source: main/trunk/greenstone2/bin/script/gsConvert.pl@ 24608

Last change on this file since 24608 was 24608, checked in by ak19, 13 years ago

New debug output statement messes up sequence of error messages in PDF tutorial.
  • Property svn:executable set to *
  • Property svn:keywords set to Author Date Id Revision
File size: 35.7 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;
93my $verbosity=0;
94
95sub main
96{
97 my (@ARGV) = @_;
98 my ($input_type,$output_type,$verbose);
99
100 # Dynamically figure out what the --type option can support, based on whether -windows_scripting
101 # is in use or not
102 my $default_type_re = "(doc|dot|pdf|ps|ppt|rtf|xls)";
103 #my $enhanced_type_re = "(docx?|dot|pdf|ps|pptx?|rtf|xlsx?)";
104 #my $enhanced_type_re = "(docx?|dot|pdf|ps|pptx?|rtf|xlsx?)";
105 # Currently only have VBA for Word and PPT(but no XLS)
106 my $enhanced_type_re = "(docx?|dot|pdf|ps|pptx?|rtf|xls)";
107
108 my $type_re = $default_type_re;
109
110 foreach my $a (@ARGV) {
111 if ($a =~ m/^windows_scripting$/i) {
112 $type_re = $enhanced_type_re;
113 }
114 }
115
116 # read command-line arguments
117 if (!parsargv::parse(\@ARGV,
118 "type/$type_re/", \$input_type,
119 '/errlog/.*/', \$faillogfile,
120 'output/(auto|html|text|pagedimg).*/', \$output_type,
121 'timeout/\d+/0',\$timeout,
122 'verbose/\d+/0', \$verbose,
123 'windows_scripting',\$windows_scripting,
124 'use_strings', \$use_strings,
125 'pdf_complex', \$pdf_complex,
126 'pdf_ignore_images', \$pdf_ignore_images,
127 'pdf_allow_images_only', \$pdf_allow_images_only,
128 'pdf_nohidden', \$pdf_nohidden,
129 'pdf_zoom/\d+/2', \$pdf_zoom
130 ))
131 {
132 print_usage();
133 }
134
135 $verbosity=$verbose if defined $verbose;
136
137 # Make sure the input file exists and can be opened for reading
138 if (scalar(@ARGV!=1)) {
139 print_usage();
140 }
141
142 my $input_filename = $ARGV[0];
143 if (!-r $input_filename) {
144 print STDERR "Error: unable to open $input_filename for reading\n";
145 exit(1);
146 }
147
148 # Deduce filenames
149 my ($tailname,$dirname,$suffix)
150 = File::Basename::fileparse($input_filename, "\\.[^\\.]+\$");
151 my $output_filestem = &util::filename_cat($dirname, "$tailname");
152
153 if ($input_type eq "")
154 {
155 $input_type = lc (substr($suffix,1,length($suffix)-1));
156 }
157
158 # Change to temporary working directory
159 my $stored_dir = cwd();
160 chdir ($dirname) || die "Unable to change to directory $dirname";
161
162 # Select convert utility
163 if (!defined $input_type) {
164 print STDERR "Error: No filename extension or input type defined\n";
165 exit(1);
166 }
167 elsif ($input_type =~ m/^docx?$/ || $input_type eq "dot") {
168 print &convertDOC($input_filename, $output_filestem, $output_type);
169 print "\n";
170 }
171 elsif ($input_type eq "rtf") {
172 print &convertRTF($input_filename, $output_filestem, $output_type);
173 print "\n";
174 }
175 elsif ($input_type eq "pdf") {
176 print &convertPDF($dirname, $input_filename, $output_filestem, $output_type);
177 print "\n";
178 }
179 elsif ($input_type eq "ps") {
180 print &convertPS($dirname, $input_filename, $output_filestem, $output_type);
181 print "\n";
182 }
183 elsif ($input_type =~ m/pptx?$/) {
184 print &convertPPT($input_filename, $output_filestem, $output_type);
185 print "\n";
186 }
187 elsif ($input_type =~ m/xlsx?$/) {
188 print &convertXLS($input_filename, $output_filestem, $output_type);
189 print "\n";
190 }
191 else {
192 print STDERR "Error: Unable to convert type '$input_type'\n";
193 exit(1);
194 }
195
196 # restore to original working directory
197 chdir ($stored_dir) || die "Unable to return to directory $stored_dir";
198
199}
200
201&main(@ARGV);
202
203
204
205# Document-type conversion functions
206#
207# The following functions attempt to convert documents from their
208# input type to the specified output type. If no output type was
209# given, then they first attempt HTML, and then TEXT.
210#
211# Each returns the output type ("html" or "text") or "fail" if no
212# conversion is possible.
213
214# Convert a Microsoft word document
215
216sub convertDOC {
217 my ($input_filename, $output_filestem, $output_type) = @_;
218
219 # Many .doc files are not in fact word documents!
220 my $realtype = &find_docfile_type($input_filename);
221
222 if ($realtype eq "word6" || $realtype eq "word7"
223 || $realtype eq "word8" || $realtype eq "docx") {
224 return &convertWord678($input_filename, $output_filestem, $output_type);
225 } elsif ($realtype eq "rtf") {
226 return &convertRTF($input_filename, $output_filestem, $output_type);
227 } else {
228 return &convertAnything($input_filename, $output_filestem, $output_type);
229 }
230}
231
232# Convert a Microsoft word 6/7/8 document
233
234sub convertWord678 {
235 my ($input_filename, $output_filestem, $output_type) = @_;
236
237 my $success = 0;
238 if (!$output_type || ($output_type =~ m/html/i)){
239 if ($windows_scripting) {
240 $success = &native_doc_to_html($input_filename, $output_filestem);
241 }
242 else {
243 $success = &doc_to_html($input_filename, $output_filestem);
244 }
245 if ($success) {
246 return "html";
247 }
248 }
249 return &convertAnything($input_filename, $output_filestem, $output_type);
250}
251
252
253# Convert a Rich Text Format (RTF) file
254
255sub convertRTF {
256 my ($input_filename, $output_filestem, $output_type) = @_;
257
258 my $success = 0;
259
260 # Attempt specialised conversion to HTML
261 if (!$output_type || ($output_type =~ m/html/i)) {
262
263 if ($windows_scripting) {
264 $success = &native_doc_to_html($input_filename, $output_filestem);
265 }
266 else {
267 $success = &rtf_to_html($input_filename, $output_filestem);
268 }
269 if ($success) {
270 return "html";
271 }
272 }
273
274# rtf is so ugly that's it's not worth running strings over.
275# One day I'll write some quick'n'dirty regexps to try to extract text - jrm21
276# return &convertAnything($input_filename, $output_filestem, $output_type);
277 return "fail";
278}
279
280
281# Convert an unidentified file
282
283sub convertAnything {
284 my ($input_filename, $output_filestem, $output_type) = @_;
285
286 my $success = 0;
287
288 # Attempt simple conversion to HTML
289 if (!$output_type || ($output_type =~ m/html/i)) {
290 $success = &any_to_html($input_filename, $output_filestem);
291 if ($success) {
292 return "html";
293 }
294 }
295
296 # Convert to text
297 if (!$output_type || ($output_type =~ m/text/i)) {
298 $success = &any_to_text($input_filename, $output_filestem);
299 if ($success) {
300 return "text";
301 }
302 }
303 return "fail";
304}
305
306
307
308# Convert an Adobe PDF document
309
310sub convertPDF {
311 my ($dirname, $input_filename, $output_filestem, $output_type) = @_;
312
313 my $success = 0;
314 $output_type =~ s/.*\-(.*)/$1/i;
315 # Attempt coversion to Image
316 if ($output_type =~ m/jp?g|gif|png/i) {
317 $success = &pdfps_to_img($dirname, $input_filename, $output_filestem, $output_type);
318 if ($success){
319 return "item";
320 }
321 }
322
323 # Attempt conversion to HTML
324 if (!$output_type || ($output_type =~ m/html/i)) {
325 $success = &pdf_to_html($dirname, $input_filename, $output_filestem);
326 if ($success) {
327 return "html";
328 }
329 }
330
331 # Attempt conversion to TEXT
332 if (!$output_type || ($output_type =~ m/text/i)) {
333 $success = &pdf_to_text($dirname, $input_filename, $output_filestem);
334 if ($success) {
335 return "text";
336 }
337 }
338
339 return "fail";
340
341}
342
343
344# Convert an Adobe PostScript document
345
346sub convertPS {
347 my ($dirname,$input_filename, $output_filestem, $output_type) = @_;
348
349 my $success = 0;
350 $output_type =~ s/.*\-(.*)/$1/i;
351 # Attempt coversion to Image
352 if ($output_type =~ m/jp?g|gif|png/i) {
353 $success = &pdfps_to_img($dirname, $input_filename, $output_filestem, $output_type);
354 if ($success){
355 return "item";
356 }
357 }
358
359 # Attempt conversion to TEXT
360 if (!$output_type || ($output_type =~ m/text/i)) {
361 $success = &ps_to_text($input_filename, $output_filestem);
362 if ($success) {
363 return "text";
364 }
365 }
366 return "fail";
367}
368
369
370sub convertPPT {
371 my ($input_filename, $output_filestem, $output_type) = @_;
372 my $success = 0;
373
374 my $ppt_convert_type = "";
375
376 #if (!$output_type || $windows_scripting || ($output_type !~ m/html/i) || ($output_type !~ m/text/i)){
377 if ($windows_scripting && ($output_type !~ m/html/i) && ($output_type !~ m/text/i)){
378 if ($output_type =~ m/gif/i) {
379 $ppt_convert_type = "-g";
380 } elsif ($output_type =~ m/jp?g/i){
381 $ppt_convert_type = "-j";
382 } elsif ($output_type =~ m/png/i){
383 $ppt_convert_type = "-p";
384 }
385 my $vbScript = &util::filename_cat($ENV{'GSDLHOME'}, "bin",
386 $ENV{'GSDLOS'}, "pptextract");
387 $vbScript = "pptextract" if ($ENV{'GSDLOS'} =~ m/^windows$/i);
388
389 my $cmd = "";
390 if ($timeout) {$cmd = "ulimit -t $timeout;";}
391 # if the converting directory already exists
392 if (-d $output_filestem) {
393 print STDERR "**The conversion directory already exists\n";
394 return "item";
395 } else {
396 $cmd .= "$vbScript $ppt_convert_type \"$input_filename\" \"$output_filestem\"";
397 $cmd .= " 2>\"$output_filestem.err\""
398 if ($ENV{'GSDLOS'} !~ m/^windows$/i || $is_winnt_2000);
399 if (system($cmd) !=0) {
400 print STDERR "Powerpoint VB Scripting convert failed\n";
401 } else {
402 return "item";
403 }
404 }
405 } elsif (!$output_type || ($output_type =~ m/html/i)) {
406 # Attempt conversion to HTML
407 #if (!$output_type || ($output_type =~ m/html/i)) {
408 # formulate the command
409 my $cmd = "";
410 my $full_perl_path = &util::get_perl_exec();
411 $cmd .= "\"$full_perl_path\" -S ppttohtml.pl ";
412 $cmd .= " \"$input_filename\" \"$output_filestem.html\"";
413 $cmd .= " 2>\"$output_filestem.err\""
414 if ($ENV{'GSDLOS'} !~ m/^windows$/i || $is_winnt_2000);
415
416 # execute the command
417 $!=0;
418 if (system($cmd)!=0)
419 {
420 print STDERR "Powerpoint 95/97 converter failed $!\n";
421 } else {
422 return "html";
423 }
424 }
425
426 $success = &any_to_text($input_filename, $output_filestem);
427 if ($success) {
428 return "text";
429 }
430
431 return "fail";
432}
433
434
435sub convertXLS {
436 my ($input_filename, $output_filestem, $output_type) = @_;
437
438 my $success = 0;
439
440 # Attempt conversion to HTML
441 if (!$output_type || ($output_type =~ m/html/i)) {
442 # formulate the command
443 my $cmd = "";
444 my $full_perl_path = &util::get_perl_exec();
445 $cmd .= "\"$full_perl_path\" -S xlstohtml.pl ";
446 $cmd .= " \"$input_filename\" \"$output_filestem.html\"";
447 $cmd .= " 2>\"$output_filestem.err\""
448 if ($ENV{'GSDLOS'} !~ m/^windows$/i || $is_winnt_2000);
449
450
451 # execute the command
452 $!=0;
453 if (system($cmd)!=0)
454 {
455 print STDERR "Excel 95/97 converter failed $!\n";
456 } else {
457 return "html";
458 }
459 }
460
461 $success = &any_to_text($input_filename, $output_filestem);
462 if ($success) {
463 return "text";
464 }
465
466 return "fail";
467}
468
469
470
471# Find the real type of a .doc file
472#
473# We seem to have a lot of files with a .doc extension that are .rtf
474# files or Word 5 files. This function attempts to tell the difference.
475sub find_docfile_type {
476 my ($input_filename) = @_;
477
478 if (($windows_scripting) && ($input_filename =~ m/\.docx$/)) {
479 return "docx";
480 }
481
482 open(CHK, "<$input_filename");
483 binmode(CHK);
484 my $line = "";
485 my $first = 1;
486
487 while (<CHK>) {
488
489 $line = $_;
490
491 if ($first) {
492 # check to see if this is an rtf file
493 if ($line =~ m/^\{\\rtf/) {
494 close(CHK);
495 return "rtf";
496 }
497 $first = 0;
498 }
499
500 # is this is a word 6/7/8 document?
501 if ($line =~ m/Word\.Document\.([678])/) {
502 close(CHK);
503
504 return "word$1";
505 }
506
507 }
508
509 return "unknown";
510}
511
512
513# Specific type-to-type conversions
514#
515# Each of the following functions attempts to convert a document from
516# a specific format to another. If they succeed they return 1 and leave
517# the output document(s) in the appropriate place; if they fail they
518# return 0 and delete any working files.
519
520
521# Attempt to convert a word document to html with the wv program
522sub doc_to_html {
523 my ($input_filename, $output_filestem) = @_;
524
525 my $wvware_status = 0;
526
527 # need to ensure that the path to perl is quoted (in case there's spaces in it)
528 my $launch_cmd = "\"".&util::get_perl_exec()."\" -S wvware.pl \"$input_filename\" \"$output_filestem\" \"$faillogfile\" $verbosity $timeout";
529
530# print STDERR "***** wvware launch cmd = $launch_cmd\n";
531
532 $wvware_status = system($launch_cmd)/256;
533 return $wvware_status;
534}
535
536# Attempt to convert a word document to html with the word2html scripting program
537sub native_doc_to_html {
538 my ($input_filename, $output_filestem) = @_;
539
540 # build up the path to the doc-to-html conversion tool we're going to use
541 my $vbScript = &util::filename_cat($ENV{'GSDLHOME'}, "bin", $ENV{'GSDLOS'});
542
543 if ($ENV{'GSDLOS'} =~ m/^windows$/i) {
544 # if windows scripting with docx input, use new VBscript to get the local Word install (if
545 # any) to do the conversion, since docX can't be processed by word2html's windows_scripting
546
547 if($input_filename =~ m/docx$/i) { # need to use full path to docx2html script,
548 # else script launch fails when there are error msgs
549 $vbScript = &util::filename_cat($vbScript, "docx2html.vbs");
550 $vbScript = "CScript //Nologo \"$vbScript\""; # launche with CScript for error output in STDERR
551 # //Nologo flag avoids Microsoft's opening/logo msgs
552 print STDERR "About to use windows scripting to process docx file $input_filename.\n";
553 print STDERR " This may take some time. Please wait...\n";
554 }
555 else { # old doc versions. use the usual VB executable word2html for the
556 # conversion. Doesn't need full path, since bin\windows is on PATH
557 $vbScript = "word2html"; #$vbScript = "\"".&util::filename_cat($vbScript, "word2html")."\"";
558 }
559 }
560 else { # not windows
561 $vbScript = "\"".&util::filename_cat($vbScript, "word2html")."\"";
562 }
563
564 if (-e "$output_filestem.html") {
565 print STDERR " The conversion file:\n";
566 print STDERR " $output_filestem.html\n";
567 print STDERR " ... already exists. Skipping\n";
568 return 1;
569 }
570
571 my $cmd = "";
572 if ($timeout) {$cmd = "ulimit -t $timeout;";}
573 #$cmd .= "$vbScript \"$input_filename\" \"$output_filestem.html\"";
574 #$cmd .= "$vbScript $input_filename $output_filestem.html";
575 $cmd .= "$vbScript \"$input_filename\" \"$output_filestem.html\"";
576
577 # redirecting STDERR
578
579 $cmd .= " 2> \"$output_filestem.err\""
580 if ($ENV {'GSDLOS'} !~ m/^windows$/i || $is_winnt_2000);
581 #print STDERR "@@@@@@@@@ cmd=$cmd\n";
582
583 # execute the command
584 $!=0;
585 if (system($cmd)!=0)
586 {
587 print STDERR "Error executing $vbScript converter:$!\n";
588 if (-s "$output_filestem.err") {
589 open (ERRFILE, "<$output_filestem.err");
590
591 my $write_to_fail_log=0;
592 if ($faillogfile ne "" && defined(open(FAILLOG,">>$faillogfile")))
593 {$write_to_fail_log=1;}
594
595 my $line;
596 while ($line=<ERRFILE>) {
597 if ($line =~ m/\w/) {
598 print STDERR "$line";
599 print FAILLOG "$line" if ($write_to_fail_log);
600 }
601 if ($line !~ m/startup error/) {next;}
602 print STDERR " (given an invalid .DOC file?)\n";
603 print FAILLOG " (given an invalid .DOC file?)\n"
604 if ($write_to_fail_log);
605
606 } # while ERRFILE
607 close FAILLOG if ($write_to_fail_log);
608 }
609 return 0; # we can try any_to_text
610 }
611
612 # Was the conversion successful?
613 if (-s "$output_filestem.html") {
614 open(TMP, "$output_filestem.html");
615 my $line = <TMP>;
616 close(TMP);
617 if ($line && $line =~ m/html/i) {
618 &util::rm("$output_filestem.err") if -e "$output_filestem.err";
619 return 1;
620 }
621 }
622
623 # If here, an error of some sort occurred
624 &util::rm("$output_filestem.html") if -e "$output_filestem.html";
625 if (-e "$output_filestem.err") {
626 if ($faillogfile ne "" && defined(open(FAILLOG,">>$faillogfile"))) {
627 open (ERRLOG,"$output_filestem.err");
628 while (<ERRLOG>) {print FAILLOG $_;}
629 close FAILLOG;
630 close ERRLOG;
631 }
632 &util::rm("$output_filestem.err");
633 }
634 return 0;
635}
636
637# Attempt to convert an RTF document to html with rtftohtml
638sub rtf_to_html {
639 my ($input_filename, $output_filestem) = @_;
640
641 # formulate the command
642 my $cmd = "";
643 if ($timeout) {$cmd = "ulimit -t $timeout;";}
644 $cmd .= "rtftohtml";
645 #$cmd .= "rtf-converter";
646
647 $cmd .= " -o \"$output_filestem.html\" \"$input_filename\"";
648
649 $cmd .= " 2>\"$output_filestem.err\""
650 if ($ENV{'GSDLOS'} !~ m/^windows$/i || $is_winnt_2000);
651
652
653 # execute the command
654 $!=0;
655 if (system($cmd)!=0)
656 {
657 print STDERR "Error executing rtf converter $!\n";
658 # don't currently bother printing out error log...
659 # keep going, in case it still created an HTML file...
660 }
661
662 # Was the conversion successful?
663 my $was_successful=0;
664 if (-s "$output_filestem.html") {
665 # make sure we have some content other than header
666 open (HTML, "$output_filestem.html"); # what to do if fail?
667 my $line;
668 my $past_header=0;
669 while ($line=<HTML>) {
670
671 if ($past_header == 0) {
672 if ($line =~ m/<body>/) {$past_header=1;}
673 next;
674 }
675
676 $line =~ s/<[^>]+>//g;
677 if ($line =~ m/\w/ && $past_header) { # we found some content...
678 $was_successful=1;
679 last;
680 }
681 }
682 close HTML;
683 }
684
685 if ($was_successful) {
686 &util::rm("$output_filestem.err")
687 if (-e "$output_filestem.err");
688 # insert the (modified) table of contents, if it exists.
689 if (-e "${output_filestem}_ToC.html") {
690 &util::mv("$output_filestem.html","$output_filestem.src");
691 my $open_failed=0;
692 open HTMLSRC, "$output_filestem.src" || ++$open_failed;
693 open TOC, "${output_filestem}_ToC.html" || ++$open_failed;
694 open HTML, ">$output_filestem.html" || ++$open_failed;
695
696 if ($open_failed) {
697 close HTMLSRC;
698 close TOC;
699 close HTML;
700 &util::mv("$output_filestem.src","$output_filestem.html");
701 return 1;
702 }
703
704 # print out header info from src html.
705 while (defined($_ = <HTMLSRC>) && $_ =~ m/\w/) {
706 print HTML "$_";
707 }
708
709 # print out table of contents, making links relative
710 <TOC>; <TOC>; # ignore first 2 lines
711 print HTML scalar(<TOC>); # line 3 = "<ol>\n"
712 my $line;
713 while ($line=<TOC>) {
714 $line =~ s@</body></html>$@@i ; # only last line has this
715 # make link relative
716 $line =~ s@href=\"[^\#]+@href=\"@i;
717 print HTML $line;
718 }
719 close TOC;
720
721 # rest of html src
722 while (<HTMLSRC>) {
723 print HTML $_;
724 }
725 close HTMLSRC;
726 close HTML;
727
728 &util::rm("${output_filestem}_ToC.html");
729 &util::rm("${output_filestem}.src");
730 }
731 # we don't yet do anything with footnotes ($output_filestem_fn.html) :(
732 return 1; # success
733 }
734
735 if (-e "$output_filestem.err") {
736 if ($faillogfile ne "" && defined(open(FAILLOG,">>$faillogfile")))
737 {
738 print FAILLOG "Error - rtftohtml - couldn't extract text\n";
739 #print FAILLOG "Error - rtf-converter - couldn't extract text\n";
740 print FAILLOG " (rtf file might be too recent):\n";
741 open (ERRLOG, "$output_filestem.err");
742 while (<ERRLOG>) {print FAILLOG $_;}
743 close ERRLOG;
744 close FAILLOG;
745 }
746 &util::rm("$output_filestem.err");
747 }
748
749 &util::rm("$output_filestem.html") if (-e "$output_filestem.html");
750
751 return 0;
752}
753
754
755# Convert a pdf file to html with the pdftohtml command
756
757sub pdf_to_html {
758 my ($dirname, $input_filename, $output_filestem) = @_;
759
760 my $cmd = "";
761 if ($timeout) {$cmd = "ulimit -t $timeout;";}
762 my $full_perl_path = &util::get_perl_exec();
763 $cmd .= "\"$full_perl_path\" -S pdftohtml.pl -zoom $pdf_zoom";
764 $cmd .= " -c" if ($pdf_complex);
765 $cmd .= " -i" if ($pdf_ignore_images);
766 $cmd .= " -a" if ($pdf_allow_images_only);
767 $cmd .= " -hidden" unless ($pdf_nohidden);
768 $cmd .= " \"$input_filename\" \"$output_filestem\"";
769
770 if ($ENV{'GSDLOS'} !~ m/^windows$/i || $is_winnt_2000) {
771 $cmd .= " > \"$output_filestem.out\" 2> \"$output_filestem.err\"";
772 } else {
773 $cmd .= " > \"$output_filestem.err\"";
774 }
775
776 $!=0;
777
778 my $retval=system($cmd);
779 if ($retval!=0)
780 {
781 print STDERR "Error executing pdftohtml.pl";
782 if ($!) {print STDERR ": $!";}
783 print STDERR "\n";
784 }
785
786 # make sure the converter made something
787 if ($retval!=0 || ! -s "$output_filestem.html")
788 {
789 &util::rm("$output_filestem.out") if (-e "$output_filestem.out");
790 # print out the converter's std err, if any
791 if (-s "$output_filestem.err") {
792 open (ERRLOG, "$output_filestem.err") || die "$!";
793 print STDERR "pdftohtml error log:\n";
794 while (<ERRLOG>) {
795 print STDERR "$_";
796 }
797 close ERRLOG;
798 }
799 #print STDERR "***********output filestem $output_filestem.html\n";
800 &util::rm("$output_filestem.html") if (-e "$output_filestem.html");
801 if (-e "$output_filestem.err") {
802 if ($faillogfile ne "" && defined(open(FAILLOG,">>$faillogfile")))
803 {
804 open (ERRLOG, "$output_filestem.err");
805 while (<ERRLOG>) {print FAILLOG $_;}
806 close ERRLOG;
807 close FAILLOG;
808 }
809 &util::rm("$output_filestem.err");
810 }
811 return 0;
812 }
813
814 &util::rm("$output_filestem.err") if (-e "$output_filestem.err");
815 &util::rm("$output_filestem.out") if (-e "$output_filestem.out");
816 return 1;
817}
818
819# Convert a pdf file to various types of image with the convert command
820
821sub pdfps_to_img {
822 my ($dirname, $input_filename, $output_filestem, $output_type) = @_;
823
824 # Check that ImageMagick is installed and available on the path (except for Windows 95/98)
825 if (!($ENV{'GSDLOS'} eq "windows" && !Win32::IsWinNT())) {
826 my $imagick_cmd = "\"".&util::get_perl_exec()."\" -S gs-magick.pl";
827 my $result = `$imagick_cmd identify 2>&1`;
828
829 # Linux and Windows return different values for "program not found".
830 # Linux returns -1 and Windows 256 for "program not found". But once they're
831 # converted to signed values, it will be -1 for Linux and 1 for Windows.
832 # Whenever we test for return values other than 0, shift by 8 and perform
833 # unsigned to signed status conversion on $? to get expected range of return vals
834 # Although gs-magick.pl already shifts its $? by 8, converts it to a signed value
835 # and then exits on that, by the time we get here, we need to do it again
836 my $status = $?;
837 $status >>= 8;
838 $status = (($status & 0x80) ? -(0x100 - ($status & 0xFF)) : $status);
839 if ($status == -1 || $status == 1) { #if ($status == -1 || $status == 256) {
840 #ImageMagick is not installed, thus the convert utility is not available.
841 print STDERR "*** ImageMagick is not installed, the convert utility is not available. Unable to convert PDF/PS to images\n";
842 return 0;
843 }
844 }
845
846 my $cmd = "";
847 if ($timeout) {$cmd = "ulimit -t $timeout;";}
848 $output_type =~ s/.*\_(.*)/$1/i;
849 my $full_perl_path = &util::get_perl_exec();
850 $cmd .= "\"$full_perl_path\" -S pdfpstoimg.pl -convert_to $output_type \"$input_filename\" \"$output_filestem\"";
851 if ($ENV{'GSDLOS'} !~ m/^windows$/i || $is_winnt_2000) {
852 $cmd .= " > \"$output_filestem.out\" 2> \"$output_filestem.err\"";
853 } else {
854 $cmd .= " > \"$output_filestem.err\"";
855 }
856
857 # don't include path on windows (to avoid having to play about
858 # with quoting when GSDLHOME might contain spaces) but assume
859 # that the PATH is set up correctly
860 $!=0;
861 my $retval=system($cmd);
862 if ($retval!=0)
863 {
864 print STDERR "Error executing pdftoimg.pl";
865 if ($!) {print STDERR ": $!";}
866 print STDERR "\n";
867 }
868
869 #make sure the converter made something
870 #if ($retval !=0) || ! -s "$output_filestem")
871 if ($retval !=0)
872 {
873 &util::rm("$output_filestem.out") if (-e "$output_filestem.out");
874 #print out the converter's std err, if any
875 if (-s "$output_filestem.err") {
876 open (ERRLOG, "$output_filestem.err") || die "$!";
877 print STDERR "pdfpstoimg error log:\n";
878 while (<ERRLOG>) {
879 print STDERR "$_";
880 }
881 close ERRLOG;
882 }
883 #&util::rm("$output_filestem.html") if (-e "$output_filestem.html");
884 if (-e "$output_filestem.err") {
885 if ($faillogfile ne "" && defined(open(FAILLOG,">>$faillogfile")))
886 {
887 open (ERRLOG, "$output_filestem.err");
888 while (<ERRLOG>) {print FAILLOG $_;}
889 close ERRLOG;
890 close FAILLOG;
891 }
892 &util::rm("$output_filestem.err");
893 }
894 return 0;
895 }
896 &util::rm("$output_filestem.err") if (-e "$output_filestem.err");
897 &util::rm("$output_filestem.out") if (-e "$output_filestem.out");
898 return 1;
899}
900
901# Convert a PDF file to text with the pdftotext command
902
903sub pdf_to_text {
904 my ($dirname, $input_filename, $output_filestem) = @_;
905
906 my $cmd = "pdftotext \"$input_filename\" \"$output_filestem.text\"";
907
908 if ($ENV{'GSDLOS'} !~ m/^windows$/i) {
909 $cmd .= " > \"$output_filestem.out\" 2> \"$output_filestem.err\"";
910 } else {
911 $cmd .= " > \"$output_filestem.err\"";
912 }
913
914 if (system($cmd)!=0)
915 {
916 print STDERR "Error executing $cmd: $!\n";
917 &util::rm("$output_filestem.text") if (-e "$output_filestem.text");
918 }
919
920 # make sure there is some extracted text.
921 if (-e "$output_filestem.text") {
922 open (EXTR_TEXT, "$output_filestem.text") || warn "open: $!";
923 binmode(EXTR_TEXT); # just in case...
924 my $line="";
925 my $seen_text=0;
926 while (($seen_text==0) && ($line=<EXTR_TEXT>)) {
927 if ($line=~ m/\w/) {$seen_text=1;}
928 }
929 close EXTR_TEXT;
930 if ($seen_text==0) { # no text was extracted
931 print STDERR "Error: pdftotext found no text\n";
932 &util::rm("$output_filestem.text");
933 }
934 }
935
936 # make sure the converter made something
937 if (! -s "$output_filestem.text")
938 {
939 # print out the converters std err, if any
940 if (-s "$output_filestem.err") {
941 open (ERRLOG, "$output_filestem.err") || die "$!";
942 print STDERR "pdftotext error log:\n";
943 while (<ERRLOG>) {
944 print STDERR "$_";
945 }
946 close ERRLOG;
947 }
948 # does this converter create a .out file?
949 &util::rm("$output_filestem.out") if (-e "$output_filestem.out");
950 &util::rm("$output_filestem.text") if (-e "$output_filestem.text");
951 if (-e "$output_filestem.err") {
952 if ($faillogfile ne "" && defined(open(FAILLOG,">>$faillogfile")))
953 {
954 open (ERRLOG,"$output_filestem.err");
955 while (<ERRLOG>) {print FAILLOG $_;}
956 close ERRLOG;
957 close FAILLOG;
958 }
959 &util::rm("$output_filestem.err");
960 }
961 return 0;
962 }
963 &util::rm("$output_filestem.err") if (-e "$output_filestem.err");
964 return 1;
965}
966
967# Convert a PostScript document to text
968# note - just using "ps2ascii" isn't good enough, as it
969# returns 0 for a postscript interpreter error. ps2ascii is just
970# a wrapper to "gs" anyway, so we use that cmd here.
971
972sub ps_to_text {
973 my ($input_filename, $output_filestem) = @_;
974
975 my $error = "";
976
977 # if we're on windows we'll fall straight through without attempting
978 # to use gs
979 if ($ENV{'GSDLOS'} =~ m/^windows$/i) {
980 $error = "Windows does not support gs";
981
982 } else {
983 my $cmd = "";
984 if ($timeout) {$cmd = "ulimit -t $timeout; ";}
985 $cmd .= "gs -q -dNODISPLAY -dNOBIND -dWRITESYSTEMDICT -dSIMPLE -c save ";
986 $cmd .= "-f ps2ascii.ps \"$input_filename\" -c quit > \"$output_filestem.text\"";
987 #$cmd .= "pstotext -output \"$output_filestem.text\" $input_filename\"";
988 $cmd .= " 2> $output_filestem.err";
989 $!=0;
990
991 my $retcode=system($cmd);
992 $retcode = $? >> 8; # see man perlfunc - system for this...
993 # if system returns -1 | 127 (couldn't start program), look at $! for message
994
995 if ($retcode!=0) {if ($!) {$error=$!;} else {$error="couldn't run.\n";}}
996 elsif (! -e "$output_filestem.text") {
997 $error="did not create output file.\n";
998 }
999 else
1000 { # make sure the interpreter didn't get an error. It is technically
1001 # possible for the actual text to start with this, but....
1002 open PSOUT, "$output_filestem.text";
1003 if (<PSOUT> =~ m/^Error: (.*)/) {
1004 $error="interpreter error - \"$1\"";
1005 }
1006 close PSOUT;
1007 }
1008 }
1009
1010 if ($error ne "")
1011 {
1012 print STDERR "Warning: Error executing gs: $error\n";
1013 &util::rm("$output_filestem.text") if (-e "$output_filestem.text");
1014
1015 if ("$faillogfile" ne "" && defined(open (FAILLOG, ">>$faillogfile")))
1016 {
1017 print FAILLOG "gs - $error\n";
1018 if (-e "$output_filestem.err") {
1019 open(ERRLOG, "$output_filestem.err");
1020 while (<ERRLOG>) {print FAILLOG $_;}
1021 close ERRLOG;
1022 }
1023 close FAILLOG;
1024 }
1025 &util::rm("$output_filestem.err") if (-e "$output_filestem.err");
1026
1027
1028 # Fine then. We'll just do a lousy job by ourselves...
1029 # Based on 5-line regexp sed script found at:
1030 # http://snark.ptc.spbu.ru/mail-archives/lout/brown/msg00003.html
1031 #
1032 print STDERR "Stripping text from postscript\n";
1033 my $errorcode=0;
1034 open (IN, "$input_filename")
1035 || ($errorcode=1, warn "Couldn't read file: $!");
1036 open (OUT, ">$output_filestem.text")
1037 || ($errorcode=1, warn "Couldn't write file: $!");
1038 if ($errorcode) {print STDERR "errors\n";return 0;}
1039
1040 my $text=""; # this is for whole .ps file...
1041 $text = join('', <IN>); # see man perlport, under "System Resources"
1042 close IN;
1043
1044 # Make sure this is a ps file...
1045 if ($text !~ m/^%!/) {
1046 print STDERR "Bad postscript header: not '%!'\n";
1047 if ($faillogfile ne "" && defined(open(FAILLOG, ">>$faillogfile")))
1048 {
1049 print FAILLOG "Bad postscript header: not '%!'\n";
1050 close FAILLOG;
1051 }
1052 return 0;
1053 }
1054
1055 # if ps has Page data, then use it to delete all stuff before it.
1056 $text =~ s/^.*?%%Page:.*?\n//s; # treat string as single line
1057
1058 # remove all leading non-data stuff
1059 $text =~ s/^.*?\(//s;
1060
1061 # remove all newline chars for easier processing
1062 $text =~ s/\n//g;
1063
1064 # Big assumption here - assume that if any co-ordinates are
1065 # given, then we are at the end of a sentence.
1066 $text =~ s/\)-?\d+\ -?\d+/\) \(\n\)/g;
1067
1068 # special characters--
1069 $text =~ s/\(\|\)/\(\ - \)/g; # j -> em-dash?
1070
1071 # ? ps text formatting (eg italics?) ?
1072 $text =~ s/Fn\(f\)/\(\{\)/g; # f -> {
1073 $text =~ s/Fn\(g\)/\(\}\)/g; # g -> }
1074 $text =~ s/Fn\(j\)/\(\|\)/g; # j -> |
1075 # default - remove the rest
1076 $text =~ s/\ ?F.\((.+?)\)/\($1\)/g;
1077
1078 # attempt to add whitespace between words...
1079 # this is based purely on observation, and may be completely wrong...
1080 $text =~ s/([^F])[defghijkuy]\(/$1 \( /g;
1081 # eg I notice "b(" is sometimes NOT a space if preceded by a
1082 # negative number.
1083 $text =~ s/\)\d+ ?b\(/\) \( /g;
1084
1085 # change quoted braces to brackets
1086 $text =~ s/([^\\])\\\(/$1\{/g;
1087 $text =~ s/([^\\])\\\)/$1\}/g ;
1088
1089 # remove everything that is not between braces
1090 $text =~ s/\)([^\(\)])+?\(//sg ;
1091
1092 # remove any Trailer eof stuff.
1093 $text =~ s/\)[^\)]*$//sg;
1094
1095 ### ligatures have special characters...
1096 $text =~ s/\\013/ff/g;
1097 $text =~ s/\\014/fi/g;
1098 $text =~ s/\\015/fl/g;
1099 $text =~ s/\\016/ffi/g;
1100 $text =~ s/\\214/fi/g;
1101 $text =~ s/\\215/fl/g;
1102 $text =~ s/\\017/\n\* /g; # asterisk?
1103 $text =~ s/\\023/\023/g; # e acute ('e)
1104 $text =~ s/\\177/\252/g; # u"
1105# $text =~ s/ ?? /\344/g; # a"
1106
1107 print OUT "$text";
1108 close OUT;
1109 }
1110 # wrap the text - use a minimum length. ie, first space after this length.
1111 my $wrap_length=72;
1112 &util::mv("$output_filestem.text", "$output_filestem.text.tmp");
1113 open INFILE, "$output_filestem.text.tmp" ||
1114 die "Couldn't open file: $!";
1115 open OUTFILE, ">$output_filestem.text" ||
1116 die "Couldn't open file for writing: $!";
1117 my $line="";
1118 while ($line=<INFILE>) {
1119 while (length($line)>0) {
1120 if (length($line)>$wrap_length) {
1121 $line =~ s/^(.{$wrap_length}[^\s]*)\s*//;
1122 print OUTFILE "$1\n";
1123 } else {
1124 print OUTFILE "$line";
1125 $line="";
1126 }
1127 }
1128 }
1129 close INFILE;
1130 close OUTFILE;
1131 &util::rm("$output_filestem.text.tmp");
1132
1133 &util::rm("$output_filestem.err") if (-e "$output_filestem.err");
1134 return 1;
1135}
1136
1137
1138# Convert any file to HTML with a crude perl implementation of the
1139# UNIX strings command.
1140
1141sub any_to_html {
1142 my ($input_filename, $output_filestem) = @_;
1143
1144 # First generate a text file
1145 return 0 unless (&any_to_text($input_filename, $output_filestem));
1146
1147 # create an HTML file from the text file
1148 open(TEXT, "<$output_filestem.text");
1149 open(HTML, ">$output_filestem.html");
1150
1151 print HTML "<html><head>\n";
1152 print HTML "<META HTTP-EQUIV=\"Content-Type\" CONTENT=\"text/html\">\n";
1153 print HTML "<META NAME=\"GENERATOR\" CONTENT=\"Greenstone any_to_html\">\n";
1154 print HTML "</head><body>\n\n";
1155
1156 my $line;
1157 while ($line=<TEXT>) {
1158 $line =~ s/</&lt;/g;
1159 $line =~ s/>/&gt;/g;
1160 if ($line =~ m/^\s*$/) {
1161 print HTML "<p>";
1162 } else {
1163 print HTML "<br> ", $line;
1164 }
1165 }
1166 print HTML "\n</body></html>\n";
1167
1168 close HTML;
1169 close TEXT;
1170
1171 &util::rm("$output_filestem.text") if (-e "$output_filestem.text");
1172 return 1;
1173}
1174
1175# Convert any file to TEXT with a crude perl implementation of the
1176# UNIX strings command.
1177# Note - this assumes ascii charsets :( (jrm21)
1178
1179sub any_to_text {
1180 my ($input_filename, $output_filestem) = @_;
1181
1182 if (!$use_strings) {
1183 return 0;
1184 }
1185
1186 print STDERR "\n**** In any to text****\n\n";
1187 open(IN, "<$input_filename") || return 0;
1188 binmode(IN);
1189 open(OUT, ">$output_filestem.text") || return 0;
1190
1191 my ($line);
1192 my $output_line_count = 0;
1193 while (<IN>) {
1194 $line = $_;
1195
1196 # delete anything that isn't a printable character
1197 $line =~ s/[^\040-\176]+/\n/sg;
1198
1199 # delete any string less than 10 characters long
1200 $line =~ s/^.{0,9}$/\n/mg;
1201 while ($line =~ m/^.{1,9}$/m) {
1202 $line =~ s/^.{0,9}$/\n/mg;
1203 $line =~ s/\n+/\n/sg;
1204 }
1205
1206 # remove extraneous whitespace
1207 $line =~ s/\n+/\n/gs;
1208 $line =~ s/^\n//gs;
1209
1210 # output whatever is left
1211 if ($line =~ m/[^\n ]/) {
1212 print OUT $line;
1213 ++$output_line_count;
1214 }
1215 }
1216
1217 close OUT;
1218 close IN;
1219
1220 if ($output_line_count) { # try to protect against binary only formats
1221 return 1;
1222 }
1223
1224 &util::rm("$output_filestem.text");
1225 return 0;
1226
1227}
Note: See TracBrowser for help on using the repository browser.