source: main/trunk/greenstone2/bin/script/gsConvert.pl@ 24362

Last change on this file since 24362 was 24362, checked in by ak19, 12 years ago

The method of locating perl has changed once more: util now defines the fuction get_perl_exec which is used by other scripts to obtain the path to the perl executable they should use.

  • Property svn:executable set to *
  • Property svn:keywords set to Author Date Id Revision
File size: 46.6 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 strict;
52
53use parsargv;
54use util;
55use Cwd;
56
57# Are we running on WinNT or Win2000 (or later)?
58my $is_winnt_2000=eval {require Win32; return (Win32::IsWinNT()); return 0;};
59if (!defined($is_winnt_2000)) {$is_winnt_2000=0;}
60
61my $use_strings;
62my $pdf_complex;
63my $pdf_nohidden;
64my $pdf_zoom;
65my $pdf_ignore_images;
66my $pdf_allow_images_only;
67my $windows_scripting;
68
69sub print_usage
70{
71 print STDERR "\n";
72 print STDERR "gsConvert.pl: Converts documents in a range of formats to html\n";
73 print STDERR " or text using third-party programs.\n\n";
74 print STDERR " usage: $0 [options] filename\n";
75 print STDERR " options:\n\t-type\tdoc|dot|pdf|ps|ppt|rtf|xls\t(input file type)\n";
76 print STDERR "\t-errlog\t<filename>\t(append err messages)\n";
77 print STDERR "\t-output\tauto|html|text|pagedimg_jpg|pagedimg_gif|pagedimg_png\t(output file type)\n";
78 print STDERR "\t-timeout\t<max cpu seconds>\t(ulimit on unix systems)\n";
79 print STDERR "\t-use_strings\tuse strings to extract text if conversion fails\n";
80 print STDERR "\t-windows_scripting\tuse windows VB script (if available) to convert Microsoft Word and PPT documents\n";
81 print STDERR "\t-pdf_complex\tuse complex output when converting PDF to HTML\n";
82 print STDERR "\t-pdf_nohidden\tDon't attempt to extract hidden text from PDF files\n";
83 print STDERR "\t-pdf_ignore_images\tdon't attempt to extract images when\n";
84 print STDERR "\t\tconverting PDF to HTML\n";
85 print STDERR "\t-pdf_allow_images_only\tallow images only (continue even if no text is present when converting to HTML)\n";
86 print STDERR "\t-pdf_zoom\tfactor by which to zoom PDF (only useful if\n";
87 print STDERR "\t\t-pdf_complex is set\n";
88 exit(1);
89}
90
91my $faillogfile="";
92my $timeout=0;
93
94sub main
95{
96 my (@ARGV) = @_;
97 my ($input_type,$output_type,$verbose);
98
99 # Dynamically figure out what the --type option can support, based on whether -windows_scripting
100 # is in use or not
101 my $default_type_re = "(doc|dot|pdf|ps|ppt|rtf|xls)";
102 #my $enhanced_type_re = "(docx?|dot|pdf|ps|pptx?|rtf|xlsx?)";
103 #my $enhanced_type_re = "(docx?|dot|pdf|ps|pptx?|rtf|xlsx?)";
104 # Currently only have VBA for Word and PPT(but no XLS)
105 my $enhanced_type_re = "(docx?|dot|pdf|ps|pptx?|rtf|xls)";
106
107 my $type_re = $default_type_re;
108
109 foreach my $a (@ARGV) {
110 if ($a =~ m/^windows_scripting$/i) {
111 $type_re = $enhanced_type_re;
112 }
113 }
114
115 # read command-line arguments
116 if (!parsargv::parse(\@ARGV,
117 "type/$type_re/", \$input_type,
118 '/errlog/.*/', \$faillogfile,
119 'output/(auto|html|text|pagedimg).*/', \$output_type,
120 'timeout/\d+/0',\$timeout,
121 'verbose/\d+/0', \$verbose,
122 'windows_scripting',\$windows_scripting,
123 'use_strings', \$use_strings,
124 'pdf_complex', \$pdf_complex,
125 'pdf_ignore_images', \$pdf_ignore_images,
126 'pdf_allow_images_only', \$pdf_allow_images_only,
127 'pdf_nohidden', \$pdf_nohidden,
128 'pdf_zoom/\d+/2', \$pdf_zoom
129 ))
130 {
131 print_usage();
132 }
133
134 # Make sure the input file exists and can be opened for reading
135 if (scalar(@ARGV!=1)) {
136 print_usage();
137 }
138
139 my $input_filename = $ARGV[0];
140 if (!-r $input_filename) {
141 print STDERR "Error: unable to open $input_filename for reading\n";
142 exit(1);
143 }
144
145 # Deduce filenames
146 my ($tailname,$dirname,$suffix)
147 = File::Basename::fileparse($input_filename, "\\.[^\\.]+\$");
148 my $output_filestem = &util::filename_cat($dirname, "$tailname");
149
150 if ($input_type eq "")
151 {
152 $input_type = lc (substr($suffix,1,length($suffix)-1));
153 }
154
155 # Change to temporary working directory
156 my $stored_dir = cwd();
157 chdir ($dirname) || die "Unable to change to directory $dirname";
158
159 # Select convert utility
160 if (!defined $input_type) {
161 print STDERR "Error: No filename extension or input type defined\n";
162 exit(1);
163 }
164 elsif ($input_type =~ m/^docx?$/ || $input_type eq "dot") {
165 print &convertDOC($input_filename, $output_filestem, $output_type);
166 print "\n";
167 }
168 elsif ($input_type eq "rtf") {
169 print &convertRTF($input_filename, $output_filestem, $output_type);
170 print "\n";
171 }
172 elsif ($input_type eq "pdf") {
173 print &convertPDF($dirname, $input_filename, $output_filestem, $output_type);
174 print "\n";
175 }
176 elsif ($input_type eq "ps") {
177 print &convertPS($dirname, $input_filename, $output_filestem, $output_type);
178 print "\n";
179 }
180 elsif ($input_type =~ m/pptx?$/) {
181 print &convertPPT($input_filename, $output_filestem, $output_type);
182 print "\n";
183 }
184 elsif ($input_type =~ m/xlsx?$/) {
185 print &convertXLS($input_filename, $output_filestem, $output_type);
186 print "\n";
187 }
188 else {
189 print STDERR "Error: Unable to convert type '$input_type'\n";
190 exit(1);
191 }
192
193 # restore to original working directory
194 chdir ($stored_dir) || die "Unable to return to directory $stored_dir";
195
196}
197
198&main(@ARGV);
199
200
201
202# Document-type conversion functions
203#
204# The following functions attempt to convert documents from their
205# input type to the specified output type. If no output type was
206# given, then they first attempt HTML, and then TEXT.
207#
208# Each returns the output type ("html" or "text") or "fail" if no
209# conversion is possible.
210
211# Convert a Microsoft word document
212
213sub convertDOC {
214 my ($input_filename, $output_filestem, $output_type) = @_;
215
216 # Many .doc files are not in fact word documents!
217 my $realtype = &find_docfile_type($input_filename);
218
219 if ($realtype eq "word6" || $realtype eq "word7"
220 || $realtype eq "word8" || $realtype eq "docx") {
221 return &convertWord678($input_filename, $output_filestem, $output_type);
222 } elsif ($realtype eq "rtf") {
223 return &convertRTF($input_filename, $output_filestem, $output_type);
224 } else {
225 return &convertAnything($input_filename, $output_filestem, $output_type);
226 }
227}
228
229# Convert a Microsoft word 6/7/8 document
230
231sub convertWord678 {
232 my ($input_filename, $output_filestem, $output_type) = @_;
233
234 my $success = 0;
235 if (!$output_type || ($output_type =~ m/html/i)){
236 if ($windows_scripting) {
237 $success = &native_doc_to_html($input_filename, $output_filestem);
238 }
239 else {
240 $success = &doc_to_html($input_filename, $output_filestem);
241 }
242 if ($success) {
243 return "html";
244 }
245 }
246 return &convertAnything($input_filename, $output_filestem, $output_type);
247}
248
249
250# Convert a Rich Text Format (RTF) file
251
252sub convertRTF {
253 my ($input_filename, $output_filestem, $output_type) = @_;
254
255 my $success = 0;
256
257 # Attempt specialised conversion to HTML
258 if (!$output_type || ($output_type =~ m/html/i)) {
259
260 if ($windows_scripting) {
261 $success = &native_doc_to_html($input_filename, $output_filestem);
262 }
263 else {
264 $success = &rtf_to_html($input_filename, $output_filestem);
265 }
266 if ($success) {
267 return "html";
268 }
269 }
270
271# rtf is so ugly that's it's not worth running strings over.
272# One day I'll write some quick'n'dirty regexps to try to extract text - jrm21
273# return &convertAnything($input_filename, $output_filestem, $output_type);
274 return "fail";
275}
276
277
278# Convert an unidentified file
279
280sub convertAnything {
281 my ($input_filename, $output_filestem, $output_type) = @_;
282
283 my $success = 0;
284
285 # Attempt simple conversion to HTML
286 if (!$output_type || ($output_type =~ m/html/i)) {
287 $success = &any_to_html($input_filename, $output_filestem);
288 if ($success) {
289 return "html";
290 }
291 }
292
293 # Convert to text
294 if (!$output_type || ($output_type =~ m/text/i)) {
295 $success = &any_to_text($input_filename, $output_filestem);
296 if ($success) {
297 return "text";
298 }
299 }
300 return "fail";
301}
302
303
304
305# Convert an Adobe PDF document
306
307sub convertPDF {
308 my ($dirname, $input_filename, $output_filestem, $output_type) = @_;
309
310 my $success = 0;
311 $output_type =~ s/.*\-(.*)/$1/i;
312 # Attempt coversion to Image
313 if ($output_type =~ m/jp?g|gif|png/i) {
314 $success = &pdfps_to_img($dirname, $input_filename, $output_filestem, $output_type);
315 if ($success){
316 return "item";
317 }
318 }
319
320 # Attempt conversion to HTML
321 if (!$output_type || ($output_type =~ m/html/i)) {
322 $success = &pdf_to_html($dirname, $input_filename, $output_filestem);
323 if ($success) {
324 return "html";
325 }
326 }
327
328 # Attempt conversion to TEXT
329 if (!$output_type || ($output_type =~ m/text/i)) {
330 $success = &pdf_to_text($dirname, $input_filename, $output_filestem);
331 if ($success) {
332 return "text";
333 }
334 }
335
336 return "fail";
337
338}
339
340
341# Convert an Adobe PostScript document
342
343sub convertPS {
344 my ($dirname,$input_filename, $output_filestem, $output_type) = @_;
345
346 my $success = 0;
347 $output_type =~ s/.*\-(.*)/$1/i;
348 # Attempt coversion to Image
349 if ($output_type =~ m/jp?g|gif|png/i) {
350 $success = &pdfps_to_img($dirname, $input_filename, $output_filestem, $output_type);
351 if ($success){
352 return "item";
353 }
354 }
355
356 # Attempt conversion to TEXT
357 if (!$output_type || ($output_type =~ m/text/i)) {
358 $success = &ps_to_text($input_filename, $output_filestem);
359 if ($success) {
360 return "text";
361 }
362 }
363 return "fail";
364}
365
366
367sub convertPPT {
368 my ($input_filename, $output_filestem, $output_type) = @_;
369 my $success = 0;
370
371 my $ppt_convert_type = "";
372
373 #if (!$output_type || $windows_scripting || ($output_type !~ m/html/i) || ($output_type !~ m/text/i)){
374 if ($windows_scripting && ($output_type !~ m/html/i) && ($output_type !~ m/text/i)){
375 if ($output_type =~ m/gif/i) {
376 $ppt_convert_type = "-g";
377 } elsif ($output_type =~ m/jp?g/i){
378 $ppt_convert_type = "-j";
379 } elsif ($output_type =~ m/png/i){
380 $ppt_convert_type = "-p";
381 }
382 my $vbScript = &util::filename_cat($ENV{'GSDLHOME'}, "bin",
383 $ENV{'GSDLOS'}, "pptextract");
384 $vbScript = "pptextract" if ($ENV{'GSDLOS'} =~ m/^windows$/i);
385
386 my $cmd = "";
387 if ($timeout) {$cmd = "ulimit -t $timeout;";}
388 # if the converting directory already exists
389 if (-d $output_filestem) {
390 print STDERR "**The conversion directory already exists\n";
391 return "item";
392 } else {
393 $cmd .= "$vbScript $ppt_convert_type \"$input_filename\" \"$output_filestem\"";
394 $cmd .= " 2>\"$output_filestem.err\""
395 if ($ENV{'GSDLOS'} !~ m/^windows$/i || $is_winnt_2000);
396 if (system($cmd) !=0) {
397 print STDERR "Powerpoint VB Scripting convert failed\n";
398 } else {
399 return "item";
400 }
401 }
402 } elsif (!$output_type || ($output_type =~ m/html/i)) {
403 # Attempt conversion to HTML
404 #if (!$output_type || ($output_type =~ m/html/i)) {
405 # formulate the command
406 my $cmd = "";
407 my $full_perl_path = &util::get_perl_exec();
408 $cmd .= "\"$full_perl_path\" -S ppttohtml.pl ";
409 $cmd .= " \"$input_filename\" \"$output_filestem.html\"";
410 $cmd .= " 2>\"$output_filestem.err\""
411 if ($ENV{'GSDLOS'} !~ m/^windows$/i || $is_winnt_2000);
412
413 # execute the command
414 $!=0;
415 if (system($cmd)!=0)
416 {
417 print STDERR "Powerpoint 95/97 converter failed $!\n";
418 } else {
419 return "html";
420 }
421 }
422
423 $success = &any_to_text($input_filename, $output_filestem);
424 if ($success) {
425 return "text";
426 }
427
428 return "fail";
429}
430
431
432sub convertXLS {
433 my ($input_filename, $output_filestem, $output_type) = @_;
434
435 my $success = 0;
436
437 # Attempt conversion to HTML
438 if (!$output_type || ($output_type =~ m/html/i)) {
439 # formulate the command
440 my $cmd = "";
441 my $full_perl_path = &util::get_perl_exec();
442 $cmd .= "\"$full_perl_path\" -S xlstohtml.pl ";
443 $cmd .= " \"$input_filename\" \"$output_filestem.html\"";
444 $cmd .= " 2>\"$output_filestem.err\""
445 if ($ENV{'GSDLOS'} !~ m/^windows$/i || $is_winnt_2000);
446
447
448 # execute the command
449 $!=0;
450 if (system($cmd)!=0)
451 {
452 print STDERR "Excel 95/97 converter failed $!\n";
453 } else {
454 return "html";
455 }
456 }
457
458 $success = &any_to_text($input_filename, $output_filestem);
459 if ($success) {
460 return "text";
461 }
462
463 return "fail";
464}
465
466
467
468# Find the real type of a .doc file
469#
470# We seem to have a lot of files with a .doc extension that are .rtf
471# files or Word 5 files. This function attempts to tell the difference.
472sub find_docfile_type {
473 my ($input_filename) = @_;
474
475 if (($windows_scripting) && ($input_filename =~ m/\.docx$/)) {
476 return "docx";
477 }
478
479 open(CHK, "<$input_filename");
480 binmode(CHK);
481 my $line = "";
482 my $first = 1;
483
484 while (<CHK>) {
485
486 $line = $_;
487
488 if ($first) {
489 # check to see if this is an rtf file
490 if ($line =~ m/^\{\\rtf/) {
491 close(CHK);
492 return "rtf";
493 }
494 $first = 0;
495 }
496
497 # is this is a word 6/7/8 document?
498 if ($line =~ m/Word\.Document\.([678])/) {
499 close(CHK);
500
501 return "word$1";
502 }
503
504 }
505
506 return "unknown";
507}
508
509
510# Specific type-to-type conversions
511#
512# Each of the following functions attempts to convert a document from
513# a specific format to another. If they succeed they return 1 and leave
514# the output document(s) in the appropriate place; if they fail they
515# return 0 and delete any working files.
516
517
518# Attempt to convert a word document to html with the wv program
519sub doc_to_html {
520 my ($input_filename, $output_filestem) = @_;
521
522 my $wvWare = &util::filename_cat($ENV{'GSDLHOME'}, "bin", $ENV{'GSDLOS'}, "wvWare");
523
524 if ( -d "$ENV{'GSDLHOME'}/bin/$ENV{'GSDLOS'}/wv" && $ENV{'GSDLOS'} eq "linux" ) {
525 $ENV{'PATH'} = "$ENV{'GSDLHOME'}/bin/$ENV{'GSDLOS'}/wv/bin:$ENV{'PATH'}";
526 $ENV{'LD_LIBRARY_PATH'} = "$ENV{'GSDLHOME'}/bin/$ENV{'GSDLOS'}/wv/lib:$ENV{'LD_LIBRARY_PATH'}";
527 $wvWare = &util::filename_cat($ENV{'GSDLHOME'}, "bin", $ENV{'GSDLOS'}, "wv", "bin", "wvWare");
528 }
529
530 # don't include path on windows (to avoid having to play about
531 # with quoting when GSDLHOME might contain spaces) but assume
532 # that the PATH is set up correctly
533 $wvWare = "wvWare" if ($ENV{'GSDLOS'} =~ m/^windows$/i);
534
535 my $wv_conf = &util::filename_cat($ENV{'GSDLHOME'}, "etc",
536 "packages", "wv", "wvHtml.xml");
537
538 # Added the following to work with replace_srcdoc_with_html.pl:
539 # Make wvWare put any associated (image) files of the word doc into
540 # folder docname-without-extention_files. This folder should be at
541 # the same level as the html file generated from the doc.
542 # wvWare will take care of proper interlinking.
543
544 # This step is necessary for replace_srcdoc_with_html.pl which will
545 # move the html and associated files into the import folder. We
546 # want to ensure that the associated files won't overwrite similarly
547 # named items already in import. Hence we put them in a folder first
548 # (to which the html links properly) and that will allow
549 # replace_srcdoc_with_html.pl to move them safely to /import.
550
551 # To do all this, we need to use wvWare's --dir and --basename options
552 # where dir is the full path to the image folder directory and
553 # basename is the full path to the image folder appended to the name
554 # which is to be prepended to every image file:
555 # eg. if the images were to have names like sample0.jpg to sampleN.jpg,
556 # then the basename is "/full/path/to/imgdir/sample".
557 # In this case, basename is the full path to and name of the document.
558 # HOWEVER: basename always takes full path, not relative url, so
559 # the greenstone browser is unable to display the images (absolute paths
560 # cause it to give an "external link" message)
561 # See http://osdir.com/ml/lib.wvware.devel/2002-11/msg00014.html
562 # and http://rpmfind.net/linux/RPM/freshmeat/rpms/wv/wv-0.5.44-1.i386.html
563 # "added --dir option to wvHtml so that pictures can be placed in
564 # a seperate directory"
565 # "running wvWare through IMP to view word documents as html. It gets
566 # invoked like this:
567 # wvWare --dir=/tmp-wvWare --basename=/tmp-wvWare/img$$- $tmp_word >$tmp_output"
568
569 # toppath is the folder where html is generated
570 # docname is the name (without extension) of the html to be generated
571 # suffix (extension) is thrown away
572 my ($docname, $toppath)
573 = &File::Basename::fileparse($input_filename, "\\.[^\\.]+\$");
574
575 # We want the image folder generated to have the same name as windows
576 # would generate ($windows_scripting) when it converts from word to html.
577 # That is, foldername=docname_files
578 my $assoc_dir = &util::filename_cat($toppath, $docname."_files");
579 #print "assoc_dir: ".$assoc_dir."\n"; # same as "$output_filestem._files"
580
581 # ensure this image directory exists
582 # if it exists already, just delete and recreate
583 if(-e $assoc_dir) {
584 &util::rm_r($assoc_dir);
585 }
586 &util::mk_dir($assoc_dir);
587
588 # the images are all going to be called image0, image1,..., imageN
589 my $img_basenames = &util::filename_cat($assoc_dir, $docname);
590
591 #print STDERR "****toppath: $toppath\n****docname: $docname\n;
592 #print STDERR "****img_basenames: $img_basenames\n" if($img_basenames);
593 #print STDERR "****assoc_dir: $assoc_dir\n" if($assoc_dir);
594
595 my $cmd = "";
596 if ($timeout) {$cmd = "ulimit -t $timeout;";}
597 # wvWare's --dir and --basename options for image directory.
598 # Replaced the next line with the *2 lines* following it:
599 # $cmd .= "$wvWare --charset utf-8 --config \"$wv_conf\"";
600 $cmd .= "$wvWare --dir \"$assoc_dir\" --basename \"$img_basenames\"";
601 $cmd .= " --charset utf-8 --config \"$wv_conf\"";
602 $cmd .= " \"$input_filename\" > \"$output_filestem.html\"";
603
604 # redirecting STDERR is a bad idea on windows 95/98
605 $cmd .= " 2> \"$output_filestem.err\""
606 if ($ENV{'GSDLOS'} !~ m/^windows$/i || $is_winnt_2000);
607 # execute the command
608 $!=0;
609 if (system($cmd)!=0)
610 {
611 print STDERR "Error executing wv converter:$!\n";
612 if (-s "$output_filestem.err") {
613 open (ERRFILE, "<$output_filestem.err");
614
615 my $write_to_fail_log=0;
616 if ($faillogfile ne "" && defined(open(FAILLOG,">>$faillogfile")))
617 {$write_to_fail_log=1;}
618
619 my $line;
620 while ($line=<ERRFILE>) {
621 if ($line =~ m/\w/) {
622 print STDERR "$line";
623 print FAILLOG "$line" if ($write_to_fail_log);
624 }
625 if ($line !~ m/startup error/) {next;}
626 print STDERR " (given an invalid .DOC file?)\n";
627 print FAILLOG " (given an invalid .DOC file?)\n"
628 if ($write_to_fail_log);
629
630 } # while ERRFILE
631 close FAILLOG if ($write_to_fail_log);
632 }
633 return 0; # we can try any_to_text
634 }
635
636 # Was the conversion successful?
637
638 if (-s "$output_filestem.html") { # if file has non-zero size (i.e. it has contents)
639 open(TMP, "$output_filestem.html");
640 my $line = <TMP>;
641 close(TMP);
642 if ($line && $line =~ m/DOCTYPE HTML/) {
643 &util::rm("$output_filestem.err") if -e "$output_filestem.err";
644
645 # Inserted this code to remove the images directory if it was still empty after
646 # the html was generated (in case there were no images in the word document)
647 if (&util::is_dir_empty($assoc_dir)) {
648 #print STDERR "***gsConvert.pl: Image dir $assoc_dir is empty, removing***\n";
649 &util::rm_r($assoc_dir);
650 } else { # there was an image folder (it was generated)
651 # Therefore, the html file generated contains absolute links to the images
652 # Replace them with relative links instead, so the folder can be moved elsewhere
653 &make_links_to_assocdir_relative($toppath, $docname, "$output_filestem.html", $assoc_dir, $docname."_files");
654 }
655 return 1;
656 }
657 }
658
659 # If here, an error of some sort occurred
660 &util::rm("$output_filestem.html") if -e "$output_filestem.html";
661 if (-e "$output_filestem.err") {
662 if ($faillogfile ne "" && defined(open(FAILLOG,">>$faillogfile"))) {
663 open (ERRLOG,"$output_filestem.err");
664 while (<ERRLOG>) {print FAILLOG $_;}
665 close FAILLOG;
666 close ERRLOG;
667 }
668 &util::rm("$output_filestem.err");
669 }
670
671 return 0;
672}
673
674# Method to work with doc_to_html - Word docs might contain images.
675# When such word docs are converted with wvWare, we make it generate a
676# <filename>_files folder with the associated images, while the html file
677# <filename> refers to the images using absolute paths to <filename>_files.
678# This method reads in that html file and replaces all the absolute paths to
679# the images in <filename>_files with the relative paths to the images from
680# that folder. (I.e. with <filename>_files/<imagename.ext>).
681sub make_links_to_assocdir_relative{
682 # toppath is the top-level folder in which the html file we're going to be fixing resides
683 # docname is just the name (without extension) of the html file
684 # html_file is the full path to the html file: /full/path/docname.html
685 # assoc_dir_path is toppath/docname_files
686 # assoc_dirname is the directory name of the folder with associated imgs: docname_files
687 my ($toppath, $docname, $html_file, $assoc_dir_path, $assoc_dirname) = @_;
688
689 # 1. Read all the contents of the html into a string
690 # open the original file for reading
691 unless(open(FIN, "<$html_file")) {
692 print STDERR "gsConvert.pl: Unable to open $html_file for reading absolute urls...ERROR: $!\n";
693 return 0;
694 }
695 # From http://perl.plover.com/local.html
696 # "It's cheaper to read the file all at once, without all the splitting and reassembling.
697 # (Some people call this slurping the file.) Perl has a special feature to support this:
698 # If the $/ variable is undefined, the <...> operator will read the entire file all at once"
699 my $html_contents;
700 {
701 local $/ = undef; # Read entire file at once
702 $html_contents = <FIN>; # Now file is read in as one single 'line'
703 }
704 close(FIN); # close the file
705 #print STDERR $html_contents;
706
707 # 2. Replace (substitute) *all* ocurrences of the assoc_dir_path in a hrefs and img src
708 # values with assoc_dirname
709 # At the end: g means substitute all occurrences (global), while s at the end means treat
710 # all new lines as a regular space. This interacts with g to consider all the lines
711 # together as a single line so that multi-occurrences can be replaced.
712
713 # we can't just replace $assoc_dir_path with $assoc_dir
714 # $assoc_dir_path represents a regular expression that needs to be replaced
715 # if it contains ., -, [, ], or Windows style backslashes in paths -- which all have special
716 # meaning in Perl regular expressions -- we need to escape these first
717 my $safe_reg_expression = $assoc_dir_path;
718 $safe_reg_expression =~ s/\\/\\\\/g;
719 $safe_reg_expression =~ s/\./\\./g;
720 $safe_reg_expression =~ s/\-/\\-/g;
721 $safe_reg_expression =~ s/\[/\\[/g;
722 $safe_reg_expression =~ s/\]/\\]/g;
723 $safe_reg_expression =~ s/ /%20/g; # wvWare put %20 in place of space, so we need to change our prefix to match
724
725 # The following regular expression substitution looks for <a or <image, followed by any other
726 # attributes and values until it comes to the FIRST (indicated by ?) href= or src=
727 # followed by " or ' no quotes at all around path, followed by the associated folder's pathname
728 # followed by characters (for the img filename), then finally the optional closing quotes
729 # in " or ' form, followed by any other attributes and values until the first > to end the tag.
730 # The substitution: all the parts preceding associated folder's pathname are retained,
731 # the associated folder path name is replaced by associated folder directory name
732 # and the rest upto and including the closing > tag is retained.
733 # The sg at the end of the pattern match treats all of html_contents as a single line (s)
734 # and performs a global replace (g) meaning that all occurrences that match in that single line
735 # are substituted.
736 $html_contents =~ s/(<(a|img).*?(href|src)=(\"|\')?)$safe_reg_expression(.*?(\"|\')?.*?>)/$1$assoc_dirname$5/sg;
737 #$html_contents =~ s/$safe_reg_expression/$assoc_dirname/gs; # this works, used as fall-back
738 # now replace any %20 chars in filenames of href or src attributes to use literal space ' '. Calls a function for this
739 $html_contents =~ s/(<(a|img).*?(href|src)=(\"|\')?)(.*)(.*?(\"|\')?.*?>)/&post_process_assocfile_urls($1, $5, $6)/sge;
740
741 #print STDERR "****assoc_dirname: $assoc_dirname***\n";
742 #print STDERR "****safe_reg_expression: $safe_reg_expression***\n";
743
744 # delete the original file and recreate it
745 my $copy_of_filename = $html_file;
746 &util::rm($copy_of_filename); # deleted the file
747
748 # Recreate the original file for writing the updated contents
749 unless(open(FOUT, ">$html_file")) { # open it as a new file for writing
750 print STDERR "gsConvert.pl: Unable to open $html_file for writing relative links...ERROR: $!\n";
751 return 0;
752 }
753
754 # write out the updated contents and close the file
755 print FOUT $html_contents;
756 close(FOUT);
757 return 1;
758}
759
760# Utility routine to make sure HTML plugin gets img src/href link pathnames that contain
761# url slashes (/) instead of windows-style backwards slashes, and to convert all %20
762# introduced in link pathnames by wvWare into space again. Converts all percent signs
763# introduced by URL encoding filenames generated into %25 in these url links referencing them
764sub post_process_assocfile_urls
765{
766 my ($pre, $text, $post) = @_;
767
768 $text =~ s/%20/ /g; # Convert %20s to space and not underscore since underscores mess with incremental rebuild
769 # $text =~ s/%20/_/g; # reinstated this line, since we no longer replace spaces with %20. We replace them with underscores
770 $text =~ s/\\/\//g;
771 $text =~ s/%/%25/g;
772
773 return "$pre$text$post";
774}
775
776# Attempt to convert a word document to html with the word2html scripting program
777sub native_doc_to_html {
778 my ($input_filename, $output_filestem) = @_;
779
780 # build up the path to the doc-to-html conversion tool we're going to use
781 my $vbScript = &util::filename_cat($ENV{'GSDLHOME'}, "bin", $ENV{'GSDLOS'});
782
783 if ($ENV{'GSDLOS'} =~ m/^windows$/i) {
784 # if windows scripting with docx input, use new VBscript to get the local Word install (if
785 # any) to do the conversion, since docX can't be processed by word2html's windows_scripting
786
787 if($input_filename =~ m/docx$/i) { # need to use full path to docx2html script,
788 # else script launch fails when there are error msgs
789 $vbScript = &util::filename_cat($vbScript, "docx2html.vbs");
790 $vbScript = "CScript //Nologo \"$vbScript\""; # launche with CScript for error output in STDERR
791 # //Nologo flag avoids Microsoft's opening/logo msgs
792 print STDERR "About to use windows scripting to process docx file $input_filename.\n";
793 print STDERR " This may take some time. Please wait...\n";
794 }
795 else { # old doc versions. use the usual VB executable word2html for the
796 # conversion. Doesn't need full path, since bin\windows is on PATH
797 $vbScript = "word2html"; #$vbScript = "\"".&util::filename_cat($vbScript, "word2html")."\"";
798 }
799 }
800 else { # not windows
801 $vbScript = "\"".&util::filename_cat($vbScript, "word2html")."\"";
802 }
803
804 if (-e "$output_filestem.html") {
805 print STDERR " The conversion file:\n";
806 print STDERR " $output_filestem.html\n";
807 print STDERR " ... already exists. Skipping\n";
808 return 1;
809 }
810
811 my $cmd = "";
812 if ($timeout) {$cmd = "ulimit -t $timeout;";}
813 #$cmd .= "$vbScript \"$input_filename\" \"$output_filestem.html\"";
814 #$cmd .= "$vbScript $input_filename $output_filestem.html";
815 $cmd .= "$vbScript \"$input_filename\" \"$output_filestem.html\"";
816
817 # redirecting STDERR
818
819 $cmd .= " 2> \"$output_filestem.err\""
820 if ($ENV {'GSDLOS'} !~ m/^windows$/i || $is_winnt_2000);
821 #print STDERR "@@@@@@@@@ cmd=$cmd\n";
822
823 # execute the command
824 $!=0;
825 if (system($cmd)!=0)
826 {
827 print STDERR "Error executing $vbScript converter:$!\n";
828 if (-s "$output_filestem.err") {
829 open (ERRFILE, "<$output_filestem.err");
830
831 my $write_to_fail_log=0;
832 if ($faillogfile ne "" && defined(open(FAILLOG,">>$faillogfile")))
833 {$write_to_fail_log=1;}
834
835 my $line;
836 while ($line=<ERRFILE>) {
837 if ($line =~ m/\w/) {
838 print STDERR "$line";
839 print FAILLOG "$line" if ($write_to_fail_log);
840 }
841 if ($line !~ m/startup error/) {next;}
842 print STDERR " (given an invalid .DOC file?)\n";
843 print FAILLOG " (given an invalid .DOC file?)\n"
844 if ($write_to_fail_log);
845
846 } # while ERRFILE
847 close FAILLOG if ($write_to_fail_log);
848 }
849 return 0; # we can try any_to_text
850 }
851
852 # Was the conversion successful?
853 if (-s "$output_filestem.html") {
854 open(TMP, "$output_filestem.html");
855 my $line = <TMP>;
856 close(TMP);
857 if ($line && $line =~ m/html/i) {
858 &util::rm("$output_filestem.err") if -e "$output_filestem.err";
859 return 1;
860 }
861 }
862
863 # If here, an error of some sort occurred
864 &util::rm("$output_filestem.html") if -e "$output_filestem.html";
865 if (-e "$output_filestem.err") {
866 if ($faillogfile ne "" && defined(open(FAILLOG,">>$faillogfile"))) {
867 open (ERRLOG,"$output_filestem.err");
868 while (<ERRLOG>) {print FAILLOG $_;}
869 close FAILLOG;
870 close ERRLOG;
871 }
872 &util::rm("$output_filestem.err");
873 }
874 return 0;
875}
876
877# Attempt to convert an RTF document to html with rtftohtml
878sub rtf_to_html {
879 my ($input_filename, $output_filestem) = @_;
880
881 # formulate the command
882 my $cmd = "";
883 if ($timeout) {$cmd = "ulimit -t $timeout;";}
884 $cmd .= "rtftohtml";
885 #$cmd .= "rtf-converter";
886
887 $cmd .= " -o \"$output_filestem.html\" \"$input_filename\"";
888
889 $cmd .= " 2>\"$output_filestem.err\""
890 if ($ENV{'GSDLOS'} !~ m/^windows$/i || $is_winnt_2000);
891
892
893 # execute the command
894 $!=0;
895 if (system($cmd)!=0)
896 {
897 print STDERR "Error executing rtf converter $!\n";
898 # don't currently bother printing out error log...
899 # keep going, in case it still created an HTML file...
900 }
901
902 # Was the conversion successful?
903 my $was_successful=0;
904 if (-s "$output_filestem.html") {
905 # make sure we have some content other than header
906 open (HTML, "$output_filestem.html"); # what to do if fail?
907 my $line;
908 my $past_header=0;
909 while ($line=<HTML>) {
910
911 if ($past_header == 0) {
912 if ($line =~ m/<body>/) {$past_header=1;}
913 next;
914 }
915
916 $line =~ s/<[^>]+>//g;
917 if ($line =~ m/\w/ && $past_header) { # we found some content...
918 $was_successful=1;
919 last;
920 }
921 }
922 close HTML;
923 }
924
925 if ($was_successful) {
926 &util::rm("$output_filestem.err")
927 if (-e "$output_filestem.err");
928 # insert the (modified) table of contents, if it exists.
929 if (-e "${output_filestem}_ToC.html") {
930 &util::mv("$output_filestem.html","$output_filestem.src");
931 my $open_failed=0;
932 open HTMLSRC, "$output_filestem.src" || ++$open_failed;
933 open TOC, "${output_filestem}_ToC.html" || ++$open_failed;
934 open HTML, ">$output_filestem.html" || ++$open_failed;
935
936 if ($open_failed) {
937 close HTMLSRC;
938 close TOC;
939 close HTML;
940 &util::mv("$output_filestem.src","$output_filestem.html");
941 return 1;
942 }
943
944 # print out header info from src html.
945 while (defined($_ = <HTMLSRC>) && $_ =~ m/\w/) {
946 print HTML "$_";
947 }
948
949 # print out table of contents, making links relative
950 <TOC>; <TOC>; # ignore first 2 lines
951 print HTML scalar(<TOC>); # line 3 = "<ol>\n"
952 my $line;
953 while ($line=<TOC>) {
954 $line =~ s@</body></html>$@@i ; # only last line has this
955 # make link relative
956 $line =~ s@href=\"[^\#]+@href=\"@i;
957 print HTML $line;
958 }
959 close TOC;
960
961 # rest of html src
962 while (<HTMLSRC>) {
963 print HTML $_;
964 }
965 close HTMLSRC;
966 close HTML;
967
968 &util::rm("${output_filestem}_ToC.html");
969 &util::rm("${output_filestem}.src");
970 }
971 # we don't yet do anything with footnotes ($output_filestem_fn.html) :(
972 return 1; # success
973 }
974
975 if (-e "$output_filestem.err") {
976 if ($faillogfile ne "" && defined(open(FAILLOG,">>$faillogfile")))
977 {
978 print FAILLOG "Error - rtftohtml - couldn't extract text\n";
979 #print FAILLOG "Error - rtf-converter - couldn't extract text\n";
980 print FAILLOG " (rtf file might be too recent):\n";
981 open (ERRLOG, "$output_filestem.err");
982 while (<ERRLOG>) {print FAILLOG $_;}
983 close ERRLOG;
984 close FAILLOG;
985 }
986 &util::rm("$output_filestem.err");
987 }
988
989 &util::rm("$output_filestem.html") if (-e "$output_filestem.html");
990
991 return 0;
992}
993
994
995# Convert a pdf file to html with the pdftohtml command
996
997sub pdf_to_html {
998 my ($dirname, $input_filename, $output_filestem) = @_;
999
1000 my $cmd = "";
1001 if ($timeout) {$cmd = "ulimit -t $timeout;";}
1002 my $full_perl_path = &util::get_perl_exec();
1003 $cmd .= "\"$full_perl_path\" -S pdftohtml.pl -zoom $pdf_zoom";
1004 $cmd .= " -c" if ($pdf_complex);
1005 $cmd .= " -i" if ($pdf_ignore_images);
1006 $cmd .= " -a" if ($pdf_allow_images_only);
1007 $cmd .= " -hidden" unless ($pdf_nohidden);
1008 $cmd .= " \"$input_filename\" \"$output_filestem\"";
1009
1010 if ($ENV{'GSDLOS'} !~ m/^windows$/i || $is_winnt_2000) {
1011 $cmd .= " > \"$output_filestem.out\" 2> \"$output_filestem.err\"";
1012 } else {
1013 $cmd .= " > \"$output_filestem.err\"";
1014 }
1015
1016 $!=0;
1017
1018 my $retval=system($cmd);
1019 if ($retval!=0)
1020 {
1021 print STDERR "Error executing pdftohtml.pl";
1022 if ($!) {print STDERR ": $!";}
1023 print STDERR "\n";
1024 }
1025
1026 # make sure the converter made something
1027 if ($retval!=0 || ! -s "$output_filestem.html")
1028 {
1029 &util::rm("$output_filestem.out") if (-e "$output_filestem.out");
1030 # print out the converter's std err, if any
1031 if (-s "$output_filestem.err") {
1032 open (ERRLOG, "$output_filestem.err") || die "$!";
1033 print STDERR "pdftohtml error log:\n";
1034 while (<ERRLOG>) {
1035 print STDERR "$_";
1036 }
1037 close ERRLOG;
1038 }
1039 print STDERR "***********output filestem $output_filestem.html\n";
1040 &util::rm("$output_filestem.html") if (-e "$output_filestem.html");
1041 if (-e "$output_filestem.err") {
1042 if ($faillogfile ne "" && defined(open(FAILLOG,">>$faillogfile")))
1043 {
1044 open (ERRLOG, "$output_filestem.err");
1045 while (<ERRLOG>) {print FAILLOG $_;}
1046 close ERRLOG;
1047 close FAILLOG;
1048 }
1049 &util::rm("$output_filestem.err");
1050 }
1051 return 0;
1052 }
1053
1054 &util::rm("$output_filestem.err") if (-e "$output_filestem.err");
1055 &util::rm("$output_filestem.out") if (-e "$output_filestem.out");
1056 return 1;
1057}
1058
1059# Convert a pdf file to various types of image with the convert command
1060
1061sub pdfps_to_img {
1062 my ($dirname, $input_filename, $output_filestem, $output_type) = @_;
1063
1064 # Check that ImageMagick is installed and available on the path (except for Windows 95/98)
1065 if (!($ENV{'GSDLOS'} eq "windows" && !Win32::IsWinNT())) {
1066 my $result = `identify 2>&1`;
1067 if ($? == -1 || $? == 256) { # Linux and Windows return different values for "program not found"
1068 #ImageMagick is not installed, thus the convert utility is not available.
1069 print STDERR "*** ImageMagick is not installed, the convert utility is not available. Unable to convert PDF/PS to images\n";
1070 return 0;
1071 }
1072 }
1073
1074 my $cmd = "";
1075 if ($timeout) {$cmd = "ulimit -t $timeout;";}
1076 $output_type =~ s/.*\_(.*)/$1/i;
1077 my $full_perl_path = &util::get_perl_exec();
1078 $cmd .= "\"$full_perl_path\" -S pdfpstoimg.pl -convert_to $output_type \"$input_filename\" \"$output_filestem\"";
1079 if ($ENV{'GSDLOS'} !~ m/^windows$/i || $is_winnt_2000) {
1080 $cmd .= " > \"$output_filestem.out\" 2> \"$output_filestem.err\"";
1081 } else {
1082 $cmd .= " > \"$output_filestem.err\"";
1083 }
1084
1085 # don't include path on windows (to avoid having to play about
1086 # with quoting when GSDLHOME might contain spaces) but assume
1087 # that the PATH is set up correctly
1088 $!=0;
1089 my $retval=system($cmd);
1090 if ($retval!=0)
1091 {
1092 print STDERR "Error executing pdftoimg.pl";
1093 if ($!) {print STDERR ": $!";}
1094 print STDERR "\n";
1095 }
1096
1097 #make sure the converter made something
1098 #if ($retval !=0) || ! -s "$output_filestem")
1099 if ($retval !=0)
1100 {
1101 &util::rm("$output_filestem.out") if (-e "$output_filestem.out");
1102 #print out the converter's std err, if any
1103 if (-s "$output_filestem.err") {
1104 open (ERRLOG, "$output_filestem.err") || die "$!";
1105 print STDERR "pdfpstoimg error log:\n";
1106 while (<ERRLOG>) {
1107 print STDERR "$_";
1108 }
1109 close ERRLOG;
1110 }
1111 #&util::rm("$output_filestem.html") if (-e "$output_filestem.html");
1112 if (-e "$output_filestem.err") {
1113 if ($faillogfile ne "" && defined(open(FAILLOG,">>$faillogfile")))
1114 {
1115 open (ERRLOG, "$output_filestem.err");
1116 while (<ERRLOG>) {print FAILLOG $_;}
1117 close ERRLOG;
1118 close FAILLOG;
1119 }
1120 &util::rm("$output_filestem.err");
1121 }
1122 return 0;
1123 }
1124 &util::rm("$output_filestem.err") if (-e "$output_filestem.err");
1125 &util::rm("$output_filestem.out") if (-e "$output_filestem.out");
1126 return 1;
1127}
1128
1129# Convert a PDF file to text with the pdftotext command
1130
1131sub pdf_to_text {
1132 my ($dirname, $input_filename, $output_filestem) = @_;
1133
1134 my $cmd = "pdftotext \"$input_filename\" \"$output_filestem.text\"";
1135
1136 if ($ENV{'GSDLOS'} !~ m/^windows$/i) {
1137 $cmd .= " > \"$output_filestem.out\" 2> \"$output_filestem.err\"";
1138 } else {
1139 $cmd .= " > \"$output_filestem.err\"";
1140 }
1141
1142 if (system($cmd)!=0)
1143 {
1144 print STDERR "Error executing $cmd: $!\n";
1145 &util::rm("$output_filestem.text") if (-e "$output_filestem.text");
1146 }
1147
1148 # make sure there is some extracted text.
1149 if (-e "$output_filestem.text") {
1150 open (EXTR_TEXT, "$output_filestem.text") || warn "open: $!";
1151 binmode(EXTR_TEXT); # just in case...
1152 my $line="";
1153 my $seen_text=0;
1154 while (($seen_text==0) && ($line=<EXTR_TEXT>)) {
1155 if ($line=~ m/\w/) {$seen_text=1;}
1156 }
1157 close EXTR_TEXT;
1158 if ($seen_text==0) { # no text was extracted
1159 print STDERR "Error: pdftotext found no text\n";
1160 &util::rm("$output_filestem.text");
1161 }
1162 }
1163
1164 # make sure the converter made something
1165 if (! -s "$output_filestem.text")
1166 {
1167 # print out the converters std err, if any
1168 if (-s "$output_filestem.err") {
1169 open (ERRLOG, "$output_filestem.err") || die "$!";
1170 print STDERR "pdftotext error log:\n";
1171 while (<ERRLOG>) {
1172 print STDERR "$_";
1173 }
1174 close ERRLOG;
1175 }
1176 # does this converter create a .out file?
1177 &util::rm("$output_filestem.out") if (-e "$output_filestem.out");
1178 &util::rm("$output_filestem.text") if (-e "$output_filestem.text");
1179 if (-e "$output_filestem.err") {
1180 if ($faillogfile ne "" && defined(open(FAILLOG,">>$faillogfile")))
1181 {
1182 open (ERRLOG,"$output_filestem.err");
1183 while (<ERRLOG>) {print FAILLOG $_;}
1184 close ERRLOG;
1185 close FAILLOG;
1186 }
1187 &util::rm("$output_filestem.err");
1188 }
1189 return 0;
1190 }
1191 &util::rm("$output_filestem.err") if (-e "$output_filestem.err");
1192 return 1;
1193}
1194
1195# Convert a PostScript document to text
1196# note - just using "ps2ascii" isn't good enough, as it
1197# returns 0 for a postscript interpreter error. ps2ascii is just
1198# a wrapper to "gs" anyway, so we use that cmd here.
1199
1200sub ps_to_text {
1201 my ($input_filename, $output_filestem) = @_;
1202
1203 my $error = "";
1204
1205 # if we're on windows we'll fall straight through without attempting
1206 # to use gs
1207 if ($ENV{'GSDLOS'} =~ m/^windows$/i) {
1208 $error = "Windows does not support gs";
1209
1210 } else {
1211 my $cmd = "";
1212 if ($timeout) {$cmd = "ulimit -t $timeout; ";}
1213 $cmd .= "gs -q -dNODISPLAY -dNOBIND -dWRITESYSTEMDICT -dSIMPLE -c save ";
1214 $cmd .= "-f ps2ascii.ps \"$input_filename\" -c quit > \"$output_filestem.text\"";
1215 #$cmd .= "pstotext -output \"$output_filestem.text\" $input_filename\"";
1216 $cmd .= " 2> $output_filestem.err";
1217 $!=0;
1218
1219 my $retcode=system($cmd);
1220 $retcode = $? >> 8; # see man perlfunc - system for this...
1221 # if system returns -1 | 127 (couldn't start program), look at $! for message
1222
1223 if ($retcode!=0) {if ($!) {$error=$!;} else {$error="couldn't run.\n";}}
1224 elsif (! -e "$output_filestem.text") {
1225 $error="did not create output file.\n";
1226 }
1227 else
1228 { # make sure the interpreter didn't get an error. It is technically
1229 # possible for the actual text to start with this, but....
1230 open PSOUT, "$output_filestem.text";
1231 if (<PSOUT> =~ m/^Error: (.*)/) {
1232 $error="interpreter error - \"$1\"";
1233 }
1234 close PSOUT;
1235 }
1236 }
1237
1238 if ($error ne "")
1239 {
1240 print STDERR "Warning: Error executing gs: $error\n";
1241 &util::rm("$output_filestem.text") if (-e "$output_filestem.text");
1242
1243 if ("$faillogfile" ne "" && defined(open (FAILLOG, ">>$faillogfile")))
1244 {
1245 print FAILLOG "gs - $error\n";
1246 if (-e "$output_filestem.err") {
1247 open(ERRLOG, "$output_filestem.err");
1248 while (<ERRLOG>) {print FAILLOG $_;}
1249 close ERRLOG;
1250 }
1251 close FAILLOG;
1252 }
1253 &util::rm("$output_filestem.err") if (-e "$output_filestem.err");
1254
1255
1256 # Fine then. We'll just do a lousy job by ourselves...
1257 # Based on 5-line regexp sed script found at:
1258 # http://snark.ptc.spbu.ru/mail-archives/lout/brown/msg00003.html
1259 #
1260 print STDERR "Stripping text from postscript\n";
1261 my $errorcode=0;
1262 open (IN, "$input_filename")
1263 || ($errorcode=1, warn "Couldn't read file: $!");
1264 open (OUT, ">$output_filestem.text")
1265 || ($errorcode=1, warn "Couldn't write file: $!");
1266 if ($errorcode) {print STDERR "errors\n";return 0;}
1267
1268 my $text=""; # this is for whole .ps file...
1269 $text = join('', <IN>); # see man perlport, under "System Resources"
1270 close IN;
1271
1272 # Make sure this is a ps file...
1273 if ($text !~ m/^%!/) {
1274 print STDERR "Bad postscript header: not '%!'\n";
1275 if ($faillogfile ne "" && defined(open(FAILLOG, ">>$faillogfile")))
1276 {
1277 print FAILLOG "Bad postscript header: not '%!'\n";
1278 close FAILLOG;
1279 }
1280 return 0;
1281 }
1282
1283 # if ps has Page data, then use it to delete all stuff before it.
1284 $text =~ s/^.*?%%Page:.*?\n//s; # treat string as single line
1285
1286 # remove all leading non-data stuff
1287 $text =~ s/^.*?\(//s;
1288
1289 # remove all newline chars for easier processing
1290 $text =~ s/\n//g;
1291
1292 # Big assumption here - assume that if any co-ordinates are
1293 # given, then we are at the end of a sentence.
1294 $text =~ s/\)-?\d+\ -?\d+/\) \(\n\)/g;
1295
1296 # special characters--
1297 $text =~ s/\(\|\)/\(\ - \)/g; # j -> em-dash?
1298
1299 # ? ps text formatting (eg italics?) ?
1300 $text =~ s/Fn\(f\)/\(\{\)/g; # f -> {
1301 $text =~ s/Fn\(g\)/\(\}\)/g; # g -> }
1302 $text =~ s/Fn\(j\)/\(\|\)/g; # j -> |
1303 # default - remove the rest
1304 $text =~ s/\ ?F.\((.+?)\)/\($1\)/g;
1305
1306 # attempt to add whitespace between words...
1307 # this is based purely on observation, and may be completely wrong...
1308 $text =~ s/([^F])[defghijkuy]\(/$1 \( /g;
1309 # eg I notice "b(" is sometimes NOT a space if preceded by a
1310 # negative number.
1311 $text =~ s/\)\d+ ?b\(/\) \( /g;
1312
1313 # change quoted braces to brackets
1314 $text =~ s/([^\\])\\\(/$1\{/g;
1315 $text =~ s/([^\\])\\\)/$1\}/g ;
1316
1317 # remove everything that is not between braces
1318 $text =~ s/\)([^\(\)])+?\(//sg ;
1319
1320 # remove any Trailer eof stuff.
1321 $text =~ s/\)[^\)]*$//sg;
1322
1323 ### ligatures have special characters...
1324 $text =~ s/\\013/ff/g;
1325 $text =~ s/\\014/fi/g;
1326 $text =~ s/\\015/fl/g;
1327 $text =~ s/\\016/ffi/g;
1328 $text =~ s/\\214/fi/g;
1329 $text =~ s/\\215/fl/g;
1330 $text =~ s/\\017/\n\* /g; # asterisk?
1331 $text =~ s/\\023/\023/g; # e acute ('e)
1332 $text =~ s/\\177/\252/g; # u"
1333# $text =~ s/ ?? /\344/g; # a"
1334
1335 print OUT "$text";
1336 close OUT;
1337 }
1338 # wrap the text - use a minimum length. ie, first space after this length.
1339 my $wrap_length=72;
1340 &util::mv("$output_filestem.text", "$output_filestem.text.tmp");
1341 open INFILE, "$output_filestem.text.tmp" ||
1342 die "Couldn't open file: $!";
1343 open OUTFILE, ">$output_filestem.text" ||
1344 die "Couldn't open file for writing: $!";
1345 my $line="";
1346 while ($line=<INFILE>) {
1347 while (length($line)>0) {
1348 if (length($line)>$wrap_length) {
1349 $line =~ s/^(.{$wrap_length}[^\s]*)\s*//;
1350 print OUTFILE "$1\n";
1351 } else {
1352 print OUTFILE "$line";
1353 $line="";
1354 }
1355 }
1356 }
1357 close INFILE;
1358 close OUTFILE;
1359 &util::rm("$output_filestem.text.tmp");
1360
1361 &util::rm("$output_filestem.err") if (-e "$output_filestem.err");
1362 return 1;
1363}
1364
1365
1366# Convert any file to HTML with a crude perl implementation of the
1367# UNIX strings command.
1368
1369sub any_to_html {
1370 my ($input_filename, $output_filestem) = @_;
1371
1372 # First generate a text file
1373 return 0 unless (&any_to_text($input_filename, $output_filestem));
1374
1375 # create an HTML file from the text file
1376 open(TEXT, "<$output_filestem.text");
1377 open(HTML, ">$output_filestem.html");
1378
1379 print HTML "<html><head>\n";
1380 print HTML "<META HTTP-EQUIV=\"Content-Type\" CONTENT=\"text/html\">\n";
1381 print HTML "<META NAME=\"GENERATOR\" CONTENT=\"Greenstone any_to_html\">\n";
1382 print HTML "</head><body>\n\n";
1383
1384 my $line;
1385 while ($line=<TEXT>) {
1386 $line =~ s/</&lt;/g;
1387 $line =~ s/>/&gt;/g;
1388 if ($line =~ m/^\s*$/) {
1389 print HTML "<p>";
1390 } else {
1391 print HTML "<br> ", $line;
1392 }
1393 }
1394 print HTML "\n</body></html>\n";
1395
1396 close HTML;
1397 close TEXT;
1398
1399 &util::rm("$output_filestem.text") if (-e "$output_filestem.text");
1400 return 1;
1401}
1402
1403# Convert any file to TEXT with a crude perl implementation of the
1404# UNIX strings command.
1405# Note - this assumes ascii charsets :( (jrm21)
1406
1407sub any_to_text {
1408 my ($input_filename, $output_filestem) = @_;
1409
1410 if (!$use_strings) {
1411 return 0;
1412 }
1413
1414 print STDERR "\n**** In any to text****\n\n";
1415 open(IN, "<$input_filename") || return 0;
1416 binmode(IN);
1417 open(OUT, ">$output_filestem.text") || return 0;
1418
1419 my ($line);
1420 my $output_line_count = 0;
1421 while (<IN>) {
1422 $line = $_;
1423
1424 # delete anything that isn't a printable character
1425 $line =~ s/[^\040-\176]+/\n/sg;
1426
1427 # delete any string less than 10 characters long
1428 $line =~ s/^.{0,9}$/\n/mg;
1429 while ($line =~ m/^.{1,9}$/m) {
1430 $line =~ s/^.{0,9}$/\n/mg;
1431 $line =~ s/\n+/\n/sg;
1432 }
1433
1434 # remove extraneous whitespace
1435 $line =~ s/\n+/\n/gs;
1436 $line =~ s/^\n//gs;
1437
1438 # output whatever is left
1439 if ($line =~ m/[^\n ]/) {
1440 print OUT $line;
1441 ++$output_line_count;
1442 }
1443 }
1444
1445 close OUT;
1446 close IN;
1447
1448 if ($output_line_count) { # try to protect against binary only formats
1449 return 1;
1450 }
1451
1452 &util::rm("$output_filestem.text");
1453 return 0;
1454
1455}
Note: See TracBrowser for help on using the repository browser.