing type is to hash on the original file (or converted file)
365# override this to return hash_on_ga_xml for filetypes where hashing on the
366# file is no good eg video
367sub get_oid_hash_type {
368
369 my $self = shift (@_);
370
371 return "hash_on_file";
372}
373
374
375# this function should be overridden to return 1
376# in recursive plugins
377sub is_recursive {
378 my $self = shift (@_);
379
380 return 0;
381}
382
383sub get_default_block_exp {
384 my $self = shift (@_);
385
386 return "";
387}
388
389sub get_default_process_exp {
390 my $self = shift (@_);
391
392 return "";
393}
394
395
396# rename imported files and assoc files using URL encoding by default
397# as this will work for most plugins and give more legible filenames
398sub get_default_file_rename_method() {
399 my $self = shift (@_);
400 return "url";
401}
402
403# returns this plugin's active (possibly user-selected) file_rename_method
404sub get_file_rename_method() {
405 my $self = shift (@_);
406 my $rename_method = $self->{'file_rename_method'};
407 if($rename_method) {
408 return $rename_method;
409 } else {
410 return $self->get_default_file_rename_method();
411 }
412}
413
414# default implementation is to do nothing
415sub store_block_files {
416
417 my $self =shift (@_);
418 my ($filename_full_path, $block_hash) = @_;
419
420}
421
422# put files to block into hash
423sub use_block_expressions {
424
425 my $self =shift (@_);
426 my ($filename_full_path, $block_hash) = @_;
427
428 $filename_full_path = &util::upgrade_if_dos_filename($filename_full_path);
429
430 if ($self->{'block_exp'} ne "" && $filename_full_path =~ /$self->{'block_exp'}/) {
431 &util::block_filename($block_hash,$filename_full_path);
432 }
433
434}
435
436#default implementation is to block a file with same name as this, but extension jpg or JPG, if cover_images is on.
437sub block_cover_image
438{
439 my $self =shift;
440 my ($filename, $block_hash) = @_;
441
442 $filename = &util::upgrade_if_dos_filename($filename);
443
444 if ($self->{'cover_image'}) {
445 my $coverfile = $filename;
446 $coverfile =~ s/\.[^\\\/\.]+$/\.jpg/;
447
448 #if there is no file extension, coverfile will be the same as filename
449 return if $coverfile eq $filename;
450
451 if (!&FileUtils::fileExists($coverfile)) {
452 $coverfile =~ s/jpg$/JPG/;
453 }
454 if (&FileUtils::fileExists($coverfile)) {
455 &util::block_filename($block_hash,$coverfile);
456 }
457 }
458
459 return;
460}
461
462
463# discover all the files that should be blocked by this plugin
464# check the args ...
465sub file_block_read {
466
467 my $self = shift (@_);
468 my ($pluginfo, $base_dir, $file, $block_hash, $metadata, $gli) = @_;
469 # Keep track of filenames with same root but different extensions
470 # Used to support -associate_ext and the more generalised
471 # -associate_tail_re
472 my ($filename_full_path, $filename_no_path) = &util::get_full_filenames($base_dir, $file);
473
474 if (!-d $filename_full_path) {
475 $block_hash->{'all_files'}->{$file} = 1;
476 }
477
478 my $associate_tail_re = $self->{'associate_tail_re'};
479 if ((defined $associate_tail_re) && ($associate_tail_re ne "")) {
480 my ($file_prefix,$file_ext)
481 = &util::get_prefix_and_tail_by_regex($filename_full_path,$associate_tail_re);
482 if ((defined $file_prefix) && (defined $file_ext)) {
483 my $shared_fileroot = $block_hash->{'shared_fileroot'};
484 if (!defined $shared_fileroot->{$file_prefix}) {
485 my $file_prefix_rec = { 'tie_to' => undef,
486 'exts' => {} };
487 $shared_fileroot->{$file_prefix} = $file_prefix_rec;
488 }
489
490 my $file_prefix_rec = $shared_fileroot->{$file_prefix};
491
492 if ($self->can_process_this_file($filename_full_path) && $file_ext !~ m/.\./) {
493 # This is the document the others should be tied to
494 $file_prefix_rec->{'tie_to'} = $file_ext;
495 }
496 else {
497 if ($file_ext =~ m/$associate_tail_re$/) {
498 # this file should be associated to the main one
499 $file_prefix_rec->{'exts'}->{$file_ext} = 1;
500 }
501 }
502
503 }
504 }
505
506 # check block expressions
507 $self->use_block_expressions($filename_full_path, $block_hash) unless $self->{'no_blocking'};
508
509 # now check whether we are actually processing this
510 if (!-f $filename_full_path || !$self->can_process_this_file($filename_full_path)) {
511 return undef; # can't recognise
512 }
513
514 # if we have a block_exp, then this overrides the normal 'smart' blocking
515 $self->store_block_files($filename_full_path, $block_hash) unless ($self->{'no_blocking'} || $self->{'block_exp'} ne "");
516
517 # block the cover image if there is one
518 if ($self->{'cover_image'}) {
519 $self->block_cover_image($filename_full_path, $block_hash);
520 }
521
522 return 1;
523}
524
525# plugins that rely on more than process_exp (eg XML plugins) can override this method
526sub can_process_this_file {
527 my $self = shift(@_);
528 my ($filename) = @_;
529
530 if (-d $filename && !$self->{'can_process_directories'}) {
531 return 0;
532 }
533
534 if ($self->{'process_exp'} ne "" && $filename =~ /$self->{'process_exp'}/) {
535 return 1;
536 }
537 return 0;
538
539}
540
541# Even if a plugin can extract metadata in its metadata_read pass,
542# make the default return 'undef' so processing of the file continues
543# down the pipeline, so other plugins can also have the opportunity to
544# locate metadata and set it up in the extrametakeys variables that
545# are passed around.
546
547sub can_process_this_file_for_metadata {
548 my $self = shift(@_);
549
550 return undef;
551}
552
553
554# just converts path as is to utf8.
555sub filepath_to_utf8 {
556 my $self = shift (@_);
557 my ($file, $file_encoding) = @_;
558 my $filemeta = $file;
559
560 my $filename_encoding = $self->{'filename_encoding'}; # filename encoding setting
561
562 # Whenever filename-encoding is set to any of the auto settings, we
563 # check if the filename is already in UTF8. If it is, then we're done.
564 if($filename_encoding =~ m/auto/) {
565 if(&unicode::check_is_utf8($filemeta))
566 {
567 $filename_encoding = "utf8";
568 return $filemeta;
569 }
570 }
571
572 # Auto setting, but filename is not utf8
573 if ($filename_encoding eq "auto")
574 {
575 # try textcat
576 $filename_encoding = $self->textcat_encoding($filemeta);
577
578 # check the locale next
579 $filename_encoding = $self->locale_encoding() if $filename_encoding eq "undefined";
580
581
582 # now try the encoding of the document, if available
583 if ($filename_encoding eq "undefined" && defined $file_encoding) {
584 $filename_encoding = $file_encoding;
585 }
586
587 }
588
589 elsif ($filename_encoding eq "auto-language-analysis")
590 {
591 $filename_encoding = $self->textcat_encoding($filemeta);
592
593 # now try the encoding of the document, if available
594 if ($filename_encoding eq "undefined" && defined $file_encoding) {
595 $filename_encoding = $file_encoding;
596 }
597 }
598
599 elsif ($filename_encoding eq "auto-filesystem-encoding")
600 {
601 # try locale
602 $filename_encoding = $self->locale_encoding();
603 }
604
605 elsif ($filename_encoding eq "auto-fl")
606 {
607 # filesystem-encoding (locale) then language-analysis (textcat)
608 $filename_encoding = $self->locale_encoding();
609
610 # try textcat
611 $filename_encoding = $self->textcat_encoding($filemeta) if $filename_encoding eq "undefined";
612
613 # else assume filename encoding is encoding of file content, if that's available
614 if ($filename_encoding eq "undefined" && defined $file_encoding) {
615 $filename_encoding = $file_encoding;
616 }
617 }
618
619 elsif ($filename_encoding eq "auto-lf")
620 {
621 # language-analysis (textcat) then filesystem-encoding (locale)
622 $filename_encoding = $self->textcat_encoding($filemeta);
623
624 # guess filename encoding from encoding of file content, if available
625 if ($filename_encoding eq "undefined" && defined $file_encoding) {
626 $filename_encoding = $file_encoding;
627 }
628
629 # try locale
630 $filename_encoding = $self->locale_encoding() if $filename_encoding eq "undefined";
631 }
632
633 # if still undefined, use utf8 as fallback
634 if ($filename_encoding eq "undefined") {
635 $filename_encoding = "utf8";
636 }
637
638 #print STDERR "**** UTF8 encoding the filename $filemeta ";
639
640 # if the filename encoding is set to utf8 but it isn't utf8 already--such as when
641 # 1. the utf8 fallback is used, or 2. if the system locale is used and happens to
642 # be always utf8 (in which case the filename's encoding is also set as utf8 even
643 # though the filename need not be if it originates from another system)--in such
644 # cases attempt to make the filename utf8 to match.
645 if($filename_encoding eq "utf8" && !&unicode::check_is_utf8($filemeta)) {
646 &unicode::ensure_utf8(\$filemeta);
647 }
648
649 # convert non-unicode encodings to utf8
650 if ($filename_encoding !~ m/(?:ascii|utf8|unicode)/) {
651 $filemeta = &unicode::unicode2utf8(
652 &unicode::convert2unicode($filename_encoding, \$filemeta)
653 );
654 }
655
656 #print STDERR " from encoding $filename_encoding -> $filemeta\n";
657 return $filemeta;
658}
659
660# gets the filename with no path, converts to utf8, and then dm safes it.
661# filename_encoding set by user
662sub filename_to_utf8_metadata
663{
664 my $self = shift (@_);
665 my ($file, $file_encoding) = @_;
666
667 my $outhandle = $self->{'outhandle'};
668
669 print $outhandle "****!!!!**** BasePlugin::filename_to_utf8_metadata now deprecated\n";
670 my ($cpackage,$cfilename,$cline,$csubr,$chas_args,$cwantarray) = caller(0);
671 print $outhandle "Calling method: $cfilename:$cline $cpackage->$csubr\n";
672
673 my ($filemeta) = $file =~ /([^\\\/]+)$/; # getting the tail of the filepath (skips all string parts containing slashes upto the end)
674 $filemeta = $self->filepath_to_utf8($filemeta, $file_encoding);
675
676 return $filemeta;
677}
678
679sub locale_encoding {
680 my $self = shift(@_);
681
682 if (!defined $self->{'filesystem_encoding'}) {
683 $self->{'filesystem_encoding'} = $self->get_filesystem_encoding();
684 }
685
686 #print STDERR "*** filename encoding determined based on locale: " . $self->{'filesystem_encoding'} . "\n";
687 return $self->{'filesystem_encoding'}; # can be the string "undefined"
688}
689
690sub textcat_encoding {
691 my $self = shift(@_);
692 my ($filemeta) = @_;
693
694 # analyse filenames without extensions and digits (and trimmed of
695 # surrounding whitespace), so that irrelevant chars don't confuse
696 # textcat
697 my $strictfilemeta = $filemeta;
698 $strictfilemeta =~ s/\.[^\.]+$//g;
699 $strictfilemeta =~ s/\d//g;
700 $strictfilemeta =~ s/^\s*//g;
701 $strictfilemeta =~ s/\s*$//g;
702
703 my $filename_encoding = $self->encoding_from_language_analysis($strictfilemeta);
704 if(!defined $filename_encoding) {
705 $filename_encoding = "undefined";
706 }
707
708 return $filename_encoding; # can be the string "undefined"
709}
710
711# performs textcat
712sub encoding_from_language_analysis {
713 my $self = shift(@_);
714 my ($text) = @_;
715
716 my $outhandle = $self->{'outhandle'};
717 my $best_encoding = undef;
718
719 # get the language/encoding of the textstring using textcat
720 require textcat; # Only load the textcat module if it is required
721 $self->{'textcat'} = new textcat() unless defined($self->{'textcat'});
722 my $results = $self->{'textcat'}->classify_cached_filename(\$text);
723
724
725 if (scalar @$results < 0) {
726 return undef;
727 }
728
729 # We have some results, we choose the first
730 my ($language, $encoding) = $results->[0] =~ /^([^-]*)(?:-(.*))?$/;
731
732 $best_encoding = $encoding;
733 if (!defined $best_encoding) {
734 return undef;
735 }
736
737 if (defined $best_encoding && $best_encoding =~ m/^iso_8859/ && &unicode::check_is_utf8($text)) {
738 # the text is valid utf8, so assume that's the real encoding (since textcat is based on probabilities)
739 $best_encoding = 'utf8';
740 }
741
742
743 # check for equivalents where textcat doesn't have some encodings...
744 # eg MS versions of standard encodings
745 if (defined $best_encoding && $best_encoding =~ /^iso_8859_(\d+)/) {
746 my $iso = $1; # which variant of the iso standard?
747 # iso-8859 sets don't use chars 0x80-0x9f, windows codepages do
748 if ($text =~ /[\x80-\x9f]/) {
749 # Western Europe
750 if ($iso == 1 or $iso == 15) { $best_encoding = 'windows_1252' }
751 elsif ($iso == 2) {$best_encoding = 'windows_1250'} # Central Europe
752 elsif ($iso == 5) {$best_encoding = 'windows_1251'} # Cyrillic
753 elsif ($iso == 6) {$best_encoding = 'windows_1256'} # Arabic
754 elsif ($iso == 7) {$best_encoding = 'windows_1253'} # Greek
755 elsif ($iso == 8) {$best_encoding = 'windows_1255'} # Hebrew
756 elsif ($iso == 9) {$best_encoding = 'windows_1254'} # Turkish
757 }
758 }
759
760 if (defined $best_encoding && $best_encoding !~ /^(ascii|utf8|unicode)$/ &&
761 !defined $encodings::encodings->{$best_encoding})
762 {
763 if ($self->{'verbosity'}) {
764 gsprintf($outhandle, "BasePlugin: {ReadTextFile.unsupported_encoding}\n", $text, $best_encoding, "undef");
765 }
766 $best_encoding = undef;
767 }
768
769 return $best_encoding;
770}
771
772# uses locale
773sub get_filesystem_encoding
774{
775
776 my $self = shift(@_);
777
778 my $outhandle = $self->{'outhandle'};
779 my $filesystem_encoding = undef;
780
781 eval {
782 # Works for Windows as well, returning the DOS code page in use
783 use POSIX qw(locale_h);
784
785 # With only one parameter, setlocale retrieves the
786 # current value
787 my $current_locale = setlocale(LC_CTYPE);
788
789 my $char_encoding = undef;
790 if ($current_locale =~ m/\./) {
791 ($char_encoding) = ($current_locale =~ m/^.*\.(.*?)$/);
792 $char_encoding = lc($char_encoding);
793 }
794 else {
795 if ($current_locale =~ m/^(posix|c)$/i) {
796 $char_encoding = "ascii";
797 }
798 }
799
800 if (defined $char_encoding) {
801 if ($char_encoding =~ m/^(iso)(8859)-?(\d{1,2})$/) {
802 $char_encoding = "$1\_$2\_$3";
803 }
804
805 $char_encoding =~ s/-/_/g;
806 $char_encoding =~ s/^utf_8$/utf8/;
807
808 if ($char_encoding =~ m/^\d+$/) {
809 if (defined $encodings::encodings->{"windows_$char_encoding"}) {
810 $char_encoding = "windows_$char_encoding";
811 }
812 elsif (defined $encodings::encodings->{"dos_$char_encoding"}) {
813 $char_encoding = "dos_$char_encoding";
814 }
815 }
816
817 if (($char_encoding =~ m/(?:ascii|utf8|unicode)/)
818 || (defined $encodings::encodings->{$char_encoding})) {
819 $filesystem_encoding = $char_encoding;
820 }
821 else {
822 print $outhandle "Warning: Unsupported character encoding '$char_encoding' from locale '$current_locale'\n";
823 }
824 }
825
826
827 };
828 if ($@) {
829 print $outhandle "$@\n";
830 print $outhandle "Warning: Unable to establish locale. Will assume filesystem is UTF-8\n";
831
832 }
833
834 return $filesystem_encoding;
835}
836
837
838sub guess_filesystem_encoding
839{
840 my $self = shift (@_);
841 # Look to file system to provide a character encoding
842 my $deduced_filename_encoding = "";
843 # If Windows NTFS, then -- assuming we work with long file names got through
844 # Win32::GetLongFilePath() -- then the underlying file system is UTF16
845
846 if (($ENV{'GSDLOS'} =~ m/^windows$/i) && ($^O ne "cygwin")) {
847 # Can do better than working with the DOS character encoding returned by locale
848 $deduced_filename_encoding = "unicode";
849 }
850 else {
851 # Unix of some form or other
852
853 # See if we can determine the file system encoding through locale
854 $deduced_filename_encoding = $self->locale_encoding();
855
856 }
857 print STDERR "guessing filesystem encoding is $deduced_filename_encoding\n";
858 return $deduced_filename_encoding;
859}
860
861
862
863sub deduce_filename_encoding
864{
865 my $self = shift (@_);
866 my ($file,$metadata,$plugin_filename_encoding) = @_;
867
868 my $gs_filename_encoding = $metadata->{"gs.filenameEncoding"};
869 my $deduced_filename_encoding = undef;
870
871 # Start by looking for manually assigned metadata
872 if (defined $gs_filename_encoding) {
873 if (ref ($gs_filename_encoding) eq "ARRAY") {
874 my $outhandle = $self->{'outhandle'};
875
876 $deduced_filename_encoding = $gs_filename_encoding->[0];
877
878 my $num_vals = scalar(@$gs_filename_encoding);
879 if ($num_vals>1) {
880 print $outhandle "Warning: gs.filenameEncoding multiply defined for $file\n";
881 print $outhandle " Selecting first value: $deduced_filename_encoding\n";
882 }
883 }
884 else {
885 $deduced_filename_encoding = $gs_filename_encoding;
886 }
887 }
888
889 if (!defined $deduced_filename_encoding || ($deduced_filename_encoding =~ m/^\s*$/)) {
890 # Look to see if plugin specifies this value
891
892 if (defined $plugin_filename_encoding) {
893 # First look to see if we're using any of the "older" (i.e. deprecated auto-... plugin options)
894 if ($plugin_filename_encoding =~ m/^auto-.*$/) {
895 my $outhandle = $self->{'outhandle'};
896 print $outhandle "Warning: $plugin_filename_encoding is no longer supported\n";
897 print $outhandle " default to 'auto'\n";
898 $self->{'filename_encoding'} = $plugin_filename_encoding = "auto";
899 }
900
901 if ($plugin_filename_encoding ne "auto") {
902 # We've been given a specific filenamne encoding
903 # => so use it!
904 $deduced_filename_encoding = $plugin_filename_encoding;
905 }
906 }
907 }
908
909 if (!defined $deduced_filename_encoding || ($deduced_filename_encoding =~ m/^\s*$/)) {
910
911 # Look to file system to provide a character encoding
912
913 # If Windows NTFS, then -- assuming we work with long file names got through
914 # Win32::GetLongFilePath() -- then the underlying file system is UTF16
915
916 if (($ENV{'GSDLOS'} =~ m/^windows$/i) && ($^O ne "cygwin")) {
917 # Can do better than working with the DOS character encoding returned by locale
918 $deduced_filename_encoding = "unicode";
919 }
920 else {
921 # Unix of some form or other
922
923 # See if we can determine the file system encoding through locale
924 $deduced_filename_encoding = $self->locale_encoding();
925
926 # if locale shows us filesystem is utf8, check to see filename is consistent
927 # => if not, then we have an "alien" filename on our hands
928
929 if (defined $deduced_filename_encoding && $deduced_filename_encoding =~ m/^utf-?8$/i) {
930 if (!&unicode::check_is_utf8($file)) {
931 # "alien" filename, so revert
932 $deduced_filename_encoding = undef;
933 }
934 }
935 }
936 }
937
938# if (!defined $deduced_filename_encoding || ($deduced_filename_encoding =~ m/^\s*$/)) {
939# # Last chance, apply textcat to deduce filename encoding
940# $deduced_filename_encoding = $self->textcat_encoding($file);
941# }
942
943 if ($self->{'verbosity'}>3) {
944 my $outhandle = $self->{'outhandle'};
945
946 if (defined $deduced_filename_encoding) {
947 print $outhandle " Deduced filename encoding as: $deduced_filename_encoding\n";
948 }
949 else {
950 print $outhandle " No filename encoding deduced\n";
951 }
952 }
953
954 return $deduced_filename_encoding;
955}
956
957
958
959
960# Notionally written to be called once for each document, it is however safe to
961# call multiple times (as in the case of ImagePlugin) which calls this later on
962# after the original image has potentially been converted to a *new* source image
963# format (e.g. TIFF to PNG)
964
965sub set_Source_metadata {
966 my $self = shift (@_);
967 my ($doc_obj, $raw_filename, $filename_encoding, $section) = @_;
968
969 # 1. Sets the filename (Source) for display encoded as Unicode if possible,
970 # and (as a fallback) using %xx if not for non-ascii chars
971 # 2. Sets the url ref (SourceFile) to the URL encoded version
972 # of filename for generated files
973
974 my ($unused_full_rf, $raw_file) = &util::get_full_filenames("", $raw_filename);
975
976 my $this_section = (defined $section)? $section : $doc_obj->get_top_section();
977
978 my $octet_file = $raw_file;
979
980 # UTF-8 version of filename
981# if ((defined $ENV{"DEBUG_UNICODE"}) && ($ENV{"DEBUG_UNICODE"})) {
982# print STDERR "**** Setting Source Metadata given: $octet_file\n";
983# }
984
985 # Deal with (on Windows) raw filenames that are in their
986 # abbreviated DOS form
987
988 if (($ENV{'GSDLOS'} =~ m/^windows$/i) && ($^O ne "cygwin")) {
989 if ((defined $filename_encoding) && ($filename_encoding eq "unicode")) {
990 if (-e $raw_filename) {
991 my $unicode_filename = Win32::GetLongPathName($raw_filename);
992
993 my $unused_full_uf;
994 ($unused_full_uf, $octet_file) = &util::get_full_filenames("", $unicode_filename);
995 }
996 }
997 }
998
999 my $url_encoded_filename;
1000 if ((defined $filename_encoding) && ($filename_encoding ne "ascii")) {
1001 # => Generate a pretty print version of filename that is mapped to Unicode
1002
1003 # Use filename_encoding to map raw filename to a Perl unicode-aware string
1004 $url_encoded_filename = decode($filename_encoding,$octet_file);
1005 }
1006 else {
1007 # otherwise generate %xx encoded version of filename for char > 127
1008 $url_encoded_filename = &unicode::raw_filename_to_url_encoded($octet_file);
1009 }
1010
1011# if ((defined $ENV{"DEBUG_UNICODE"}) && ($ENV{"DEBUG_UNICODE"})) {
1012# print STDERR "****** saving Source as: $url_encoded_filename\n";
1013# }
1014
1015 # In the case of converted files and (generalized) exploded documents, there
1016 # will already be a source filename => store as OrigSource before overriding
1017 my $orig_source = $doc_obj->get_metadata_element ($this_section, "Source");
1018 if ((defined $orig_source) && ($orig_source !~ m/^\s*$/)) {
1019 $doc_obj->set_utf8_metadata_element($this_section, "OrigSource", $orig_source);
1020 }
1021
1022 # Source is the UTF8 display name - not necessarily the name of the file on the system
1023 if ($ENV{'GSDLOS'} =~ m/^darwin$/i) {
1024 # on Darwin want all display strings to be in composed form, then can search on that
1025 $url_encoded_filename = normalize('C', $url_encoded_filename); # Normalisation Form 'C' (composition)
1026 }
1027 # set_utf8_metadata actually sets perl unicode aware strings. not utf8
1028 $doc_obj->set_utf8_metadata_element($this_section, "Source", $url_encoded_filename);
1029
1030
1031 my $renamed_raw_file = &util::rename_file($raw_file, $self->{'file_rename_method'});
1032 # If using URL encoding, then SourceFile is the url-reference to url-encoded
1033 # renamed_raw_url: it's a url that refers to the actual file on the system
1034 # this call just replaces % with %25
1035 my $renamed_raw_url = &unicode::filename_to_url($renamed_raw_file);
1036
1037 $doc_obj->set_utf8_metadata_element($this_section, "SourceFile",
1038 $renamed_raw_url);
1039
1040# if ((defined $ENV{"DEBUG_UNICODE"}) && ($ENV{"DEBUG_UNICODE"})) {
1041# print STDERR "****** saving SourceFile as: $renamed_raw_url\n";
1042# }
1043}
1044
1045# this should be called by all plugins to set the oid of the doc obj, rather
1046# than calling doc_obj->set_OID directly
1047sub add_OID {
1048 my $self = shift (@_);
1049 my ($doc_obj, $force) = @_;
1050
1051 # don't add one if there is one already set, unless we are forced to do so
1052 return unless ($doc_obj->get_OID() =~ /^NULL$/ || $force);
1053 $doc_obj->set_OIDtype($self->{'OIDtype'}, $self->{'OIDmetadata'});
1054
1055 # see if there is a plugin specific set_OID function
1056 if (defined ($self->can('set_OID'))) {
1057 $self->set_OID(@_); # pass through doc_obj and any extra arguments
1058 }
1059 else {
1060 # use the default set_OID() in doc.pm
1061 $doc_obj->set_OID();
1062 }
1063
1064}
1065
1066# The BasePlugin read_into_doc_obj() function. This function does all the
1067# right things to make general options work for a given plugin. It doesn't do anything with the file other than setting reads in
1068# a file and sets up a slew of metadata all saved in doc_obj, which
1069# it then returns as part of a tuple (process_status,doc_obj)
1070#
1071# Much of this functionality used to reside in read, but it was broken
1072# down into a supporting routine to make the code more flexible.
1073#
1074# recursive plugins (e.g. RecPlug) and specialized plugins like those
1075# capable of processing many documents within a single file (e.g.
1076# GMLPlug) will normally want to implement their own version of
1077# read_into_doc_obj()
1078#
1079# Note that $base_dir might be "" and that $file might
1080# include directories
1081
1082# currently blocking has been done before it gets here - does this affect secondary plugin stuff??
1083sub read_into_doc_obj {
1084 my $self = shift (@_);
1085 my ($pluginfo, $base_dir, $file, $block_hash, $metadata, $processor, $maxdocs, $total_count, $gli) = @_;
1086
1087 my $outhandle = $self->{'outhandle'};
1088
1089 # should we move this to read? What about secondary plugins?
1090 my $pp_file = &util::prettyprint_file($base_dir,$file,$gli);
1091 print STDERR "<Processing n='$file' p='$self->{'plugin_type'}'>\n" if ($gli);
1092 print $outhandle "$self->{'plugin_type'} processing $pp_file\n"
1093 if $self->{'verbosity'} > 1;
1094
1095 my ($filename_full_path, $filename_no_path) = &util::get_full_filenames($base_dir, $file);
1096
1097 # create a new document
1098 my $doc_obj = new doc ($filename_full_path, "indexed_doc", $self->{'file_rename_method'});
1099 my $top_section = $doc_obj->get_top_section();
1100
1101 $doc_obj->add_utf8_metadata($top_section, "Plugin", "$self->{'plugin_type'}");
1102 $doc_obj->add_utf8_metadata($top_section, "FileSize", (-s $filename_full_path));
1103
1104
1105 my $plugin_filename_encoding = $self->{'filename_encoding'};
1106 my $filename_encoding = $self->deduce_filename_encoding($file,$metadata,$plugin_filename_encoding);
1107 $self->set_Source_metadata($doc_obj,$filename_full_path,$filename_encoding,$top_section);
1108
1109 # plugin specific stuff - what args do we need here??
1110 unless (defined ($self->process($pluginfo, $base_dir, $file, $metadata, $doc_obj, $gli))) {
1111 print STDERR "<ProcessingError n='$file'>\n" if ($gli);
1112 return -1;
1113 }
1114
1115 # include any metadata passed in from previous plugins
1116 # note that this metadata is associated with the top level section
1117 my $section = $doc_obj->get_top_section();
1118 # can we merge these two methods??
1119 $self->add_associated_files($doc_obj, $filename_full_path);
1120 $self->extra_metadata ($doc_obj, $section, $metadata);
1121 $self->auto_extract_metadata($doc_obj);
1122
1123 # if we haven't found any Title so far, assign one
1124 # this was shifted to here from inside read()
1125 $self->title_fallback($doc_obj,$section,$filename_no_path);
1126
1127 $self->add_OID($doc_obj);
1128
1129 $self->post_process_doc_obj($pluginfo, $base_dir, $file, $metadata, $doc_obj, $gli);
1130 return (1,$doc_obj);
1131}
1132
1133sub post_process_doc_obj {
1134 my $self = shift (@_);
1135 my ($pluginfo, $base_dir, $file, $metadata, $doc_obj, $gli) = @_;
1136
1137 return 1;
1138}
1139
1140sub add_dummy_text {
1141 my $self = shift(@_);
1142 my ($doc_obj, $section) = @_;
1143
1144 # add NoText metadata so we can hide this dummy text in format statements
1145 $doc_obj->add_metadata($section, "NoText", "1");
1146
1147 # lookup_string with extra '1' arg returns perl internal unicode aware text, so we use add_utf8_text so no encoding is done on it.
1148 $doc_obj->add_utf8_text($section, &gsprintf::lookup_string("{BasePlugin.dummy_text}",1));
1149 #$doc_obj->add_text($section, &gsprintf::lookup_string("{BasePlugin.dummy_text}",1));
1150
1151
1152}
1153
1154# does nothing. Can be overridden by subclass
1155sub auto_extract_metadata {
1156 my $self = shift(@_);
1157 my ($doc_obj) = @_;
1158}
1159
1160# adds cover image, associate_file options stuff. Should be called by sub class
1161# read_into_doc_obj
1162sub add_associated_files {
1163 my $self = shift(@_);
1164 # whatis filename??
1165 my ($doc_obj, $filename) = @_;
1166
1167 # add in the cover image
1168 if ($self->{'cover_image'}) {
1169 $self->associate_cover_image($doc_obj, $filename);
1170 }
1171 # store the original (used for eg TextPlugin to store the original for OAI)
1172 if ($self->{'store_original_file'}) {
1173 $self->associate_source_file($doc_obj, $filename);
1174 }
1175
1176
1177}
1178
1179# implement this if you are extracting metadata for other documents
1180sub metadata_read {
1181 my $self = shift (@_);
1182 my ($pluginfo, $base_dir, $file, $block_hash,
1183 $extrametakeys, $extrametadata, $extrametafile,
1184 $processor, $gli, $aux) = @_;
1185
1186 # can we process this file??
1187 my ($filename_full_path, $filename_no_path) = &util::get_full_filenames($base_dir, $file);
1188 return undef unless $self->can_process_this_file_for_metadata($filename_full_path);
1189
1190 return 1; # we recognise the file, but don't actually do anything with it
1191}
1192
1193
1194# The BasePlugin read() function. This function calls read_into_doc_obj()
1195# to ensure all the right things to make general options work for a
1196# given plugin are done. It then calls the process() function which
1197# does all the work specific to a plugin (like the old read functions
1198# used to do). Most plugins should define their own process() function
1199# and let this read() function keep control.
1200#
1201# recursive plugins (e.g. RecPlug) and specialized plugins like those
1202# capable of processing many documents within a single file (e.g.
1203# GMLPlug) might want to implement their own version of read(), but
1204# more likely need to implement their own version of read_into_doc_obj()
1205#
1206# Return number of files processed, undef if can't recognise, -1 if can't
1207# process
1208
1209sub read {
1210 my $self = shift (@_);
1211 my ($pluginfo, $base_dir, $file, $block_hash, $metadata, $processor, $maxdocs, $total_count, $gli) = @_;
1212
1213 # can we process this file??
1214 my ($filename_full_path, $filename_no_path) = &util::get_full_filenames($base_dir, $file);
1215
1216 return undef unless $self->can_process_this_file($filename_full_path);
1217
1218 #print STDERR "**** BEFORE READ INTO DOC OBJ: $file\n";
1219 my ($process_status,$doc_obj) = $self->read_into_doc_obj(@_);
1220 #print STDERR "**** AFTER READ INTO DOC OBJ: $file\n";
1221
1222 if ((defined $process_status) && ($process_status == 1)) {
1223
1224 # process the document
1225 $processor->process($doc_obj);
1226
1227 $self->{'num_processed'} ++;
1228 undef $doc_obj;
1229 }
1230 # delete any temp files that we may have created
1231 $self->clean_up_after_doc_obj_processing();
1232
1233
1234 # if process_status == 1, then the file has been processed.
1235 return $process_status;
1236
1237}
1238
1239# returns undef if file is rejected by the plugin
1240sub process {
1241 my $self = shift (@_);
1242 my ($textref, $pluginfo, $base_dir, $file, $metadata, $doc_obj, $gli) = @_;
1243
1244 gsprintf(STDERR, "BasePlugin::process {common.must_be_implemented}\n");
1245
1246 my ($cpackage,$cfilename,$cline,$csubr,$chas_args,$cwantarray) = caller(1);
1247 print STDERR "Calling method: $cfilename:$cline $cpackage->$csubr\n";
1248
1249 die "\n";
1250
1251 return undef; # never gets here
1252}
1253
1254# overwrite this method to delete any temp files that we have created
1255sub clean_up_after_doc_obj_processing {
1256 my $self = shift(@_);
1257
1258}
1259
1260# write_file -- used by ConvertToPlug, for example in post processing
1261#
1262# where should this go, is here the best place??
1263sub utf8_write_file {
1264 my $self = shift (@_);
1265 my ($textref, $filename) = @_;
1266
1267 if (!open (FILE, ">:utf8", $filename)) {
1268 gsprintf(STDERR, "ConvertToPlug::write_file {ConvertToPlug.could_not_open_for_writing} ($!)\n", $filename);
1269 die "\n";
1270 }
1271 print FILE $$textref;
1272
1273 close FILE;
1274}
1275
1276
1277sub filename_based_title
1278{
1279 my $self = shift (@_);
1280 my ($file) = @_;
1281
1282 my $file_derived_title = $file;
1283 $file_derived_title =~ s/_/ /g;
1284 $file_derived_title =~ s/\.[^.]+$//;
1285
1286 return $file_derived_title;
1287}
1288
1289
1290sub title_fallback
1291{
1292 my $self = shift (@_);
1293 my ($doc_obj,$section,$file) = @_;
1294
1295 if (!defined $doc_obj->get_metadata_element ($section, "Title")
1296 || $doc_obj->get_metadata_element($section, "Title") eq "") {
1297
1298 my $source_file = $doc_obj->get_metadata_element($section, "Source");
1299 my $file_derived_title;
1300 if (defined $source_file) {
1301 $file_derived_title = $self->filename_based_title($source_file);
1302 }
1303 else {
1304 # pp = pretty print
1305 my $pp_file = (defined $source_file) ? $source_file : $file;
1306
1307 my $raw_title = $self->filename_based_title($file);
1308 my $file_derived_title = &unicode::raw_filename_to_url_encoded($raw_title);
1309 }
1310
1311
1312 if (!defined $doc_obj->get_metadata_element ($section, "Title")) {
1313 $doc_obj->add_utf8_metadata ($section, "Title", $file_derived_title);
1314 }
1315 else {
1316 $doc_obj->set_utf8_metadata_element ($section, "Title", $file_derived_title);
1317 }
1318 }
1319
1320}
1321
1322# add any extra metadata that's been passed around from one
1323# plugin to another.
1324# extra_metadata uses add_utf8_metadata so it expects metadata values
1325# to already be in utf8
1326sub extra_metadata {
1327 my $self = shift (@_);
1328 my ($doc_obj, $cursection, $metadata) = @_;
1329
1330 my $associate_tail_re = $self->{'associate_tail_re'};
1331
1332# Sort the extra metadata for diffcol so these meta appear in a consistent order
1333# in doc.xml. Necessary for the ex.PDF.* and ex.File.* meta that's extracted in
1334# the PDFBox collection, as the order of these varies between CentOS and Ubuntu.
1335 foreach my $field (sort keys(%$metadata)) {
1336# foreach my $field (keys(%$metadata)) {
1337 # $metadata->{$field} may be an array reference
1338 if ($field eq "gsdlassocfile_tobe") {
1339 # 'gsdlassocfile_tobe' is artificially introduced metadata
1340 # that is used to signal that certain additional files should
1341 # be tied to this document. Useful in situations where a
1342 # metadata pass in the plugin pipeline works out some files
1343 # need to be associated with a document, but the document hasn't
1344 # been formed yet.
1345 my $equiv_form = "";
1346 foreach my $gaf (@{$metadata->{$field}}) {
1347 my ($full_filename,$mimetype) = ($gaf =~ m/^(.*):(.*):$/);
1348 my ($tail_filename) = ($full_filename =~ /^.*[\/\\](.+?)$/);
1349
1350 # we need to make sure the filename is valid utf-8 - we do
1351 # this by url or base64 encoding it
1352 # $tail_filename is the name that we store the file as
1353 $tail_filename = &util::rename_file($tail_filename, $self->{'file_rename_method'});
1354 $doc_obj->associate_file($full_filename,$tail_filename,$mimetype);
1355 $doc_obj->associate_source_file($full_filename);
1356 # If the filename is url_encoded, we need to encode the % signs
1357 # in the filename, so that it works in a url
1358 my $url_tail_filename = &unicode::filename_to_url($tail_filename);
1359 # work out extended tail extension (i.e. matching tail re)
1360
1361 my ($file_prefix,$file_extended_ext)
1362 = &util::get_prefix_and_tail_by_regex($tail_filename,$associate_tail_re);
1363 my ($pre_doc_ext) = ($file_extended_ext =~ m/^(.*)\..*$/);
1364 my ($doc_ext) = ($tail_filename =~ m/^.*\.(.*)$/);
1365
1366 # the greenstone 2 stuff
1367 my $start_doclink = "<a href=\"_httpprefix_/collect/[collection]/index/assoc/{Or}{[parent(Top):assocfilepath],[assocfilepath]}/$url_tail_filename\">";
1368 #my $start_doclink = "<a href=\'_httpprefix_/collect/[collection]/index/assoc/[assocfilepath]/$url_tail_filename\'>";
1369 my $start_doclink_gs3 = "<a href=\'_httpprefix_/collect/[collection]/index/assoc/[assocfilepath]/$url_tail_filename\'>";
1370
1371 my $srcicon = "_icon".$doc_ext."_";
1372 my $end_doclink = "</a>";
1373
1374 my $assoc_form = "$start_doclink\{If\}{$srcicon,$srcicon,$doc_ext\}$end_doclink";
1375
1376
1377 if (defined $pre_doc_ext && $pre_doc_ext ne "") {
1378 # for metadata such as [mp3._edited] [mp3._full] ...
1379 $doc_obj->add_utf8_metadata ($cursection, "$doc_ext.$pre_doc_ext", $assoc_form);
1380 }
1381
1382 # for multiple metadata such as [mp3.assoclink]
1383 $doc_obj->add_utf8_metadata ($cursection, "$doc_ext.assoclink", $assoc_form);
1384
1385 $equiv_form .= " $assoc_form";
1386
1387 # following are used for greenstone 3,
1388 $doc_obj->add_utf8_metadata ($cursection, "equivDocLink", $start_doclink_gs3);
1389 $doc_obj->add_utf8_metadata ($cursection, "equivDocIcon", $srcicon);
1390 $doc_obj->add_utf8_metadata ($cursection, "/equivDocLink", $end_doclink);
1391
1392 }
1393 $doc_obj->add_utf8_metadata ($cursection, "equivlink", $equiv_form);
1394 }
1395 elsif ($field eq "gsdlzipfilename") {
1396 # special case for when files have come out of a zip. source_path
1397 # (used for archives dbs and keeping track for incremental import)
1398 # must be set to the zip file name
1399 my $zip_filename = $metadata->{$field};
1400 # overwrite the source_path
1401 $doc_obj->set_source_path($zip_filename);
1402 # and set the metadata
1403 $zip_filename = &util::filename_within_collection($zip_filename);
1404 $zip_filename = $doc_obj->encode_filename($zip_filename, $self->{'file_rename_method'});
1405 $doc_obj->add_utf8_metadata ($cursection, $field, $zip_filename);
1406 }
1407 elsif (ref ($metadata->{$field}) eq "ARRAY") {
1408 map {
1409 $doc_obj->add_utf8_metadata ($cursection, $field, $_);
1410 } @{$metadata->{$field}};
1411 } else {
1412 $doc_obj->add_utf8_metadata ($cursection, $field, $metadata->{$field});
1413 }
1414 }
1415}
1416
1417
1418sub compile_stats {
1419 my $self = shift(@_);
1420 my ($stats) = @_;
1421
1422 $stats->{'num_processed'} += $self->{'num_processed'};
1423 $stats->{'num_not_processed'} += $self->{'num_not_processed'};
1424 $stats->{'num_archives'} += $self->{'num_archives'};
1425
1426}
1427sub associate_source_file {
1428 my $self = shift(@_);
1429
1430 my ($doc_obj, $filename) = @_;
1431 my $cursection = $doc_obj->get_top_section();
1432 my $assocfilename = $doc_obj->get_assocfile_from_sourcefile();
1433
1434 $doc_obj->associate_file($filename, $assocfilename, undef, $cursection);
1435 # srclink_file is now deprecated because of the "_" in the metadataname. Use srclinkFile
1436 $doc_obj->add_utf8_metadata ($cursection, "srclink_file", $doc_obj->get_sourcefile());
1437 $doc_obj->add_utf8_metadata ($cursection, "srclinkFile", $doc_obj->get_sourcefile());
1438}
1439
1440sub associate_cover_image {
1441 my $self = shift(@_);
1442 my ($doc_obj, $filename) = @_;
1443
1444 my $upgraded_filename = &util::upgrade_if_dos_filename($filename);
1445
1446 $filename =~ s/\.[^\\\/\.]+$/\.jpg/;
1447 $upgraded_filename =~ s/\.[^\\\/\.]+$/\.jpg/;
1448
1449 if (exists $self->{'covers_missing_cache'}->{$upgraded_filename}) {
1450 # don't stat() for existence e.g. for multiple document input files
1451 # (eg SplitPlug)
1452 return;
1453 }
1454
1455 my $top_section=$doc_obj->get_top_section();
1456
1457 if (&FileUtils::fileExists($upgraded_filename)) {
1458 $doc_obj->associate_source_file($filename);
1459 $doc_obj->associate_file($filename, "cover.jpg", "image/jpeg");
1460 $doc_obj->add_utf8_metadata($top_section, "hascover", 1);
1461 } else {
1462 my $upper_filename = $filename;
1463 my $upgraded_upper_filename = $upgraded_filename;
1464
1465 $upper_filename =~ s/jpg$/JPG/;
1466 $upgraded_upper_filename =~ s/jpg$/JPG/;
1467
1468 if (&FileUtils::fileExists($upgraded_upper_filename)) {
1469 $doc_obj->associate_source_file($upper_filename);
1470 $doc_obj->associate_file($upper_filename, "cover.jpg",
1471 "image/jpeg");
1472 $doc_obj->add_utf8_metadata($top_section, "hascover", 1);
1473 } else {
1474 # file doesn't exist, so record the fact that it's missing so
1475 # we don't stat() again (stat is slow)
1476 $self->{'covers_missing_cache'}->{$upgraded_filename} = 1;
1477 }
1478 }
1479
1480}
1481
1482
1483# Overridden by exploding plugins (eg. ISISPlug)
1484sub clean_up_after_exploding
1485{
1486 my $self = shift(@_);
1487}
1488
1489
1490
14911;
Note: See TracBrowser for help on using the repository browser.