source: main/trunk/greenstone2/bin/script/gsConvert.pl@ 24223

Last change on this file since 24223 was 24223, checked in by ak19, 13 years ago

More changes to do with obtaining the perlpath using Config: moved the Use Config statement into the Begin block (it cannot be a Require statement for some reason).
  • Property svn:executable set to *
  • Property svn:keywords set to Author Date Id Revision
File size: 47.2 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 if(!$ENV{'PERLPATH'}) {
51 use Config; # for some reason, this can't be a require statement
52 my $full_perl_exec = $Config{perlpath}; #$^X;
53 require File::Basename;
54 my $perl_path = File::Basename::dirname($full_perl_exec);
55 $ENV{'PERLPATH'} = $perl_path;
56 }
57
58}
59
60use strict;
61
62use parsargv;
63use util;
64use Cwd;
65#use Config; # for getting the perlpath in the recommended way
66# Note: even though this (and other) use statement comes after its
67# usage in BEGIN, the use statement is in fact executed before BEGIN.
68# Doesn't work here for Greenstone 3.
69
70
71# Are we running on WinNT or Win2000 (or later)?
72my $is_winnt_2000=eval {require Win32; return (Win32::IsWinNT()); return 0;};
73if (!defined($is_winnt_2000)) {$is_winnt_2000=0;}
74
75my $use_strings;
76my $pdf_complex;
77my $pdf_nohidden;
78my $pdf_zoom;
79my $pdf_ignore_images;
80my $pdf_allow_images_only;
81my $windows_scripting;
82
83sub print_usage
84{
85 print STDERR "\n";
86 print STDERR "gsConvert.pl: Converts documents in a range of formats to html\n";
87 print STDERR " or text using third-party programs.\n\n";
88 print STDERR " usage: $0 [options] filename\n";
89 print STDERR " options:\n\t-type\tdoc|dot|pdf|ps|ppt|rtf|xls\t(input file type)\n";
90 print STDERR "\t-errlog\t<filename>\t(append err messages)\n";
91 print STDERR "\t-output\tauto|html|text|pagedimg_jpg|pagedimg_gif|pagedimg_png\t(output file type)\n";
92 print STDERR "\t-timeout\t<max cpu seconds>\t(ulimit on unix systems)\n";
93 print STDERR "\t-use_strings\tuse strings to extract text if conversion fails\n";
94 print STDERR "\t-windows_scripting\tuse windows VB script (if available) to convert Microsoft Word and PPT documents\n";
95 print STDERR "\t-pdf_complex\tuse complex output when converting PDF to HTML\n";
96 print STDERR "\t-pdf_nohidden\tDon't attempt to extract hidden text from PDF files\n";
97 print STDERR "\t-pdf_ignore_images\tdon't attempt to extract images when\n";
98 print STDERR "\t\tconverting PDF to HTML\n";
99 print STDERR "\t-pdf_allow_images_only\tallow images only (continue even if no text is present when converting to HTML)\n";
100 print STDERR "\t-pdf_zoom\tfactor by which to zoom PDF (only useful if\n";
101 print STDERR "\t\t-pdf_complex is set\n";
102 exit(1);
103}
104
105my $faillogfile="";
106my $timeout=0;
107
108sub main
109{
110 my (@ARGV) = @_;
111 my ($input_type,$output_type,$verbose);
112
113 # Dynamically figure out what the --type option can support, based on whether -windows_scripting
114 # is in use or not
115 my $default_type_re = "(doc|dot|pdf|ps|ppt|rtf|xls)";
116 #my $enhanced_type_re = "(docx?|dot|pdf|ps|pptx?|rtf|xlsx?)";
117 #my $enhanced_type_re = "(docx?|dot|pdf|ps|pptx?|rtf|xlsx?)";
118 # Currently only have VBA for Word and PPT(but no XLS)
119 my $enhanced_type_re = "(docx?|dot|pdf|ps|pptx?|rtf|xls)";
120
121 my $type_re = $default_type_re;
122
123 foreach my $a (@ARGV) {
124 if ($a =~ m/^windows_scripting$/i) {
125 $type_re = $enhanced_type_re;
126 }
127 }
128
129 # read command-line arguments
130 if (!parsargv::parse(\@ARGV,
131 "type/$type_re/", \$input_type,
132 '/errlog/.*/', \$faillogfile,
133 'output/(auto|html|text|pagedimg).*/', \$output_type,
134 'timeout/\d+/0',\$timeout,
135 'verbose/\d+/0', \$verbose,
136 'windows_scripting',\$windows_scripting,
137 'use_strings', \$use_strings,
138 'pdf_complex', \$pdf_complex,
139 'pdf_ignore_images', \$pdf_ignore_images,
140 'pdf_allow_images_only', \$pdf_allow_images_only,
141 'pdf_nohidden', \$pdf_nohidden,
142 'pdf_zoom/\d+/2', \$pdf_zoom
143 ))
144 {
145 print_usage();
146 }
147
148 # Make sure the input file exists and can be opened for reading
149 if (scalar(@ARGV!=1)) {
150 print_usage();
151 }
152
153 my $input_filename = $ARGV[0];
154 if (!-r $input_filename) {
155 print STDERR "Error: unable to open $input_filename for reading\n";
156 exit(1);
157 }
158
159 # Deduce filenames
160 my ($tailname,$dirname,$suffix)
161 = File::Basename::fileparse($input_filename, "\\.[^\\.]+\$");
162 my $output_filestem = &util::filename_cat($dirname, "$tailname");
163
164 if ($input_type eq "")
165 {
166 $input_type = lc (substr($suffix,1,length($suffix)-1));
167 }
168
169 # Change to temporary working directory
170 my $stored_dir = cwd();
171 chdir ($dirname) || die "Unable to change to directory $dirname";
172
173 # Select convert utility
174 if (!defined $input_type) {
175 print STDERR "Error: No filename extension or input type defined\n";
176 exit(1);
177 }
178 elsif ($input_type =~ m/^docx?$/ || $input_type eq "dot") {
179 print &convertDOC($input_filename, $output_filestem, $output_type);
180 print "\n";
181 }
182 elsif ($input_type eq "rtf") {
183 print &convertRTF($input_filename, $output_filestem, $output_type);
184 print "\n";
185 }
186 elsif ($input_type eq "pdf") {
187 print &convertPDF($dirname, $input_filename, $output_filestem, $output_type);
188 print "\n";
189 }
190 elsif ($input_type eq "ps") {
191 print &convertPS($dirname, $input_filename, $output_filestem, $output_type);
192 print "\n";
193 }
194 elsif ($input_type =~ m/pptx?$/) {
195 print &convertPPT($input_filename, $output_filestem, $output_type);
196 print "\n";
197 }
198 elsif ($input_type =~ m/xlsx?$/) {
199 print &convertXLS($input_filename, $output_filestem, $output_type);
200 print "\n";
201 }
202 else {
203 print STDERR "Error: Unable to convert type '$input_type'\n";
204 exit(1);
205 }
206
207 # restore to original working directory
208 chdir ($stored_dir) || die "Unable to return to directory $stored_dir";
209
210}
211
212&main(@ARGV);
213
214
215
216# Document-type conversion functions
217#
218# The following functions attempt to convert documents from their
219# input type to the specified output type. If no output type was
220# given, then they first attempt HTML, and then TEXT.
221#
222# Each returns the output type ("html" or "text") or "fail" if no
223# conversion is possible.
224
225# Convert a Microsoft word document
226
227sub convertDOC {
228 my ($input_filename, $output_filestem, $output_type) = @_;
229
230 # Many .doc files are not in fact word documents!
231 my $realtype = &find_docfile_type($input_filename);
232
233 if ($realtype eq "word6" || $realtype eq "word7"
234 || $realtype eq "word8" || $realtype eq "docx") {
235 return &convertWord678($input_filename, $output_filestem, $output_type);
236 } elsif ($realtype eq "rtf") {
237 return &convertRTF($input_filename, $output_filestem, $output_type);
238 } else {
239 return &convertAnything($input_filename, $output_filestem, $output_type);
240 }
241}
242
243# Convert a Microsoft word 6/7/8 document
244
245sub convertWord678 {
246 my ($input_filename, $output_filestem, $output_type) = @_;
247
248 my $success = 0;
249 if (!$output_type || ($output_type =~ m/html/i)){
250 if ($windows_scripting) {
251 $success = &native_doc_to_html($input_filename, $output_filestem);
252 }
253 else {
254 $success = &doc_to_html($input_filename, $output_filestem);
255 }
256 if ($success) {
257 return "html";
258 }
259 }
260 return &convertAnything($input_filename, $output_filestem, $output_type);
261}
262
263
264# Convert a Rich Text Format (RTF) file
265
266sub convertRTF {
267 my ($input_filename, $output_filestem, $output_type) = @_;
268
269 my $success = 0;
270
271 # Attempt specialised conversion to HTML
272 if (!$output_type || ($output_type =~ m/html/i)) {
273
274 if ($windows_scripting) {
275 $success = &native_doc_to_html($input_filename, $output_filestem);
276 }
277 else {
278 $success = &rtf_to_html($input_filename, $output_filestem);
279 }
280 if ($success) {
281 return "html";
282 }
283 }
284
285# rtf is so ugly that's it's not worth running strings over.
286# One day I'll write some quick'n'dirty regexps to try to extract text - jrm21
287# return &convertAnything($input_filename, $output_filestem, $output_type);
288 return "fail";
289}
290
291
292# Convert an unidentified file
293
294sub convertAnything {
295 my ($input_filename, $output_filestem, $output_type) = @_;
296
297 my $success = 0;
298
299 # Attempt simple conversion to HTML
300 if (!$output_type || ($output_type =~ m/html/i)) {
301 $success = &any_to_html($input_filename, $output_filestem);
302 if ($success) {
303 return "html";
304 }
305 }
306
307 # Convert to text
308 if (!$output_type || ($output_type =~ m/text/i)) {
309 $success = &any_to_text($input_filename, $output_filestem);
310 if ($success) {
311 return "text";
312 }
313 }
314 return "fail";
315}
316
317
318
319# Convert an Adobe PDF document
320
321sub convertPDF {
322 my ($dirname, $input_filename, $output_filestem, $output_type) = @_;
323
324 my $success = 0;
325 $output_type =~ s/.*\-(.*)/$1/i;
326 # Attempt coversion to Image
327 if ($output_type =~ m/jp?g|gif|png/i) {
328 $success = &pdfps_to_img($dirname, $input_filename, $output_filestem, $output_type);
329 if ($success){
330 return "item";
331 }
332 }
333
334 # Attempt conversion to HTML
335 if (!$output_type || ($output_type =~ m/html/i)) {
336 $success = &pdf_to_html($dirname, $input_filename, $output_filestem);
337 if ($success) {
338 return "html";
339 }
340 }
341
342 # Attempt conversion to TEXT
343 if (!$output_type || ($output_type =~ m/text/i)) {
344 $success = &pdf_to_text($dirname, $input_filename, $output_filestem);
345 if ($success) {
346 return "text";
347 }
348 }
349
350 return "fail";
351
352}
353
354
355# Convert an Adobe PostScript document
356
357sub convertPS {
358 my ($dirname,$input_filename, $output_filestem, $output_type) = @_;
359
360 my $success = 0;
361 $output_type =~ s/.*\-(.*)/$1/i;
362 # Attempt coversion to Image
363 if ($output_type =~ m/jp?g|gif|png/i) {
364 $success = &pdfps_to_img($dirname, $input_filename, $output_filestem, $output_type);
365 if ($success){
366 return "item";
367 }
368 }
369
370 # Attempt conversion to TEXT
371 if (!$output_type || ($output_type =~ m/text/i)) {
372 $success = &ps_to_text($input_filename, $output_filestem);
373 if ($success) {
374 return "text";
375 }
376 }
377 return "fail";
378}
379
380
381sub convertPPT {
382 my ($input_filename, $output_filestem, $output_type) = @_;
383 my $success = 0;
384
385 my $ppt_convert_type = "";
386
387 #if (!$output_type || $windows_scripting || ($output_type !~ m/html/i) || ($output_type !~ m/text/i)){
388 if ($windows_scripting && ($output_type !~ m/html/i) && ($output_type !~ m/text/i)){
389 if ($output_type =~ m/gif/i) {
390 $ppt_convert_type = "-g";
391 } elsif ($output_type =~ m/jp?g/i){
392 $ppt_convert_type = "-j";
393 } elsif ($output_type =~ m/png/i){
394 $ppt_convert_type = "-p";
395 }
396 my $vbScript = &util::filename_cat($ENV{'GSDLHOME'}, "bin",
397 $ENV{'GSDLOS'}, "pptextract");
398 $vbScript = "pptextract" if ($ENV{'GSDLOS'} =~ m/^windows$/i);
399
400 my $cmd = "";
401 if ($timeout) {$cmd = "ulimit -t $timeout;";}
402 # if the converting directory already exists
403 if (-d $output_filestem) {
404 print STDERR "**The conversion directory already exists\n";
405 return "item";
406 } else {
407 $cmd .= "$vbScript $ppt_convert_type \"$input_filename\" \"$output_filestem\"";
408 $cmd .= " 2>\"$output_filestem.err\""
409 if ($ENV{'GSDLOS'} !~ m/^windows$/i || $is_winnt_2000);
410 if (system($cmd) !=0) {
411 print STDERR "Powerpoint VB Scripting convert failed\n";
412 } else {
413 return "item";
414 }
415 }
416 } elsif (!$output_type || ($output_type =~ m/html/i)) {
417 # Attempt conversion to HTML
418 #if (!$output_type || ($output_type =~ m/html/i)) {
419 # formulate the command
420 my $cmd = "";
421 my $full_perl_path = &util::filename_cat($ENV{'PERLPATH'},"perl");
422 $cmd .= "\"$full_perl_path\" -S ppttohtml.pl ";
423 $cmd .= " \"$input_filename\" \"$output_filestem.html\"";
424 $cmd .= " 2>\"$output_filestem.err\""
425 if ($ENV{'GSDLOS'} !~ m/^windows$/i || $is_winnt_2000);
426
427 # execute the command
428 $!=0;
429 if (system($cmd)!=0)
430 {
431 print STDERR "Powerpoint 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
446sub convertXLS {
447 my ($input_filename, $output_filestem, $output_type) = @_;
448
449 my $success = 0;
450
451 # Attempt conversion to HTML
452 if (!$output_type || ($output_type =~ m/html/i)) {
453 # formulate the command
454 my $cmd = "";
455 my $full_perl_path = &util::filename_cat($ENV{'PERLPATH'},"perl");
456 $cmd .= "\"$full_perl_path\" -S xlstohtml.pl ";
457 $cmd .= " \"$input_filename\" \"$output_filestem.html\"";
458 $cmd .= " 2>\"$output_filestem.err\""
459 if ($ENV{'GSDLOS'} !~ m/^windows$/i || $is_winnt_2000);
460
461
462 # execute the command
463 $!=0;
464 if (system($cmd)!=0)
465 {
466 print STDERR "Excel 95/97 converter failed $!\n";
467 } else {
468 return "html";
469 }
470 }
471
472 $success = &any_to_text($input_filename, $output_filestem);
473 if ($success) {
474 return "text";
475 }
476
477 return "fail";
478}
479
480
481
482# Find the real type of a .doc file
483#
484# We seem to have a lot of files with a .doc extension that are .rtf
485# files or Word 5 files. This function attempts to tell the difference.
486sub find_docfile_type {
487 my ($input_filename) = @_;
488
489 if (($windows_scripting) && ($input_filename =~ m/\.docx$/)) {
490 return "docx";
491 }
492
493 open(CHK, "<$input_filename");
494 binmode(CHK);
495 my $line = "";
496 my $first = 1;
497
498 while (<CHK>) {
499
500 $line = $_;
501
502 if ($first) {
503 # check to see if this is an rtf file
504 if ($line =~ m/^\{\\rtf/) {
505 close(CHK);
506 return "rtf";
507 }
508 $first = 0;
509 }
510
511 # is this is a word 6/7/8 document?
512 if ($line =~ m/Word\.Document\.([678])/) {
513 close(CHK);
514
515 return "word$1";
516 }
517
518 }
519
520 return "unknown";
521}
522
523
524# Specific type-to-type conversions
525#
526# Each of the following functions attempts to convert a document from
527# a specific format to another. If they succeed they return 1 and leave
528# the output document(s) in the appropriate place; if they fail they
529# return 0 and delete any working files.
530
531
532# Attempt to convert a word document to html with the wv program
533sub doc_to_html {
534 my ($input_filename, $output_filestem) = @_;
535
536 my $wvWare = &util::filename_cat($ENV{'GSDLHOME'}, "bin", $ENV{'GSDLOS'}, "wvWare");
537
538 if ( -d "$ENV{'GSDLHOME'}/bin/$ENV{'GSDLOS'}/wv" && $ENV{'GSDLOS'} eq "linux" ) {
539 $ENV{'PATH'} = "$ENV{'GSDLHOME'}/bin/$ENV{'GSDLOS'}/wv/bin:$ENV{'PATH'}";
540 $ENV{'LD_LIBRARY_PATH'} = "$ENV{'GSDLHOME'}/bin/$ENV{'GSDLOS'}/wv/lib:$ENV{'LD_LIBRARY_PATH'}";
541 $wvWare = &util::filename_cat($ENV{'GSDLHOME'}, "bin", $ENV{'GSDLOS'}, "wv", "bin", "wvWare");
542 }
543
544 # don't include path on windows (to avoid having to play about
545 # with quoting when GSDLHOME might contain spaces) but assume
546 # that the PATH is set up correctly
547 $wvWare = "wvWare" if ($ENV{'GSDLOS'} =~ m/^windows$/i);
548
549 my $wv_conf = &util::filename_cat($ENV{'GSDLHOME'}, "etc",
550 "packages", "wv", "wvHtml.xml");
551
552 # Added the following to work with replace_srcdoc_with_html.pl:
553 # Make wvWare put any associated (image) files of the word doc into
554 # folder docname-without-extention_files. This folder should be at
555 # the same level as the html file generated from the doc.
556 # wvWare will take care of proper interlinking.
557
558 # This step is necessary for replace_srcdoc_with_html.pl which will
559 # move the html and associated files into the import folder. We
560 # want to ensure that the associated files won't overwrite similarly
561 # named items already in import. Hence we put them in a folder first
562 # (to which the html links properly) and that will allow
563 # replace_srcdoc_with_html.pl to move them safely to /import.
564
565 # To do all this, we need to use wvWare's --dir and --basename options
566 # where dir is the full path to the image folder directory and
567 # basename is the full path to the image folder appended to the name
568 # which is to be prepended to every image file:
569 # eg. if the images were to have names like sample0.jpg to sampleN.jpg,
570 # then the basename is "/full/path/to/imgdir/sample".
571 # In this case, basename is the full path to and name of the document.
572 # HOWEVER: basename always takes full path, not relative url, so
573 # the greenstone browser is unable to display the images (absolute paths
574 # cause it to give an "external link" message)
575 # See http://osdir.com/ml/lib.wvware.devel/2002-11/msg00014.html
576 # and http://rpmfind.net/linux/RPM/freshmeat/rpms/wv/wv-0.5.44-1.i386.html
577 # "added --dir option to wvHtml so that pictures can be placed in
578 # a seperate directory"
579 # "running wvWare through IMP to view word documents as html. It gets
580 # invoked like this:
581 # wvWare --dir=/tmp-wvWare --basename=/tmp-wvWare/img$$- $tmp_word >$tmp_output"
582
583 # toppath is the folder where html is generated
584 # docname is the name (without extension) of the html to be generated
585 # suffix (extension) is thrown away
586 my ($docname, $toppath)
587 = &File::Basename::fileparse($input_filename, "\\.[^\\.]+\$");
588
589 # We want the image folder generated to have the same name as windows
590 # would generate ($windows_scripting) when it converts from word to html.
591 # That is, foldername=docname_files
592 my $assoc_dir = &util::filename_cat($toppath, $docname."_files");
593 #print "assoc_dir: ".$assoc_dir."\n"; # same as "$output_filestem._files"
594
595 # ensure this image directory exists
596 # if it exists already, just delete and recreate
597 if(-e $assoc_dir) {
598 &util::rm_r($assoc_dir);
599 }
600 &util::mk_dir($assoc_dir);
601
602 # the images are all going to be called image0, image1,..., imageN
603 my $img_basenames = &util::filename_cat($assoc_dir, $docname);
604
605 #print STDERR "****toppath: $toppath\n****docname: $docname\n;
606 #print STDERR "****img_basenames: $img_basenames\n" if($img_basenames);
607 #print STDERR "****assoc_dir: $assoc_dir\n" if($assoc_dir);
608
609 my $cmd = "";
610 if ($timeout) {$cmd = "ulimit -t $timeout;";}
611 # wvWare's --dir and --basename options for image directory.
612 # Replaced the next line with the *2 lines* following it:
613 # $cmd .= "$wvWare --charset utf-8 --config \"$wv_conf\"";
614 $cmd .= "$wvWare --dir \"$assoc_dir\" --basename \"$img_basenames\"";
615 $cmd .= " --charset utf-8 --config \"$wv_conf\"";
616 $cmd .= " \"$input_filename\" > \"$output_filestem.html\"";
617
618 # redirecting STDERR is a bad idea on windows 95/98
619 $cmd .= " 2> \"$output_filestem.err\""
620 if ($ENV{'GSDLOS'} !~ m/^windows$/i || $is_winnt_2000);
621 # execute the command
622 $!=0;
623 if (system($cmd)!=0)
624 {
625 print STDERR "Error executing wv converter:$!\n";
626 if (-s "$output_filestem.err") {
627 open (ERRFILE, "<$output_filestem.err");
628
629 my $write_to_fail_log=0;
630 if ($faillogfile ne "" && defined(open(FAILLOG,">>$faillogfile")))
631 {$write_to_fail_log=1;}
632
633 my $line;
634 while ($line=<ERRFILE>) {
635 if ($line =~ m/\w/) {
636 print STDERR "$line";
637 print FAILLOG "$line" if ($write_to_fail_log);
638 }
639 if ($line !~ m/startup error/) {next;}
640 print STDERR " (given an invalid .DOC file?)\n";
641 print FAILLOG " (given an invalid .DOC file?)\n"
642 if ($write_to_fail_log);
643
644 } # while ERRFILE
645 close FAILLOG if ($write_to_fail_log);
646 }
647 return 0; # we can try any_to_text
648 }
649
650 # Was the conversion successful?
651
652 if (-s "$output_filestem.html") { # if file has non-zero size (i.e. it has contents)
653 open(TMP, "$output_filestem.html");
654 my $line = <TMP>;
655 close(TMP);
656 if ($line && $line =~ m/DOCTYPE HTML/) {
657 &util::rm("$output_filestem.err") if -e "$output_filestem.err";
658
659 # Inserted this code to remove the images directory if it was still empty after
660 # the html was generated (in case there were no images in the word document)
661 if (&util::is_dir_empty($assoc_dir)) {
662 #print STDERR "***gsConvert.pl: Image dir $assoc_dir is empty, removing***\n";
663 &util::rm_r($assoc_dir);
664 } else { # there was an image folder (it was generated)
665 # Therefore, the html file generated contains absolute links to the images
666 # Replace them with relative links instead, so the folder can be moved elsewhere
667 &make_links_to_assocdir_relative($toppath, $docname, "$output_filestem.html", $assoc_dir, $docname."_files");
668 }
669 return 1;
670 }
671 }
672
673 # If here, an error of some sort occurred
674 &util::rm("$output_filestem.html") if -e "$output_filestem.html";
675 if (-e "$output_filestem.err") {
676 if ($faillogfile ne "" && defined(open(FAILLOG,">>$faillogfile"))) {
677 open (ERRLOG,"$output_filestem.err");
678 while (<ERRLOG>) {print FAILLOG $_;}
679 close FAILLOG;
680 close ERRLOG;
681 }
682 &util::rm("$output_filestem.err");
683 }
684
685 return 0;
686}
687
688# Method to work with doc_to_html - Word docs might contain images.
689# When such word docs are converted with wvWare, we make it generate a
690# <filename>_files folder with the associated images, while the html file
691# <filename> refers to the images using absolute paths to <filename>_files.
692# This method reads in that html file and replaces all the absolute paths to
693# the images in <filename>_files with the relative paths to the images from
694# that folder. (I.e. with <filename>_files/<imagename.ext>).
695sub make_links_to_assocdir_relative{
696 # toppath is the top-level folder in which the html file we're going to be fixing resides
697 # docname is just the name (without extension) of the html file
698 # html_file is the full path to the html file: /full/path/docname.html
699 # assoc_dir_path is toppath/docname_files
700 # assoc_dirname is the directory name of the folder with associated imgs: docname_files
701 my ($toppath, $docname, $html_file, $assoc_dir_path, $assoc_dirname) = @_;
702
703 # 1. Read all the contents of the html into a string
704 # open the original file for reading
705 unless(open(FIN, "<$html_file")) {
706 print STDERR "gsConvert.pl: Unable to open $html_file for reading absolute urls...ERROR: $!\n";
707 return 0;
708 }
709 # From http://perl.plover.com/local.html
710 # "It's cheaper to read the file all at once, without all the splitting and reassembling.
711 # (Some people call this slurping the file.) Perl has a special feature to support this:
712 # If the $/ variable is undefined, the <...> operator will read the entire file all at once"
713 my $html_contents;
714 {
715 local $/ = undef; # Read entire file at once
716 $html_contents = <FIN>; # Now file is read in as one single 'line'
717 }
718 close(FIN); # close the file
719 #print STDERR $html_contents;
720
721 # 2. Replace (substitute) *all* ocurrences of the assoc_dir_path in a hrefs and img src
722 # values with assoc_dirname
723 # At the end: g means substitute all occurrences (global), while s at the end means treat
724 # all new lines as a regular space. This interacts with g to consider all the lines
725 # together as a single line so that multi-occurrences can be replaced.
726
727 # we can't just replace $assoc_dir_path with $assoc_dir
728 # $assoc_dir_path represents a regular expression that needs to be replaced
729 # if it contains ., -, [, ], or Windows style backslashes in paths -- which all have special
730 # meaning in Perl regular expressions -- we need to escape these first
731 my $safe_reg_expression = $assoc_dir_path;
732 $safe_reg_expression =~ s/\\/\\\\/g;
733 $safe_reg_expression =~ s/\./\\./g;
734 $safe_reg_expression =~ s/\-/\\-/g;
735 $safe_reg_expression =~ s/\[/\\[/g;
736 $safe_reg_expression =~ s/\]/\\]/g;
737 $safe_reg_expression =~ s/ /%20/g; # wvWare put %20 in place of space, so we need to change our prefix to match
738
739 # The following regular expression substitution looks for <a or <image, followed by any other
740 # attributes and values until it comes to the FIRST (indicated by ?) href= or src=
741 # followed by " or ' no quotes at all around path, followed by the associated folder's pathname
742 # followed by characters (for the img filename), then finally the optional closing quotes
743 # in " or ' form, followed by any other attributes and values until the first > to end the tag.
744 # The substitution: all the parts preceding associated folder's pathname are retained,
745 # the associated folder path name is replaced by associated folder directory name
746 # and the rest upto and including the closing > tag is retained.
747 # The sg at the end of the pattern match treats all of html_contents as a single line (s)
748 # and performs a global replace (g) meaning that all occurrences that match in that single line
749 # are substituted.
750 $html_contents =~ s/(<(a|img).*?(href|src)=(\"|\')?)$safe_reg_expression(.*?(\"|\')?.*?>)/$1$assoc_dirname$5/sg;
751 #$html_contents =~ s/$safe_reg_expression/$assoc_dirname/gs; # this works, used as fall-back
752 # now replace any %20 chars in filenames of href or src attributes to use literal space ' '. Calls a function for this
753 $html_contents =~ s/(<(a|img).*?(href|src)=(\"|\')?)(.*)(.*?(\"|\')?.*?>)/&post_process_assocfile_urls($1, $5, $6)/sge;
754
755 #print STDERR "****assoc_dirname: $assoc_dirname***\n";
756 #print STDERR "****safe_reg_expression: $safe_reg_expression***\n";
757
758 # delete the original file and recreate it
759 my $copy_of_filename = $html_file;
760 &util::rm($copy_of_filename); # deleted the file
761
762 # Recreate the original file for writing the updated contents
763 unless(open(FOUT, ">$html_file")) { # open it as a new file for writing
764 print STDERR "gsConvert.pl: Unable to open $html_file for writing relative links...ERROR: $!\n";
765 return 0;
766 }
767
768 # write out the updated contents and close the file
769 print FOUT $html_contents;
770 close(FOUT);
771 return 1;
772}
773
774# Utility routine to make sure HTML plugin gets img src/href link pathnames that contain
775# url slashes (/) instead of windows-style backwards slashes, and to convert all %20
776# introduced in link pathnames by wvWare into space again. Converts all percent signs
777# introduced by URL encoding filenames generated into %25 in these url links referencing them
778sub post_process_assocfile_urls
779{
780 my ($pre, $text, $post) = @_;
781
782 $text =~ s/%20/ /g; # Convert %20s to space and not underscore since underscores mess with incremental rebuild
783 # $text =~ s/%20/_/g; # reinstated this line, since we no longer replace spaces with %20. We replace them with underscores
784 $text =~ s/\\/\//g;
785 $text =~ s/%/%25/g;
786
787 return "$pre$text$post";
788}
789
790# Attempt to convert a word document to html with the word2html scripting program
791sub native_doc_to_html {
792 my ($input_filename, $output_filestem) = @_;
793
794 # build up the path to the doc-to-html conversion tool we're going to use
795 my $vbScript = &util::filename_cat($ENV{'GSDLHOME'}, "bin", $ENV{'GSDLOS'});
796
797 if ($ENV{'GSDLOS'} =~ m/^windows$/i) {
798 # if windows scripting with docx input, use new VBscript to get the local Word install (if
799 # any) to do the conversion, since docX can't be processed by word2html's windows_scripting
800
801 if($input_filename =~ m/docx$/i) { # need to use full path to docx2html script,
802 # else script launch fails when there are error msgs
803 $vbScript = &util::filename_cat($vbScript, "docx2html.vbs");
804 $vbScript = "CScript //Nologo \"$vbScript\""; # launche with CScript for error output in STDERR
805 # //Nologo flag avoids Microsoft's opening/logo msgs
806 print STDERR "About to use windows scripting to process docx file $input_filename.\n";
807 print STDERR " This may take some time. Please wait...\n";
808 }
809 else { # old doc versions. use the usual VB executable word2html for the
810 # conversion. Doesn't need full path, since bin\windows is on PATH
811 $vbScript = "word2html"; #$vbScript = "\"".&util::filename_cat($vbScript, "word2html")."\"";
812 }
813 }
814 else { # not windows
815 $vbScript = "\"".&util::filename_cat($vbScript, "word2html")."\"";
816 }
817
818 if (-e "$output_filestem.html") {
819 print STDERR " The conversion file:\n";
820 print STDERR " $output_filestem.html\n";
821 print STDERR " ... already exists. Skipping\n";
822 return 1;
823 }
824
825 my $cmd = "";
826 if ($timeout) {$cmd = "ulimit -t $timeout;";}
827 #$cmd .= "$vbScript \"$input_filename\" \"$output_filestem.html\"";
828 #$cmd .= "$vbScript $input_filename $output_filestem.html";
829 $cmd .= "$vbScript \"$input_filename\" \"$output_filestem.html\"";
830
831 # redirecting STDERR
832
833 $cmd .= " 2> \"$output_filestem.err\""
834 if ($ENV {'GSDLOS'} !~ m/^windows$/i || $is_winnt_2000);
835 #print STDERR "@@@@@@@@@ cmd=$cmd\n";
836
837 # execute the command
838 $!=0;
839 if (system($cmd)!=0)
840 {
841 print STDERR "Error executing $vbScript converter:$!\n";
842 if (-s "$output_filestem.err") {
843 open (ERRFILE, "<$output_filestem.err");
844
845 my $write_to_fail_log=0;
846 if ($faillogfile ne "" && defined(open(FAILLOG,">>$faillogfile")))
847 {$write_to_fail_log=1;}
848
849 my $line;
850 while ($line=<ERRFILE>) {
851 if ($line =~ m/\w/) {
852 print STDERR "$line";
853 print FAILLOG "$line" if ($write_to_fail_log);
854 }
855 if ($line !~ m/startup error/) {next;}
856 print STDERR " (given an invalid .DOC file?)\n";
857 print FAILLOG " (given an invalid .DOC file?)\n"
858 if ($write_to_fail_log);
859
860 } # while ERRFILE
861 close FAILLOG if ($write_to_fail_log);
862 }
863 return 0; # we can try any_to_text
864 }
865
866 # Was the conversion successful?
867 if (-s "$output_filestem.html") {
868 open(TMP, "$output_filestem.html");
869 my $line = <TMP>;
870 close(TMP);
871 if ($line && $line =~ m/html/i) {
872 &util::rm("$output_filestem.err") if -e "$output_filestem.err";
873 return 1;
874 }
875 }
876
877 # If here, an error of some sort occurred
878 &util::rm("$output_filestem.html") if -e "$output_filestem.html";
879 if (-e "$output_filestem.err") {
880 if ($faillogfile ne "" && defined(open(FAILLOG,">>$faillogfile"))) {
881 open (ERRLOG,"$output_filestem.err");
882 while (<ERRLOG>) {print FAILLOG $_;}
883 close FAILLOG;
884 close ERRLOG;
885 }
886 &util::rm("$output_filestem.err");
887 }
888 return 0;
889}
890
891# Attempt to convert an RTF document to html with rtftohtml
892sub rtf_to_html {
893 my ($input_filename, $output_filestem) = @_;
894
895 # formulate the command
896 my $cmd = "";
897 if ($timeout) {$cmd = "ulimit -t $timeout;";}
898 $cmd .= "rtftohtml";
899 #$cmd .= "rtf-converter";
900
901 $cmd .= " -o \"$output_filestem.html\" \"$input_filename\"";
902
903 $cmd .= " 2>\"$output_filestem.err\""
904 if ($ENV{'GSDLOS'} !~ m/^windows$/i || $is_winnt_2000);
905
906
907 # execute the command
908 $!=0;
909 if (system($cmd)!=0)
910 {
911 print STDERR "Error executing rtf converter $!\n";
912 # don't currently bother printing out error log...
913 # keep going, in case it still created an HTML file...
914 }
915
916 # Was the conversion successful?
917 my $was_successful=0;
918 if (-s "$output_filestem.html") {
919 # make sure we have some content other than header
920 open (HTML, "$output_filestem.html"); # what to do if fail?
921 my $line;
922 my $past_header=0;
923 while ($line=<HTML>) {
924
925 if ($past_header == 0) {
926 if ($line =~ m/<body>/) {$past_header=1;}
927 next;
928 }
929
930 $line =~ s/<[^>]+>//g;
931 if ($line =~ m/\w/ && $past_header) { # we found some content...
932 $was_successful=1;
933 last;
934 }
935 }
936 close HTML;
937 }
938
939 if ($was_successful) {
940 &util::rm("$output_filestem.err")
941 if (-e "$output_filestem.err");
942 # insert the (modified) table of contents, if it exists.
943 if (-e "${output_filestem}_ToC.html") {
944 &util::mv("$output_filestem.html","$output_filestem.src");
945 my $open_failed=0;
946 open HTMLSRC, "$output_filestem.src" || ++$open_failed;
947 open TOC, "${output_filestem}_ToC.html" || ++$open_failed;
948 open HTML, ">$output_filestem.html" || ++$open_failed;
949
950 if ($open_failed) {
951 close HTMLSRC;
952 close TOC;
953 close HTML;
954 &util::mv("$output_filestem.src","$output_filestem.html");
955 return 1;
956 }
957
958 # print out header info from src html.
959 while (defined($_ = <HTMLSRC>) && $_ =~ m/\w/) {
960 print HTML "$_";
961 }
962
963 # print out table of contents, making links relative
964 <TOC>; <TOC>; # ignore first 2 lines
965 print HTML scalar(<TOC>); # line 3 = "<ol>\n"
966 my $line;
967 while ($line=<TOC>) {
968 $line =~ s@</body></html>$@@i ; # only last line has this
969 # make link relative
970 $line =~ s@href=\"[^\#]+@href=\"@i;
971 print HTML $line;
972 }
973 close TOC;
974
975 # rest of html src
976 while (<HTMLSRC>) {
977 print HTML $_;
978 }
979 close HTMLSRC;
980 close HTML;
981
982 &util::rm("${output_filestem}_ToC.html");
983 &util::rm("${output_filestem}.src");
984 }
985 # we don't yet do anything with footnotes ($output_filestem_fn.html) :(
986 return 1; # success
987 }
988
989 if (-e "$output_filestem.err") {
990 if ($faillogfile ne "" && defined(open(FAILLOG,">>$faillogfile")))
991 {
992 print FAILLOG "Error - rtftohtml - couldn't extract text\n";
993 #print FAILLOG "Error - rtf-converter - couldn't extract text\n";
994 print FAILLOG " (rtf file might be too recent):\n";
995 open (ERRLOG, "$output_filestem.err");
996 while (<ERRLOG>) {print FAILLOG $_;}
997 close ERRLOG;
998 close FAILLOG;
999 }
1000 &util::rm("$output_filestem.err");
1001 }
1002
1003 &util::rm("$output_filestem.html") if (-e "$output_filestem.html");
1004
1005 return 0;
1006}
1007
1008
1009# Convert a pdf file to html with the pdftohtml command
1010
1011sub pdf_to_html {
1012 my ($dirname, $input_filename, $output_filestem) = @_;
1013
1014 my $cmd = "";
1015 if ($timeout) {$cmd = "ulimit -t $timeout;";}
1016 my $full_perl_path = &util::filename_cat($ENV{'PERLPATH'},"perl");
1017 $cmd .= "\"$full_perl_path\" -S pdftohtml.pl -zoom $pdf_zoom";
1018 $cmd .= " -c" if ($pdf_complex);
1019 $cmd .= " -i" if ($pdf_ignore_images);
1020 $cmd .= " -a" if ($pdf_allow_images_only);
1021 $cmd .= " -hidden" unless ($pdf_nohidden);
1022 $cmd .= " \"$input_filename\" \"$output_filestem\"";
1023
1024 if ($ENV{'GSDLOS'} !~ m/^windows$/i || $is_winnt_2000) {
1025 $cmd .= " > \"$output_filestem.out\" 2> \"$output_filestem.err\"";
1026 } else {
1027 $cmd .= " > \"$output_filestem.err\"";
1028 }
1029
1030 $!=0;
1031
1032 my $retval=system($cmd);
1033 if ($retval!=0)
1034 {
1035 print STDERR "Error executing pdftohtml.pl";
1036 if ($!) {print STDERR ": $!";}
1037 print STDERR "\n";
1038 }
1039
1040 # make sure the converter made something
1041 if ($retval!=0 || ! -s "$output_filestem.html")
1042 {
1043 &util::rm("$output_filestem.out") if (-e "$output_filestem.out");
1044 # print out the converter's std err, if any
1045 if (-s "$output_filestem.err") {
1046 open (ERRLOG, "$output_filestem.err") || die "$!";
1047 print STDERR "pdftohtml error log:\n";
1048 while (<ERRLOG>) {
1049 print STDERR "$_";
1050 }
1051 close ERRLOG;
1052 }
1053 print STDERR "***********output filestem $output_filestem.html\n";
1054 &util::rm("$output_filestem.html") if (-e "$output_filestem.html");
1055 if (-e "$output_filestem.err") {
1056 if ($faillogfile ne "" && defined(open(FAILLOG,">>$faillogfile")))
1057 {
1058 open (ERRLOG, "$output_filestem.err");
1059 while (<ERRLOG>) {print FAILLOG $_;}
1060 close ERRLOG;
1061 close FAILLOG;
1062 }
1063 &util::rm("$output_filestem.err");
1064 }
1065 return 0;
1066 }
1067
1068 &util::rm("$output_filestem.err") if (-e "$output_filestem.err");
1069 &util::rm("$output_filestem.out") if (-e "$output_filestem.out");
1070 return 1;
1071}
1072
1073# Convert a pdf file to various types of image with the convert command
1074
1075sub pdfps_to_img {
1076 my ($dirname, $input_filename, $output_filestem, $output_type) = @_;
1077
1078 # Check that ImageMagick is installed and available on the path (except for Windows 95/98)
1079 if (!($ENV{'GSDLOS'} eq "windows" && !Win32::IsWinNT())) {
1080 my $result = `identify 2>&1`;
1081 if ($? == -1 || $? == 256) { # Linux and Windows return different values for "program not found"
1082 #ImageMagick is not installed, thus the convert utility is not available.
1083 print STDERR "*** ImageMagick is not installed, the convert utility is not available. Unable to convert PDF/PS to images\n";
1084 return 0;
1085 }
1086 }
1087
1088 my $cmd = "";
1089 if ($timeout) {$cmd = "ulimit -t $timeout;";}
1090 $output_type =~ s/.*\_(.*)/$1/i;
1091 my $full_perl_path = &util::filename_cat($ENV{'PERLPATH'},"perl");
1092 $cmd .= "\"$full_perl_path\" -S pdfpstoimg.pl -convert_to $output_type \"$input_filename\" \"$output_filestem\"";
1093 if ($ENV{'GSDLOS'} !~ m/^windows$/i || $is_winnt_2000) {
1094 $cmd .= " > \"$output_filestem.out\" 2> \"$output_filestem.err\"";
1095 } else {
1096 $cmd .= " > \"$output_filestem.err\"";
1097 }
1098
1099 # don't include path on windows (to avoid having to play about
1100 # with quoting when GSDLHOME might contain spaces) but assume
1101 # that the PATH is set up correctly
1102 $!=0;
1103 my $retval=system($cmd);
1104 if ($retval!=0)
1105 {
1106 print STDERR "Error executing pdftoimg.pl";
1107 if ($!) {print STDERR ": $!";}
1108 print STDERR "\n";
1109 }
1110
1111 #make sure the converter made something
1112 #if ($retval !=0) || ! -s "$output_filestem")
1113 if ($retval !=0)
1114 {
1115 &util::rm("$output_filestem.out") if (-e "$output_filestem.out");
1116 #print out the converter's std err, if any
1117 if (-s "$output_filestem.err") {
1118 open (ERRLOG, "$output_filestem.err") || die "$!";
1119 print STDERR "pdfpstoimg error log:\n";
1120 while (<ERRLOG>) {
1121 print STDERR "$_";
1122 }
1123 close ERRLOG;
1124 }
1125 #&util::rm("$output_filestem.html") if (-e "$output_filestem.html");
1126 if (-e "$output_filestem.err") {
1127 if ($faillogfile ne "" && defined(open(FAILLOG,">>$faillogfile")))
1128 {
1129 open (ERRLOG, "$output_filestem.err");
1130 while (<ERRLOG>) {print FAILLOG $_;}
1131 close ERRLOG;
1132 close FAILLOG;
1133 }
1134 &util::rm("$output_filestem.err");
1135 }
1136 return 0;
1137 }
1138 &util::rm("$output_filestem.err") if (-e "$output_filestem.err");
1139 &util::rm("$output_filestem.out") if (-e "$output_filestem.out");
1140 return 1;
1141}
1142
1143# Convert a PDF file to text with the pdftotext command
1144
1145sub pdf_to_text {
1146 my ($dirname, $input_filename, $output_filestem) = @_;
1147
1148 my $cmd = "pdftotext \"$input_filename\" \"$output_filestem.text\"";
1149
1150 if ($ENV{'GSDLOS'} !~ m/^windows$/i) {
1151 $cmd .= " > \"$output_filestem.out\" 2> \"$output_filestem.err\"";
1152 } else {
1153 $cmd .= " > \"$output_filestem.err\"";
1154 }
1155
1156 if (system($cmd)!=0)
1157 {
1158 print STDERR "Error executing $cmd: $!\n";
1159 &util::rm("$output_filestem.text") if (-e "$output_filestem.text");
1160 }
1161
1162 # make sure there is some extracted text.
1163 if (-e "$output_filestem.text") {
1164 open (EXTR_TEXT, "$output_filestem.text") || warn "open: $!";
1165 binmode(EXTR_TEXT); # just in case...
1166 my $line="";
1167 my $seen_text=0;
1168 while (($seen_text==0) && ($line=<EXTR_TEXT>)) {
1169 if ($line=~ m/\w/) {$seen_text=1;}
1170 }
1171 close EXTR_TEXT;
1172 if ($seen_text==0) { # no text was extracted
1173 print STDERR "Error: pdftotext found no text\n";
1174 &util::rm("$output_filestem.text");
1175 }
1176 }
1177
1178 # make sure the converter made something
1179 if (! -s "$output_filestem.text")
1180 {
1181 # print out the converters std err, if any
1182 if (-s "$output_filestem.err") {
1183 open (ERRLOG, "$output_filestem.err") || die "$!";
1184 print STDERR "pdftotext error log:\n";
1185 while (<ERRLOG>) {
1186 print STDERR "$_";
1187 }
1188 close ERRLOG;
1189 }
1190 # does this converter create a .out file?
1191 &util::rm("$output_filestem.out") if (-e "$output_filestem.out");
1192 &util::rm("$output_filestem.text") if (-e "$output_filestem.text");
1193 if (-e "$output_filestem.err") {
1194 if ($faillogfile ne "" && defined(open(FAILLOG,">>$faillogfile")))
1195 {
1196 open (ERRLOG,"$output_filestem.err");
1197 while (<ERRLOG>) {print FAILLOG $_;}
1198 close ERRLOG;
1199 close FAILLOG;
1200 }
1201 &util::rm("$output_filestem.err");
1202 }
1203 return 0;
1204 }
1205 &util::rm("$output_filestem.err") if (-e "$output_filestem.err");
1206 return 1;
1207}
1208
1209# Convert a PostScript document to text
1210# note - just using "ps2ascii" isn't good enough, as it
1211# returns 0 for a postscript interpreter error. ps2ascii is just
1212# a wrapper to "gs" anyway, so we use that cmd here.
1213
1214sub ps_to_text {
1215 my ($input_filename, $output_filestem) = @_;
1216
1217 my $error = "";
1218
1219 # if we're on windows we'll fall straight through without attempting
1220 # to use gs
1221 if ($ENV{'GSDLOS'} =~ m/^windows$/i) {
1222 $error = "Windows does not support gs";
1223
1224 } else {
1225 my $cmd = "";
1226 if ($timeout) {$cmd = "ulimit -t $timeout; ";}
1227 $cmd .= "gs -q -dNODISPLAY -dNOBIND -dWRITESYSTEMDICT -dSIMPLE -c save ";
1228 $cmd .= "-f ps2ascii.ps \"$input_filename\" -c quit > \"$output_filestem.text\"";
1229 #$cmd .= "pstotext -output \"$output_filestem.text\" $input_filename\"";
1230 $cmd .= " 2> $output_filestem.err";
1231 $!=0;
1232
1233 my $retcode=system($cmd);
1234 $retcode = $? >> 8; # see man perlfunc - system for this...
1235 # if system returns -1 | 127 (couldn't start program), look at $! for message
1236
1237 if ($retcode!=0) {if ($!) {$error=$!;} else {$error="couldn't run.\n";}}
1238 elsif (! -e "$output_filestem.text") {
1239 $error="did not create output file.\n";
1240 }
1241 else
1242 { # make sure the interpreter didn't get an error. It is technically
1243 # possible for the actual text to start with this, but....
1244 open PSOUT, "$output_filestem.text";
1245 if (<PSOUT> =~ m/^Error: (.*)/) {
1246 $error="interpreter error - \"$1\"";
1247 }
1248 close PSOUT;
1249 }
1250 }
1251
1252 if ($error ne "")
1253 {
1254 print STDERR "Warning: Error executing gs: $error\n";
1255 &util::rm("$output_filestem.text") if (-e "$output_filestem.text");
1256
1257 if ("$faillogfile" ne "" && defined(open (FAILLOG, ">>$faillogfile")))
1258 {
1259 print FAILLOG "gs - $error\n";
1260 if (-e "$output_filestem.err") {
1261 open(ERRLOG, "$output_filestem.err");
1262 while (<ERRLOG>) {print FAILLOG $_;}
1263 close ERRLOG;
1264 }
1265 close FAILLOG;
1266 }
1267 &util::rm("$output_filestem.err") if (-e "$output_filestem.err");
1268
1269
1270 # Fine then. We'll just do a lousy job by ourselves...
1271 # Based on 5-line regexp sed script found at:
1272 # http://snark.ptc.spbu.ru/mail-archives/lout/brown/msg00003.html
1273 #
1274 print STDERR "Stripping text from postscript\n";
1275 my $errorcode=0;
1276 open (IN, "$input_filename")
1277 || ($errorcode=1, warn "Couldn't read file: $!");
1278 open (OUT, ">$output_filestem.text")
1279 || ($errorcode=1, warn "Couldn't write file: $!");
1280 if ($errorcode) {print STDERR "errors\n";return 0;}
1281
1282 my $text=""; # this is for whole .ps file...
1283 $text = join('', <IN>); # see man perlport, under "System Resources"
1284 close IN;
1285
1286 # Make sure this is a ps file...
1287 if ($text !~ m/^%!/) {
1288 print STDERR "Bad postscript header: not '%!'\n";
1289 if ($faillogfile ne "" && defined(open(FAILLOG, ">>$faillogfile")))
1290 {
1291 print FAILLOG "Bad postscript header: not '%!'\n";
1292 close FAILLOG;
1293 }
1294 return 0;
1295 }
1296
1297 # if ps has Page data, then use it to delete all stuff before it.
1298 $text =~ s/^.*?%%Page:.*?\n//s; # treat string as single line
1299
1300 # remove all leading non-data stuff
1301 $text =~ s/^.*?\(//s;
1302
1303 # remove all newline chars for easier processing
1304 $text =~ s/\n//g;
1305
1306 # Big assumption here - assume that if any co-ordinates are
1307 # given, then we are at the end of a sentence.
1308 $text =~ s/\)-?\d+\ -?\d+/\) \(\n\)/g;
1309
1310 # special characters--
1311 $text =~ s/\(\|\)/\(\ - \)/g; # j -> em-dash?
1312
1313 # ? ps text formatting (eg italics?) ?
1314 $text =~ s/Fn\(f\)/\(\{\)/g; # f -> {
1315 $text =~ s/Fn\(g\)/\(\}\)/g; # g -> }
1316 $text =~ s/Fn\(j\)/\(\|\)/g; # j -> |
1317 # default - remove the rest
1318 $text =~ s/\ ?F.\((.+?)\)/\($1\)/g;
1319
1320 # attempt to add whitespace between words...
1321 # this is based purely on observation, and may be completely wrong...
1322 $text =~ s/([^F])[defghijkuy]\(/$1 \( /g;
1323 # eg I notice "b(" is sometimes NOT a space if preceded by a
1324 # negative number.
1325 $text =~ s/\)\d+ ?b\(/\) \( /g;
1326
1327 # change quoted braces to brackets
1328 $text =~ s/([^\\])\\\(/$1\{/g;
1329 $text =~ s/([^\\])\\\)/$1\}/g ;
1330
1331 # remove everything that is not between braces
1332 $text =~ s/\)([^\(\)])+?\(//sg ;
1333
1334 # remove any Trailer eof stuff.
1335 $text =~ s/\)[^\)]*$//sg;
1336
1337 ### ligatures have special characters...
1338 $text =~ s/\\013/ff/g;
1339 $text =~ s/\\014/fi/g;
1340 $text =~ s/\\015/fl/g;
1341 $text =~ s/\\016/ffi/g;
1342 $text =~ s/\\214/fi/g;
1343 $text =~ s/\\215/fl/g;
1344 $text =~ s/\\017/\n\* /g; # asterisk?
1345 $text =~ s/\\023/\023/g; # e acute ('e)
1346 $text =~ s/\\177/\252/g; # u"
1347# $text =~ s/ ?? /\344/g; # a"
1348
1349 print OUT "$text";
1350 close OUT;
1351 }
1352 # wrap the text - use a minimum length. ie, first space after this length.
1353 my $wrap_length=72;
1354 &util::mv("$output_filestem.text", "$output_filestem.text.tmp");
1355 open INFILE, "$output_filestem.text.tmp" ||
1356 die "Couldn't open file: $!";
1357 open OUTFILE, ">$output_filestem.text" ||
1358 die "Couldn't open file for writing: $!";
1359 my $line="";
1360 while ($line=<INFILE>) {
1361 while (length($line)>0) {
1362 if (length($line)>$wrap_length) {
1363 $line =~ s/^(.{$wrap_length}[^\s]*)\s*//;
1364 print OUTFILE "$1\n";
1365 } else {
1366 print OUTFILE "$line";
1367 $line="";
1368 }
1369 }
1370 }
1371 close INFILE;
1372 close OUTFILE;
1373 &util::rm("$output_filestem.text.tmp");
1374
1375 &util::rm("$output_filestem.err") if (-e "$output_filestem.err");
1376 return 1;
1377}
1378
1379
1380# Convert any file to HTML with a crude perl implementation of the
1381# UNIX strings command.
1382
1383sub any_to_html {
1384 my ($input_filename, $output_filestem) = @_;
1385
1386 # First generate a text file
1387 return 0 unless (&any_to_text($input_filename, $output_filestem));
1388
1389 # create an HTML file from the text file
1390 open(TEXT, "<$output_filestem.text");
1391 open(HTML, ">$output_filestem.html");
1392
1393 print HTML "<html><head>\n";
1394 print HTML "<META HTTP-EQUIV=\"Content-Type\" CONTENT=\"text/html\">\n";
1395 print HTML "<META NAME=\"GENERATOR\" CONTENT=\"Greenstone any_to_html\">\n";
1396 print HTML "</head><body>\n\n";
1397
1398 my $line;
1399 while ($line=<TEXT>) {
1400 $line =~ s/</&lt;/g;
1401 $line =~ s/>/&gt;/g;
1402 if ($line =~ m/^\s*$/) {
1403 print HTML "<p>";
1404 } else {
1405 print HTML "<br> ", $line;
1406 }
1407 }
1408 print HTML "\n</body></html>\n";
1409
1410 close HTML;
1411 close TEXT;
1412
1413 &util::rm("$output_filestem.text") if (-e "$output_filestem.text");
1414 return 1;
1415}
1416
1417# Convert any file to TEXT with a crude perl implementation of the
1418# UNIX strings command.
1419# Note - this assumes ascii charsets :( (jrm21)
1420
1421sub any_to_text {
1422 my ($input_filename, $output_filestem) = @_;
1423
1424 if (!$use_strings) {
1425 return 0;
1426 }
1427
1428 print STDERR "\n**** In any to text****\n\n";
1429 open(IN, "<$input_filename") || return 0;
1430 binmode(IN);
1431 open(OUT, ">$output_filestem.text") || return 0;
1432
1433 my ($line);
1434 my $output_line_count = 0;
1435 while (<IN>) {
1436 $line = $_;
1437
1438 # delete anything that isn't a printable character
1439 $line =~ s/[^\040-\176]+/\n/sg;
1440
1441 # delete any string less than 10 characters long
1442 $line =~ s/^.{0,9}$/\n/mg;
1443 while ($line =~ m/^.{1,9}$/m) {
1444 $line =~ s/^.{0,9}$/\n/mg;
1445 $line =~ s/\n+/\n/sg;
1446 }
1447
1448 # remove extraneous whitespace
1449 $line =~ s/\n+/\n/gs;
1450 $line =~ s/^\n//gs;
1451
1452 # output whatever is left
1453 if ($line =~ m/[^\n ]/) {
1454 print OUT $line;
1455 ++$output_line_count;
1456 }
1457 }
1458
1459 close OUT;
1460 close IN;
1461
1462 if ($output_line_count) { # try to protect against binary only formats
1463 return 1;
1464 }
1465
1466 &util::rm("$output_filestem.text");
1467 return 0;
1468
1469}
Note: See TracBrowser for help on using the repository browser.