source: main/trunk/greenstone2/bin/script/gsConvert.pl@ 28355

Last change on this file since 28355 was 28355, checked in by ak19, 8 years ago
  1. Now gsConvert.pl calls the new pptextract.vbs VBScript (which creates .item files and ppt slide.txt files in utf-8) instead of the older VB pptextract.exe executable which created .item and slide.txt files in windows default utf-16 LE. 2. PagedImagePlugin.pm::tidy_item_file now reads in the .item files in utf-8 mode, so that its strings are unicode aware. Substitutions are of unicode code points instead of byte sequences, since the strings in the file are now unicode aware.
  • Property svn:executable set to *
  • Property svn:keywords set to Author Date Id Revision
File size: 36.6 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 FileUtils;
56use Cwd;
57
58# Are we running on WinNT or Win2000 (or later)?
59my $is_winnt_2000=eval {require Win32; return (Win32::IsWinNT()); return 0;};
60if (!defined($is_winnt_2000)) {$is_winnt_2000=0;}
61
62my $use_strings;
63my $pdf_complex;
64my $pdf_nohidden;
65my $pdf_zoom;
66my $pdf_ignore_images;
67my $pdf_allow_images_only;
68my $windows_scripting;
69
70sub print_usage
71{
72 print STDERR "\n";
73 print STDERR "gsConvert.pl: Converts documents in a range of formats to html\n";
74 print STDERR " or text using third-party programs.\n\n";
75 print STDERR " usage: $0 [options] filename\n";
76 print STDERR " options:\n\t-type\tdoc|dot|pdf|ps|ppt|rtf|xls\t(input file type)\n";
77 print STDERR "\t-errlog\t<filename>\t(append err messages)\n";
78 print STDERR "\t-output\tauto|html|text|pagedimg_jpg|pagedimg_gif|pagedimg_png\t(output file type)\n";
79 print STDERR "\t-timeout\t<max cpu seconds>\t(ulimit on unix systems)\n";
80 print STDERR "\t-use_strings\tuse strings to extract text if conversion fails\n";
81 print STDERR "\t-windows_scripting\tuse windows VB script (if available) to convert Microsoft Word and PPT documents\n";
82 print STDERR "\t-pdf_complex\tuse complex output when converting PDF to HTML\n";
83 print STDERR "\t-pdf_nohidden\tDon't attempt to extract hidden text from PDF files\n";
84 print STDERR "\t-pdf_ignore_images\tdon't attempt to extract images when\n";
85 print STDERR "\t\tconverting PDF to HTML\n";
86 print STDERR "\t-pdf_allow_images_only\tallow images only (continue even if no text is present when converting to HTML)\n";
87 print STDERR "\t-pdf_zoom\tfactor by which to zoom PDF (only useful if\n";
88 print STDERR "\t\t-pdf_complex is set\n";
89 exit(1);
90}
91
92my $faillogfile="";
93my $timeout=0;
94my $verbosity=0;
95
96sub main
97{
98 my (@ARGV) = @_;
99 my ($input_type,$output_type,$verbose);
100
101 # Dynamically figure out what the --type option can support, based on whether -windows_scripting
102 # is in use or not
103 my $default_type_re = "(doc|dot|pdf|ps|ppt|rtf|xls)";
104 #my $enhanced_type_re = "(docx?|dot|pdf|ps|pptx?|rtf|xlsx?)";
105 #my $enhanced_type_re = "(docx?|dot|pdf|ps|pptx?|rtf|xlsx?)";
106 # Currently only have VBA for Word and PPT(but no XLS)
107 my $enhanced_type_re = "(docx?|dot|pdf|ps|pptx?|rtf|xls)";
108
109 my $type_re = $default_type_re;
110
111 foreach my $a (@ARGV) {
112 if ($a =~ m/^windows_scripting$/i) {
113 $type_re = $enhanced_type_re;
114 }
115 }
116
117 # read command-line arguments
118 if (!parsargv::parse(\@ARGV,
119 "type/$type_re/", \$input_type,
120 '/errlog/.*/', \$faillogfile,
121 'output/(auto|html|text|pagedimg).*/', \$output_type,
122 'timeout/\d+/0',\$timeout,
123 'verbose/\d+/0', \$verbose,
124 'windows_scripting',\$windows_scripting,
125 'use_strings', \$use_strings,
126 'pdf_complex', \$pdf_complex,
127 'pdf_ignore_images', \$pdf_ignore_images,
128 'pdf_allow_images_only', \$pdf_allow_images_only,
129 'pdf_nohidden', \$pdf_nohidden,
130 'pdf_zoom/\d+/2', \$pdf_zoom
131 ))
132 {
133 print_usage();
134 }
135
136 $verbosity=$verbose if defined $verbose;
137
138 # Make sure the input file exists and can be opened for reading
139 if (scalar(@ARGV!=1)) {
140 print_usage();
141 }
142
143 my $input_filename = $ARGV[0];
144 if (!-r $input_filename) {
145 print STDERR "Error: unable to open $input_filename for reading\n";
146 exit(1);
147 }
148
149 # Deduce filenames
150 my ($tailname,$dirname,$suffix)
151 = File::Basename::fileparse($input_filename, "\\.[^\\.]+\$");
152 my $output_filestem = &FileUtils::filenameConcatenate($dirname, "$tailname");
153
154 if ($input_type eq "")
155 {
156 $input_type = lc (substr($suffix,1,length($suffix)-1));
157 }
158
159 # Change to temporary working directory
160 my $stored_dir = cwd();
161 chdir ($dirname) || die "Unable to change to directory $dirname";
162
163 # Select convert utility
164 if (!defined $input_type) {
165 print STDERR "Error: No filename extension or input type defined\n";
166 exit(1);
167 }
168 elsif ($input_type =~ m/^docx?$/ || $input_type eq "dot") {
169 print &convertDOC($input_filename, $output_filestem, $output_type);
170 print "\n";
171 }
172 elsif ($input_type eq "rtf") {
173 print &convertRTF($input_filename, $output_filestem, $output_type);
174 print "\n";
175 }
176 elsif ($input_type eq "pdf") {
177 print &convertPDF($dirname, $input_filename, $output_filestem, $output_type);
178 print "\n";
179 }
180 elsif ($input_type eq "ps") {
181 print &convertPS($dirname, $input_filename, $output_filestem, $output_type);
182 print "\n";
183 }
184 elsif ($input_type =~ m/pptx?$/) {
185 print &convertPPT($input_filename, $output_filestem, $output_type);
186 print "\n";
187 }
188 elsif ($input_type =~ m/xlsx?$/) {
189 print &convertXLS($input_filename, $output_filestem, $output_type);
190 print "\n";
191 }
192 else {
193 print STDERR "Error: Unable to convert type '$input_type'\n";
194 exit(1);
195 }
196
197 # restore to original working directory
198 chdir ($stored_dir) || die "Unable to return to directory $stored_dir";
199
200}
201
202&main(@ARGV);
203
204
205
206# Document-type conversion functions
207#
208# The following functions attempt to convert documents from their
209# input type to the specified output type. If no output type was
210# given, then they first attempt HTML, and then TEXT.
211#
212# Each returns the output type ("html" or "text") or "fail" if no
213# conversion is possible.
214
215# Convert a Microsoft word document
216
217sub convertDOC {
218 my ($input_filename, $output_filestem, $output_type) = @_;
219
220 # Many .doc files are not in fact word documents!
221 my $realtype = &find_docfile_type($input_filename);
222
223 if ($realtype eq "word6" || $realtype eq "word7"
224 || $realtype eq "word8" || $realtype eq "docx") {
225 return &convertWord678($input_filename, $output_filestem, $output_type);
226 } elsif ($realtype eq "rtf") {
227 return &convertRTF($input_filename, $output_filestem, $output_type);
228 } else {
229 return &convertAnything($input_filename, $output_filestem, $output_type);
230 }
231}
232
233# Convert a Microsoft word 6/7/8 document
234
235sub convertWord678 {
236 my ($input_filename, $output_filestem, $output_type) = @_;
237
238 my $success = 0;
239 if (!$output_type || ($output_type =~ m/html/i)){
240 if ($windows_scripting) {
241 $success = &native_doc_to_html($input_filename, $output_filestem);
242 }
243 else {
244 $success = &doc_to_html($input_filename, $output_filestem);
245 }
246 if ($success) {
247 return "html";
248 }
249 }
250 return &convertAnything($input_filename, $output_filestem, $output_type);
251}
252
253
254# Convert a Rich Text Format (RTF) file
255
256sub convertRTF {
257 my ($input_filename, $output_filestem, $output_type) = @_;
258
259 my $success = 0;
260
261 # Attempt specialised conversion to HTML
262 if (!$output_type || ($output_type =~ m/html/i)) {
263
264 if ($windows_scripting) {
265 $success = &native_doc_to_html($input_filename, $output_filestem);
266 }
267 else {
268 $success = &rtf_to_html($input_filename, $output_filestem);
269 }
270 if ($success) {
271 return "html";
272 }
273 }
274
275# rtf is so ugly that's it's not worth running strings over.
276# One day I'll write some quick'n'dirty regexps to try to extract text - jrm21
277# return &convertAnything($input_filename, $output_filestem, $output_type);
278 return "fail";
279}
280
281
282# Convert an unidentified file
283
284sub convertAnything {
285 my ($input_filename, $output_filestem, $output_type) = @_;
286
287 my $success = 0;
288
289 # Attempt simple conversion to HTML
290 if (!$output_type || ($output_type =~ m/html/i)) {
291 $success = &any_to_html($input_filename, $output_filestem);
292 if ($success) {
293 return "html";
294 }
295 }
296
297 # Convert to text
298 if (!$output_type || ($output_type =~ m/text/i)) {
299 $success = &any_to_text($input_filename, $output_filestem);
300 if ($success) {
301 return "text";
302 }
303 }
304 return "fail";
305}
306
307
308
309# Convert an Adobe PDF document
310
311sub convertPDF {
312 my ($dirname, $input_filename, $output_filestem, $output_type) = @_;
313
314 my $success = 0;
315 $output_type =~ s/.*\-(.*)/$1/i;
316 # Attempt coversion to Image
317 if ($output_type =~ m/jp?g|gif|png/i) {
318 $success = &pdfps_to_img($dirname, $input_filename, $output_filestem, $output_type);
319 if ($success){
320 return "item";
321 }
322 }
323
324 # Attempt conversion to HTML
325 if (!$output_type || ($output_type =~ m/html/i)) {
326 $success = &pdf_to_html($dirname, $input_filename, $output_filestem);
327 if ($success) {
328 return "html";
329 }
330 }
331
332 # Attempt conversion to TEXT
333 if (!$output_type || ($output_type =~ m/text/i)) {
334 $success = &pdf_to_text($dirname, $input_filename, $output_filestem);
335 if ($success) {
336 return "text";
337 }
338 }
339
340 return "fail";
341
342}
343
344
345# Convert an Adobe PostScript document
346
347sub convertPS {
348 my ($dirname,$input_filename, $output_filestem, $output_type) = @_;
349
350 my $success = 0;
351 $output_type =~ s/.*\-(.*)/$1/i;
352 # Attempt coversion to Image
353 if ($output_type =~ m/jp?g|gif|png/i) {
354 $success = &pdfps_to_img($dirname, $input_filename, $output_filestem, $output_type);
355 if ($success){
356 return "item";
357 }
358 }
359
360 # Attempt conversion to TEXT
361 if (!$output_type || ($output_type =~ m/text/i)) {
362 $success = &ps_to_text($input_filename, $output_filestem);
363 if ($success) {
364 return "text";
365 }
366 }
367 return "fail";
368}
369
370
371sub convertPPT {
372 my ($input_filename, $output_filestem, $output_type) = @_;
373 my $success = 0;
374
375 my $ppt_convert_type = "";
376
377 #if (!$output_type || $windows_scripting || ($output_type !~ m/html/i) || ($output_type !~ m/text/i)){
378 if ($windows_scripting && ($output_type !~ m/html/i) && ($output_type !~ m/text/i)){
379 if ($output_type =~ m/gif/i) {
380 $ppt_convert_type = "-g";
381 } elsif ($output_type =~ m/jp?g/i){
382 $ppt_convert_type = "-j";
383 } elsif ($output_type =~ m/png/i){
384 $ppt_convert_type = "-p";
385 }
386 my $vbScript = &FileUtils::filenameConcatenate($ENV{'GSDLHOME'}, "bin",
387 $ENV{'GSDLOS'}, "pptextract");
388 $vbScript = "CScript //Nologo \"".$vbScript.".vbs\"" if ($ENV{'GSDLOS'} =~ m/^windows$/i); # now we use the .vbs VBScript
389 # $vbScript = "pptextract" if ($ENV{'GSDLOS'} =~ m/^windows$/i); # back when the pptextract.exe VB executable was used
390
391 my $cmd = "";
392 if ($timeout) {$cmd = "ulimit -t $timeout;";}
393 # if the converting directory already exists
394 if (-d $output_filestem) {
395 print STDERR "**The conversion directory already exists\n";
396 return "item";
397 } else {
398 $cmd .= "$vbScript $ppt_convert_type \"$input_filename\" \"$output_filestem\"";
399 $cmd .= " 2>\"$output_filestem.err\""
400 if ($ENV{'GSDLOS'} !~ m/^windows$/i || $is_winnt_2000);
401
402 if (system($cmd) !=0) {
403 print STDERR "Powerpoint VB Scripting convert failed\n";
404 } else {
405 return "item";
406 }
407 }
408 } elsif (!$output_type || ($output_type =~ m/html/i)) {
409 # Attempt conversion to HTML
410 #if (!$output_type || ($output_type =~ m/html/i)) {
411 # formulate the command
412 my $cmd = "";
413 my $full_perl_path = &util::get_perl_exec();
414 $cmd .= "\"$full_perl_path\" -S ppttohtml.pl ";
415 $cmd .= " \"$input_filename\" \"$output_filestem.html\"";
416 $cmd .= " 2>\"$output_filestem.err\""
417 if ($ENV{'GSDLOS'} !~ m/^windows$/i || $is_winnt_2000);
418
419 # execute the command
420 $!=0;
421 if (system($cmd)!=0)
422 {
423 print STDERR "Powerpoint 95/97 converter failed $!\n";
424 } else {
425 return "html";
426 }
427 }
428
429 $success = &any_to_text($input_filename, $output_filestem);
430 if ($success) {
431 return "text";
432 }
433
434 return "fail";
435}
436
437
438sub convertXLS {
439 my ($input_filename, $output_filestem, $output_type) = @_;
440
441 my $success = 0;
442
443 # Attempt conversion to HTML
444 if (!$output_type || ($output_type =~ m/html/i)) {
445 # formulate the command
446 my $cmd = "";
447 my $full_perl_path = &util::get_perl_exec();
448 $cmd .= "\"$full_perl_path\" -S xlstohtml.pl ";
449 $cmd .= " \"$input_filename\" \"$output_filestem.html\"";
450 $cmd .= " 2>\"$output_filestem.err\""
451 if ($ENV{'GSDLOS'} !~ m/^windows$/i || $is_winnt_2000);
452
453
454 # execute the command
455 $!=0;
456 if (system($cmd)!=0)
457 {
458 print STDERR "Excel 95/97 converter failed $!\n";
459 } else {
460 return "html";
461 }
462 }
463
464 $success = &any_to_text($input_filename, $output_filestem);
465 if ($success) {
466 return "text";
467 }
468
469 return "fail";
470}
471
472
473
474# Find the real type of a .doc file
475#
476# We seem to have a lot of files with a .doc extension that are .rtf
477# files or Word 5 files. This function attempts to tell the difference.
478sub find_docfile_type {
479 my ($input_filename) = @_;
480
481 if (($windows_scripting) && ($input_filename =~ m/\.docx$/)) {
482 return "docx";
483 }
484
485 open(CHK, "<$input_filename");
486 binmode(CHK);
487 my $line = "";
488 my $first = 1;
489
490 while (<CHK>) {
491
492 $line = $_;
493
494 if ($first) {
495 # check to see if this is an rtf file
496 if ($line =~ m/^\{\\rtf/) {
497 close(CHK);
498 return "rtf";
499 }
500 $first = 0;
501 }
502
503 # is this is a word 6/7/8 document?
504 if ($line =~ m/Word\.Document\.([678])/) {
505 close(CHK);
506
507 return "word$1";
508 }
509
510 }
511
512 return "unknown";
513}
514
515
516# Specific type-to-type conversions
517#
518# Each of the following functions attempts to convert a document from
519# a specific format to another. If they succeed they return 1 and leave
520# the output document(s) in the appropriate place; if they fail they
521# return 0 and delete any working files.
522
523
524# Attempt to convert a word document to html with the wv program
525sub doc_to_html {
526 my ($input_filename, $output_filestem) = @_;
527
528 my $wvware_status = 0;
529
530 # need to ensure that the path to perl is quoted (in case there's spaces in it)
531 my $launch_cmd = "\"".&util::get_perl_exec()."\" -S wvware.pl \"$input_filename\" \"$output_filestem\" \"$faillogfile\" $verbosity $timeout";
532
533# print STDERR "***** wvware launch cmd = $launch_cmd\n";
534
535 $wvware_status = system($launch_cmd)/256;
536 return $wvware_status;
537}
538
539# Attempt to convert a word document to html with the word2html scripting program
540sub native_doc_to_html {
541 my ($input_filename, $output_filestem) = @_;
542
543 # build up the path to the doc-to-html conversion tool we're going to use
544 my $vbScript = &FileUtils::filenameConcatenate($ENV{'GSDLHOME'}, "bin", $ENV{'GSDLOS'});
545
546 if ($ENV{'GSDLOS'} =~ m/^windows$/i) {
547 # if windows scripting with docx input, use new VBscript to get the local Word install (if
548 # any) to do the conversion, since docX can't be processed by word2html's windows_scripting
549
550 if($input_filename =~ m/docx$/i) { # need to use full path to docx2html script,
551 # else script launch fails when there are error msgs
552 $vbScript = &FileUtils::filenameConcatenate($vbScript, "docx2html.vbs");
553 $vbScript = "CScript //Nologo \"$vbScript\""; # launch with CScript for error output in STDERR
554 # //Nologo flag avoids Microsoft's opening/logo msgs
555 print STDERR "About to use windows scripting to process docx file $input_filename.\n";
556 print STDERR " This may take some time. Please wait...\n";
557 }
558 else { # old doc versions. use the usual VB executable word2html for the
559 # conversion. Doesn't need full path, since bin\windows is on PATH
560 $vbScript = "word2html"; #$vbScript = "\"".&FileUtils::filenameConcatenate($vbScript, "word2html")."\"";
561 }
562 }
563 else { # not windows
564 $vbScript = "\"".&FileUtils::filenameConcatenate($vbScript, "word2html")."\"";
565 }
566
567 if (-e "$output_filestem.html") {
568 print STDERR " The conversion file:\n";
569 print STDERR " $output_filestem.html\n";
570 print STDERR " ... already exists. Skipping\n";
571 return 1;
572 }
573
574 my $cmd = "";
575 if ($timeout) {$cmd = "ulimit -t $timeout;";}
576 #$cmd .= "$vbScript \"$input_filename\" \"$output_filestem.html\"";
577 #$cmd .= "$vbScript $input_filename $output_filestem.html";
578 $cmd .= "$vbScript \"$input_filename\" \"$output_filestem.html\"";
579
580 # redirecting STDERR
581
582 $cmd .= " 2> \"$output_filestem.err\""
583 if ($ENV {'GSDLOS'} !~ m/^windows$/i || $is_winnt_2000);
584 #print STDERR "@@@@@@@@@ cmd=$cmd\n";
585
586 # execute the command
587 $!=0;
588 if (system($cmd)!=0)
589 {
590 print STDERR "Error executing $vbScript converter:$!\n";
591 if (-s "$output_filestem.err") {
592 open (ERRFILE, "<$output_filestem.err");
593
594 my $write_to_fail_log=0;
595 if ($faillogfile ne "" && defined(open(FAILLOG,">>$faillogfile")))
596 {$write_to_fail_log=1;}
597
598 my $line;
599 while ($line=<ERRFILE>) {
600 if ($line =~ m/\w/) {
601 print STDERR "$line";
602 print FAILLOG "$line" if ($write_to_fail_log);
603 }
604 if ($line !~ m/startup error/) {next;}
605 print STDERR " (given an invalid .DOC file?)\n";
606 print FAILLOG " (given an invalid .DOC file?)\n"
607 if ($write_to_fail_log);
608
609 } # while ERRFILE
610 close FAILLOG if ($write_to_fail_log);
611 }
612 return 0; # we can try any_to_text
613 }
614
615 # Was the conversion successful?
616 if (-s "$output_filestem.html") {
617 open(TMP, "$output_filestem.html");
618 my $line = <TMP>;
619 close(TMP);
620 if ($line && $line =~ m/html/i) {
621 &FileUtils::removeFiles("$output_filestem.err") if -e "$output_filestem.err";
622 return 1;
623 }
624 }
625
626 # If here, an error of some sort occurred
627 &FileUtils::removeFiles("$output_filestem.html") if -e "$output_filestem.html";
628 if (-e "$output_filestem.err") {
629 if ($faillogfile ne "" && defined(open(FAILLOG,">>$faillogfile"))) {
630 open (ERRLOG,"$output_filestem.err");
631 while (<ERRLOG>) {print FAILLOG $_;}
632 close FAILLOG;
633 close ERRLOG;
634 }
635 &FileUtils::removeFiles("$output_filestem.err");
636 }
637 return 0;
638}
639
640# Attempt to convert an RTF document to html with rtftohtml
641sub rtf_to_html {
642 my ($input_filename, $output_filestem) = @_;
643
644 # formulate the command
645 my $cmd = "";
646 if ($timeout) {$cmd = "ulimit -t $timeout;";}
647 $cmd .= "rtftohtml";
648 #$cmd .= "rtf-converter";
649
650 $cmd .= " -o \"$output_filestem.html\" \"$input_filename\"";
651
652 $cmd .= " 2>\"$output_filestem.err\""
653 if ($ENV{'GSDLOS'} !~ m/^windows$/i || $is_winnt_2000);
654
655
656 # execute the command
657 $!=0;
658 if (system($cmd)!=0)
659 {
660 print STDERR "Error executing rtf converter $!\n";
661 # don't currently bother printing out error log...
662 # keep going, in case it still created an HTML file...
663 }
664
665 # Was the conversion successful?
666 my $was_successful=0;
667 if (-s "$output_filestem.html") {
668 # make sure we have some content other than header
669 open (HTML, "$output_filestem.html"); # what to do if fail?
670 my $line;
671 my $past_header=0;
672 while ($line=<HTML>) {
673
674 if ($past_header == 0) {
675 if ($line =~ m/<body>/) {$past_header=1;}
676 next;
677 }
678
679 $line =~ s/<[^>]+>//g;
680 if ($line =~ m/\w/ && $past_header) { # we found some content...
681 $was_successful=1;
682 last;
683 }
684 }
685 close HTML;
686 }
687
688 if ($was_successful) {
689 &FileUtils::removeFiles("$output_filestem.err")
690 if (-e "$output_filestem.err");
691 # insert the (modified) table of contents, if it exists.
692 if (-e "${output_filestem}_ToC.html") {
693 &FileUtils::moveFiles("$output_filestem.html","$output_filestem.src");
694 my $open_failed=0;
695 open HTMLSRC, "$output_filestem.src" || ++$open_failed;
696 open TOC, "${output_filestem}_ToC.html" || ++$open_failed;
697 open HTML, ">$output_filestem.html" || ++$open_failed;
698
699 if ($open_failed) {
700 close HTMLSRC;
701 close TOC;
702 close HTML;
703 &FileUtils::moveFiles("$output_filestem.src","$output_filestem.html");
704 return 1;
705 }
706
707 # print out header info from src html.
708 while (defined($_ = <HTMLSRC>) && $_ =~ m/\w/) {
709 print HTML "$_";
710 }
711
712 # print out table of contents, making links relative
713 <TOC>; <TOC>; # ignore first 2 lines
714 print HTML scalar(<TOC>); # line 3 = "<ol>\n"
715 my $line;
716 while ($line=<TOC>) {
717 $line =~ s@</body></html>$@@i ; # only last line has this
718 # make link relative
719 $line =~ s@href=\"[^\#]+@href=\"@i;
720 print HTML $line;
721 }
722 close TOC;
723
724 # rest of html src
725 while (<HTMLSRC>) {
726 print HTML $_;
727 }
728 close HTMLSRC;
729 close HTML;
730
731 &FileUtils::removeFiles("${output_filestem}_ToC.html");
732 &FileUtils::removeFiles("${output_filestem}.src");
733 }
734 # we don't yet do anything with footnotes ($output_filestem_fn.html) :(
735 return 1; # success
736 }
737
738 if (-e "$output_filestem.err") {
739 if ($faillogfile ne "" && defined(open(FAILLOG,">>$faillogfile")))
740 {
741 print FAILLOG "Error - rtftohtml - couldn't extract text\n";
742 #print FAILLOG "Error - rtf-converter - couldn't extract text\n";
743 print FAILLOG " (rtf file might be too recent):\n";
744 open (ERRLOG, "$output_filestem.err");
745 while (<ERRLOG>) {print FAILLOG $_;}
746 close ERRLOG;
747 close FAILLOG;
748 }
749 &FileUtils::removeFiles("$output_filestem.err");
750 }
751
752 &FileUtils::removeFiles("$output_filestem.html") if (-e "$output_filestem.html");
753
754 return 0;
755}
756
757
758# Convert a pdf file to html with the pdftohtml command
759
760sub pdf_to_html {
761 my ($dirname, $input_filename, $output_filestem) = @_;
762
763 my $cmd = "";
764 if ($timeout) {$cmd = "ulimit -t $timeout;";}
765 my $full_perl_path = &util::get_perl_exec();
766 $cmd .= "\"$full_perl_path\" -S pdftohtml.pl -zoom $pdf_zoom";
767 $cmd .= " -c" if ($pdf_complex);
768 $cmd .= " -i" if ($pdf_ignore_images);
769 $cmd .= " -a" if ($pdf_allow_images_only);
770 $cmd .= " -hidden" unless ($pdf_nohidden);
771 $cmd .= " \"$input_filename\" \"$output_filestem\"";
772
773 if ($ENV{'GSDLOS'} !~ m/^windows$/i || $is_winnt_2000) {
774 $cmd .= " > \"$output_filestem.out\" 2> \"$output_filestem.err\"";
775 } else {
776 $cmd .= " > \"$output_filestem.err\"";
777 }
778
779 $!=0;
780
781 my $retval=system($cmd);
782 if ($retval!=0)
783 {
784 print STDERR "Error executing pdftohtml.pl";
785 if ($!) {print STDERR ": $!";}
786 print STDERR "\n";
787 }
788
789 # make sure the converter made something
790 if ($retval!=0 || ! -s "$output_filestem.html")
791 {
792 &FileUtils::removeFiles("$output_filestem.out") if (-e "$output_filestem.out");
793 # print out the converter's std err, if any
794 if (-s "$output_filestem.err") {
795 open (ERRLOG, "$output_filestem.err") || die "$!";
796 print STDERR "pdftohtml error log:\n";
797 while (<ERRLOG>) {
798 print STDERR "$_";
799 }
800 close ERRLOG;
801 }
802 #print STDERR "***********output filestem $output_filestem.html\n";
803 &FileUtils::removeFiles("$output_filestem.html") if (-e "$output_filestem.html");
804 if (-e "$output_filestem.err") {
805 if ($faillogfile ne "" && defined(open(FAILLOG,">>$faillogfile")))
806 {
807 open (ERRLOG, "$output_filestem.err");
808 while (<ERRLOG>) {print FAILLOG $_;}
809 close ERRLOG;
810 close FAILLOG;
811 }
812 &FileUtils::removeFiles("$output_filestem.err");
813 }
814 return 0;
815 }
816
817 &FileUtils::removeFiles("$output_filestem.err") if (-e "$output_filestem.err");
818 &FileUtils::removeFiles("$output_filestem.out") if (-e "$output_filestem.out");
819 return 1;
820}
821
822# Convert a pdf file to various types of image with the convert command
823
824sub pdfps_to_img {
825 my ($dirname, $input_filename, $output_filestem, $output_type) = @_;
826
827 # Check that ImageMagick is installed and available on the path (except for Windows 95/98)
828 if (!($ENV{'GSDLOS'} eq "windows" && !Win32::IsWinNT())) {
829 my $imagick_cmd = "\"".&util::get_perl_exec()."\" -S gs-magick.pl";
830 $imagick_cmd = $imagick_cmd." --verbosity=$verbosity" if defined $verbosity;
831 my $result = `$imagick_cmd identify 2>&1`;
832
833 # Linux and Windows return different values for "program not found".
834 # Linux returns -1 and Windows 256 for "program not found". But once they're
835 # converted to signed values, it will be -1 for Linux and 1 for Windows.
836 # Whenever we test for return values other than 0, shift by 8 and perform
837 # unsigned to signed status conversion on $? to get expected range of return vals
838 # Although gs-magick.pl already shifts its $? by 8, converts it to a signed value
839 # and then exits on that, by the time we get here, we need to do it again
840 my $status = $?;
841 $status >>= 8;
842 $status = (($status & 0x80) ? -(0x100 - ($status & 0xFF)) : $status);
843 if (($ENV{'GSDLOS'} ne "windows" && $status == -1) || ($ENV{'GSDLOS'} eq "windows" && $status == 1)) {
844 # if ($status == -1 || $status == 1) #if ($status == -1 || $status == 256) {
845 #ImageMagick is not installed, thus the convert utility is not available.
846 print STDERR "*** ImageMagick is not installed, the convert utility is not available. Unable to convert PDF/PS to images. Status: $status\n";
847 return 0;
848 }
849 }
850
851 my $cmd = "";
852 if ($timeout) {$cmd = "ulimit -t $timeout;";}
853 $output_type =~ s/.*\_(.*)/$1/i;
854 my $full_perl_path = &util::get_perl_exec();
855 $cmd .= "\"$full_perl_path\" -S pdfpstoimg.pl -convert_to $output_type \"$input_filename\" \"$output_filestem\"";
856 if ($ENV{'GSDLOS'} !~ m/^windows$/i || $is_winnt_2000) {
857 $cmd .= " > \"$output_filestem.out\" 2> \"$output_filestem.err\"";
858 } else {
859 $cmd .= " > \"$output_filestem.err\"";
860 }
861
862 # don't include path on windows (to avoid having to play about
863 # with quoting when GSDLHOME might contain spaces) but assume
864 # that the PATH is set up correctly
865 $!=0;
866 my $retval=system($cmd);
867 if ($retval!=0)
868 {
869 print STDERR "Error executing pdfpstoimg.pl";
870 if ($!) {print STDERR ": $!";}
871 print STDERR "\n";
872 }
873
874 #make sure the converter made something
875 #if ($retval !=0) || ! -s "$output_filestem")
876 if ($retval !=0)
877 {
878 &FileUtils::removeFiles("$output_filestem.out") if (-e "$output_filestem.out");
879 #print out the converter's std err, if any
880 if (-s "$output_filestem.err") {
881 open (ERRLOG, "$output_filestem.err") || die "$!";
882 print STDERR "pdfpstoimg error log:\n";
883 while (<ERRLOG>) {
884 print STDERR "$_";
885 }
886 close ERRLOG;
887 }
888 #&FileUtils::removeFiles("$output_filestem.html") if (-e "$output_filestem.html");
889 if (-e "$output_filestem.err") {
890 if ($faillogfile ne "" && defined(open(FAILLOG,">>$faillogfile")))
891 {
892 open (ERRLOG, "$output_filestem.err");
893 while (<ERRLOG>) {print FAILLOG $_;}
894 close ERRLOG;
895 close FAILLOG;
896 }
897 &FileUtils::removeFiles("$output_filestem.err");
898 }
899 return 0;
900 }
901 &FileUtils::removeFiles("$output_filestem.err") if (-e "$output_filestem.err");
902 &FileUtils::removeFiles("$output_filestem.out") if (-e "$output_filestem.out");
903 return 1;
904}
905
906# Convert a PDF file to text with the pdftotext command
907
908sub pdf_to_text {
909 my ($dirname, $input_filename, $output_filestem) = @_;
910
911 my $cmd = "pdftotext \"$input_filename\" \"$output_filestem.text\"";
912
913 if ($ENV{'GSDLOS'} !~ m/^windows$/i) {
914 $cmd .= " > \"$output_filestem.out\" 2> \"$output_filestem.err\"";
915 } else {
916 $cmd .= " > \"$output_filestem.err\"";
917 }
918
919 if (system($cmd)!=0)
920 {
921 print STDERR "Error executing $cmd: $!\n";
922 &FileUtils::removeFiles("$output_filestem.text") if (-e "$output_filestem.text");
923 }
924
925 # make sure there is some extracted text.
926 if (-e "$output_filestem.text") {
927 open (EXTR_TEXT, "$output_filestem.text") || warn "open: $!";
928 binmode(EXTR_TEXT); # just in case...
929 my $line="";
930 my $seen_text=0;
931 while (($seen_text==0) && ($line=<EXTR_TEXT>)) {
932 if ($line=~ m/\w/) {$seen_text=1;}
933 }
934 close EXTR_TEXT;
935 if ($seen_text==0) { # no text was extracted
936 print STDERR "Error: pdftotext found no text\n";
937 &FileUtils::removeFiles("$output_filestem.text");
938 }
939 }
940
941 # make sure the converter made something
942 if (! -s "$output_filestem.text")
943 {
944 # print out the converters std err, if any
945 if (-s "$output_filestem.err") {
946 open (ERRLOG, "$output_filestem.err") || die "$!";
947 print STDERR "pdftotext error log:\n";
948 while (<ERRLOG>) {
949 print STDERR "$_";
950 }
951 close ERRLOG;
952 }
953 # does this converter create a .out file?
954 &FileUtils::removeFiles("$output_filestem.out") if (-e "$output_filestem.out");
955 &FileUtils::removeFiles("$output_filestem.text") if (-e "$output_filestem.text");
956 if (-e "$output_filestem.err") {
957 if ($faillogfile ne "" && defined(open(FAILLOG,">>$faillogfile")))
958 {
959 open (ERRLOG,"$output_filestem.err");
960 while (<ERRLOG>) {print FAILLOG $_;}
961 close ERRLOG;
962 close FAILLOG;
963 }
964 &FileUtils::removeFiles("$output_filestem.err");
965 }
966 return 0;
967 }
968 &FileUtils::removeFiles("$output_filestem.err") if (-e "$output_filestem.err");
969 return 1;
970}
971
972# Convert a PostScript document to text
973# note - just using "ps2ascii" isn't good enough, as it
974# returns 0 for a postscript interpreter error. ps2ascii is just
975# a wrapper to "gs" anyway, so we use that cmd here.
976
977sub ps_to_text {
978 my ($input_filename, $output_filestem) = @_;
979
980 my $error = "";
981
982 # if we're on windows we'll fall straight through without attempting
983 # to use gs
984 if ($ENV{'GSDLOS'} =~ m/^windows$/i) {
985 $error = "Windows does not support gs";
986
987 } else {
988 my $cmd = "";
989 if ($timeout) {$cmd = "ulimit -t $timeout; ";}
990 $cmd .= "gs -q -dNODISPLAY -dNOBIND -dWRITESYSTEMDICT -dSIMPLE -c save ";
991 $cmd .= "-f ps2ascii.ps \"$input_filename\" -c quit > \"$output_filestem.text\"";
992 #$cmd .= "pstotext -output \"$output_filestem.text\" $input_filename\"";
993 $cmd .= " 2> $output_filestem.err";
994 $!=0;
995
996 my $retcode=system($cmd);
997 $retcode = $? >> 8; # see man perlfunc - system for this...
998 # if system returns -1 | 127 (couldn't start program), look at $! for message
999
1000 if ($retcode!=0) {if ($!) {$error=$!;} else {$error="couldn't run.\n";}}
1001 elsif (! -e "$output_filestem.text") {
1002 $error="did not create output file.\n";
1003 }
1004 else
1005 { # make sure the interpreter didn't get an error. It is technically
1006 # possible for the actual text to start with this, but....
1007 open PSOUT, "$output_filestem.text";
1008 if (<PSOUT> =~ m/^Error: (.*)/) {
1009 $error="interpreter error - \"$1\"";
1010 }
1011 close PSOUT;
1012 }
1013 }
1014
1015 if ($error ne "")
1016 {
1017 print STDERR "Warning: Error executing gs: $error\n";
1018 &FileUtils::removeFiles("$output_filestem.text") if (-e "$output_filestem.text");
1019
1020 if ("$faillogfile" ne "" && defined(open (FAILLOG, ">>$faillogfile")))
1021 {
1022 print FAILLOG "gs - $error\n";
1023 if (-e "$output_filestem.err") {
1024 open(ERRLOG, "$output_filestem.err");
1025 while (<ERRLOG>) {print FAILLOG $_;}
1026 close ERRLOG;
1027 }
1028 close FAILLOG;
1029 }
1030 &FileUtils::removeFiles("$output_filestem.err") if (-e "$output_filestem.err");
1031
1032
1033 # Fine then. We'll just do a lousy job by ourselves...
1034 # Based on 5-line regexp sed script found at:
1035 # http://snark.ptc.spbu.ru/mail-archives/lout/brown/msg00003.html
1036 #
1037 print STDERR "Stripping text from postscript\n";
1038 my $errorcode=0;
1039 open (IN, "$input_filename")
1040 || ($errorcode=1, warn "Couldn't read file: $!");
1041 open (OUT, ">$output_filestem.text")
1042 || ($errorcode=1, warn "Couldn't write file: $!");
1043 if ($errorcode) {print STDERR "errors\n";return 0;}
1044
1045 my $text=""; # this is for whole .ps file...
1046 $text = join('', <IN>); # see man perlport, under "System Resources"
1047 close IN;
1048
1049 # Make sure this is a ps file...
1050 if ($text !~ m/^%!/) {
1051 print STDERR "Bad postscript header: not '%!'\n";
1052 if ($faillogfile ne "" && defined(open(FAILLOG, ">>$faillogfile")))
1053 {
1054 print FAILLOG "Bad postscript header: not '%!'\n";
1055 close FAILLOG;
1056 }
1057 return 0;
1058 }
1059
1060 # if ps has Page data, then use it to delete all stuff before it.
1061 $text =~ s/^.*?%%Page:.*?\n//s; # treat string as single line
1062
1063 # remove all leading non-data stuff
1064 $text =~ s/^.*?\(//s;
1065
1066 # remove all newline chars for easier processing
1067 $text =~ s/\n//g;
1068
1069 # Big assumption here - assume that if any co-ordinates are
1070 # given, then we are at the end of a sentence.
1071 $text =~ s/\)-?\d+\ -?\d+/\) \(\n\)/g;
1072
1073 # special characters--
1074 $text =~ s/\(\|\)/\(\ - \)/g; # j -> em-dash?
1075
1076 # ? ps text formatting (eg italics?) ?
1077 $text =~ s/Fn\(f\)/\(\{\)/g; # f -> {
1078 $text =~ s/Fn\(g\)/\(\}\)/g; # g -> }
1079 $text =~ s/Fn\(j\)/\(\|\)/g; # j -> |
1080 # default - remove the rest
1081 $text =~ s/\ ?F.\((.+?)\)/\($1\)/g;
1082
1083 # attempt to add whitespace between words...
1084 # this is based purely on observation, and may be completely wrong...
1085 $text =~ s/([^F])[defghijkuy]\(/$1 \( /g;
1086 # eg I notice "b(" is sometimes NOT a space if preceded by a
1087 # negative number.
1088 $text =~ s/\)\d+ ?b\(/\) \( /g;
1089
1090 # change quoted braces to brackets
1091 $text =~ s/([^\\])\\\(/$1\{/g;
1092 $text =~ s/([^\\])\\\)/$1\}/g ;
1093
1094 # remove everything that is not between braces
1095 $text =~ s/\)([^\(\)])+?\(//sg ;
1096
1097 # remove any Trailer eof stuff.
1098 $text =~ s/\)[^\)]*$//sg;
1099
1100 ### ligatures have special characters...
1101 $text =~ s/\\013/ff/g;
1102 $text =~ s/\\014/fi/g;
1103 $text =~ s/\\015/fl/g;
1104 $text =~ s/\\016/ffi/g;
1105 $text =~ s/\\214/fi/g;
1106 $text =~ s/\\215/fl/g;
1107 $text =~ s/\\017/\n\* /g; # asterisk?
1108 $text =~ s/\\023/\023/g; # e acute ('e)
1109 $text =~ s/\\177/\252/g; # u"
1110# $text =~ s/ ?? /\344/g; # a"
1111
1112 print OUT "$text";
1113 close OUT;
1114 }
1115 # wrap the text - use a minimum length. ie, first space after this length.
1116 my $wrap_length=72;
1117 &FileUtils::moveFiles("$output_filestem.text", "$output_filestem.text.tmp");
1118 open INFILE, "$output_filestem.text.tmp" ||
1119 die "Couldn't open file: $!";
1120 open OUTFILE, ">$output_filestem.text" ||
1121 die "Couldn't open file for writing: $!";
1122 my $line="";
1123 while ($line=<INFILE>) {
1124 while (length($line)>0) {
1125 if (length($line)>$wrap_length) {
1126 $line =~ s/^(.{$wrap_length}[^\s]*)\s*//;
1127 print OUTFILE "$1\n";
1128 } else {
1129 print OUTFILE "$line";
1130 $line="";
1131 }
1132 }
1133 }
1134 close INFILE;
1135 close OUTFILE;
1136 &FileUtils::removeFiles("$output_filestem.text.tmp");
1137
1138 &FileUtils::removeFiles("$output_filestem.err") if (-e "$output_filestem.err");
1139 return 1;
1140}
1141
1142
1143# Convert any file to HTML with a crude perl implementation of the
1144# UNIX strings command.
1145
1146sub any_to_html {
1147 my ($input_filename, $output_filestem) = @_;
1148
1149 # First generate a text file
1150 return 0 unless (&any_to_text($input_filename, $output_filestem));
1151
1152 # create an HTML file from the text file
1153 open(TEXT, "<$output_filestem.text");
1154 open(HTML, ">$output_filestem.html");
1155
1156 print HTML "<html><head>\n";
1157 print HTML "<META HTTP-EQUIV=\"Content-Type\" CONTENT=\"text/html\">\n";
1158 print HTML "<META NAME=\"GENERATOR\" CONTENT=\"Greenstone any_to_html\">\n";
1159 print HTML "</head><body>\n\n";
1160
1161 my $line;
1162 while ($line=<TEXT>) {
1163 $line =~ s/</&lt;/g;
1164 $line =~ s/>/&gt;/g;
1165 if ($line =~ m/^\s*$/) {
1166 print HTML "<p>";
1167 } else {
1168 print HTML "<br> ", $line;
1169 }
1170 }
1171 print HTML "\n</body></html>\n";
1172
1173 close HTML;
1174 close TEXT;
1175
1176 &FileUtils::removeFiles("$output_filestem.text") if (-e "$output_filestem.text");
1177 return 1;
1178}
1179
1180# Convert any file to TEXT with a crude perl implementation of the
1181# UNIX strings command.
1182# Note - this assumes ascii charsets :( (jrm21)
1183
1184sub any_to_text {
1185 my ($input_filename, $output_filestem) = @_;
1186
1187 if (!$use_strings) {
1188 return 0;
1189 }
1190
1191 print STDERR "\n**** In any to text****\n\n";
1192 open(IN, "<$input_filename") || return 0;
1193 binmode(IN);
1194 open(OUT, ">$output_filestem.text") || return 0;
1195
1196 my ($line);
1197 my $output_line_count = 0;
1198 while (<IN>) {
1199 $line = $_;
1200
1201 # delete anything that isn't a printable character
1202 $line =~ s/[^\040-\176]+/\n/sg;
1203
1204 # delete any string less than 10 characters long
1205 $line =~ s/^.{0,9}$/\n/mg;
1206 while ($line =~ m/^.{1,9}$/m) {
1207 $line =~ s/^.{0,9}$/\n/mg;
1208 $line =~ s/\n+/\n/sg;
1209 }
1210
1211 # remove extraneous whitespace
1212 $line =~ s/\n+/\n/gs;
1213 $line =~ s/^\n//gs;
1214
1215 # output whatever is left
1216 if ($line =~ m/[^\n ]/) {
1217 print OUT $line;
1218 ++$output_line_count;
1219 }
1220 }
1221
1222 close OUT;
1223 close IN;
1224
1225 if ($output_line_count) { # try to protect against binary only formats
1226 return 1;
1227 }
1228
1229 &FileUtils::removeFiles("$output_filestem.text");
1230 return 0;
1231
1232}
Note: See TracBrowser for help on using the repository browser.