source: gsdl/trunk/bin/script/gsConvert.pl@ 15120

Last change on this file since 15120 was 15120, checked in by ak19, 16 years ago

Changes to work with new script replace_srcdoc_with_html.pl: added to subroutine doc_to_html, created new subs make_links_to_assocdir_relative and is_dir_empty
  • Property svn:executable set to *
  • Property svn:keywords set to Author Date Id Revision
File size: 44.8 KB
Line 
1#!/usr/bin/perl -w
2
3###########################################################################
4#
5# gsConvert.pl -- convert documents to HTML or TEXT format
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) 1999-2002 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# gsConvert.pl converts documents in a range of formats to HTML or TEXT
30# by exploiting third-party programs. The sources of these are usually found
31# in the $GSDLHOME/packages directory, and the executables should live in
32# $GSDLHOME/bin/$GSDLOS (which is on the search path).
33#
34# Currently, we can convert the following formats by using external
35# conversion utilities:
36# Microsoft Word (versions 2,6,7 [==95?], 8[==97?], 9[==2000?]), RTF,
37# Adobe PDF, PostScript, MS PowerPoint (95 and 97), and MS Excel (95 and 97).
38#
39# We can try to convert any file to text with a perl implementation of the
40# UNIX strings command.
41#
42# We try to convert Postscript files to text using "gs" which is often on
43# *nix machines. We fall back to performing weak text extraction by using
44# regular expressions.
45
46BEGIN {
47 die "GSDLHOME not set\n" unless defined $ENV{'GSDLHOME'};
48 unshift (@INC, "$ENV{'GSDLHOME'}/perllib");
49}
50
51use parsargv;
52use util;
53use Cwd;
54use File::Basename;
55
56# Are we running on WinNT or Win2000 (or later)?
57my $is_winnt_2000=eval {require Win32; return (Win32::IsWinNT()); return 0;};
58if (!defined($is_winnt_2000)) {$is_winnt_2000=0;}
59
60my $use_strings;
61my $pdf_complex;
62my $pdf_nohidden;
63my $pdf_zoom;
64my $pdf_ignore_images;
65my $pdf_allow_images_only;
66my $windows_scripting;
67
68sub print_usage
69{
70 print STDERR "\n";
71 print STDERR "gsConvert.pl: Converts documents in a range of formats to html\n";
72 print STDERR " or text using third-party programs.\n\n";
73 print STDERR " usage: $0 [options] filename\n";
74 print STDERR " options:\n\t-type\tdoc|dot|pdf|ps|ppt|rtf|xls\t(input file type)\n";
75 print STDERR "\t-errlog\t<filename>\t(append err messages)\n";
76 print STDERR "\t-output\tauto|html|text|pagedimg-jpg|pagedimg-gif|pagedimg-png\t(output file type)\n";
77 print STDERR "\t-timeout\t<max cpu seconds>\t(ulimit on unix systems)\n";
78 print STDERR "\t-use_strings\tuse strings to extract text if conversion fails\n";
79 print STDERR "\t-windows_scripting\tuse windows script when converting Microsoft Word and PPT via VB script\n";
80 print STDERR "\t-pdf_complex\tuse complex output when converting PDF to HTML\n";
81 print STDERR "\t-pdf_nohidden\tDon't attempt to extract hidden text from PDF files\n";
82 print STDERR "\t-pdf_ignore_images\tdon't attempt to extract images when\n";
83 print STDERR "\t\tconverting PDF to HTML\n";
84 print STDERR "\t-pdf_allow_images_only\tallow images only (continue even if no text is present when converting to HTML)\n";
85 print STDERR "\t-pdf_zoom\tfactor by which to zoom PDF (only useful if\n";
86 print STDERR "\t\t-pdf_complex is set\n";
87 exit(1);
88}
89
90my $faillogfile="";
91my $timeout=0;
92
93sub main
94{
95 my (@ARGV) = @_;
96 my ($input_type,$output_type,$verbose);
97
98 # read command-line arguments
99 if (!parsargv::parse(\@ARGV,
100 'type/(doc|dot|pdf|ps|ppt|rtf|xls)/', \$input_type,
101 '/errlog/.*/', \$faillogfile,
102 'output/(auto|html|text|pagedimg).*/', \$output_type,
103 'timeout/\d+/0',\$timeout,
104 'verbose/\d+/0', \$verbose,
105 'use_strings', \$use_strings,
106 'windows_scripting',\$windows_scripting,
107 'pdf_complex', \$pdf_complex,
108 'pdf_ignore_images', \$pdf_ignore_images,
109 'pdf_allow_images_only', \$pdf_allow_images_only,
110 'pdf_nohidden', \$pdf_nohidden,
111 'pdf_zoom/\d+/2', \$pdf_zoom
112 ))
113 {
114 print_usage();
115 }
116
117 # Make sure the input file exists and can be opened for reading
118 if (scalar(@ARGV!=1)) {
119 print_usage();
120 }
121
122 my $input_filename = $ARGV[0];
123 if (!-r $input_filename) {
124 print STDERR "Error: unable to open $input_filename for reading\n";
125 exit(1);
126 }
127
128 # Deduce filenames
129 my ($tailname,$dirname,$suffix)
130 = File::Basename::fileparse($input_filename, "\\.[^\\.]+\$");
131 my $output_filestem = &util::filename_cat($dirname, "$tailname");
132
133 if ($input_type eq "")
134 {
135 $input_type = lc (substr($suffix,1,length($suffix)-1));
136 }
137
138 # Change to temporary working directory
139 my $stored_dir = cwd();
140 chdir ($dirname) || die "Unable to change to directory $dirname";
141
142 # Select convert utility
143 if (!defined $input_type) {
144 print STDERR "Error: No filename extension or input type defined\n";
145 exit(1);
146 }
147 elsif ($input_type eq "doc" || $input_type eq "dot") {
148 print &convertDOC($input_filename, $output_filestem, $output_type);
149 print "\n";
150 }
151 elsif ($input_type eq "rtf") {
152 print &convertRTF($input_filename, $output_filestem, $output_type);
153 print "\n";
154 }
155 elsif ($input_type eq "pdf") {
156 print &convertPDF($dirname, $input_filename, $output_filestem, $output_type);
157 print "\n";
158 }
159 elsif ($input_type eq "ps") {
160 print &convertPS($input_filename, $output_filestem, $output_type);
161 print "\n";
162 }
163 elsif ($input_type eq "ppt") {
164 print &convertPPT($input_filename, $output_filestem, $output_type);
165 print "\n";
166 }
167 elsif ($input_type eq "xls") {
168 print &convertXLS($input_filename, $output_filestem, $output_type);
169 print "\n";
170 }
171 else {
172 print STDERR "Error: Unable to convert type '$input_type'\n";
173 exit(1);
174 }
175
176 # restore to original working directory
177 chdir ($stored_dir) || die "Unable to return to directory $stored_dir";
178
179}
180
181&main(@ARGV);
182
183
184
185# Document-type conversion functions
186#
187# The following functions attempt to convert documents from their
188# input type to the specified output type. If no output type was
189# given, then they first attempt HTML, and then TEXT.
190#
191# Each returns the output type ("html" or "text") or "fail" if no
192# conversion is possible.
193
194# Convert a Microsoft word document
195
196sub convertDOC {
197 ($input_filename, $output_filestem, $output_type) = @_;
198
199 # Many .doc files are not in fact word documents!
200 my $realtype = &find_docfile_type($input_filename);
201
202 if ($realtype eq "word6" || $realtype eq "word7" || $realtype eq "word8") {
203 return &convertWord678($input_filename, $output_filestem, $output_type);
204 } elsif ($realtype eq "rtf") {
205 return &convertRTF($input_filename, $output_filestem, $output_type);
206 } else {
207 return &convertAnything($input_filename, $output_filestem, $output_type);
208 }
209}
210
211# Convert a Microsoft word 6/7/8 document
212
213sub convertWord678 {
214 ($input_filename, $output_filestem, $output_type) = @_;
215
216 my $success = 0;
217 if (!$output_type || ($output_type =~ /html/i)){
218 if ($windows_scripting) {
219 $success = &native_doc_to_html($input_filename, $output_filestem);
220 }
221 else {
222 $success = &doc_to_html($input_filename, $output_filestem);
223 }
224 if ($success) {
225 return "html";
226 }
227 }
228 return &convertAnything($input_filename, $output_filestem, $output_type);
229}
230
231
232# Convert a Rich Text Format (RTF) file
233
234sub convertRTF {
235 ($input_filename, $output_filestem, $output_type) = @_;
236
237 my $success = 0;
238
239 # Attempt specialised conversion to HTML
240 if (!$output_type || ($output_type =~ /html/i)) {
241
242 if ($windows_scripting) {
243 $success = &native_doc_to_html($input_filename, $output_filestem);
244 }
245 else {
246 $success = &rtf_to_html($input_filename, $output_filestem);
247 }
248 if ($success) {
249 return "html";
250 }
251 }
252
253# rtf is so ugly that's it's not worth running strings over.
254# One day I'll write some quick'n'dirty regexps to try to extract text - jrm21
255# return &convertAnything($input_filename, $output_filestem, $output_type);
256 return "fail";
257}
258
259
260# Convert an unidentified file
261
262sub convertAnything {
263 ($input_filename, $output_filestem, $output_type) = @_;
264
265 my $success = 0;
266
267 # Attempt simple conversion to HTML
268 if (!$output_type || ($output_type =~ /html/i)) {
269 $success = &any_to_html($input_filename, $output_filestem);
270 if ($success) {
271 return "html";
272 }
273 }
274
275 # Convert to text
276 if (!$output_type || ($output_type =~ /text/i)) {
277 $success = &any_to_text($input_filename, $output_filestem);
278 if ($success) {
279 return "text";
280 }
281 }
282 return "fail";
283}
284
285
286
287# Convert an Adobe PDF document
288
289sub convertPDF {
290 my ($dirname, $input_filename, $output_filestem, $output_type) = @_;
291
292 my $success = 0;
293 $output_type =~ s/.*\-(.*)/$1/i;
294 # Attempt coversion to Image
295 if ($output_type =~ /jp?g|gif|png/i) {
296 $success = &pdf_to_img($dirname, $input_filename, $output_filestem, $output_type);
297 if ($success){
298 return "item";
299 }
300 }
301
302 # Attempt conversion to HTML
303 if (!$output_type || ($output_type =~ /html/i)) {
304 $success = &pdf_to_html($dirname, $input_filename, $output_filestem);
305 if ($success) {
306 return "html";
307 }
308 }
309
310 # Attempt conversion to TEXT
311 if (!$output_type || ($output_type =~ /text/i)) {
312 $success = &pdf_to_text($dirname, $input_filename, $output_filestem);
313 if ($success) {
314 return "text";
315 }
316 }
317
318 return "fail";
319
320}
321
322
323# Convert an Adobe PostScript document
324
325sub convertPS {
326 ($input_filename, $output_filestem, $output_type) = @_;
327
328 my $success = 0;
329 $output_type =~ s/.*\-(.*)/$1/i;
330 # Attempt coversion to Image
331 if ($output_type =~ /jp?g|gif|png/i) {
332 $success = &ps_to_img($dirname, $input_filename, $output_filestem, $output_type);
333 if ($success){
334 return "item";
335 }
336 }
337
338 # Attempt conversion to TEXT
339 if (!$output_type || ($output_type =~ /text/i)) {
340 $success = &ps_to_text($input_filename, $output_filestem);
341 if ($success) {
342 return "text";
343 }
344 }
345 return "fail";
346}
347
348
349sub convertPPT {
350 my ($input_filename, $output_filestem, $output_type) = @_;
351 my $success = 0;
352
353 my $ppt_convert_type = "";
354 #if (!$output_type || $windows_scripting ||($output_type !~ /html/i) ||($output_type !~ /text/i)){
355 if ($windows_scripting && ($output_type !~ /html/i) && ($output_type !~ /text/i)){
356 if ($output_type =~ /gif/i) {
357 $ppt_convert_type = "-g";
358 } elsif ($output_type =~ /jp?g/i){
359 $ppt_convert_type = "-j";
360 } elsif ($output_type =~ /png/i){
361 $ppt_convert_type = "-p";
362 }
363 my $vbScript = &util::filename_cat($ENV{'GSDLHOME'}, "bin",
364 $ENV{'GSDLOS'}, "pptextract");
365 $vbScript = "pptextract" if ($ENV{'GSDLOS'} =~ /^windows$/i);
366
367 $cmd = "";
368 if ($timeout) {$cmd = "ulimit -t $timeout;";}
369 # if the converting directory has already existed
370 if (-d $output_filestem) {
371 print STDERR "**The conversion directory has existed\n";
372 return "item";
373 } else {
374 $cmd .= "$vbScript $ppt_convert_type \"$input_filename\" \"$output_filestem\"";
375 $cmd .= " 2>\"$output_filestem.err\""
376 if ($ENV{'GSDLOS'} !~ /^windows$/i || $is_winnt_2000);
377 if (system($cmd) !=0) {
378 print STDERR "Powerpoint VB Scripting convert failed\n";
379 } else {
380 return "item";
381 }
382 }
383 } elsif (!$output_type || ($output_type =~ /html/i)) {
384 # Attempt conversion to HTML
385 #if (!$output_type || ($output_type =~ /html/i)) {
386 # formulate the command
387 $cmd = "";
388 $cmd .= "perl -S ppttohtml.pl ";
389 $cmd .= " \"$input_filename\" \"$output_filestem.html\"";
390 $cmd .= " 2>\"$output_filestem.err\""
391 if ($ENV{'GSDLOS'} !~ /^windows$/i || $is_winnt_2000);
392
393 # execute the command
394 $!=0;
395 if (system($cmd)!=0)
396 {
397 print STDERR "Powerpoint 95/97 converter failed $!\n";
398 } else {
399 return "html";
400 }
401 }
402
403 $success = &any_to_text($input_filename, $output_filestem);
404 if ($success) {
405 return "text";
406 }
407
408 return "fail";
409}
410
411
412sub convertXLS {
413 my ($input_filename, $output_filestem, $output_type) = @_;
414
415 my $success = 0;
416
417 # Attempt conversion to HTML
418 if (!$output_type || ($output_type =~ /html/i)) {
419 # formulate the command
420 $cmd = "";
421 $cmd .= "perl -S xlstohtml.pl ";
422 $cmd .= " \"$input_filename\" \"$output_filestem.html\"";
423 $cmd .= " 2>\"$output_filestem.err\""
424 if ($ENV{'GSDLOS'} !~ /^windows$/i || $is_winnt_2000);
425
426
427 # execute the command
428 $!=0;
429 if (system($cmd)!=0)
430 {
431 print STDERR "Excel 95/97 converter failed $!\n";
432 } else {
433 return "html";
434 }
435 }
436
437 $success = &any_to_text($input_filename, $output_filestem);
438 if ($success) {
439 return "text";
440 }
441
442 return "fail";
443}
444
445
446
447# Find the real type of a .doc file
448#
449# We seem to have a lot of files with a .doc extension that are .rtf
450# files or Word 5 files. This function attempts to tell the difference.
451sub find_docfile_type {
452 ($input_filename) = @_;
453
454 open(CHK, "<$input_filename");
455 binmode(CHK);
456 my $line = "";
457 my $first = 1;
458
459 while (<CHK>) {
460
461 $line = $_;
462
463 if ($first) {
464 # check to see if this is an rtf file
465 if ($line =~ /^\{\\rtf/) {
466 close(CHK);
467 return "rtf";
468 }
469 $first = 0;
470 }
471
472 # is this is a word 6/7/8 document?
473 if ($line =~ /Word\.Document\.([678])/) {
474 close(CHK);
475 return "word$1";
476 }
477
478 }
479
480 return "unknown";
481}
482
483
484# Specific type-to-type conversions
485#
486# Each of the following functions attempts to convert a document from
487# a specific format to another. If they succeed they return 1 and leave
488# the output document(s) in the appropriate place; if they fail they
489# return 0 and delete any working files.
490
491
492# Attempt to convert a word document to html with the wv program
493sub doc_to_html {
494 ($input_filename, $output_filestem) = @_;
495
496 my $wvWare = &util::filename_cat($ENV{'GSDLHOME'}, "bin",
497 $ENV{'GSDLOS'}, "wvWare");
498
499 # don't include path on windows (to avoid having to play about
500 # with quoting when GSDLHOME might contain spaces) but assume
501 # that the PATH is set up correctly
502 $wvWare = "wvWare" if ($ENV{'GSDLOS'} =~ /^windows$/i);
503
504 my $wv_conf = &util::filename_cat($ENV{'GSDLHOME'}, "etc",
505 "packages", "wv", "wvHtml.xml");
506
507 # Added the following to work with replace_srcdoc_with_html.pl:
508 # Make wvWare put any associated (image) files of the word doc into
509 # folder docname-without-extention_files. This folder should be at
510 # the same level as the html file generated from the doc.
511 # wvWare will take care of proper interlinking.
512
513 # This step is necessary for replace_srcdoc_with_html.pl which will
514 # move the html and associated files into the import folder. We
515 # want to ensure that the associated files won't overwrite similarly
516 # named items already in import. Hence we put them in a folder first
517 # (to which the html links properly) and that will allow
518 # replace_srcdoc_with_html.pl to move them safely to /import.
519
520 # To do all this, we need to use wvWare's --dir and --basename options
521 # where dir is the full path to the image folder directory and
522 # basename is the full path to the image folder appended to the name
523 # which is to be prepended to every image file:
524 # eg. if the images were to have names like sample0.jpg to sampleN.jpg,
525 # then the basename is "/full/path/to/imgdir/sample".
526 # In this case, basename is the full path to and name of the document.
527 # HOWEVER: basename always takes full path, not relative url, so
528 # the greenstone browser is unable to display the images (absolute paths
529 # cause it to give an "external link" message)
530 # See http://osdir.com/ml/lib.wvware.devel/2002-11/msg00014.html
531 # and http://rpmfind.net/linux/RPM/freshmeat/rpms/wv/wv-0.5.44-1.i386.html
532 # "added --dir option to wvHtml so that pictures can be placed in
533 # a seperate directory"
534 # "running wvWare through IMP to view word documents as html. It gets
535 # invoked like this:
536 # wvWare --dir=/tmp-wvWare --basename=/tmp-wvWare/img$$- $tmp_word >$tmp_output"
537
538 # toppath is the folder where html is generated
539 # docname is the name (without extension) of the html to be generated
540 # suffix (extension) is thrown away
541 my ($docname, $toppath)
542 = &File::Basename::fileparse($input_filename, "\\.[^\\.]+\$");
543
544 # We want the image folder generated to have the same name as windows
545 # would generate ($windows_scripting) when it converts from word to html.
546 # That is, foldername=docname_files
547 my $assoc_dir = &util::filename_cat($toppath, $docname."_files");
548 #print "assoc_dir: ".$assoc_dir."\n"; # same as "$output_filestem._files"
549
550 # ensure this image directory exists
551 # if it exists already, just delete and recreate
552 if(-e $assoc_dir) {
553 &util::rm_r($assoc_dir);
554 }
555 &util::mk_dir($assoc_dir);
556
557 # the images are all going to be called image0, image1,..., imageN
558 my $img_basenames = &util::filename_cat($assoc_dir, $docname);
559
560 #print STDERR "****toppath: $toppath\n****docname: $docname\n;
561 #print STDERR "****img_basenames: $img_basenames\n" if($img_basenames);
562 #print STDERR "****assoc_dir: $assoc_dir\n" if($assoc_dir);
563
564 my $cmd = "";
565 if ($timeout) {$cmd = "ulimit -t $timeout;";}
566 # wvWare's --dir and --basename options for image directory.
567 # Replaced the next line with the *2 lines* following it:
568 # $cmd .= "$wvWare --charset utf-8 --config \"$wv_conf\"";
569 $cmd .= "$wvWare --dir \"$assoc_dir\" --basename \"$img_basenames\"";
570 $cmd .= " --charset utf-8 --config \"$wv_conf\"";
571 $cmd .= " \"$input_filename\" > \"$output_filestem.html\"";
572
573 # redirecting STDERR is a bad idea on windows 95/98
574 $cmd .= " 2> \"$output_filestem.err\""
575 if ($ENV{'GSDLOS'} !~ /^windows$/i || $is_winnt_2000);
576 # execute the command
577 $!=0;
578 if (system($cmd)!=0)
579 {
580 print STDERR "Error executing wv converter:$!\n";
581 if (-s "$output_filestem.err") {
582 open (ERRFILE, "<$output_filestem.err");
583
584 my $write_to_fail_log=0;
585 if ($faillogfile ne "" && defined(open(FAILLOG,">>$faillogfile")))
586 {$write_to_fail_log=1;}
587
588 my $line;
589 while ($line=<ERRFILE>) {
590 if ($line =~ /\w/) {
591 print STDERR "$line";
592 print FAILLOG "$line" if ($write_to_fail_log);
593 }
594 if ($line !~ m/startup error/) {next;}
595 print STDERR " (given an invalid .DOC file?)\n";
596 print FAILLOG " (given an invalid .DOC file?)\n"
597 if ($write_to_fail_log);
598
599 } # while ERRFILE
600 close FAILLOG if ($write_to_fail_log);
601 }
602 return 0; # we can try any_to_text
603 }
604
605 # Was the conversion successful?
606
607 if (-s "$output_filestem.html") { # if file has non-zero size (i.e. it has contents)
608 open(TMP, "$output_filestem.html");
609 $line = <TMP>;
610 close(TMP);
611 if ($line && $line =~ /DOCTYPE HTML/) {
612 &util::rm("$output_filestem.err") if -e "$output_filestem.err";
613
614 # Inserted this code to remove the images directory if it was still empty after
615 # the html was generated (in case there were no images in the word document)
616 if(is_dir_empty($assoc_dir)) {
617 print STDERR "***gsConvert.pl: Image dir $assoc_dir is empty, removing***\n";
618 &util::rm_r($assoc_dir);
619 } else { # there was an image folder (it was generated)
620 # Therefore, the html file generated contains absolute links to the images
621 # If the folder contains images
622 # Replace them with relative links instead, so it can be moved elsewhere
623 make_links_to_assocdir_relative($toppath, $docname, "$output_filestem.html", $assoc_dir, $docname."_files");
624 }
625 return 1;
626 }
627 }
628
629 # If here, an error of some sort occurred
630 &util::rm("$output_filestem.html") if -e "$output_filestem.html";
631 if (-e "$output_filestem.err") {
632 if ($faillogfile ne "" && defined(open(FAILLOG,">>$faillogfile"))) {
633 open (ERRLOG,"$output_filestem.err");
634 while (<ERRLOG>) {print FAILLOG $_;}
635 close FAILLOG;
636 close ERRLOG;
637 }
638 &util::rm("$output_filestem.err");
639 }
640
641 return 0;
642}
643
644# Method to work with doc_to_html - Word docs might contain images.
645# When such word docs are converted with wvWare, we make it generate a
646# <filename>_files folder with the associated images, while the html file
647# <filename> refers to the images using absolute paths to <filename>_files.
648# This method reads in that html file and replaces all the absolute paths to
649# the images in <filename>_files with the relative paths to the images from
650# that folder. (I.e. with <filename>_files/<imagename.ext>).
651sub make_links_to_assocdir_relative{
652 # toppath is the top-level folder in which the html file we're going to be fixing resides
653 # docname is just the name (without extension) of the html file
654 # html_file is the full path to the html file: /full/path/docname.html
655 # assoc_dir_path is toppath/docname_files
656 # assoc_dirname is the directory name of the folder with associated imgs: docname_files
657 my ($toppath, $docname, $html_file, $assoc_dir_path, $assoc_dirname) = @_;
658
659 # 1. Read all the contents of the html into a string
660 # open the original file for reading
661 unless(open(FIN, "<$html_file")) {
662 print STDERR "gsConvert.pl: Unable to open $html_file for reading absolute urls...ERROR\n";
663 return;
664 }
665 # From http://perl.plover.com/local.html
666 # "It's cheaper to read the file all at once, without all the splitting and reassembling.
667 # (Some people call this slurping the file.) Perl has a special feature to support this:
668 # If the $/ variable is undefined, the <...> operator will read the entire file all at once"
669 $/ = undef; # Read entire file at once
670 my $html_contents = <FIN>; # Now file is read in as one single 'line'
671 close(FIN); # close the file
672 print STDERR $html_contents;
673
674 # 2. Replace (substitute) *all* ocurrences of the assoc_dir_path in a hrefs and img src
675 # values with assoc_dirname
676 # At the end: g means substitute all occurrences (global), while s at the end means treat
677 # all new lines as a regular space. This interacts with g to consider all the lines
678 # together as a single line so that multi-occurrences can be replaced.
679
680 # The following regular expression substitution looks for <a or <image, followed by any other
681 # attributes and values until it comes to the FIRST (indicated by ?) href= or src=
682 # followed by " or ' no quotes at all around path, followed by the associated folder's pathname
683 # followed by characters (for the img filename), then finally the optional closing quotes
684 # in " or ' form, followed by any other attributes and values until the first > to end the tag.
685 # The substitution: all the parts preceding associated folder's pathname are retained,
686 # the associated folder path name is replaced by associated folder directory name
687 # and the rest upto and including the closing > tag is retained.
688 # The sg at the end of the pattern match treats all of html_contents as a single line (s)
689 # and performs a global replace (g) meaning that all occurrences that match in that single line
690 # are substituted.
691 $html_contents =~ s/(<(a|img).*?(href|src)=(\"|\')?)$assoc_dir_path(.*?(\"|\')?.*?>)/$1$assoc_dirname$5/sg;
692 #$html_contents =~ s/$assoc_dir_path/$assoc_dirname/gs; # this works, used as fall-back
693
694 # delete the original file and recreate it
695 my $copy_of_filename = $html_file;
696 &util::rm($copy_of_filename); # deleted the file
697
698 # Recreate the original file for writing the updated contents
699 unless(open(FOUT, ">$html_file")) { # open it as a new file for writing
700 print STDERR "gsConvert.pl: Unable to open $html_file for writing relative links...ERROR\n";
701 return;
702 }
703 # write out the updated contents and close the file
704 print FOUT $html_contents;
705 close(FOUT);
706
707}
708
709# A method to check if directory is empty (note that an empty directory still has non-zero size!!!)
710# Code is from http://episteme.arstechnica.com/eve/forums/a/tpc/f/6330927813/m/436007700831
711sub is_dir_empty
712{
713 my ($path) = @_;
714 opendir DIR, $path;
715 while(my $entry = readdir DIR) {
716 next if($entry =~ /^\.\.?$/);
717 closedir DIR;
718 return 0;
719 }
720 closedir DIR;
721 return 1;
722}
723
724
725# Attempt to convert a word document to html with the word2html scripting program
726sub native_doc_to_html {
727 ($input_filename, $output_filestem) = @_;
728
729 my $vbScript = &util::filename_cat($ENV{'GSDLHOME'}, "bin",
730 $ENV{'GSDLOS'}, "word2html");
731
732 $vbScript = "word2html" if ($ENV{'GSDLOS'} =~ /^windows$/i);
733 if (-e "$output_filestem.html") {
734 print STDERR "*** The conversion file has existed\n";
735 return 1;
736 }
737
738 my $cmd = "";
739 if ($timeout) {$cmd = "ulimit -t $timeout;";}
740 #$cmd .= "$vbScript \"$input_filename\" \"$output_filestem.html\"";
741 #$cmd .= "$vbScript $input_filename $output_filestem.html";
742 $cmd .= "$vbScript \"$input_filename\" \"$output_filestem.html\"";
743
744 # redirecting STDERR
745 $cmd .= " 2> \"$output_filestem.err\""
746 if ($ENV {'GSDLOS'} !~ /^windows$/i || $is_winnt_2000);
747
748 # execute the command
749 $!=0;
750 if (system($cmd)!=0)
751 {
752 print STDERR "Error executing word2Html converter:$!\n";
753 if (-s "$output_filestem.err") {
754 open (ERRFILE, "<$output_filestem.err");
755
756 my $write_to_fail_log=0;
757 if ($faillogfile ne "" && defined(open(FAILLOG,">>$faillogfile")))
758 {$write_to_fail_log=1;}
759
760 my $line;
761 while ($line=<ERRFILE>) {
762 if ($line =~ /\w/) {
763 print STDERR "$line";
764 print FAILLOG "$line" if ($write_to_fail_log);
765 }
766 if ($line !~ m/startup error/) {next;}
767 print STDERR " (given an invalid .DOC file?)\n";
768 print FAILLOG " (given an invalid .DOC file?)\n"
769 if ($write_to_fail_log);
770
771 } # while ERRFILE
772 close FAILLOG if ($write_to_fail_log);
773 }
774 return 0; # we can try any_to_text
775 }
776
777 # Was the conversion successful?
778 if (-s "$output_filestem.html") {
779 open(TMP, "$output_filestem.html");
780 $line = <TMP>;
781 close(TMP);
782 if ($line && $line =~ /html/) {
783 &util::rm("$output_filestem.err") if -e "$output_filestem.err";
784 return 1;
785 }
786 }
787
788 # If here, an error of some sort occurred
789 &util::rm("$output_filestem.html") if -e "$output_filestem.html";
790 if (-e "$output_filestem.err") {
791 if ($faillogfile ne "" && defined(open(FAILLOG,">>$faillogfile"))) {
792 open (ERRLOG,"$output_filestem.err");
793 while (<ERRLOG>) {print FAILLOG $_;}
794 close FAILLOG;
795 close ERRLOG;
796 }
797 &util::rm("$output_filestem.err");
798 }
799 return 0;
800}
801
802# Attempt to convert an RTF document to html with rtftohtml
803
804sub rtf_to_html {
805 my ($input_filename, $output_filestem) = @_;
806
807 # formulate the command
808 $cmd = "";
809 if ($timeout) {$cmd = "ulimit -t $timeout;";}
810 $cmd .= "rtftohtml";
811 #$cmd .= "rtf-converter";
812
813 $cmd .= " -o \"$output_filestem.html\" \"$input_filename\"";
814
815 $cmd .= " 2>\"$output_filestem.err\""
816 if ($ENV{'GSDLOS'} !~ /^windows$/i || $is_winnt_2000);
817
818
819 # execute the command
820 $!=0;
821 if (system($cmd)!=0)
822 {
823 print STDERR "Error executing rtf converter $!\n";
824 # don't currently bother printing out error log...
825 # keep going, in case it still created an HTML file...
826 }
827
828 # Was the conversion successful?
829 my $was_successful=0;
830 if (-s "$output_filestem.html") {
831 # make sure we have some content other than header
832 open (HTML, "$output_filestem.html"); # what to do if fail?
833 my $line;
834 my $past_header=0;
835 while ($line=<HTML>) {
836
837 if ($past_header == 0) {
838 if ($line =~ /<body>/) {$past_header=1;}
839 next;
840 }
841
842 $line =~ s/<[^>]+>//g;
843 if ($line =~ /\w/ && $past_header) { # we found some content...
844 $was_successful=1;
845 last;
846 }
847 }
848 close HTML;
849 }
850
851 if ($was_successful) {
852 &util::rm("$output_filestem.err")
853 if (-e "$output_filestem.err");
854 # insert the (modified) table of contents, if it exists.
855 if (-e "${output_filestem}_ToC.html") {
856 &util::mv("$output_filestem.html","$output_filestem.src");
857 my $open_failed=0;
858 open HTMLSRC, "$output_filestem.src" || ++$open_failed;
859 open TOC, "${output_filestem}_ToC.html" || ++$open_failed;
860 open HTML, ">$output_filestem.html" || ++$open_failed;
861
862 if ($open_failed) {
863 close HTMLSRC;
864 close TOC;
865 close HTML;
866 &util::mv("$output_filestem.src","$output_filestem.html");
867 return 1;
868 }
869
870 # print out header info from src html.
871 while (defined($_ = <HTMLSRC>) && $_ =~ /\w/) {
872 print HTML "$_";
873 }
874
875 # print out table of contents, making links relative
876 <TOC>; <TOC>; # ignore first 2 lines
877 print HTML scalar(<TOC>); # line 3 = "<ol>\n"
878 my $line;
879 while ($line=<TOC>) {
880 $line =~ s@</body></html>$@@ ; # only last line has this
881 # make link relative
882 $line =~ s@href=\"[^\#]+@href=\"@;
883 print HTML $line;
884 }
885 close TOC;
886
887 # rest of html src
888 while (<HTMLSRC>) {
889 print HTML $_;
890 }
891 close HTMLSRC;
892 close HTML;
893
894 &util::rm("${output_filestem}_ToC.html");
895 &util::rm("${output_filestem}.src");
896 }
897 # we don't yet do anything with footnotes ($output_filestem_fn.html) :(
898 return 1; # success
899 }
900
901 if (-e "$output_filestem.err") {
902 if ($faillogfile ne "" && defined(open(FAILLOG,">>$faillogfile")))
903 {
904 print FAILLOG "Error - rtftohtml - couldn't extract text\n";
905 #print FAILLOG "Error - rtf-converter - couldn't extract text\n";
906 print FAILLOG " (rtf file might be too recent):\n";
907 open (ERRLOG, "$output_filestem.err");
908 while (<ERRLOG>) {print FAILLOG $_;}
909 close ERRLOG;
910 close FAILLOG;
911 }
912 &util::rm("$output_filestem.err");
913 }
914
915 &util::rm("$output_filestem.html") if (-e "$output_filestem.html");
916
917 return 0;
918}
919
920
921# Convert a pdf file to html with the pdftohtml command
922
923sub pdf_to_html {
924 my ($dirname, $input_filename, $output_filestem) = @_;
925
926 $cmd = "";
927 if ($timeout) {$cmd = "ulimit -t $timeout;";}
928 $cmd .= "perl -S pdftohtml.pl -zoom $pdf_zoom";
929 $cmd .= " -c" if ($pdf_complex);
930 $cmd .= " -i" if ($pdf_ignore_images);
931 $cmd .= " -a" if ($pdf_allow_images_only);
932 $cmd .= " -hidden" unless ($pdf_nohidden);
933 $cmd .= " \"$input_filename\" \"$output_filestem\"";
934
935 if ($ENV{'GSDLOS'} !~ /^windows$/i || $is_winnt_2000) {
936 $cmd .= " > \"$output_filestem.out\" 2> \"$output_filestem.err\"";
937 } else {
938 $cmd .= " > \"$output_filestem.err\"";
939 }
940
941 $!=0;
942
943 my $retval=system($cmd);
944 if ($retval!=0)
945 {
946 print STDERR "Error executing pdftohtml.pl";
947 if ($!) {print STDERR ": $!";}
948 print STDERR "\n";
949 }
950
951 # make sure the converter made something
952 if ($retval!=0 || ! -s "$output_filestem.html")
953 {
954 &util::rm("$output_filestem.out") if (-e "$output_filestem.out");
955 # print out the converter's std err, if any
956 if (-s "$output_filestem.err") {
957 open (ERRLOG, "$output_filestem.err") || die "$!";
958 print STDERR "pdftohtml error log:\n";
959 while (<ERRLOG>) {
960 print STDERR "$_";
961 }
962 close ERRLOG;
963 }
964 &util::rm("$output_filestem.html") if (-e "$output_filestem.html");
965 if (-e "$output_filestem.err") {
966 if ($faillogfile ne "" && defined(open(FAILLOG,">>$faillogfile")))
967 {
968 open (ERRLOG, "$output_filestem.err");
969 while (<ERRLOG>) {print FAILLOG $_;}
970 close ERRLOG;
971 close FAILLOG;
972 }
973 &util::rm("$output_filestem.err");
974 }
975 return 0;
976 }
977
978 &util::rm("$output_filestem.err") if (-e "$output_filestem.err");
979 &util::rm("$output_filestem.out") if (-e "$output_filestem.out");
980 return 1;
981}
982
983# Convert a pdf file to various types of image with the convert command
984
985sub pdf_to_img {
986 my ($dirname, $input_filename, $output_filestem, $output_type) = @_;
987
988 # Check that ImageMagick is installed and available on the path (except for Windows 95/98)
989 if (!($ENV{'GSDLOS'} eq "windows" && !Win32::IsWinNT())) {
990 my $result = `identify 2>&1`;
991 if ($? == -1 || $? == 256) { # Linux and Windows return different values for "program not found"
992 #ImageMagick is not installed, thus the convert utility is not available.
993 print STDERR "*** ImageMagick is not installed, the convert utility is not available\n";
994 return 0;
995 }
996 }
997
998 $cmd = "";
999 if ($timeout) {$cmd = "ulimit -t $timeout;";}
1000 $output_type =~ s/.*\_(.*)/$1/i;
1001 $cmd .= "perl -S pdftoimg.pl -convert_to $output_type \"$input_filename\" \"$output_filestem\"";
1002 if ($ENV{'GSDLOS'} !~ /^windows$/i || $is_winnt_2000) {
1003 $cmd .= " > \"$output_filestem.out\" 2> \"$output_filestem.err\"";
1004 } else {
1005 $cmd .= " > \"$output_filestem.err\"";
1006 }
1007
1008 # don't include path on windows (to avoid having to play about
1009 # with quoting when GSDLHOME might contain spaces) but assume
1010 # that the PATH is set up correctly
1011 $!=0;
1012 my $retval=system($cmd);
1013 if ($retval!=0)
1014 {
1015 print STDERR "Error executing pdftoimg.pl";
1016 if ($!) {print STDERR ": $!";}
1017 print STDERR "\n";
1018 }
1019
1020 #make sure the converter made something
1021 #if ($retval !=0) || ! -s "$output_filestem")
1022 if ($retval !=0)
1023 {
1024 &util::rm("$output_filestem.out") if (-e "$output_filestem.out");
1025 #print out the converter's std err, if any
1026 if (-s "$output_filestem.err") {
1027 open (ERRLOG, "$output_filestem.err") || die "$!";
1028 print STDERR "pdftoimg error log:\n";
1029 while (<ERRLOG>) {
1030 print STDERR "$_";
1031 }
1032 close ERRLOG;
1033 }
1034 #&util::rm("$output_filestem.html") if (-e "$output_filestem.html");
1035 if (-e "$output_filestem.err") {
1036 if ($faillogfile ne "" && defined(open(FAILLOG,">>$faillogfile")))
1037 {
1038 open (ERRLOG, "$output_filestem.err");
1039 while (<ERRLOG>) {print FAILLOG $_;}
1040 close ERRLOG;
1041 close FAILLOG;
1042 }
1043 &util::rm("$output_filestem.err");
1044 }
1045 return 0;
1046 }
1047 &util::rm("$output_filestem.err") if (-e "$output_filestem.err");
1048 &util::rm("$output_filestem.out") if (-e "$output_filestem.out");
1049 return 1;
1050}
1051
1052# Convert a PDF file to text with the pdftotext command
1053
1054sub pdf_to_text {
1055 my ($dirname, $input_filename, $output_filestem) = @_;
1056
1057 my $cmd = "pdftotext \"$input_filename\" \"$output_filestem.text\"";
1058
1059 if ($ENV{'GSDLOS'} !~ /^windows$/i) {
1060 $cmd .= " > \"$output_filestem.out\" 2> \"$output_filestem.err\"";
1061 } else {
1062 $cmd .= " > \"$output_filestem.err\"";
1063 }
1064
1065 if (system($cmd)!=0)
1066 {
1067 print STDERR "Error executing $cmd: $!\n";
1068 &util::rm("$output_filestem.text") if (-e "$output_filestem.text");
1069 }
1070
1071 # make sure there is some extracted text.
1072 if (-e "$output_filestem.text") {
1073 open (EXTR_TEXT, "$output_filestem.text") || warn "open: $!";
1074 binmode(EXTR_TEXT); # just in case...
1075 my $line="";
1076 my $seen_text=0;
1077 while (($seen_text==0) && ($line=<EXTR_TEXT>)) {
1078 if ($line=~ /\w/) {$seen_text=1;}
1079 }
1080 close EXTR_TEXT;
1081 if ($seen_text==0) { # no text was extracted
1082 print STDERR "Error: pdftotext found no text\n";
1083 &util::rm("$output_filestem.text");
1084 }
1085 }
1086
1087 # make sure the converter made something
1088 if (! -s "$output_filestem.text")
1089 {
1090 # print out the converters std err, if any
1091 if (-s "$output_filestem.err") {
1092 open (ERRLOG, "$output_filestem.err") || die "$!";
1093 print STDERR "pdftotext error log:\n";
1094 while (<ERRLOG>) {
1095 print STDERR "$_";
1096 }
1097 close ERRLOG;
1098 }
1099 # does this converter create a .out file?
1100 &util::rm("$output_filestem.out") if (-e "$output_filestem.out");
1101 &util::rm("$output_filestem.text") if (-e "$output_filestem.text");
1102 if (-e "$output_filestem.err") {
1103 if ($faillogfile ne "" && defined(open(FAILLOG,">>$faillogfile")))
1104 {
1105 open (ERRLOG,"$output_filestem.err");
1106 while (<ERRLOG>) {print FAILLOG $_;}
1107 close ERRLOG;
1108 close FAILLOG;
1109 }
1110 &util::rm("$output_filestem.err");
1111 }
1112 return 0;
1113 }
1114 &util::rm("$output_filestem.err") if (-e "$output_filestem.err");
1115 return 1;
1116}
1117
1118# Convert a PostScript document to text
1119# note - just using "ps2ascii" isn't good enough, as it
1120# returns 0 for a postscript interpreter error. ps2ascii is just
1121# a wrapper to "gs" anyway, so we use that cmd here.
1122
1123sub ps_to_text {
1124 my ($input_filename, $output_filestem) = @_;
1125
1126 my $error = "";
1127
1128 # if we're on windows we'll fall straight through without attempting
1129 # to use gs
1130 if ($ENV{'GSDLOS'} =~ /^windows$/i) {
1131 $error = "Windows does not support gs";
1132
1133 } else {
1134 my $cmd = "";
1135 if ($timeout) {$cmd = "ulimit -t $timeout; ";}
1136 $cmd .= "gs -q -dNODISPLAY -dNOBIND -dWRITESYSTEMDICT -dSIMPLE -c save ";
1137 $cmd .= "-f ps2ascii.ps \"$input_filename\" -c quit > \"$output_filestem.text\"";
1138 #$cmd .= "pstotext -output \"$output_filestem.text\" $input_filename\"";
1139 $cmd .= " 2> $output_filestem.err";
1140 $!=0;
1141
1142 my $retcode=system($cmd);
1143 $retcode = $? >> 8; # see man perlfunc - system for this...
1144 # if system returns -1 | 127 (couldn't start program), look at $! for message
1145
1146 if ($retcode!=0) {if ($!) {$error=$!;} else {$error="couldn't run.\n";}}
1147 elsif (! -e "$output_filestem.text") {
1148 $error="did not create output file.\n";
1149 }
1150 else
1151 { # make sure the interpreter didn't get an error. It is technically
1152 # possible for the actual text to start with this, but....
1153 open PSOUT, "$output_filestem.text";
1154 if (<PSOUT> =~ /^Error: (.*)/) {
1155 $error="interpreter error - \"$1\"";
1156 }
1157 close PSOUT;
1158 }
1159 }
1160
1161 if ($error ne "")
1162 {
1163 print STDERR "Warning: Error executing gs: $error\n";
1164 &util::rm("$output_filestem.text") if (-e "$output_filestem.text");
1165
1166 if ("$faillogfile" ne "" && defined(open (FAILLOG, ">>$faillogfile")))
1167 {
1168 print FAILLOG "gs - $error\n";
1169 if (-e "$output_filestem.err") {
1170 open(ERRLOG, "$output_filestem.err");
1171 while (<ERRLOG>) {print FAILLOG $_;}
1172 close ERRLOG;
1173 }
1174 close FAILLOG;
1175 }
1176 &util::rm("$output_filestem.err") if (-e "$output_filestem.err");
1177
1178
1179 # Fine then. We'll just do a lousy job by ourselves...
1180 # Based on 5-line regexp sed script found at:
1181 # http://snark.ptc.spbu.ru/mail-archives/lout/brown/msg00003.html
1182 #
1183 print STDERR "Stripping text from postscript\n";
1184 my $errorcode=0;
1185 open (IN, "$input_filename")
1186 || ($errorcode=1, warn "Couldn't read file: $!");
1187 open (OUT, ">$output_filestem.text")
1188 || ($errorcode=1, warn "Couldn't write file: $!");
1189 if ($errorcode) {print STDERR "errors\n";return 0;}
1190
1191 my $text=""; # this is for whole .ps file...
1192 $text = join('', <IN>); # see man perlport, under "System Resources"
1193 close IN;
1194
1195 # Make sure this is a ps file...
1196 if ($text !~ /^%!/) {
1197 print STDERR "Bad postscript header: not '%!'\n";
1198 if ($faillogfile ne "" && defined(open(FAILLOG, ">>$faillogfile")))
1199 {
1200 print FAILLOG "Bad postscript header: not '%!'\n";
1201 close FAILLOG;
1202 }
1203 return 0;
1204 }
1205
1206 # if ps has Page data, then use it to delete all stuff before it.
1207 $text =~ s/^.*?%%Page:.*?\n//s; # treat string as single line
1208
1209 # remove all leading non-data stuff
1210 $text =~ s/^.*?\(//s;
1211
1212 # remove all newline chars for easier processing
1213 $text =~ s/\n//g;
1214
1215 # Big assumption here - assume that if any co-ordinates are
1216 # given, then we are at the end of a sentence.
1217 $text =~ s/\)-?\d+\ -?\d+/\) \(\n\)/g;
1218
1219 # special characters--
1220 $text =~ s/\(\|\)/\(\ - \)/g; # j -> em-dash?
1221
1222 # ? ps text formatting (eg italics?) ?
1223 $text =~ s/Fn\(f\)/\(\{\)/g; # f -> {
1224 $text =~ s/Fn\(g\)/\(\}\)/g; # g -> }
1225 $text =~ s/Fn\(j\)/\(\|\)/g; # j -> |
1226 # default - remove the rest
1227 $text =~ s/\ ?F.\((.+?)\)/\($1\)/g;
1228
1229 # attempt to add whitespace between words...
1230 # this is based purely on observation, and may be completely wrong...
1231 $text =~ s/([^F])[defghijkuy]\(/$1 \( /g;
1232 # eg I notice "b(" is sometimes NOT a space if preceded by a
1233 # negative number.
1234 $text =~ s/\)\d+ ?b\(/\) \( /g;
1235
1236 # change quoted braces to brackets
1237 $text =~ s/([^\\])\\\(/$1\{/g;
1238 $text =~ s/([^\\])\\\)/$1\}/g ;
1239
1240 # remove everything that is not between braces
1241 $text =~ s/\)([^\(\)])+?\(//sg ;
1242
1243 # remove any Trailer eof stuff.
1244 $text =~ s/\)[^\)]*$//sg;
1245
1246 ### ligatures have special characters...
1247 $text =~ s/\\013/ff/g;
1248 $text =~ s/\\014/fi/g;
1249 $text =~ s/\\015/fl/g;
1250 $text =~ s/\\016/ffi/g;
1251 $text =~ s/\\214/fi/g;
1252 $text =~ s/\\215/fl/g;
1253 $text =~ s/\\017/\n\* /g; # asterisk?
1254 $text =~ s/\\023/\023/g; # e acute ('e)
1255 $text =~ s/\\177/\252/g; # u"
1256# $text =~ s/ ?? /\344/g; # a"
1257
1258 print OUT "$text";
1259 close OUT;
1260 }
1261 # wrap the text - use a minimum length. ie, first space after this length.
1262 my $wrap_length=72;
1263 &util::mv("$output_filestem.text", "$output_filestem.text.tmp");
1264 open INFILE, "$output_filestem.text.tmp" ||
1265 die "Couldn't open file: $!";
1266 open OUTFILE, ">$output_filestem.text" ||
1267 die "Couldn't open file for writing: $!";
1268 my $line="";
1269 while ($line=<INFILE>) {
1270 while (length($line)>0) {
1271 if (length($line)>$wrap_length) {
1272 $line =~ s/^(.{$wrap_length}[^\s]*)\s*//;
1273 print OUTFILE "$1\n";
1274 } else {
1275 print OUTFILE "$line";
1276 $line="";
1277 }
1278 }
1279 }
1280 close INFILE;
1281 close OUTFILE;
1282 &util::rm("$output_filestem.text.tmp");
1283
1284 &util::rm("$output_filestem.err") if (-e "$output_filestem.err");
1285 return 1;
1286}
1287
1288
1289# Convert a PS file to various types of image with the convert utility
1290sub ps_to_img {
1291 my ($dirname, $input_filename, $output_filestem, $output_type) = @_;
1292
1293 # Check that ImageMagick is installed and available on the path (except for Windows 95/98)
1294 if (!($ENV{'GSDLOS'} eq "windows" && !Win32::IsWinNT())) {
1295 my $result = `identify 2>&1`;
1296 if ($? == -1 || $? == 256) { # Linux and Windows return different values for "program not found"
1297 #ImageMagick is not installed, thus the convert utility is not available.
1298 print STDERR "*** ImageMagick is not installed, the convert utility is not available\n";
1299 return 0;
1300 }
1301 }
1302
1303 $cmd = "";
1304 if ($timeout) {$cmd = "ulimit -t $timeout;";}
1305 $output_type =~ s/.*\_(.*)/$1/i;
1306 $cmd .= "perl -S pstoimg.pl -convert_to $output_type \"$input_filename\" \"$output_filestem\"";
1307 if ($ENV{'GSDLOS'} !~ /^windows$/i || $is_winnt_2000) {
1308 $cmd .= " > \"$output_filestem.out\" 2> \"$output_filestem.err\"";
1309 } else {
1310 $cmd .= " > \"$output_filestem.err\"";
1311 }
1312
1313 # don't include path on windows (to avoid having to play about
1314 # with quoting when GSDLHOME might contain spaces) but assume
1315 # that the PATH is set up correctly
1316 $!=0;
1317 my $retval=system($cmd);
1318 if ($retval!=0)
1319 {
1320 print STDERR "Error executing pstoimg.pl";
1321 if ($!) {print STDERR ": $!";}
1322 print STDERR "\n";
1323 }
1324
1325 #make sure the converter made something
1326 #if ($retval !=0) || ! -s "$output_filestem")
1327 if ($retval !=0)
1328 {
1329 &util::rm("$output_filestem.out") if (-e "$output_filestem.out");
1330 #print out the converter's std err, if any
1331 if (-s "$output_filestem.err") {
1332 open (ERRLOG, "$output_filestem.err") || die "$!";
1333 print STDERR "pstoimg error log:\n";
1334 while (<ERRLOG>) {
1335 print STDERR "$_";
1336 }
1337 close ERRLOG;
1338 }
1339 #&util::rm("$output_filestem.html") if (-e "$output_filestem.html");
1340 if (-e "$output_filestem.err") {
1341 if ($faillogfile ne "" && defined(open(FAILLOG,">>$faillogfile")))
1342 {
1343 open (ERRLOG, "$output_filestem.err");
1344 while (<ERRLOG>) {print FAILLOG $_;}
1345 close ERRLOG;
1346 close FAILLOG;
1347 }
1348 &util::rm("$output_filestem.err");
1349 }
1350 return 0;
1351 }
1352 &util::rm("$output_filestem.err") if (-e "$output_filestem.err");
1353 &util::rm("$output_filestem.out") if (-e "$output_filestem.out");
1354 return 1;
1355}
1356
1357# Convert any file to HTML with a crude perl implementation of the
1358# UNIX strings command.
1359
1360sub any_to_html {
1361 ($input_filename, $output_filestem) = @_;
1362
1363 # First generate a text file
1364 return 0 unless (&any_to_text($input_filename, $output_filestem));
1365
1366 # create an HTML file from the text file
1367 open(TEXT, "<$output_filestem.text");
1368 open(HTML, ">$output_filestem.html");
1369
1370 print HTML "<html><head>\n";
1371 print HTML "<META HTTP-EQUIV=\"Content-Type\" CONTENT=\"text/html\">\n";
1372 print HTML "<META NAME=\"GENERATOR\" CONTENT=\"Greenstone any_to_html\">\n";
1373 print HTML "</head><body>\n\n";
1374
1375 my $line;
1376 while ($line=<TEXT>) {
1377 $line =~ s/</&lt;/g;
1378 $line =~ s/>/&gt;/g;
1379 if ($line =~ /^\s*$/) {
1380 print HTML "<p>";
1381 } else {
1382 print HTML "<br> ", $line;
1383 }
1384 }
1385 print HTML "\n</body></html>\n";
1386
1387 close HTML;
1388 close TEXT;
1389
1390 &util::rm("$output_filestem.text") if (-e "$output_filestem.text");
1391 return 1;
1392}
1393
1394# Convert any file to TEXT with a crude perl implementation of the
1395# UNIX strings command.
1396# Note - this assumes ascii charsets :( (jrm21)
1397
1398sub any_to_text {
1399 ($input_filename, $output_filestem) = @_;
1400
1401 if (!$use_strings) {
1402 return 0;
1403 }
1404
1405 print STDERR "\n**** In any to text****\n\n";
1406 open(IN, "<$input_filename") || return 0;
1407 binmode(IN);
1408 open(OUT, ">$output_filestem.text") || return 0;
1409
1410 my ($line);
1411 my $output_line_count = 0;
1412 while (<IN>) {
1413 $line = $_;
1414
1415 # delete anything that isn't a printable character
1416 $line =~ s/[^\040-\176]+/\n/sg;
1417
1418 # delete any string less than 10 characters long
1419 $line =~ s/^.{0,9}$/\n/mg;
1420 while ($line =~ /^.{1,9}$/m) {
1421 $line =~ s/^.{0,9}$/\n/mg;
1422 $line =~ s/\n+/\n/sg;
1423 }
1424
1425 # remove extraneous whitespace
1426 $line =~ s/\n+/\n/gs;
1427 $line =~ s/^\n//gs;
1428
1429 # output whatever is left
1430 if ($line =~ /[^\n ]/) {
1431 print OUT $line;
1432 ++$output_line_count;
1433 }
1434 }
1435
1436 close OUT;
1437 close IN;
1438
1439 if ($output_line_count) { # try to protect against binary only formats
1440 return 1;
1441 }
1442
1443 &util::rm("$output_filestem.text");
1444 return 0;
1445
1446}
Note: See TracBrowser for help on using the repository browser.